id
large_stringlengths 9
16
| submitter
large_stringlengths 3
64
⌀ | authors
large_stringlengths 5
20.2k
| title
large_stringlengths 5
245
| comments
large_stringlengths 1
482
⌀ | journal-ref
large_stringlengths 4
382
⌀ | doi
large_stringlengths 9
151
⌀ | report-no
large_stringlengths 2
509
⌀ | categories
large_stringlengths 5
78
| license
large_stringclasses 9
values | abstract
large_stringlengths 83
3.55k
| versions
large listlengths 1
21
| update_date
timestamp[ms]date 2007-05-23 00:00:00
2025-05-30 00:00:00
| authors_parsed
large listlengths 1
1.35k
| prompt
large_stringlengths 166
3.64k
| label
stringclasses 2
values | prob
float64 0.5
0.73
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2305.05627 | Ilias Chalkidis | Yova Kementchedjhieva and Ilias Chalkidis | An Exploration of Encoder-Decoder Approaches to Multi-Label
Classification for Legal and Biomedical Text | 9 pages, long paper at ACL 2023 Findings | null | null | null | cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | Standard methods for multi-label text classification largely rely on
encoder-only pre-trained language models, whereas encoder-decoder models have
proven more effective in other classification tasks. In this study, we compare
four methods for multi-label classification, two based on an encoder only, and
two based on an encoder-decoder. We carry out experiments on four datasets --
two in the legal domain and two in the biomedical domain, each with two levels
of label granularity -- and always depart from the same pre-trained model, T5.
Our results show that encoder-decoder methods outperform encoder-only methods,
with a growing advantage on more complex datasets and labeling schemes of finer
granularity. Using encoder-decoder models in a non-autoregressive fashion, in
particular, yields the best performance overall, so we further study this
approach through ablations to better understand its strengths.
| [
{
"version": "v1",
"created": "Tue, 9 May 2023 17:13:53 GMT"
}
] | 2023-05-10T00:00:00 | [
[
"Kementchedjhieva",
"Yova",
""
],
[
"Chalkidis",
"Ilias",
""
]
] | TITLE: An Exploration of Encoder-Decoder Approaches to Multi-Label
Classification for Legal and Biomedical Text
ABSTRACT: Standard methods for multi-label text classification largely rely on
encoder-only pre-trained language models, whereas encoder-decoder models have
proven more effective in other classification tasks. In this study, we compare
four methods for multi-label classification, two based on an encoder only, and
two based on an encoder-decoder. We carry out experiments on four datasets --
two in the legal domain and two in the biomedical domain, each with two levels
of label granularity -- and always depart from the same pre-trained model, T5.
Our results show that encoder-decoder methods outperform encoder-only methods,
with a growing advantage on more complex datasets and labeling schemes of finer
granularity. Using encoder-decoder models in a non-autoregressive fashion, in
particular, yields the best performance overall, so we further study this
approach through ablations to better understand its strengths.
| no_new_dataset | 0.710653 |
2305.05652 | Jinfeng Liu | Long Wu, Xunyuan Yin, Lei Pan, Jinfeng Liu (University of Alberta) | Distributed economic predictive control of integrated energy systems for
enhanced synergy and grid response: A decomposition and cooperation strategy | null | null | null | null | eess.SY cs.SY math.DS | http://creativecommons.org/licenses/by/4.0/ | The close integration of increasing operating units into an integrated energy
system (IES) results in complex interconnections between these units. The
strong dynamic interactions create barriers to designing a successful
distributed coordinated controller to achieve synergy between all the units and
unlock the potential for grid response. To address these challenges, we
introduce a directed graph representation of IESs using an augmented Jacobian
matrix to depict their underlying dynamics topology. By utilizing this
representation, a generic subsystem decomposition method is proposed to
partition the entire IES vertically based on the dynamic time scale and
horizontally based on the closeness of interconnections between the operating
units. Exploiting the decomposed subsystems, we develop a cooperative
distributed economic model predictive control (DEMPC) with multiple global
objectives that regulate the generated power at the grid's requests and satisfy
the customers cooling and system economic requirements. In the DEMPC, multiple
local decision-making agents cooperate sequentially and iteratively to leverage
the potential across all the units for system-wide dynamic synergy.
Furthermore, we discuss how subsystem decomposition impacts the design of
distributed cooperation schemes for IESs and provide a control-oriented basic
guideline on the optimal decomposition of complex energy systems. Extensive
simulations demonstrate that the control strategies with different levels of
decomposition and collaboration will lead to marked differences in the overall
performance of IES. The standard control scheme based on the proposed subsystem
configuration outperforms the empirical decomposition-based control benchmark
by about 20%. The DEMPC architecture further improves the overall performance
of the IES by about 55% compared to the benchmark.
| [
{
"version": "v1",
"created": "Tue, 9 May 2023 17:49:28 GMT"
}
] | 2023-05-10T00:00:00 | [
[
"Wu",
"Long",
"",
"University of Alberta"
],
[
"Yin",
"Xunyuan",
"",
"University of Alberta"
],
[
"Pan",
"Lei",
"",
"University of Alberta"
],
[
"Liu",
"Jinfeng",
"",
"University of Alberta"
]
] | TITLE: Distributed economic predictive control of integrated energy systems for
enhanced synergy and grid response: A decomposition and cooperation strategy
ABSTRACT: The close integration of increasing operating units into an integrated energy
system (IES) results in complex interconnections between these units. The
strong dynamic interactions create barriers to designing a successful
distributed coordinated controller to achieve synergy between all the units and
unlock the potential for grid response. To address these challenges, we
introduce a directed graph representation of IESs using an augmented Jacobian
matrix to depict their underlying dynamics topology. By utilizing this
representation, a generic subsystem decomposition method is proposed to
partition the entire IES vertically based on the dynamic time scale and
horizontally based on the closeness of interconnections between the operating
units. Exploiting the decomposed subsystems, we develop a cooperative
distributed economic model predictive control (DEMPC) with multiple global
objectives that regulate the generated power at the grid's requests and satisfy
the customers cooling and system economic requirements. In the DEMPC, multiple
local decision-making agents cooperate sequentially and iteratively to leverage
the potential across all the units for system-wide dynamic synergy.
Furthermore, we discuss how subsystem decomposition impacts the design of
distributed cooperation schemes for IESs and provide a control-oriented basic
guideline on the optimal decomposition of complex energy systems. Extensive
simulations demonstrate that the control strategies with different levels of
decomposition and collaboration will lead to marked differences in the overall
performance of IES. The standard control scheme based on the proposed subsystem
configuration outperforms the empirical decomposition-based control benchmark
by about 20%. The DEMPC architecture further improves the overall performance
of the IES by about 55% compared to the benchmark.
| no_new_dataset | 0.708389 |
2305.05661 | R. Kenny Jones | R. Kenny Jones and Paul Guerrero and Niloy J. Mitra and Daniel Ritchie | ShapeCoder: Discovering Abstractions for Visual Programs from
Unstructured Primitives | SIGGRAPH 2023 | null | null | null | cs.GR cs.AI cs.CV cs.LG cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Programs are an increasingly popular representation for visual data, exposing
compact, interpretable structure that supports manipulation. Visual programs
are usually written in domain-specific languages (DSLs). Finding "good"
programs, that only expose meaningful degrees of freedom, requires access to a
DSL with a "good" library of functions, both of which are typically authored by
domain experts. We present ShapeCoder, the first system capable of taking a
dataset of shapes, represented with unstructured primitives, and jointly
discovering (i) useful abstraction functions and (ii) programs that use these
abstractions to explain the input shapes. The discovered abstractions capture
common patterns (both structural and parametric) across the dataset, so that
programs rewritten with these abstractions are more compact, and expose fewer
degrees of freedom. ShapeCoder improves upon previous abstraction discovery
methods, finding better abstractions, for more complex inputs, under less
stringent input assumptions. This is principally made possible by two
methodological advancements: (a) a shape to program recognition network that
learns to solve sub-problems and (b) the use of e-graphs, augmented with a
conditional rewrite scheme, to determine when abstractions with complex
parametric expressions can be applied, in a tractable manner. We evaluate
ShapeCoder on multiple datasets of 3D shapes, where primitive decompositions
are either parsed from manual annotations or produced by an unsupervised cuboid
abstraction method. In all domains, ShapeCoder discovers a library of
abstractions that capture high-level relationships, remove extraneous degrees
of freedom, and achieve better dataset compression compared with alternative
approaches. Finally, we investigate how programs rewritten to use discovered
abstractions prove useful for downstream tasks.
| [
{
"version": "v1",
"created": "Tue, 9 May 2023 17:55:48 GMT"
}
] | 2023-05-10T00:00:00 | [
[
"Jones",
"R. Kenny",
""
],
[
"Guerrero",
"Paul",
""
],
[
"Mitra",
"Niloy J.",
""
],
[
"Ritchie",
"Daniel",
""
]
] | TITLE: ShapeCoder: Discovering Abstractions for Visual Programs from
Unstructured Primitives
ABSTRACT: Programs are an increasingly popular representation for visual data, exposing
compact, interpretable structure that supports manipulation. Visual programs
are usually written in domain-specific languages (DSLs). Finding "good"
programs, that only expose meaningful degrees of freedom, requires access to a
DSL with a "good" library of functions, both of which are typically authored by
domain experts. We present ShapeCoder, the first system capable of taking a
dataset of shapes, represented with unstructured primitives, and jointly
discovering (i) useful abstraction functions and (ii) programs that use these
abstractions to explain the input shapes. The discovered abstractions capture
common patterns (both structural and parametric) across the dataset, so that
programs rewritten with these abstractions are more compact, and expose fewer
degrees of freedom. ShapeCoder improves upon previous abstraction discovery
methods, finding better abstractions, for more complex inputs, under less
stringent input assumptions. This is principally made possible by two
methodological advancements: (a) a shape to program recognition network that
learns to solve sub-problems and (b) the use of e-graphs, augmented with a
conditional rewrite scheme, to determine when abstractions with complex
parametric expressions can be applied, in a tractable manner. We evaluate
ShapeCoder on multiple datasets of 3D shapes, where primitive decompositions
are either parsed from manual annotations or produced by an unsupervised cuboid
abstraction method. In all domains, ShapeCoder discovers a library of
abstractions that capture high-level relationships, remove extraneous degrees
of freedom, and achieve better dataset compression compared with alternative
approaches. Finally, we investigate how programs rewritten to use discovered
abstractions prove useful for downstream tasks.
| no_new_dataset | 0.708584 |
1810.00946 | Makoto Onizuka | Seiji Maekawa, Koh Takeuch, Makoto Onizuka | Non-linear Attributed Graph Clustering by Symmetric NMF with PU Learning | null | null | 10.2197/ipsjjip.28.427 | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the clustering problem of attributed graphs. Our challenge is how
we can design an effective and efficient clustering method that precisely
captures the hidden relationship between the topology and the attributes in
real-world graphs. We propose Non-linear Attributed Graph Clustering by
Symmetric Non-negative Matrix Factorization with Positive Unlabeled Learning.
The features of our method are three holds. 1) it learns a non-linear
projection function between the different cluster assignments of the topology
and the attributes of graphs so as to capture the complicated relationship
between the topology and the attributes in real-world graphs, 2) it leverages
the positive unlabeled learning to take the effect of partially observed
positive edges into the cluster assignment, and 3) it achieves efficient
computational complexity, $O((n^2+mn)kt)$, where $n$ is the vertex size, $m$ is
the attribute size, $k$ is the number of clusters, and $t$ is the number of
iterations for learning the cluster assignment. We conducted experiments
extensively for various clustering methods with various real datasets to
validate that our method outperforms the former clustering methods regarding
the clustering quality.
| [
{
"version": "v1",
"created": "Fri, 21 Sep 2018 12:05:43 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Maekawa",
"Seiji",
""
],
[
"Takeuch",
"Koh",
""
],
[
"Onizuka",
"Makoto",
""
]
] | TITLE: Non-linear Attributed Graph Clustering by Symmetric NMF with PU Learning
ABSTRACT: We consider the clustering problem of attributed graphs. Our challenge is how
we can design an effective and efficient clustering method that precisely
captures the hidden relationship between the topology and the attributes in
real-world graphs. We propose Non-linear Attributed Graph Clustering by
Symmetric Non-negative Matrix Factorization with Positive Unlabeled Learning.
The features of our method are three holds. 1) it learns a non-linear
projection function between the different cluster assignments of the topology
and the attributes of graphs so as to capture the complicated relationship
between the topology and the attributes in real-world graphs, 2) it leverages
the positive unlabeled learning to take the effect of partially observed
positive edges into the cluster assignment, and 3) it achieves efficient
computational complexity, $O((n^2+mn)kt)$, where $n$ is the vertex size, $m$ is
the attribute size, $k$ is the number of clusters, and $t$ is the number of
iterations for learning the cluster assignment. We conducted experiments
extensively for various clustering methods with various real datasets to
validate that our method outperforms the former clustering methods regarding
the clustering quality.
| no_new_dataset | 0.710434 |
1812.05380 | Abhishek Tiwari | Abhishek Tiwari, Sascha Gro{\ss}, Christian Hammer | IIFA: Modular Inter-app Intent Information Flow Analysis of Android
Applications | null | null | null | null | cs.SE cs.CR cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Android apps cooperate through message passing via intents. However, when
apps do not have identical sets of privileges inter-app communication (IAC) can
accidentally or maliciously be misused, e.g., to leak sensitive information
contrary to users expectations. Recent research considered static program
analysis to detect dangerous data leaks due to inter-component communication
(ICC) or IAC, but suffers from shortcomings with respect to precision,
soundness, and scalability. To solve these issues we propose a novel approach
for static ICC/IAC analysis. We perform a fixed-point iteration of ICC/IAC
summary information to precisely resolve intent communication with more than
two apps involved. We integrate these results with information flows generated
by a baseline (i.e. not considering intents) information flow analysis, and
resolve if sensitive data is flowing (transitively) through components/apps in
order to be ultimately leaked. Our main contribution is the first fully
automatic sound and precise ICC/IAC information flow analysis that is scalable
for realistic apps due to modularity, avoiding combinatorial explosion: Our
approach determines communicating apps using short summaries rather than
inlining intent calls, which often requires simultaneously analyzing all tuples
of apps. We evaluated our tool IIFA in terms of scalability, precision, and
recall. Using benchmarks we establish that precision and recall of our
algorithm are considerably better than prominent state-of-the-art analyses for
IAC. But foremost, applied to the 90 most popular applications from the Google
Playstore, IIFA demonstrated its scalability to a large corpus of real-world
apps. IIFA reports 62 problematic ICC-/IAC-related information flows via two or
more apps/components.
| [
{
"version": "v1",
"created": "Thu, 13 Dec 2018 12:27:11 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Tiwari",
"Abhishek",
""
],
[
"Groß",
"Sascha",
""
],
[
"Hammer",
"Christian",
""
]
] | TITLE: IIFA: Modular Inter-app Intent Information Flow Analysis of Android
Applications
ABSTRACT: Android apps cooperate through message passing via intents. However, when
apps do not have identical sets of privileges inter-app communication (IAC) can
accidentally or maliciously be misused, e.g., to leak sensitive information
contrary to users expectations. Recent research considered static program
analysis to detect dangerous data leaks due to inter-component communication
(ICC) or IAC, but suffers from shortcomings with respect to precision,
soundness, and scalability. To solve these issues we propose a novel approach
for static ICC/IAC analysis. We perform a fixed-point iteration of ICC/IAC
summary information to precisely resolve intent communication with more than
two apps involved. We integrate these results with information flows generated
by a baseline (i.e. not considering intents) information flow analysis, and
resolve if sensitive data is flowing (transitively) through components/apps in
order to be ultimately leaked. Our main contribution is the first fully
automatic sound and precise ICC/IAC information flow analysis that is scalable
for realistic apps due to modularity, avoiding combinatorial explosion: Our
approach determines communicating apps using short summaries rather than
inlining intent calls, which often requires simultaneously analyzing all tuples
of apps. We evaluated our tool IIFA in terms of scalability, precision, and
recall. Using benchmarks we establish that precision and recall of our
algorithm are considerably better than prominent state-of-the-art analyses for
IAC. But foremost, applied to the 90 most popular applications from the Google
Playstore, IIFA demonstrated its scalability to a large corpus of real-world
apps. IIFA reports 62 problematic ICC-/IAC-related information flows via two or
more apps/components.
| no_new_dataset | 0.707967 |
2002.02572 | Enmao Diao | Enmao Diao, Jie Ding, Vahid Tarokh | Multimodal Controller for Generative Models | null | null | 10.1007/978-981-19-7867-8_10 | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Class-conditional generative models are crucial tools for data generation
from user-specified class labels. Existing approaches for class-conditional
generative models require nontrivial modifications of backbone generative
architectures to model conditional information fed into the model. This paper
introduces a plug-and-play module named `multimodal controller' to generate
multimodal data without introducing additional learning parameters. In the
absence of the controllers, our model reduces to non-conditional generative
models. We test the efficacy of multimodal controllers on CIFAR10, COIL100, and
Omniglot benchmark datasets. We demonstrate that multimodal controlled
generative models (including VAE, PixelCNN, Glow, and GAN) can generate
class-conditional images of significantly better quality when compared with
conditional generative models. Moreover, we show that multimodal controlled
models can also create novel modalities of images.
| [
{
"version": "v1",
"created": "Fri, 7 Feb 2020 00:55:10 GMT"
},
{
"version": "v2",
"created": "Thu, 11 Jun 2020 17:24:54 GMT"
},
{
"version": "v3",
"created": "Fri, 12 Jun 2020 15:04:30 GMT"
},
{
"version": "v4",
"created": "Fri, 11 Sep 2020 22:44:57 GMT"
},
{
"version": "v5",
"created": "Fri, 16 Oct 2020 14:43:11 GMT"
},
{
"version": "v6",
"created": "Mon, 29 Mar 2021 00:08:28 GMT"
},
{
"version": "v7",
"created": "Wed, 3 Aug 2022 05:33:45 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Diao",
"Enmao",
""
],
[
"Ding",
"Jie",
""
],
[
"Tarokh",
"Vahid",
""
]
] | TITLE: Multimodal Controller for Generative Models
ABSTRACT: Class-conditional generative models are crucial tools for data generation
from user-specified class labels. Existing approaches for class-conditional
generative models require nontrivial modifications of backbone generative
architectures to model conditional information fed into the model. This paper
introduces a plug-and-play module named `multimodal controller' to generate
multimodal data without introducing additional learning parameters. In the
absence of the controllers, our model reduces to non-conditional generative
models. We test the efficacy of multimodal controllers on CIFAR10, COIL100, and
Omniglot benchmark datasets. We demonstrate that multimodal controlled
generative models (including VAE, PixelCNN, Glow, and GAN) can generate
class-conditional images of significantly better quality when compared with
conditional generative models. Moreover, we show that multimodal controlled
models can also create novel modalities of images.
| no_new_dataset | 0.709856 |
2005.00356 | Nagabhushan Somraj | Nagabhushan Somraj, Manoj Surya Kashi, S. P. Arun and Rajiv
Soundararajan | Understanding the Perceived Quality of Video Predictions | Project website:
https://nagabhushansn95.github.io/publications/2020/pvqa.html. Accepted in
SPIC | Signal Processing: Image Communication. 102 (2022) 116626 | 10.1016/j.image.2021.116626 | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The study of video prediction models is believed to be a fundamental approach
to representation learning for videos. While a plethora of generative models
for predicting the future frame pixel values given the past few frames exist,
the quantitative evaluation of the predicted frames has been found to be
extremely challenging. In this context, we study the problem of quality
assessment of predicted videos. We create the Indian Institute of Science
Predicted Videos Quality Assessment (IISc PVQA) Database consisting of 300
videos, obtained by applying different prediction models on different datasets,
and accompanying human opinion scores. We collected subjective ratings of
quality from 50 human participants for these videos. Our subjective study
reveals that human observers were highly consistent in their judgments of
quality of predicted videos. We benchmark several popularly used measures for
evaluating video prediction and show that they do not adequately correlate with
these subjective scores. We introduce two new features to effectively capture
the quality of predicted videos, motion-compensated cosine similarities of deep
features of predicted frames with past frames, and deep features extracted from
rescaled frame differences. We show that our feature design leads to state of
the art quality prediction in accordance with human judgments on our IISc PVQA
Database. The database and code are publicly available on our project website:
https://nagabhushansn95.github.io/publications/2020/pvqa
| [
{
"version": "v1",
"created": "Fri, 1 May 2020 13:05:22 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Dec 2020 03:31:08 GMT"
},
{
"version": "v3",
"created": "Wed, 16 Jun 2021 10:37:35 GMT"
},
{
"version": "v4",
"created": "Tue, 22 Jun 2021 08:29:16 GMT"
},
{
"version": "v5",
"created": "Sat, 25 Dec 2021 05:33:19 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Somraj",
"Nagabhushan",
""
],
[
"Kashi",
"Manoj Surya",
""
],
[
"Arun",
"S. P.",
""
],
[
"Soundararajan",
"Rajiv",
""
]
] | TITLE: Understanding the Perceived Quality of Video Predictions
ABSTRACT: The study of video prediction models is believed to be a fundamental approach
to representation learning for videos. While a plethora of generative models
for predicting the future frame pixel values given the past few frames exist,
the quantitative evaluation of the predicted frames has been found to be
extremely challenging. In this context, we study the problem of quality
assessment of predicted videos. We create the Indian Institute of Science
Predicted Videos Quality Assessment (IISc PVQA) Database consisting of 300
videos, obtained by applying different prediction models on different datasets,
and accompanying human opinion scores. We collected subjective ratings of
quality from 50 human participants for these videos. Our subjective study
reveals that human observers were highly consistent in their judgments of
quality of predicted videos. We benchmark several popularly used measures for
evaluating video prediction and show that they do not adequately correlate with
these subjective scores. We introduce two new features to effectively capture
the quality of predicted videos, motion-compensated cosine similarities of deep
features of predicted frames with past frames, and deep features extracted from
rescaled frame differences. We show that our feature design leads to state of
the art quality prediction in accordance with human judgments on our IISc PVQA
Database. The database and code are publicly available on our project website:
https://nagabhushansn95.github.io/publications/2020/pvqa
| new_dataset | 0.636099 |
2009.04448 | Jie-Neng Chen | Xiangde Luo, Jieneng Chen, Tao Song, Yinan Chen, Guotai Wang, Shaoting
Zhang | Semi-supervised Medical Image Segmentation through Dual-task Consistency | AAAI2021; Code is available at https://github.com/HiLab-git/DTC | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning-based semi-supervised learning (SSL) algorithms have led to
promising results in medical images segmentation and can alleviate doctors'
expensive annotations by leveraging unlabeled data. However, most of the
existing SSL algorithms in literature tend to regularize the model training by
perturbing networks and/or data. Observing that multi/dual-task learning
attends to various levels of information which have inherent prediction
perturbation, we ask the question in this work: can we explicitly build
task-level regularization rather than implicitly constructing networks- and/or
data-level perturbation-and-transformation for SSL? To answer this question, we
propose a novel dual-task-consistency semi-supervised framework for the first
time. Concretely, we use a dual-task deep network that jointly predicts a
pixel-wise segmentation map and a geometry-aware level set representation of
the target. The level set representation is converted to an approximated
segmentation map through a differentiable task transform layer. Simultaneously,
we introduce a dual-task consistency regularization between the level
set-derived segmentation maps and directly predicted segmentation maps for both
labeled and unlabeled data. Extensive experiments on two public datasets show
that our method can largely improve the performance by incorporating the
unlabeled data. Meanwhile, our framework outperforms the state-of-the-art
semi-supervised medical image segmentation methods. Code is available at:
https://github.com/Luoxd1996/DTC
| [
{
"version": "v1",
"created": "Wed, 9 Sep 2020 17:49:21 GMT"
},
{
"version": "v2",
"created": "Sat, 6 Mar 2021 08:49:39 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 06:36:37 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Luo",
"Xiangde",
""
],
[
"Chen",
"Jieneng",
""
],
[
"Song",
"Tao",
""
],
[
"Chen",
"Yinan",
""
],
[
"Wang",
"Guotai",
""
],
[
"Zhang",
"Shaoting",
""
]
] | TITLE: Semi-supervised Medical Image Segmentation through Dual-task Consistency
ABSTRACT: Deep learning-based semi-supervised learning (SSL) algorithms have led to
promising results in medical images segmentation and can alleviate doctors'
expensive annotations by leveraging unlabeled data. However, most of the
existing SSL algorithms in literature tend to regularize the model training by
perturbing networks and/or data. Observing that multi/dual-task learning
attends to various levels of information which have inherent prediction
perturbation, we ask the question in this work: can we explicitly build
task-level regularization rather than implicitly constructing networks- and/or
data-level perturbation-and-transformation for SSL? To answer this question, we
propose a novel dual-task-consistency semi-supervised framework for the first
time. Concretely, we use a dual-task deep network that jointly predicts a
pixel-wise segmentation map and a geometry-aware level set representation of
the target. The level set representation is converted to an approximated
segmentation map through a differentiable task transform layer. Simultaneously,
we introduce a dual-task consistency regularization between the level
set-derived segmentation maps and directly predicted segmentation maps for both
labeled and unlabeled data. Extensive experiments on two public datasets show
that our method can largely improve the performance by incorporating the
unlabeled data. Meanwhile, our framework outperforms the state-of-the-art
semi-supervised medical image segmentation methods. Code is available at:
https://github.com/Luoxd1996/DTC
| no_new_dataset | 0.709818 |
2012.13805 | Kunpeng Zhang | Dongcheng Zhang, Kunpeng Zhang | Weighting-Based Treatment Effect Estimation via Distribution Learning | There are some fatal errors in the paper. Please allow us to withdraw
to improve the paper. Thank you | null | null | null | cs.LG econ.EM stat.ML | http://creativecommons.org/licenses/by/4.0/ | Existing weighting methods for treatment effect estimation are often built
upon the idea of propensity scores or covariate balance. They usually impose
strong assumptions on treatment assignment or outcome model to obtain unbiased
estimation, such as linearity or specific functional forms, which easily leads
to the major drawback of model mis-specification. In this paper, we aim to
alleviate these issues by developing a distribution learning-based weighting
method. We first learn the true underlying distribution of covariates
conditioned on treatment assignment, then leverage the ratio of covariates'
density in the treatment group to that of the control group as the weight for
estimating treatment effects. Specifically, we propose to approximate the
distribution of covariates in both treatment and control groups through
invertible transformations via change of variables. To demonstrate the
superiority, robustness, and generalizability of our method, we conduct
extensive experiments using synthetic and real data. From the experiment
results, we find that our method for estimating average treatment effect on
treated (ATT) with observational data outperforms several cutting-edge
weighting-only benchmarking methods, and it maintains its advantage under a
doubly-robust estimation framework that combines weighting with some advanced
outcome modeling methods.
| [
{
"version": "v1",
"created": "Sat, 26 Dec 2020 20:15:44 GMT"
},
{
"version": "v2",
"created": "Tue, 5 Jan 2021 18:36:16 GMT"
},
{
"version": "v3",
"created": "Wed, 6 Jan 2021 04:23:34 GMT"
},
{
"version": "v4",
"created": "Mon, 8 May 2023 15:57:32 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Zhang",
"Dongcheng",
""
],
[
"Zhang",
"Kunpeng",
""
]
] | TITLE: Weighting-Based Treatment Effect Estimation via Distribution Learning
ABSTRACT: Existing weighting methods for treatment effect estimation are often built
upon the idea of propensity scores or covariate balance. They usually impose
strong assumptions on treatment assignment or outcome model to obtain unbiased
estimation, such as linearity or specific functional forms, which easily leads
to the major drawback of model mis-specification. In this paper, we aim to
alleviate these issues by developing a distribution learning-based weighting
method. We first learn the true underlying distribution of covariates
conditioned on treatment assignment, then leverage the ratio of covariates'
density in the treatment group to that of the control group as the weight for
estimating treatment effects. Specifically, we propose to approximate the
distribution of covariates in both treatment and control groups through
invertible transformations via change of variables. To demonstrate the
superiority, robustness, and generalizability of our method, we conduct
extensive experiments using synthetic and real data. From the experiment
results, we find that our method for estimating average treatment effect on
treated (ATT) with observational data outperforms several cutting-edge
weighting-only benchmarking methods, and it maintains its advantage under a
doubly-robust estimation framework that combines weighting with some advanced
outcome modeling methods.
| no_new_dataset | 0.7116 |
2103.04526 | Pengbo Liu | Pengbo Liu, Li Xiao, S. Kevin Zhou | Incremental Learning for Multi-organ Segmentation with Partially Labeled
Datasets | null | Medical Image Computing and Computer Assisted Intervention--MICCAI
2022: 25th International Conference, Singapore, September 18--22, 2022,
Proceedings, Part IV | 10.1007/978-3-031-16440-8_68 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There exists a large number of datasets for organ segmentation, which are
partially annotated, and sequentially constructed. A typical dataset is
constructed at a certain time by curating medical images and annotating the
organs of interest. In other words, new datasets with annotations of new organ
categories are built over time. To unleash the potential behind these partially
labeled, sequentially-constructed datasets, we propose to learn a multi-organ
segmentation model through incremental learning (IL). In each IL stage, we lose
access to the previous annotations, whose knowledge is assumingly captured by
the current model, and gain the access to a new dataset with annotations of new
organ categories, from which we learn to update the organ segmentation model to
include the new organs. We give the first attempt to conjecture that the
different distribution is the key reason for 'catastrophic forgetting' that
commonly exists in IL methods, and verify that IL has the natural adaptability
to medical image scenarios. Extensive experiments on five open-sourced datasets
are conducted to prove the effectiveness of our method and the conjecture
mentioned above.
| [
{
"version": "v1",
"created": "Mon, 8 Mar 2021 03:15:59 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Liu",
"Pengbo",
""
],
[
"Xiao",
"Li",
""
],
[
"Zhou",
"S. Kevin",
""
]
] | TITLE: Incremental Learning for Multi-organ Segmentation with Partially Labeled
Datasets
ABSTRACT: There exists a large number of datasets for organ segmentation, which are
partially annotated, and sequentially constructed. A typical dataset is
constructed at a certain time by curating medical images and annotating the
organs of interest. In other words, new datasets with annotations of new organ
categories are built over time. To unleash the potential behind these partially
labeled, sequentially-constructed datasets, we propose to learn a multi-organ
segmentation model through incremental learning (IL). In each IL stage, we lose
access to the previous annotations, whose knowledge is assumingly captured by
the current model, and gain the access to a new dataset with annotations of new
organ categories, from which we learn to update the organ segmentation model to
include the new organs. We give the first attempt to conjecture that the
different distribution is the key reason for 'catastrophic forgetting' that
commonly exists in IL methods, and verify that IL has the natural adaptability
to medical image scenarios. Extensive experiments on five open-sourced datasets
are conducted to prove the effectiveness of our method and the conjecture
mentioned above.
| no_new_dataset | 0.706646 |
2107.00932 | Conghao Wong | Conghao Wong, Beihao Xia, Qinmu Peng, Wei Yuan and Xinge You | MSN: Multi-Style Network for Trajectory Prediction | Accepted by IEEE Transactions on Intelligent Transportation Systems | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Trajectory prediction aims to forecast agents' possible future locations
considering their observations along with the video context. It is strongly
needed by many autonomous platforms like tracking, detection, robot navigation,
and self-driving cars. Whether it is agents' internal personality factors,
interactive behaviors with the neighborhood, or the influence of surroundings,
they all impact agents' future planning. However, many previous methods model
and predict agents' behaviors with the same strategy or feature distribution,
making them challenging to make predictions with sufficient style differences.
This paper proposes the Multi-Style Network (MSN), which utilizes style
proposal and stylized prediction using two sub-networks, to provide multi-style
predictions in a novel categorical way adaptively. The proposed network
contains a series of style channels, and each channel is bound to a unique and
specific behavior style. We use agents' end-point plannings and their
interaction context as the basis for the behavior classification, so as to
adaptively learn multiple diverse behavior styles through these channels. Then,
we assume that the target agents may plan their future behaviors according to
each of these categorized styles, thus utilizing different style channels to
make predictions with significant style differences in parallel. Experiments
show that the proposed MSN outperforms current state-of-the-art methods up to
10% quantitatively on two widely used datasets, and presents better multi-style
characteristics qualitatively.
| [
{
"version": "v1",
"created": "Fri, 2 Jul 2021 09:43:59 GMT"
},
{
"version": "v2",
"created": "Fri, 17 Sep 2021 08:57:41 GMT"
},
{
"version": "v3",
"created": "Mon, 15 Nov 2021 08:59:14 GMT"
},
{
"version": "v4",
"created": "Fri, 1 Jul 2022 03:30:05 GMT"
},
{
"version": "v5",
"created": "Mon, 8 May 2023 07:30:35 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Wong",
"Conghao",
""
],
[
"Xia",
"Beihao",
""
],
[
"Peng",
"Qinmu",
""
],
[
"Yuan",
"Wei",
""
],
[
"You",
"Xinge",
""
]
] | TITLE: MSN: Multi-Style Network for Trajectory Prediction
ABSTRACT: Trajectory prediction aims to forecast agents' possible future locations
considering their observations along with the video context. It is strongly
needed by many autonomous platforms like tracking, detection, robot navigation,
and self-driving cars. Whether it is agents' internal personality factors,
interactive behaviors with the neighborhood, or the influence of surroundings,
they all impact agents' future planning. However, many previous methods model
and predict agents' behaviors with the same strategy or feature distribution,
making them challenging to make predictions with sufficient style differences.
This paper proposes the Multi-Style Network (MSN), which utilizes style
proposal and stylized prediction using two sub-networks, to provide multi-style
predictions in a novel categorical way adaptively. The proposed network
contains a series of style channels, and each channel is bound to a unique and
specific behavior style. We use agents' end-point plannings and their
interaction context as the basis for the behavior classification, so as to
adaptively learn multiple diverse behavior styles through these channels. Then,
we assume that the target agents may plan their future behaviors according to
each of these categorized styles, thus utilizing different style channels to
make predictions with significant style differences in parallel. Experiments
show that the proposed MSN outperforms current state-of-the-art methods up to
10% quantitatively on two widely used datasets, and presents better multi-style
characteristics qualitatively.
| no_new_dataset | 0.709837 |
2110.10895 | Jingshuang Chen | Zhiqiang Cai, Jingshuang Chen, Min Liu | Least-Squares Neural Network (LSNN) Method For Scalar Nonlinear
Hyperbolic Conservation Laws: Discrete Divergence Operator | Published on Journal of Computational and Applied Mathematics | null | null | null | math.NA cs.LG cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A least-squares neural network (LSNN) method was introduced for solving
scalar linear and nonlinear hyperbolic conservation laws (HCLs) in [7, 6]. This
method is based on an equivalent least-squares (LS) formulation and uses ReLU
neural network as approximating functions, making it ideal for approximating
discontinuous functions with unknown interface location. In the design of the
LSNN method for HCLs, the numerical approximation of differential operators is
a critical factor, and standard numerical or automatic differentiation along
coordinate directions can often lead to a failed NN-based method. To overcome
this challenge, this paper rewrites HCLs in their divergence form of space and
time and introduces a new discrete divergence operator. As a result, the
proposed LSNN method is free of penalization of artificial viscosity.
Theoretically, the accuracy of the discrete divergence operator is estimated
even for discontinuous solutions. Numerically, the LSNN method with the new
discrete divergence operator was tested for several benchmark problems with
both convex and non-convex fluxes, and was able to compute the correct physical
solution for problems with rarefaction, shock or compound waves. The method is
capable of capturing the shock of the underlying problem without oscillation or
smearing, even without any penalization of the entropy condition, total
variation, and/or artificial viscosity.
| [
{
"version": "v1",
"created": "Thu, 21 Oct 2021 04:50:57 GMT"
},
{
"version": "v2",
"created": "Mon, 12 Sep 2022 23:13:45 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 06:12:16 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Cai",
"Zhiqiang",
""
],
[
"Chen",
"Jingshuang",
""
],
[
"Liu",
"Min",
""
]
] | TITLE: Least-Squares Neural Network (LSNN) Method For Scalar Nonlinear
Hyperbolic Conservation Laws: Discrete Divergence Operator
ABSTRACT: A least-squares neural network (LSNN) method was introduced for solving
scalar linear and nonlinear hyperbolic conservation laws (HCLs) in [7, 6]. This
method is based on an equivalent least-squares (LS) formulation and uses ReLU
neural network as approximating functions, making it ideal for approximating
discontinuous functions with unknown interface location. In the design of the
LSNN method for HCLs, the numerical approximation of differential operators is
a critical factor, and standard numerical or automatic differentiation along
coordinate directions can often lead to a failed NN-based method. To overcome
this challenge, this paper rewrites HCLs in their divergence form of space and
time and introduces a new discrete divergence operator. As a result, the
proposed LSNN method is free of penalization of artificial viscosity.
Theoretically, the accuracy of the discrete divergence operator is estimated
even for discontinuous solutions. Numerically, the LSNN method with the new
discrete divergence operator was tested for several benchmark problems with
both convex and non-convex fluxes, and was able to compute the correct physical
solution for problems with rarefaction, shock or compound waves. The method is
capable of capturing the shock of the underlying problem without oscillation or
smearing, even without any penalization of the entropy condition, total
variation, and/or artificial viscosity.
| no_new_dataset | 0.710057 |
2201.01811 | Arash Nasr-Esfahany | Abdullah Alomar, Pouya Hamadanian, Arash Nasr-Esfahany, Anish Agarwal,
Mohammad Alizadeh, Devavrat Shah | CausalSim: A Causal Framework for Unbiased Trace-Driven Simulation | NSDI'23 Best Paper Award | 20th USENIX Symposium on Networked Systems Design and
Implementation (2023) 1115--1147 | null | null | cs.LG cs.AI cs.NI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present CausalSim, a causal framework for unbiased trace-driven
simulation. Current trace-driven simulators assume that the interventions being
simulated (e.g., a new algorithm) would not affect the validity of the traces.
However, real-world traces are often biased by the choices algorithms make
during trace collection, and hence replaying traces under an intervention may
lead to incorrect results. CausalSim addresses this challenge by learning a
causal model of the system dynamics and latent factors capturing the underlying
system conditions during trace collection. It learns these models using an
initial randomized control trial (RCT) under a fixed set of algorithms, and
then applies them to remove biases from trace data when simulating new
algorithms.
Key to CausalSim is mapping unbiased trace-driven simulation to a tensor
completion problem with extremely sparse observations. By exploiting a basic
distributional invariance property present in RCT data, CausalSim enables a
novel tensor completion method despite the sparsity of observations. Our
extensive evaluation of CausalSim on both real and synthetic datasets,
including more than ten months of real data from the Puffer video streaming
system shows it improves simulation accuracy, reducing errors by 53% and 61% on
average compared to expert-designed and supervised learning baselines.
Moreover, CausalSim provides markedly different insights about ABR algorithms
compared to the biased baseline simulator, which we validate with a real
deployment.
| [
{
"version": "v1",
"created": "Wed, 5 Jan 2022 20:45:24 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Feb 2022 23:36:54 GMT"
},
{
"version": "v3",
"created": "Tue, 27 Sep 2022 22:39:31 GMT"
},
{
"version": "v4",
"created": "Fri, 5 May 2023 23:14:14 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Alomar",
"Abdullah",
""
],
[
"Hamadanian",
"Pouya",
""
],
[
"Nasr-Esfahany",
"Arash",
""
],
[
"Agarwal",
"Anish",
""
],
[
"Alizadeh",
"Mohammad",
""
],
[
"Shah",
"Devavrat",
""
]
] | TITLE: CausalSim: A Causal Framework for Unbiased Trace-Driven Simulation
ABSTRACT: We present CausalSim, a causal framework for unbiased trace-driven
simulation. Current trace-driven simulators assume that the interventions being
simulated (e.g., a new algorithm) would not affect the validity of the traces.
However, real-world traces are often biased by the choices algorithms make
during trace collection, and hence replaying traces under an intervention may
lead to incorrect results. CausalSim addresses this challenge by learning a
causal model of the system dynamics and latent factors capturing the underlying
system conditions during trace collection. It learns these models using an
initial randomized control trial (RCT) under a fixed set of algorithms, and
then applies them to remove biases from trace data when simulating new
algorithms.
Key to CausalSim is mapping unbiased trace-driven simulation to a tensor
completion problem with extremely sparse observations. By exploiting a basic
distributional invariance property present in RCT data, CausalSim enables a
novel tensor completion method despite the sparsity of observations. Our
extensive evaluation of CausalSim on both real and synthetic datasets,
including more than ten months of real data from the Puffer video streaming
system shows it improves simulation accuracy, reducing errors by 53% and 61% on
average compared to expert-designed and supervised learning baselines.
Moreover, CausalSim provides markedly different insights about ABR algorithms
compared to the biased baseline simulator, which we validate with a real
deployment.
| no_new_dataset | 0.709221 |
2201.12482 | Feng Li | Feng Li, Xuyang Yuan, Lina Wang, Huan Yang, Dongxiao Yu, Weifeng Lv,
Xiuzhen Cheng | Collaborative Learning in General Graphs with Limited Memorization:
Complexity, Learnability, and Reliability | null | null | null | null | cs.LG cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider a K-armed bandit problem in general graphs where agents are
arbitrarily connected and each of them has limited memorizing capabilities and
communication bandwidth. The goal is to let each of the agents eventually learn
the best arm. It is assumed in these studies that the communication graph
should be complete or well-structured, whereas such an assumption is not always
valid in practice. Furthermore, limited memorization and communication
bandwidth also restrict the collaborations of the agents, since the agents
memorize and communicate very few experiences. Additionally, an agent may be
corrupted to share falsified experiences to its peers, while the resource limit
in terms of memorization and communication may considerably restrict the
reliability of the learning process. To address the above issues, we propose a
three-staged collaborative learning algorithm. In each step, the agents share
their latest experiences with each other through light-weight random walks in a
general communication graph, and then make decisions on which arms to pull
according to the recommendations received from their peers. The agents finally
update their adoptions (i.e., preferences to the arms) based on the reward
obtained by pulling the arms. Our theoretical analysis shows that, when there
are a sufficient number of agents participating in the collaborative learning
process, all the agents eventually learn the best arm with high probability,
even with limited memorizing capabilities and light-weight communications. We
also reveal in our theoretical analysis the upper bound on the number of
corrupted agents our algorithm can tolerate. The efficacy of our proposed
three-staged collaborative learning algorithm is finally verified by extensive
experiments on both synthetic and real datasets.
| [
{
"version": "v1",
"created": "Sat, 29 Jan 2022 02:42:25 GMT"
},
{
"version": "v2",
"created": "Wed, 3 May 2023 12:03:20 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 01:32:59 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Li",
"Feng",
""
],
[
"Yuan",
"Xuyang",
""
],
[
"Wang",
"Lina",
""
],
[
"Yang",
"Huan",
""
],
[
"Yu",
"Dongxiao",
""
],
[
"Lv",
"Weifeng",
""
],
[
"Cheng",
"Xiuzhen",
""
]
] | TITLE: Collaborative Learning in General Graphs with Limited Memorization:
Complexity, Learnability, and Reliability
ABSTRACT: We consider a K-armed bandit problem in general graphs where agents are
arbitrarily connected and each of them has limited memorizing capabilities and
communication bandwidth. The goal is to let each of the agents eventually learn
the best arm. It is assumed in these studies that the communication graph
should be complete or well-structured, whereas such an assumption is not always
valid in practice. Furthermore, limited memorization and communication
bandwidth also restrict the collaborations of the agents, since the agents
memorize and communicate very few experiences. Additionally, an agent may be
corrupted to share falsified experiences to its peers, while the resource limit
in terms of memorization and communication may considerably restrict the
reliability of the learning process. To address the above issues, we propose a
three-staged collaborative learning algorithm. In each step, the agents share
their latest experiences with each other through light-weight random walks in a
general communication graph, and then make decisions on which arms to pull
according to the recommendations received from their peers. The agents finally
update their adoptions (i.e., preferences to the arms) based on the reward
obtained by pulling the arms. Our theoretical analysis shows that, when there
are a sufficient number of agents participating in the collaborative learning
process, all the agents eventually learn the best arm with high probability,
even with limited memorizing capabilities and light-weight communications. We
also reveal in our theoretical analysis the upper bound on the number of
corrupted agents our algorithm can tolerate. The efficacy of our proposed
three-staged collaborative learning algorithm is finally verified by extensive
experiments on both synthetic and real datasets.
| no_new_dataset | 0.712201 |
2201.12745 | Iurii Nagornov S. | Iurii S. Nagornov | Approximate Bayesian Computation Based on Maxima Weighted Isolation
Kernel Mapping | null | New Frontiers in Artificial Intelligence. JSAI-isAI 2022. Lecture
Notes in Computer Science(), vol 13859. Springer, Cham | 10.1007/978-3-031-29168-5_18 | null | stat.ML cs.LG q-bio.GN q-bio.PE | http://creativecommons.org/licenses/by/4.0/ | Motivation: A branching processes model yields an unevenly stochastically
distributed dataset that consists of sparse and dense regions. This work
addresses the problem of precisely evaluating parameters for such a model.
Applying a branching processes model to an area such as cancer cell evolution
faces a number of obstacles, including high dimensionality and the rare
appearance of a result of interest. We take on the ambitious task of obtaining
the coefficients of a model that reflects the relationship of driver gene
mutations and cancer hallmarks on the basis of personal data regarding variant
allele frequencies. Results: An approximate Bayesian computation method based
on Isolation Kernel is developed. The method involves the transformation of row
data to a Hilbert space (mapping) and the measurement of the similarity between
simulated points and maxima weighted Isolation Kernel mapping related to the
observation point. We also design a heuristic algorithm for parameter
estimation that requires no calculation and is dimension independent. The
advantages of the proposed machine learning method are illustrated using
multidimensional test data as well as a specific example focused on cancer cell
evolution.
| [
{
"version": "v1",
"created": "Sun, 30 Jan 2022 07:11:57 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Feb 2022 16:45:39 GMT"
},
{
"version": "v3",
"created": "Thu, 3 Mar 2022 12:49:25 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Nagornov",
"Iurii S.",
""
]
] | TITLE: Approximate Bayesian Computation Based on Maxima Weighted Isolation
Kernel Mapping
ABSTRACT: Motivation: A branching processes model yields an unevenly stochastically
distributed dataset that consists of sparse and dense regions. This work
addresses the problem of precisely evaluating parameters for such a model.
Applying a branching processes model to an area such as cancer cell evolution
faces a number of obstacles, including high dimensionality and the rare
appearance of a result of interest. We take on the ambitious task of obtaining
the coefficients of a model that reflects the relationship of driver gene
mutations and cancer hallmarks on the basis of personal data regarding variant
allele frequencies. Results: An approximate Bayesian computation method based
on Isolation Kernel is developed. The method involves the transformation of row
data to a Hilbert space (mapping) and the measurement of the similarity between
simulated points and maxima weighted Isolation Kernel mapping related to the
observation point. We also design a heuristic algorithm for parameter
estimation that requires no calculation and is dimension independent. The
advantages of the proposed machine learning method are illustrated using
multidimensional test data as well as a specific example focused on cancer cell
evolution.
| no_new_dataset | 0.710672 |
2202.10821 | J\k{e}drzej Kozal | J\k{e}drzej Kozal, Micha{\l} Wo\'zniak | Increasing Depth of Neural Networks for Life-long Learning | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Purpose: We propose a novel method for continual learning based on the
increasing depth of neural networks. This work explores whether extending
neural network depth may be beneficial in a life-long learning setting.
Methods: We propose a novel approach based on adding new layers on top of
existing ones to enable the forward transfer of knowledge and adapting
previously learned representations. We employ a method of determining the most
similar tasks for selecting the best location in our network to add new nodes
with trainable parameters. This approach allows for creating a tree-like model,
where each node is a set of neural network parameters dedicated to a specific
task. The Progressive Neural Network concept inspires the proposed method.
Therefore, it benefits from dynamic changes in network structure. However,
Progressive Neural Network allocates a lot of memory for the whole network
structure during the learning process. The proposed method alleviates this by
adding only part of a network for a new task and utilizing a subset of
previously trained weights. At the same time, we may retain the benefit of PNN,
such as no forgetting guaranteed by design, without needing a memory buffer.
Results: Experiments on Split CIFAR and Split Tiny ImageNet show that the
proposed algorithm is on par with other continual learning methods. In a more
challenging setup with a single computer vision dataset as a separate task, our
method outperforms Experience Replay.
Conclusion: It is compatible with commonly used computer vision architectures
and does not require a custom network structure. As an adaptation to changing
data distribution is made by expanding the architecture, there is no need to
utilize a rehearsal buffer. For this reason, our method could be used for
sensitive applications where data privacy must be considered.
| [
{
"version": "v1",
"created": "Tue, 22 Feb 2022 11:21:41 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 16:17:40 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Kozal",
"Jędrzej",
""
],
[
"Woźniak",
"Michał",
""
]
] | TITLE: Increasing Depth of Neural Networks for Life-long Learning
ABSTRACT: Purpose: We propose a novel method for continual learning based on the
increasing depth of neural networks. This work explores whether extending
neural network depth may be beneficial in a life-long learning setting.
Methods: We propose a novel approach based on adding new layers on top of
existing ones to enable the forward transfer of knowledge and adapting
previously learned representations. We employ a method of determining the most
similar tasks for selecting the best location in our network to add new nodes
with trainable parameters. This approach allows for creating a tree-like model,
where each node is a set of neural network parameters dedicated to a specific
task. The Progressive Neural Network concept inspires the proposed method.
Therefore, it benefits from dynamic changes in network structure. However,
Progressive Neural Network allocates a lot of memory for the whole network
structure during the learning process. The proposed method alleviates this by
adding only part of a network for a new task and utilizing a subset of
previously trained weights. At the same time, we may retain the benefit of PNN,
such as no forgetting guaranteed by design, without needing a memory buffer.
Results: Experiments on Split CIFAR and Split Tiny ImageNet show that the
proposed algorithm is on par with other continual learning methods. In a more
challenging setup with a single computer vision dataset as a separate task, our
method outperforms Experience Replay.
Conclusion: It is compatible with commonly used computer vision architectures
and does not require a custom network structure. As an adaptation to changing
data distribution is made by expanding the architecture, there is no need to
utilize a rehearsal buffer. For this reason, our method could be used for
sensitive applications where data privacy must be considered.
| no_new_dataset | 0.708427 |
2203.06967 | Zejin Wang | Zejin Wang, Jiazheng Liu, Guoqing Li, Hua Han | Blind2Unblind: Self-Supervised Image Denoising with Visible Blind Spots | Accepted to CVPR2022 | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Real noisy-clean pairs on a large scale are costly and difficult to obtain.
Meanwhile, supervised denoisers trained on synthetic data perform poorly in
practice. Self-supervised denoisers, which learn only from single noisy images,
solve the data collection problem. However, self-supervised denoising methods,
especially blindspot-driven ones, suffer sizable information loss during input
or network design. The absence of valuable information dramatically reduces the
upper bound of denoising performance. In this paper, we propose a simple yet
efficient approach called Blind2Unblind to overcome the information loss in
blindspot-driven denoising methods. First, we introduce a global-aware mask
mapper that enables global perception and accelerates training. The mask mapper
samples all pixels at blind spots on denoised volumes and maps them to the same
channel, allowing the loss function to optimize all blind spots at once.
Second, we propose a re-visible loss to train the denoising network and make
blind spots visible. The denoiser can learn directly from raw noise images
without losing information or being trapped in identity mapping. We also
theoretically analyze the convergence of the re-visible loss. Extensive
experiments on synthetic and real-world datasets demonstrate the superior
performance of our approach compared to previous work. Code is available at
https://github.com/demonsjin/Blind2Unblind.
| [
{
"version": "v1",
"created": "Mon, 14 Mar 2022 10:07:42 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Mar 2022 04:23:31 GMT"
},
{
"version": "v3",
"created": "Mon, 8 May 2023 02:39:55 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Wang",
"Zejin",
""
],
[
"Liu",
"Jiazheng",
""
],
[
"Li",
"Guoqing",
""
],
[
"Han",
"Hua",
""
]
] | TITLE: Blind2Unblind: Self-Supervised Image Denoising with Visible Blind Spots
ABSTRACT: Real noisy-clean pairs on a large scale are costly and difficult to obtain.
Meanwhile, supervised denoisers trained on synthetic data perform poorly in
practice. Self-supervised denoisers, which learn only from single noisy images,
solve the data collection problem. However, self-supervised denoising methods,
especially blindspot-driven ones, suffer sizable information loss during input
or network design. The absence of valuable information dramatically reduces the
upper bound of denoising performance. In this paper, we propose a simple yet
efficient approach called Blind2Unblind to overcome the information loss in
blindspot-driven denoising methods. First, we introduce a global-aware mask
mapper that enables global perception and accelerates training. The mask mapper
samples all pixels at blind spots on denoised volumes and maps them to the same
channel, allowing the loss function to optimize all blind spots at once.
Second, we propose a re-visible loss to train the denoising network and make
blind spots visible. The denoiser can learn directly from raw noise images
without losing information or being trapped in identity mapping. We also
theoretically analyze the convergence of the re-visible loss. Extensive
experiments on synthetic and real-world datasets demonstrate the superior
performance of our approach compared to previous work. Code is available at
https://github.com/demonsjin/Blind2Unblind.
| no_new_dataset | 0.710415 |
2204.03511 | Shounak Datta | Shounak Datta, Sankha Subhra Mullick, Anish Chakrabarty, Swagatam Das | Interval Bound Interpolation for Few-shot Learning with Few Tasks | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Few-shot learning aims to transfer the knowledge acquired from training on a
diverse set of tasks to unseen tasks from the same task distribution with a
limited amount of labeled data. The underlying requirement for effective
few-shot generalization is to learn a good representation of the task manifold.
This becomes more difficult when only a limited number of tasks are available
for training. In such a few-task few-shot setting, it is beneficial to
explicitly preserve the local neighborhoods from the task manifold and exploit
this to generate artificial tasks for training. To this end, we introduce the
notion of interval bounds from the provably robust training literature to
few-shot learning. The interval bounds are used to characterize neighborhoods
around the training tasks. These neighborhoods can then be preserved by
minimizing the distance between a task and its respective bounds. We then use a
novel strategy to artificially form new tasks for training by interpolating
between the available tasks and their respective interval bounds. We apply our
framework to both model-agnostic meta-learning as well as prototype-based
metric-learning paradigms. The efficacy of our proposed approach is evident
from the improved performance on several datasets from diverse domains compared
to current methods.
| [
{
"version": "v1",
"created": "Thu, 7 Apr 2022 15:29:27 GMT"
},
{
"version": "v2",
"created": "Fri, 8 Apr 2022 15:26:02 GMT"
},
{
"version": "v3",
"created": "Wed, 3 May 2023 23:11:04 GMT"
},
{
"version": "v4",
"created": "Sun, 7 May 2023 18:15:26 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Datta",
"Shounak",
""
],
[
"Mullick",
"Sankha Subhra",
""
],
[
"Chakrabarty",
"Anish",
""
],
[
"Das",
"Swagatam",
""
]
] | TITLE: Interval Bound Interpolation for Few-shot Learning with Few Tasks
ABSTRACT: Few-shot learning aims to transfer the knowledge acquired from training on a
diverse set of tasks to unseen tasks from the same task distribution with a
limited amount of labeled data. The underlying requirement for effective
few-shot generalization is to learn a good representation of the task manifold.
This becomes more difficult when only a limited number of tasks are available
for training. In such a few-task few-shot setting, it is beneficial to
explicitly preserve the local neighborhoods from the task manifold and exploit
this to generate artificial tasks for training. To this end, we introduce the
notion of interval bounds from the provably robust training literature to
few-shot learning. The interval bounds are used to characterize neighborhoods
around the training tasks. These neighborhoods can then be preserved by
minimizing the distance between a task and its respective bounds. We then use a
novel strategy to artificially form new tasks for training by interpolating
between the available tasks and their respective interval bounds. We apply our
framework to both model-agnostic meta-learning as well as prototype-based
metric-learning paradigms. The efficacy of our proposed approach is evident
from the improved performance on several datasets from diverse domains compared
to current methods.
| no_new_dataset | 0.708427 |
2204.06292 | Javier Morlana | Javier Morlana and J.M.M. Montiel | Reuse your features: unifying retrieval and feature-metric alignment | ICRA 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We propose a compact pipeline to unify all the steps of Visual Localization:
image retrieval, candidate re-ranking and initial pose estimation, and camera
pose refinement. Our key assumption is that the deep features used for these
individual tasks share common characteristics, so we should reuse them in all
the procedures of the pipeline. Our DRAN (Deep Retrieval and image Alignment
Network) is able to extract global descriptors for efficient image retrieval,
use intermediate hierarchical features to re-rank the retrieval list and
produce an initial pose guess, which is finally refined by means of a
feature-metric optimization based on learned deep multi-scale dense features.
DRAN is the first single network able to produce the features for the three
steps of visual localization. DRAN achieves competitive performance in terms of
robustness and accuracy under challenging conditions in public benchmarks,
outperforming other unified approaches and consuming lower computational and
memory cost than its counterparts using multiple networks. Code and models will
be publicly available at https://github.com/jmorlana/DRAN.
| [
{
"version": "v1",
"created": "Wed, 13 Apr 2022 10:42:00 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 12:10:26 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Morlana",
"Javier",
""
],
[
"Montiel",
"J. M. M.",
""
]
] | TITLE: Reuse your features: unifying retrieval and feature-metric alignment
ABSTRACT: We propose a compact pipeline to unify all the steps of Visual Localization:
image retrieval, candidate re-ranking and initial pose estimation, and camera
pose refinement. Our key assumption is that the deep features used for these
individual tasks share common characteristics, so we should reuse them in all
the procedures of the pipeline. Our DRAN (Deep Retrieval and image Alignment
Network) is able to extract global descriptors for efficient image retrieval,
use intermediate hierarchical features to re-rank the retrieval list and
produce an initial pose guess, which is finally refined by means of a
feature-metric optimization based on learned deep multi-scale dense features.
DRAN is the first single network able to produce the features for the three
steps of visual localization. DRAN achieves competitive performance in terms of
robustness and accuracy under challenging conditions in public benchmarks,
outperforming other unified approaches and consuming lower computational and
memory cost than its counterparts using multiple networks. Code and models will
be publicly available at https://github.com/jmorlana/DRAN.
| no_new_dataset | 0.71 |
2204.08154 | Jinwei Ren | Jinwei Ren, Jianke Zhu, and Jialiang Zhang | End-to-end Weakly-supervised Single-stage Multiple 3D Hand Mesh
Reconstruction from a Single RGB Image | Computer Vision and Image Understanding | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we consider the challenging task of simultaneously locating
and recovering multiple hands from a single 2D image. Previous studies either
focus on single hand reconstruction or solve this problem in a multi-stage way.
Moreover, the conventional two-stage pipeline firstly detects hand areas, and
then estimates 3D hand pose from each cropped patch. To reduce the
computational redundancy in preprocessing and feature extraction, for the first
time, we propose a concise but efficient single-stage pipeline for multi-hand
reconstruction. Specifically, we design a multi-head auto-encoder structure,
where each head network shares the same feature map and outputs the hand
center, pose and texture, respectively. Besides, we adopt a weakly-supervised
scheme to alleviate the burden of expensive 3D real-world data annotations. To
this end, we propose a series of losses optimized by a stage-wise training
scheme, where a multi-hand dataset with 2D annotations is generated based on
the publicly available single hand datasets. In order to further improve the
accuracy of the weakly supervised model, we adopt several feature consistency
constraints in both single and multiple hand settings. Specifically, the
keypoints of each hand estimated from local features should be consistent with
the re-projected points predicted from global features. Extensive experiments
on public benchmarks including FreiHAND, HO3D, InterHand2.6M and RHD
demonstrate that our method outperforms the state-of-the-art model-based
methods in both weakly-supervised and fully-supervised manners. The code and
models are available at {https://github.com/zijinxuxu/SMHR}.
| [
{
"version": "v1",
"created": "Mon, 18 Apr 2022 03:57:14 GMT"
},
{
"version": "v2",
"created": "Fri, 19 Aug 2022 08:39:46 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 08:38:24 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Ren",
"Jinwei",
""
],
[
"Zhu",
"Jianke",
""
],
[
"Zhang",
"Jialiang",
""
]
] | TITLE: End-to-end Weakly-supervised Single-stage Multiple 3D Hand Mesh
Reconstruction from a Single RGB Image
ABSTRACT: In this paper, we consider the challenging task of simultaneously locating
and recovering multiple hands from a single 2D image. Previous studies either
focus on single hand reconstruction or solve this problem in a multi-stage way.
Moreover, the conventional two-stage pipeline firstly detects hand areas, and
then estimates 3D hand pose from each cropped patch. To reduce the
computational redundancy in preprocessing and feature extraction, for the first
time, we propose a concise but efficient single-stage pipeline for multi-hand
reconstruction. Specifically, we design a multi-head auto-encoder structure,
where each head network shares the same feature map and outputs the hand
center, pose and texture, respectively. Besides, we adopt a weakly-supervised
scheme to alleviate the burden of expensive 3D real-world data annotations. To
this end, we propose a series of losses optimized by a stage-wise training
scheme, where a multi-hand dataset with 2D annotations is generated based on
the publicly available single hand datasets. In order to further improve the
accuracy of the weakly supervised model, we adopt several feature consistency
constraints in both single and multiple hand settings. Specifically, the
keypoints of each hand estimated from local features should be consistent with
the re-projected points predicted from global features. Extensive experiments
on public benchmarks including FreiHAND, HO3D, InterHand2.6M and RHD
demonstrate that our method outperforms the state-of-the-art model-based
methods in both weakly-supervised and fully-supervised manners. The code and
models are available at {https://github.com/zijinxuxu/SMHR}.
| no_new_dataset | 0.708055 |
2204.11164 | Jiaxin Liu | Jiaxin Liu, Yuefei Lyu, Xi Zhang, Sihong Xie | Are Your Reviewers Being Treated Equally? Discovering Subgroup
Structures to Improve Fairness in Spam Detection | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | User-generated reviews of products are vital assets of online commerce, such
as Amazon and Yelp, while fake reviews are prevalent to mislead customers. GNN
is the state-of-the-art method that detects suspicious reviewers by exploiting
the topologies of the graph connecting reviewers, reviews, and target products.
However, the discrepancy in the detection accuracy over different groups of
reviewers can degrade reviewer engagement and customer trust in the review
websites. Unlike the previous belief that the difference between the groups
causes unfairness, we study the subgroup structures within the groups that can
also cause discrepancies in treating different groups. This paper addresses the
challenges of defining, approximating, and utilizing a new subgroup structure
for fair spam detection. We first identify subgroup structures in the review
graph that lead to discrepant accuracy in the groups. The complex dependencies
over the review graph create difficulties in teasing out subgroups hidden
within larger groups. We design a model that can be trained to jointly infer
the hidden subgroup memberships and exploits the membership for calibrating the
detection accuracy across groups. Comprehensive comparisons against baselines
on three large Yelp review datasets demonstrate that the subgroup membership
can be identified and exploited for group fairness.
| [
{
"version": "v1",
"created": "Sun, 24 Apr 2022 02:19:22 GMT"
},
{
"version": "v2",
"created": "Fri, 5 May 2023 23:22:32 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Liu",
"Jiaxin",
""
],
[
"Lyu",
"Yuefei",
""
],
[
"Zhang",
"Xi",
""
],
[
"Xie",
"Sihong",
""
]
] | TITLE: Are Your Reviewers Being Treated Equally? Discovering Subgroup
Structures to Improve Fairness in Spam Detection
ABSTRACT: User-generated reviews of products are vital assets of online commerce, such
as Amazon and Yelp, while fake reviews are prevalent to mislead customers. GNN
is the state-of-the-art method that detects suspicious reviewers by exploiting
the topologies of the graph connecting reviewers, reviews, and target products.
However, the discrepancy in the detection accuracy over different groups of
reviewers can degrade reviewer engagement and customer trust in the review
websites. Unlike the previous belief that the difference between the groups
causes unfairness, we study the subgroup structures within the groups that can
also cause discrepancies in treating different groups. This paper addresses the
challenges of defining, approximating, and utilizing a new subgroup structure
for fair spam detection. We first identify subgroup structures in the review
graph that lead to discrepant accuracy in the groups. The complex dependencies
over the review graph create difficulties in teasing out subgroups hidden
within larger groups. We design a model that can be trained to jointly infer
the hidden subgroup memberships and exploits the membership for calibrating the
detection accuracy across groups. Comprehensive comparisons against baselines
on three large Yelp review datasets demonstrate that the subgroup membership
can be identified and exploited for group fairness.
| no_new_dataset | 0.710823 |
2205.03747 | Zhiwei Zhang | Anastasios Kyrillidis, Moshe Y. Vardi, Zhiwei Zhang | DPMS: An ADD-Based Symbolic Approach for Generalized MaxSAT Solving | null | null | null | null | cs.AI cs.LO math.OC | http://creativecommons.org/licenses/by/4.0/ | Boolean MaxSAT, as well as generalized formulations such as Min-MaxSAT and
Max-hybrid-SAT, are fundamental optimization problems in Boolean reasoning.
Existing methods for MaxSAT have been successful in solving benchmarks in CNF
format. They lack, however, the ability to handle 1) (non-CNF) hybrid
constraints, such as XORs and 2) generalized MaxSAT problems natively. To
address this issue, we propose a novel dynamic-programming approach for solving
generalized MaxSAT problems with hybrid constraints -- called
\emph{Dynamic-Programming-MaxSAT} or DPMS for short -- based on Algebraic
Decision Diagrams (ADDs). With the power of ADDs and the (graded)
project-join-tree builder, our versatile framework admits many generalizations
of CNF-MaxSAT, such as MaxSAT, Min-MaxSAT, and MinSAT with hybrid constraints.
Moreover, DPMS scales provably well on instances with low width. Empirical
results indicate that DPMS is able to solve certain problems quickly, where
other algorithms based on various techniques all fail. Hence, DPMS is a
promising framework and opens a new line of research that invites more
investigation in the future.
| [
{
"version": "v1",
"created": "Sun, 8 May 2022 00:31:53 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 02:52:07 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Kyrillidis",
"Anastasios",
""
],
[
"Vardi",
"Moshe Y.",
""
],
[
"Zhang",
"Zhiwei",
""
]
] | TITLE: DPMS: An ADD-Based Symbolic Approach for Generalized MaxSAT Solving
ABSTRACT: Boolean MaxSAT, as well as generalized formulations such as Min-MaxSAT and
Max-hybrid-SAT, are fundamental optimization problems in Boolean reasoning.
Existing methods for MaxSAT have been successful in solving benchmarks in CNF
format. They lack, however, the ability to handle 1) (non-CNF) hybrid
constraints, such as XORs and 2) generalized MaxSAT problems natively. To
address this issue, we propose a novel dynamic-programming approach for solving
generalized MaxSAT problems with hybrid constraints -- called
\emph{Dynamic-Programming-MaxSAT} or DPMS for short -- based on Algebraic
Decision Diagrams (ADDs). With the power of ADDs and the (graded)
project-join-tree builder, our versatile framework admits many generalizations
of CNF-MaxSAT, such as MaxSAT, Min-MaxSAT, and MinSAT with hybrid constraints.
Moreover, DPMS scales provably well on instances with low width. Empirical
results indicate that DPMS is able to solve certain problems quickly, where
other algorithms based on various techniques all fail. Hence, DPMS is a
promising framework and opens a new line of research that invites more
investigation in the future.
| no_new_dataset | 0.706823 |
2205.12302 | William Jurayj | William Jurayj, William Rudman, Carsten Eickhoff | Garden-Path Traversal in GPT-2 | 9 pages, 6 figures, Accepted to EMNLP BlackBox NLP 2022 | ACL 2022, BlackBox NLP, pages 305-313
https://aclanthology.org/2022.blackboxnlp-1.25/ | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | In recent years, large-scale transformer decoders such as the GPT-x family of
models have become increasingly popular. Studies examining the behavior of
these models tend to focus only on the output of the language modeling head and
avoid analysis of the internal states of the transformer decoder. In this
study, we present a collection of methods to analyze the hidden states of GPT-2
and use the model's navigation of garden path sentences as a case study. To
enable this, we compile the largest currently available dataset of garden path
sentences. We show that Manhattan distances and cosine similarities provide
more reliable insights compared to established surprisal methods that analyze
next-token probabilities computed by a language modeling head. Using these
methods, we find that negating tokens have minimal impacts on the model's
representations for unambiguous forms of sentences with ambiguity solely over
what the object of a verb is, but have a more substantial impact of
representations for unambiguous sentences whose ambiguity would stem from the
voice of a verb. Further, we find that analyzing the decoder model's hidden
states reveals periods of ambiguity that might conclude in a garden path effect
but happen not to, whereas surprisal analyses routinely miss this detail.
| [
{
"version": "v1",
"created": "Tue, 24 May 2022 18:21:58 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Oct 2022 16:21:06 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Jurayj",
"William",
""
],
[
"Rudman",
"William",
""
],
[
"Eickhoff",
"Carsten",
""
]
] | TITLE: Garden-Path Traversal in GPT-2
ABSTRACT: In recent years, large-scale transformer decoders such as the GPT-x family of
models have become increasingly popular. Studies examining the behavior of
these models tend to focus only on the output of the language modeling head and
avoid analysis of the internal states of the transformer decoder. In this
study, we present a collection of methods to analyze the hidden states of GPT-2
and use the model's navigation of garden path sentences as a case study. To
enable this, we compile the largest currently available dataset of garden path
sentences. We show that Manhattan distances and cosine similarities provide
more reliable insights compared to established surprisal methods that analyze
next-token probabilities computed by a language modeling head. Using these
methods, we find that negating tokens have minimal impacts on the model's
representations for unambiguous forms of sentences with ambiguity solely over
what the object of a verb is, but have a more substantial impact of
representations for unambiguous sentences whose ambiguity would stem from the
voice of a verb. Further, we find that analyzing the decoder model's hidden
states reveals periods of ambiguity that might conclude in a garden path effect
but happen not to, whereas surprisal analyses routinely miss this detail.
| new_dataset | 0.71398 |
2206.14245 | Alexander Black | Alexander Black, Tu Bui, Simon Jenni, Zhifei Zhang, Viswanathan
Swaminanthan, John Collomosse | SImProv: Scalable Image Provenance Framework for Robust Content
Attribution | Under consideration at Computer Vision and Image Understanding | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We present SImProv - a scalable image provenance framework to match a query
image back to a trusted database of originals and identify possible
manipulations on the query. SImProv consists of three stages: a scalable search
stage for retrieving top-k most similar images; a re-ranking and
near-duplicated detection stage for identifying the original among the
candidates; and finally a manipulation detection and visualization stage for
localizing regions within the query that may have been manipulated to differ
from the original. SImProv is robust to benign image transformations that
commonly occur during online redistribution, such as artifacts due to noise and
recompression degradation, as well as out-of-place transformations due to image
padding, warping, and changes in size and shape. Robustness towards
out-of-place transformations is achieved via the end-to-end training of a
differentiable warping module within the comparator architecture. We
demonstrate effective retrieval and manipulation detection over a dataset of
100 million images.
| [
{
"version": "v1",
"created": "Tue, 28 Jun 2022 18:42:36 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 12:05:05 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Black",
"Alexander",
""
],
[
"Bui",
"Tu",
""
],
[
"Jenni",
"Simon",
""
],
[
"Zhang",
"Zhifei",
""
],
[
"Swaminanthan",
"Viswanathan",
""
],
[
"Collomosse",
"John",
""
]
] | TITLE: SImProv: Scalable Image Provenance Framework for Robust Content
Attribution
ABSTRACT: We present SImProv - a scalable image provenance framework to match a query
image back to a trusted database of originals and identify possible
manipulations on the query. SImProv consists of three stages: a scalable search
stage for retrieving top-k most similar images; a re-ranking and
near-duplicated detection stage for identifying the original among the
candidates; and finally a manipulation detection and visualization stage for
localizing regions within the query that may have been manipulated to differ
from the original. SImProv is robust to benign image transformations that
commonly occur during online redistribution, such as artifacts due to noise and
recompression degradation, as well as out-of-place transformations due to image
padding, warping, and changes in size and shape. Robustness towards
out-of-place transformations is achieved via the end-to-end training of a
differentiable warping module within the comparator architecture. We
demonstrate effective retrieval and manipulation detection over a dataset of
100 million images.
| no_new_dataset | 0.707436 |
2207.01071 | Xiaoke Shen | Xiaoke Shen and Ioannis Stamos | Unified Object Detector for Different Modalities based on Vision
Transformers | arXiv admin note: text overlap with arXiv:2203.10456 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Traditional systems typically require different models for processing
different modalities, such as one model for RGB images and another for depth
images. Recent research has demonstrated that a single model for one modality
can be adapted for another using cross-modality transfer learning. In this
paper, we extend this approach by combining cross/inter-modality transfer
learning with a vision transformer to develop a unified detector that achieves
superior performance across diverse modalities. Our research envisions an
application scenario for robotics, where the unified system seamlessly switches
between RGB cameras and depth sensors in varying lighting conditions.
Importantly, the system requires no model architecture or weight updates to
enable this smooth transition. Specifically, the system uses the depth sensor
during low-lighting conditions (night time) and both the RGB camera and depth
sensor or RGB caemra only in well-lit environments. We evaluate our unified
model on the SUN RGB-D dataset, and demonstrate that it achieves similar or
better performance in terms of mAP50 compared to state-of-the-art methods in
the SUNRGBD16 category, and comparable performance in point cloud only mode. We
also introduce a novel inter-modality mixing method that enables our model to
achieve significantly better results than previous methods. We provide our
code, including training/inference logs and model checkpoints, to facilitate
reproducibility and further research.
\url{https://github.com/liketheflower/UODDM}
| [
{
"version": "v1",
"created": "Sun, 3 Jul 2022 16:01:04 GMT"
},
{
"version": "v2",
"created": "Fri, 5 May 2023 23:20:33 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Shen",
"Xiaoke",
""
],
[
"Stamos",
"Ioannis",
""
]
] | TITLE: Unified Object Detector for Different Modalities based on Vision
Transformers
ABSTRACT: Traditional systems typically require different models for processing
different modalities, such as one model for RGB images and another for depth
images. Recent research has demonstrated that a single model for one modality
can be adapted for another using cross-modality transfer learning. In this
paper, we extend this approach by combining cross/inter-modality transfer
learning with a vision transformer to develop a unified detector that achieves
superior performance across diverse modalities. Our research envisions an
application scenario for robotics, where the unified system seamlessly switches
between RGB cameras and depth sensors in varying lighting conditions.
Importantly, the system requires no model architecture or weight updates to
enable this smooth transition. Specifically, the system uses the depth sensor
during low-lighting conditions (night time) and both the RGB camera and depth
sensor or RGB caemra only in well-lit environments. We evaluate our unified
model on the SUN RGB-D dataset, and demonstrate that it achieves similar or
better performance in terms of mAP50 compared to state-of-the-art methods in
the SUNRGBD16 category, and comparable performance in point cloud only mode. We
also introduce a novel inter-modality mixing method that enables our model to
achieve significantly better results than previous methods. We provide our
code, including training/inference logs and model checkpoints, to facilitate
reproducibility and further research.
\url{https://github.com/liketheflower/UODDM}
| no_new_dataset | 0.710635 |
2207.11007 | V\'ictor Gallego-Fontenla | Victor Gallego-Fontenla, Juan C. Vidal, Manuel Lama | Gradual Drift Detection in Process Models Using Conformance Metrics | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Changes, planned or unexpected, are common during the execution of real-life
processes. Detecting these changes is a must for optimizing the performance of
organizations running such processes. Most of the algorithms present in the
state-of-the-art focus on the detection of sudden changes, leaving aside other
types of changes. In this paper, we will focus on the automatic detection of
gradual drifts, a special type of change, in which the cases of two models
overlap during a period of time. The proposed algorithm relies on conformance
checking metrics to carry out the automatic detection of the changes,
performing also a fully automatic classification of these changes into sudden
or gradual. The approach has been validated with a synthetic dataset consisting
of 120 logs with different distributions of changes, getting better results in
terms of detection and classification accuracy, delay and change region
overlapping than the main state-of-the-art algorithms.
| [
{
"version": "v1",
"created": "Fri, 22 Jul 2022 10:56:35 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 08:40:44 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Gallego-Fontenla",
"Victor",
""
],
[
"Vidal",
"Juan C.",
""
],
[
"Lama",
"Manuel",
""
]
] | TITLE: Gradual Drift Detection in Process Models Using Conformance Metrics
ABSTRACT: Changes, planned or unexpected, are common during the execution of real-life
processes. Detecting these changes is a must for optimizing the performance of
organizations running such processes. Most of the algorithms present in the
state-of-the-art focus on the detection of sudden changes, leaving aside other
types of changes. In this paper, we will focus on the automatic detection of
gradual drifts, a special type of change, in which the cases of two models
overlap during a period of time. The proposed algorithm relies on conformance
checking metrics to carry out the automatic detection of the changes,
performing also a fully automatic classification of these changes into sudden
or gradual. The approach has been validated with a synthetic dataset consisting
of 120 logs with different distributions of changes, getting better results in
terms of detection and classification accuracy, delay and change region
overlapping than the main state-of-the-art algorithms.
| new_dataset | 0.712595 |
2207.11177 | Rem Yang | Rem Yang, Jacob Laurel, Sasa Misailovic, Gagandeep Singh | Provable Defense Against Geometric Transformations | ICLR 2023 notable-top-25% | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | Geometric image transformations that arise in the real world, such as scaling
and rotation, have been shown to easily deceive deep neural networks (DNNs).
Hence, training DNNs to be certifiably robust to these perturbations is
critical. However, no prior work has been able to incorporate the objective of
deterministic certified robustness against geometric transformations into the
training procedure, as existing verifiers are exceedingly slow. To address
these challenges, we propose the first provable defense for deterministic
certified geometric robustness. Our framework leverages a novel GPU-optimized
verifier that can certify images between 60$\times$ to 42,600$\times$ faster
than existing geometric robustness verifiers, and thus unlike existing works,
is fast enough for use in training. Across multiple datasets, our results show
that networks trained via our framework consistently achieve state-of-the-art
deterministic certified geometric robustness and clean accuracy. Furthermore,
for the first time, we verify the geometric robustness of a neural network for
the challenging, real-world setting of autonomous driving.
| [
{
"version": "v1",
"created": "Fri, 22 Jul 2022 16:40:03 GMT"
},
{
"version": "v2",
"created": "Fri, 30 Sep 2022 17:55:52 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 17:10:23 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Yang",
"Rem",
""
],
[
"Laurel",
"Jacob",
""
],
[
"Misailovic",
"Sasa",
""
],
[
"Singh",
"Gagandeep",
""
]
] | TITLE: Provable Defense Against Geometric Transformations
ABSTRACT: Geometric image transformations that arise in the real world, such as scaling
and rotation, have been shown to easily deceive deep neural networks (DNNs).
Hence, training DNNs to be certifiably robust to these perturbations is
critical. However, no prior work has been able to incorporate the objective of
deterministic certified robustness against geometric transformations into the
training procedure, as existing verifiers are exceedingly slow. To address
these challenges, we propose the first provable defense for deterministic
certified geometric robustness. Our framework leverages a novel GPU-optimized
verifier that can certify images between 60$\times$ to 42,600$\times$ faster
than existing geometric robustness verifiers, and thus unlike existing works,
is fast enough for use in training. Across multiple datasets, our results show
that networks trained via our framework consistently achieve state-of-the-art
deterministic certified geometric robustness and clean accuracy. Furthermore,
for the first time, we verify the geometric robustness of a neural network for
the challenging, real-world setting of autonomous driving.
| no_new_dataset | 0.710176 |
2208.03232 | Samuel Joutard | Samuel Joutard, Reuben Dorent, Sebastien Ourselin, Tom Vercauteren,
Marc Modat | Driving Points Prediction For Abdominal Probabilistic Registration | null | null | 10.1007/978-3-031-21014-3_30 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Inter-patient abdominal registration has various applications, from
pharmakinematic studies to anatomy modeling. Yet, it remains a challenging
application due to the morphological heterogeneity and variability of the human
abdomen. Among the various registration methods proposed for this task,
probabilistic displacement registration models estimate displacement
distribution for a subset of points by comparing feature vectors of points from
the two images. These probabilistic models are informative and robust while
allowing large displacements by design. As the displacement distributions are
typically estimated on a subset of points (which we refer to as driving
points), due to computational requirements, we propose in this work to learn a
driving points predictor. Compared to previously proposed methods, the driving
points predictor is optimized in an end-to-end fashion to infer driving points
tailored for a specific registration pipeline. We evaluate the impact of our
contribution on two different datasets corresponding to different modalities.
Specifically, we compared the performances of 6 different probabilistic
displacement registration models when using a driving points predictor or one
of 2 other standard driving points selection methods. The proposed method
improved performances in 11 out of 12 experiments.
| [
{
"version": "v1",
"created": "Fri, 5 Aug 2022 15:28:01 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Joutard",
"Samuel",
""
],
[
"Dorent",
"Reuben",
""
],
[
"Ourselin",
"Sebastien",
""
],
[
"Vercauteren",
"Tom",
""
],
[
"Modat",
"Marc",
""
]
] | TITLE: Driving Points Prediction For Abdominal Probabilistic Registration
ABSTRACT: Inter-patient abdominal registration has various applications, from
pharmakinematic studies to anatomy modeling. Yet, it remains a challenging
application due to the morphological heterogeneity and variability of the human
abdomen. Among the various registration methods proposed for this task,
probabilistic displacement registration models estimate displacement
distribution for a subset of points by comparing feature vectors of points from
the two images. These probabilistic models are informative and robust while
allowing large displacements by design. As the displacement distributions are
typically estimated on a subset of points (which we refer to as driving
points), due to computational requirements, we propose in this work to learn a
driving points predictor. Compared to previously proposed methods, the driving
points predictor is optimized in an end-to-end fashion to infer driving points
tailored for a specific registration pipeline. We evaluate the impact of our
contribution on two different datasets corresponding to different modalities.
Specifically, we compared the performances of 6 different probabilistic
displacement registration models when using a driving points predictor or one
of 2 other standard driving points selection methods. The proposed method
improved performances in 11 out of 12 experiments.
| no_new_dataset | 0.711437 |
2208.08041 | Cody Reading | John Willes, Cody Reading, Steven L. Waslander | InterTrack: Interaction Transformer for 3D Multi-Object Tracking | Accepted to CRV 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | 3D multi-object tracking (MOT) is a key problem for autonomous vehicles,
required to perform well-informed motion planning in dynamic environments.
Particularly for densely occupied scenes, associating existing tracks to new
detections remains challenging as existing systems tend to omit critical
contextual information. Our proposed solution, InterTrack, introduces the
Interaction Transformer for 3D MOT to generate discriminative object
representations for data association. We extract state and shape features for
each track and detection, and efficiently aggregate global information via
attention. We then perform a learned regression on each track/detection feature
pair to estimate affinities, and use a robust two-stage data association and
track management approach to produce the final tracks. We validate our approach
on the nuScenes 3D MOT benchmark, where we observe significant improvements,
particularly on classes with small physical sizes and clustered objects. As of
submission, InterTrack ranks 1st in overall AMOTA among methods using
CenterPoint detections.
| [
{
"version": "v1",
"created": "Wed, 17 Aug 2022 03:24:36 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 14:03:57 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Willes",
"John",
""
],
[
"Reading",
"Cody",
""
],
[
"Waslander",
"Steven L.",
""
]
] | TITLE: InterTrack: Interaction Transformer for 3D Multi-Object Tracking
ABSTRACT: 3D multi-object tracking (MOT) is a key problem for autonomous vehicles,
required to perform well-informed motion planning in dynamic environments.
Particularly for densely occupied scenes, associating existing tracks to new
detections remains challenging as existing systems tend to omit critical
contextual information. Our proposed solution, InterTrack, introduces the
Interaction Transformer for 3D MOT to generate discriminative object
representations for data association. We extract state and shape features for
each track and detection, and efficiently aggregate global information via
attention. We then perform a learned regression on each track/detection feature
pair to estimate affinities, and use a robust two-stage data association and
track management approach to produce the final tracks. We validate our approach
on the nuScenes 3D MOT benchmark, where we observe significant improvements,
particularly on classes with small physical sizes and clustered objects. As of
submission, InterTrack ranks 1st in overall AMOTA among methods using
CenterPoint detections.
| no_new_dataset | 0.706368 |
2208.09606 | Rui Meng | Rui Meng, Tong Wang, Xingdi Yuan, Yingbo Zhou, Daqing He | General-to-Specific Transfer Labeling for Domain Adaptable Keyphrase
Generation | The submission has been accepted to the Findings of ACL 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | Training keyphrase generation (KPG) models require a large amount of
annotated data, which can be prohibitively expensive and often limited to
specific domains. In this study, we first demonstrate that large distribution
shifts among different domains severely hinder the transferability of KPG
models. We then propose a three-stage pipeline, which gradually guides KPG
models' learning focus from general syntactical features to domain-related
semantics, in a data-efficient manner. With Domain-general Phrase pre-training,
we pre-train Sequence-to-Sequence models with generic phrase annotations that
are widely available on the web, which enables the models to generate phrases
in a wide range of domains. The resulting model is then applied in the Transfer
Labeling stage to produce domain-specific pseudo keyphrases, which help adapt
models to a new domain. Finally, we fine-tune the model with limited data with
true labels to fully adapt it to the target domain. Our experiment results show
that the proposed process can produce good-quality keyphrases in new domains
and achieve consistent improvements after adaptation with limited in-domain
annotated data. All code and datasets are available at
https://github.com/memray/OpenNMT-kpg-release.
| [
{
"version": "v1",
"created": "Sat, 20 Aug 2022 04:43:01 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 19:53:04 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Meng",
"Rui",
""
],
[
"Wang",
"Tong",
""
],
[
"Yuan",
"Xingdi",
""
],
[
"Zhou",
"Yingbo",
""
],
[
"He",
"Daqing",
""
]
] | TITLE: General-to-Specific Transfer Labeling for Domain Adaptable Keyphrase
Generation
ABSTRACT: Training keyphrase generation (KPG) models require a large amount of
annotated data, which can be prohibitively expensive and often limited to
specific domains. In this study, we first demonstrate that large distribution
shifts among different domains severely hinder the transferability of KPG
models. We then propose a three-stage pipeline, which gradually guides KPG
models' learning focus from general syntactical features to domain-related
semantics, in a data-efficient manner. With Domain-general Phrase pre-training,
we pre-train Sequence-to-Sequence models with generic phrase annotations that
are widely available on the web, which enables the models to generate phrases
in a wide range of domains. The resulting model is then applied in the Transfer
Labeling stage to produce domain-specific pseudo keyphrases, which help adapt
models to a new domain. Finally, we fine-tune the model with limited data with
true labels to fully adapt it to the target domain. Our experiment results show
that the proposed process can produce good-quality keyphrases in new domains
and achieve consistent improvements after adaptation with limited in-domain
annotated data. All code and datasets are available at
https://github.com/memray/OpenNMT-kpg-release.
| no_new_dataset | 0.710804 |
2208.09626 | Yaman Kumar Singla | Yaman Kumar Singla, Rajat Jha, Arunim Gupta, Milan Aggarwal, Aditya
Garg, Tushar Malyan, Ayush Bhardwaj, Rajiv Ratn Shah, Balaji Krishnamurthy,
and Changyou Chen | Persuasion Strategies in Advertisements | Accepted at AAAI-23 | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modeling what makes an advertisement persuasive, i.e., eliciting the desired
response from consumer, is critical to the study of propaganda, social
psychology, and marketing. Despite its importance, computational modeling of
persuasion in computer vision is still in its infancy, primarily due to the
lack of benchmark datasets that can provide persuasion-strategy labels
associated with ads. Motivated by persuasion literature in social psychology
and marketing, we introduce an extensive vocabulary of persuasion strategies
and build the first ad image corpus annotated with persuasion strategies. We
then formulate the task of persuasion strategy prediction with multi-modal
learning, where we design a multi-task attention fusion model that can leverage
other ad-understanding tasks to predict persuasion strategies. Further, we
conduct a real-world case study on 1600 advertising campaigns of 30 Fortune-500
companies where we use our model's predictions to analyze which strategies work
with different demographics (age and gender). The dataset also provides image
segmentation masks, which labels persuasion strategies in the corresponding ad
images on the test split. We publicly release our code and dataset
https://midas-research.github.io/persuasion-advertisements/.
| [
{
"version": "v1",
"created": "Sat, 20 Aug 2022 07:33:13 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 07:28:51 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Singla",
"Yaman Kumar",
""
],
[
"Jha",
"Rajat",
""
],
[
"Gupta",
"Arunim",
""
],
[
"Aggarwal",
"Milan",
""
],
[
"Garg",
"Aditya",
""
],
[
"Malyan",
"Tushar",
""
],
[
"Bhardwaj",
"Ayush",
""
],
[
"Shah",
"Rajiv Ratn",
""
],
[
"Krishnamurthy",
"Balaji",
""
],
[
"Chen",
"Changyou",
""
]
] | TITLE: Persuasion Strategies in Advertisements
ABSTRACT: Modeling what makes an advertisement persuasive, i.e., eliciting the desired
response from consumer, is critical to the study of propaganda, social
psychology, and marketing. Despite its importance, computational modeling of
persuasion in computer vision is still in its infancy, primarily due to the
lack of benchmark datasets that can provide persuasion-strategy labels
associated with ads. Motivated by persuasion literature in social psychology
and marketing, we introduce an extensive vocabulary of persuasion strategies
and build the first ad image corpus annotated with persuasion strategies. We
then formulate the task of persuasion strategy prediction with multi-modal
learning, where we design a multi-task attention fusion model that can leverage
other ad-understanding tasks to predict persuasion strategies. Further, we
conduct a real-world case study on 1600 advertising campaigns of 30 Fortune-500
companies where we use our model's predictions to analyze which strategies work
with different demographics (age and gender). The dataset also provides image
segmentation masks, which labels persuasion strategies in the corresponding ad
images on the test split. We publicly release our code and dataset
https://midas-research.github.io/persuasion-advertisements/.
| new_dataset | 0.717173 |
2208.10930 | Yunqing Zhao | Yunqing Zhao and Ngai-Man Cheung | FS-BAN: Born-Again Networks for Domain Generalization Few-Shot
Classification | 15 pages, 9 figures, 15 tables. IEEE Transactions on Image Processing
(TIP), 2023 | null | 10.1109/TIP.2023.3266172 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conventional Few-shot classification (FSC) aims to recognize samples from
novel classes given limited labeled data. Recently, domain generalization FSC
(DG-FSC) has been proposed with the goal to recognize novel class samples from
unseen domains. DG-FSC poses considerable challenges to many models due to the
domain shift between base classes (used in training) and novel classes
(encountered in evaluation). In this work, we make two novel contributions to
tackle DG-FSC. Our first contribution is to propose Born-Again Network (BAN)
episodic training and comprehensively investigate its effectiveness for DG-FSC.
As a specific form of knowledge distillation, BAN has been shown to achieve
improved generalization in conventional supervised classification with a
closed-set setup. This improved generalization motivates us to study BAN for
DG-FSC, and we show that BAN is promising to address the domain shift
encountered in DG-FSC. Building on the encouraging findings, our second (major)
contribution is to propose Few-Shot BAN (FS-BAN), a novel BAN approach for
DG-FSC. Our proposed FS-BAN includes novel multi-task learning objectives:
Mutual Regularization, Mismatched Teacher, and Meta-Control Temperature, each
of these is specifically designed to overcome central and unique challenges in
DG-FSC, namely overfitting and domain discrepancy. We analyze different design
choices of these techniques. We conduct comprehensive quantitative and
qualitative analysis and evaluation over six datasets and three baseline
models. The results suggest that our proposed FS-BAN consistently improves the
generalization performance of baseline models and achieves state-of-the-art
accuracy for DG-FSC. Project Page: https://yunqing-me.github.io/Born-Again-FS/.
| [
{
"version": "v1",
"created": "Tue, 23 Aug 2022 13:02:14 GMT"
},
{
"version": "v2",
"created": "Wed, 24 Aug 2022 05:39:17 GMT"
},
{
"version": "v3",
"created": "Tue, 21 Mar 2023 08:57:23 GMT"
},
{
"version": "v4",
"created": "Mon, 8 May 2023 15:57:07 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Zhao",
"Yunqing",
""
],
[
"Cheung",
"Ngai-Man",
""
]
] | TITLE: FS-BAN: Born-Again Networks for Domain Generalization Few-Shot
Classification
ABSTRACT: Conventional Few-shot classification (FSC) aims to recognize samples from
novel classes given limited labeled data. Recently, domain generalization FSC
(DG-FSC) has been proposed with the goal to recognize novel class samples from
unseen domains. DG-FSC poses considerable challenges to many models due to the
domain shift between base classes (used in training) and novel classes
(encountered in evaluation). In this work, we make two novel contributions to
tackle DG-FSC. Our first contribution is to propose Born-Again Network (BAN)
episodic training and comprehensively investigate its effectiveness for DG-FSC.
As a specific form of knowledge distillation, BAN has been shown to achieve
improved generalization in conventional supervised classification with a
closed-set setup. This improved generalization motivates us to study BAN for
DG-FSC, and we show that BAN is promising to address the domain shift
encountered in DG-FSC. Building on the encouraging findings, our second (major)
contribution is to propose Few-Shot BAN (FS-BAN), a novel BAN approach for
DG-FSC. Our proposed FS-BAN includes novel multi-task learning objectives:
Mutual Regularization, Mismatched Teacher, and Meta-Control Temperature, each
of these is specifically designed to overcome central and unique challenges in
DG-FSC, namely overfitting and domain discrepancy. We analyze different design
choices of these techniques. We conduct comprehensive quantitative and
qualitative analysis and evaluation over six datasets and three baseline
models. The results suggest that our proposed FS-BAN consistently improves the
generalization performance of baseline models and achieves state-of-the-art
accuracy for DG-FSC. Project Page: https://yunqing-me.github.io/Born-Again-FS/.
| no_new_dataset | 0.710019 |
2208.11857 | Mengnan Du | Mengnan Du, Fengxiang He, Na Zou, Dacheng Tao and Xia Hu | Shortcut Learning of Large Language Models in Natural Language
Understanding | Accepted by Communications of the ACM (CACM), Review Article | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have achieved state-of-the-art performance on a
series of natural language understanding tasks. However, these LLMs might rely
on dataset bias and artifacts as shortcuts for prediction. This has
significantly affected their generalizability and adversarial robustness. In
this paper, we provide a review of recent developments that address the
shortcut learning and robustness challenge of LLMs. We first introduce the
concepts of shortcut learning of language models. We then introduce methods to
identify shortcut learning behavior in language models, characterize the
reasons for shortcut learning, as well as introduce mitigation solutions.
Finally, we discuss key research challenges and potential research directions
in order to advance the field of LLMs.
| [
{
"version": "v1",
"created": "Thu, 25 Aug 2022 03:51:39 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 23:55:09 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Du",
"Mengnan",
""
],
[
"He",
"Fengxiang",
""
],
[
"Zou",
"Na",
""
],
[
"Tao",
"Dacheng",
""
],
[
"Hu",
"Xia",
""
]
] | TITLE: Shortcut Learning of Large Language Models in Natural Language
Understanding
ABSTRACT: Large language models (LLMs) have achieved state-of-the-art performance on a
series of natural language understanding tasks. However, these LLMs might rely
on dataset bias and artifacts as shortcuts for prediction. This has
significantly affected their generalizability and adversarial robustness. In
this paper, we provide a review of recent developments that address the
shortcut learning and robustness challenge of LLMs. We first introduce the
concepts of shortcut learning of language models. We then introduce methods to
identify shortcut learning behavior in language models, characterize the
reasons for shortcut learning, as well as introduce mitigation solutions.
Finally, we discuss key research challenges and potential research directions
in order to advance the field of LLMs.
| no_new_dataset | 0.710842 |
2209.10307 | Dong Zhang | Dong Zhang, Yi Lin, Hao Chen, Zhuotao Tian, Xin Yang, Jinhui Tang,
Kwang Ting Cheng | Understanding the Tricks of Deep Learning in Medical Image Segmentation:
Challenges and Future Directions | Under submission | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Over the past few years, the rapid development of deep learning technologies
for computer vision has significantly improved the performance of medical image
segmentation (MedISeg). However, the diverse implementation strategies of
various models have led to an extremely complex MedISeg system, resulting in a
potential problem of unfair result comparisons. In this paper, we collect a
series of MedISeg tricks for different model implementation phases (i.e.,
pre-training model, data pre-processing, data augmentation, model
implementation, model inference, and result post-processing), and
experimentally explore the effectiveness of these tricks on consistent
baselines. With the extensive experimental results on both the representative
2D and 3D medical image datasets, we explicitly clarify the effect of these
tricks. Moreover, based on the surveyed tricks, we also open-sourced a strong
MedISeg repository, where each component has the advantage of plug-and-play. We
believe that this milestone work not only completes a comprehensive and
complementary survey of the state-of-the-art MedISeg approaches, but also
offers a practical guide for addressing the future medical image processing
challenges including but not limited to small dataset, class imbalance
learning, multi-modality learning, and domain adaptation. The code and training
weights have been released at: https://github.com/hust-linyi/seg_trick.
| [
{
"version": "v1",
"created": "Wed, 21 Sep 2022 12:30:05 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 10:23:24 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Zhang",
"Dong",
""
],
[
"Lin",
"Yi",
""
],
[
"Chen",
"Hao",
""
],
[
"Tian",
"Zhuotao",
""
],
[
"Yang",
"Xin",
""
],
[
"Tang",
"Jinhui",
""
],
[
"Cheng",
"Kwang Ting",
""
]
] | TITLE: Understanding the Tricks of Deep Learning in Medical Image Segmentation:
Challenges and Future Directions
ABSTRACT: Over the past few years, the rapid development of deep learning technologies
for computer vision has significantly improved the performance of medical image
segmentation (MedISeg). However, the diverse implementation strategies of
various models have led to an extremely complex MedISeg system, resulting in a
potential problem of unfair result comparisons. In this paper, we collect a
series of MedISeg tricks for different model implementation phases (i.e.,
pre-training model, data pre-processing, data augmentation, model
implementation, model inference, and result post-processing), and
experimentally explore the effectiveness of these tricks on consistent
baselines. With the extensive experimental results on both the representative
2D and 3D medical image datasets, we explicitly clarify the effect of these
tricks. Moreover, based on the surveyed tricks, we also open-sourced a strong
MedISeg repository, where each component has the advantage of plug-and-play. We
believe that this milestone work not only completes a comprehensive and
complementary survey of the state-of-the-art MedISeg approaches, but also
offers a practical guide for addressing the future medical image processing
challenges including but not limited to small dataset, class imbalance
learning, multi-modality learning, and domain adaptation. The code and training
weights have been released at: https://github.com/hust-linyi/seg_trick.
| no_new_dataset | 0.708458 |
2210.00434 | Zhihuan Kuang | Zhihuan Kuang, Shi Zong, Jianbing Zhang, Jiajun Chen, Hongfu Liu | Music-to-Text Synaesthesia: Generating Descriptive Text from Music
Recordings | null | null | null | null | eess.AS cs.AI cs.CL cs.MM cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we consider a novel research problem: music-to-text
synaesthesia. Different from the classical music tagging problem that
classifies a music recording into pre-defined categories, music-to-text
synaesthesia aims to generate descriptive texts from music recordings with the
same sentiment for further understanding. As existing music-related datasets do
not contain the semantic descriptions on music recordings, we collect a new
dataset that contains 1,955 aligned pairs of classical music recordings and
text descriptions. Based on this, we build a computational model to generate
sentences that can describe the content of the music recording. To tackle the
highly non-discriminative classical music, we design a group
topology-preservation loss, which considers more samples as a group reference
and preserves the relative topology among different samples. Extensive
experimental results qualitatively and quantitatively demonstrate the
effectiveness of our proposed model over five heuristics or pre-trained
competitive methods and their variants on our collected dataset.
| [
{
"version": "v1",
"created": "Sun, 2 Oct 2022 06:06:55 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 03:09:27 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Kuang",
"Zhihuan",
""
],
[
"Zong",
"Shi",
""
],
[
"Zhang",
"Jianbing",
""
],
[
"Chen",
"Jiajun",
""
],
[
"Liu",
"Hongfu",
""
]
] | TITLE: Music-to-Text Synaesthesia: Generating Descriptive Text from Music
Recordings
ABSTRACT: In this paper, we consider a novel research problem: music-to-text
synaesthesia. Different from the classical music tagging problem that
classifies a music recording into pre-defined categories, music-to-text
synaesthesia aims to generate descriptive texts from music recordings with the
same sentiment for further understanding. As existing music-related datasets do
not contain the semantic descriptions on music recordings, we collect a new
dataset that contains 1,955 aligned pairs of classical music recordings and
text descriptions. Based on this, we build a computational model to generate
sentences that can describe the content of the music recording. To tackle the
highly non-discriminative classical music, we design a group
topology-preservation loss, which considers more samples as a group reference
and preserves the relative topology among different samples. Extensive
experimental results qualitatively and quantitatively demonstrate the
effectiveness of our proposed model over five heuristics or pre-trained
competitive methods and their variants on our collected dataset.
| new_dataset | 0.714622 |
2210.00858 | Georgios Tziafas | Georgios Tziafas, Hamidreza Kasaei | Enhancing Interpretability and Interactivity in Robot Manipulation: A
Neurosymbolic Approach | Submitted T-RO | null | null | null | cs.RO cs.AI cs.HC | http://creativecommons.org/licenses/by/4.0/ | In this paper we present a neurosymbolic architecture for coupling
language-guided visual reasoning with robot manipulation. A non-expert human
user can prompt the robot using unconstrained natural language, providing a
referring expression (REF), a question (VQA), or a grasp action instruction.
The system tackles all cases in a task-agnostic fashion through the utilization
of a shared library of primitive skills. Each primitive handles an independent
sub-task, such as reasoning about visual attributes, spatial relation
comprehension, logic and enumeration, as well as arm control. A language parser
maps the input query to an executable program composed of such primitives,
depending on the context. While some primitives are purely symbolic operations
(e.g. counting), others are trainable neural functions (e.g. visual grounding),
therefore marrying the interpretability and systematic generalization benefits
of discrete symbolic approaches with the scalability and representational power
of deep networks. We generate a 3D vision-and-language synthetic dataset of
tabletop scenes in a simulation environment to train our approach and perform
extensive evaluations in both synthetic and real-world scenes. Results showcase
the benefits of our approach in terms of accuracy, sample-efficiency, and
robustness to the user's vocabulary, while being transferable to real-world
scenes with few-shot visual fine-tuning. Finally, we integrate our method with
a robot framework and demonstrate how it can serve as an interpretable solution
for an interactive object-picking task, both in simulation and with a real
robot. We make our datasets available in
https://gtziafas.github.io/neurosymbolic-manipulation.
| [
{
"version": "v1",
"created": "Mon, 3 Oct 2022 12:21:45 GMT"
},
{
"version": "v2",
"created": "Tue, 10 Jan 2023 12:47:28 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 17:06:49 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Tziafas",
"Georgios",
""
],
[
"Kasaei",
"Hamidreza",
""
]
] | TITLE: Enhancing Interpretability and Interactivity in Robot Manipulation: A
Neurosymbolic Approach
ABSTRACT: In this paper we present a neurosymbolic architecture for coupling
language-guided visual reasoning with robot manipulation. A non-expert human
user can prompt the robot using unconstrained natural language, providing a
referring expression (REF), a question (VQA), or a grasp action instruction.
The system tackles all cases in a task-agnostic fashion through the utilization
of a shared library of primitive skills. Each primitive handles an independent
sub-task, such as reasoning about visual attributes, spatial relation
comprehension, logic and enumeration, as well as arm control. A language parser
maps the input query to an executable program composed of such primitives,
depending on the context. While some primitives are purely symbolic operations
(e.g. counting), others are trainable neural functions (e.g. visual grounding),
therefore marrying the interpretability and systematic generalization benefits
of discrete symbolic approaches with the scalability and representational power
of deep networks. We generate a 3D vision-and-language synthetic dataset of
tabletop scenes in a simulation environment to train our approach and perform
extensive evaluations in both synthetic and real-world scenes. Results showcase
the benefits of our approach in terms of accuracy, sample-efficiency, and
robustness to the user's vocabulary, while being transferable to real-world
scenes with few-shot visual fine-tuning. Finally, we integrate our method with
a robot framework and demonstrate how it can serve as an interpretable solution
for an interactive object-picking task, both in simulation and with a real
robot. We make our datasets available in
https://gtziafas.github.io/neurosymbolic-manipulation.
| new_dataset | 0.713407 |
2210.03151 | Satrajit Chakrabarty | Satrajit Chakrabarty, Syed Amaan Abidi, Mina Mousa, Mahati Mokkarala,
Isabelle Hren, Divya Yadav, Matthew Kelsey, Pamela LaMontagne, John Wood,
Michael Adams, Yuzhuo Su, Sherry Thorpe, Caroline Chung, Aristeidis Sotiras,
and Daniel S. Marcus | Integrative Imaging Informatics for Cancer Research: Workflow Automation
for Neuro-oncology (I3CR-WANO) | null | null | 10.1200/cci.22.00177 | null | eess.IV cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Efforts to utilize growing volumes of clinical imaging data to generate tumor
evaluations continue to require significant manual data wrangling owing to the
data heterogeneity. Here, we propose an artificial intelligence-based solution
for the aggregation and processing of multisequence neuro-oncology MRI data to
extract quantitative tumor measurements. Our end-to-end framework i) classifies
MRI sequences using an ensemble classifier, ii) preprocesses the data in a
reproducible manner, iii) delineates tumor tissue subtypes using convolutional
neural networks, and iv) extracts diverse radiomic features. Moreover, it is
robust to missing sequences and adopts an expert-in-the-loop approach, where
the segmentation results may be manually refined by radiologists. Following the
implementation of the framework in Docker containers, it was applied to two
retrospective glioma datasets collected from the Washington University School
of Medicine (WUSM; n = 384) and the M.D. Anderson Cancer Center (MDA; n = 30)
comprising preoperative MRI scans from patients with pathologically confirmed
gliomas. The scan-type classifier yielded an accuracy of over 99%, correctly
identifying sequences from 380/384 and 30/30 sessions from the WUSM and MDA
datasets, respectively. Segmentation performance was quantified using the Dice
Similarity Coefficient between the predicted and expert-refined tumor masks.
Mean Dice scores were 0.882 ($\pm$0.244) and 0.977 ($\pm$0.04) for whole tumor
segmentation for WUSM and MDA, respectively. This streamlined framework
automatically curated, processed, and segmented raw MRI data of patients with
varying grades of gliomas, enabling the curation of large-scale neuro-oncology
datasets and demonstrating a high potential for integration as an assistive
tool in clinical practice.
| [
{
"version": "v1",
"created": "Thu, 6 Oct 2022 18:23:42 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Chakrabarty",
"Satrajit",
""
],
[
"Abidi",
"Syed Amaan",
""
],
[
"Mousa",
"Mina",
""
],
[
"Mokkarala",
"Mahati",
""
],
[
"Hren",
"Isabelle",
""
],
[
"Yadav",
"Divya",
""
],
[
"Kelsey",
"Matthew",
""
],
[
"LaMontagne",
"Pamela",
""
],
[
"Wood",
"John",
""
],
[
"Adams",
"Michael",
""
],
[
"Su",
"Yuzhuo",
""
],
[
"Thorpe",
"Sherry",
""
],
[
"Chung",
"Caroline",
""
],
[
"Sotiras",
"Aristeidis",
""
],
[
"Marcus",
"Daniel S.",
""
]
] | TITLE: Integrative Imaging Informatics for Cancer Research: Workflow Automation
for Neuro-oncology (I3CR-WANO)
ABSTRACT: Efforts to utilize growing volumes of clinical imaging data to generate tumor
evaluations continue to require significant manual data wrangling owing to the
data heterogeneity. Here, we propose an artificial intelligence-based solution
for the aggregation and processing of multisequence neuro-oncology MRI data to
extract quantitative tumor measurements. Our end-to-end framework i) classifies
MRI sequences using an ensemble classifier, ii) preprocesses the data in a
reproducible manner, iii) delineates tumor tissue subtypes using convolutional
neural networks, and iv) extracts diverse radiomic features. Moreover, it is
robust to missing sequences and adopts an expert-in-the-loop approach, where
the segmentation results may be manually refined by radiologists. Following the
implementation of the framework in Docker containers, it was applied to two
retrospective glioma datasets collected from the Washington University School
of Medicine (WUSM; n = 384) and the M.D. Anderson Cancer Center (MDA; n = 30)
comprising preoperative MRI scans from patients with pathologically confirmed
gliomas. The scan-type classifier yielded an accuracy of over 99%, correctly
identifying sequences from 380/384 and 30/30 sessions from the WUSM and MDA
datasets, respectively. Segmentation performance was quantified using the Dice
Similarity Coefficient between the predicted and expert-refined tumor masks.
Mean Dice scores were 0.882 ($\pm$0.244) and 0.977 ($\pm$0.04) for whole tumor
segmentation for WUSM and MDA, respectively. This streamlined framework
automatically curated, processed, and segmented raw MRI data of patients with
varying grades of gliomas, enabling the curation of large-scale neuro-oncology
datasets and demonstrating a high potential for integration as an assistive
tool in clinical practice.
| no_new_dataset | 0.71086 |
2210.07039 | Giambattista Albora | Giambattista Albora, Lavinia Rossi-Mori, Andrea Zaccaria | Sapling Similarity: a performing and interpretable memory-based tool for
recommendation | 27 pages, 11 figures | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many bipartite networks describe systems where an edge represents a relation
between a user and an item. Measuring the similarity between either users or
items is the basis of memory-based collaborative filtering, a widely used
method to build a recommender system with the purpose of proposing items to
users. When the edges of the network are unweighted, the popular common
neighbors-based approaches, allowing only positive similarity values, neglect
the possibility and the effect of two users (or two items) being very
dissimilar. Moreover, they underperform with respect to model-based (machine
learning) approaches, although providing higher interpretability. Inspired by
the functioning of Decision Trees, we propose a method to compute similarity
that allows also negative values, the Sapling Similarity. The key idea is to
look at how the information that a user is connected to an item influences our
prior estimation of the probability that another user is connected to the same
item: if it is reduced, then the similarity between the two users will be
negative, otherwise, it will be positive. We show that, when used to build
memory-based collaborative filtering, Sapling Similarity provides better
recommendations than existing similarity metrics. Then we compare the Sapling
Similarity Collaborative Filtering (SSCF, a hybrid of the item-based and the
user-based) with state-of-the-art models using standard datasets. Even if SSCF
depends on only one straightforward hyperparameter, it has comparable or higher
recommending accuracy, and outperforms all other models on the Amazon-Book
dataset, while retaining the high explainability of memory-based approaches.
| [
{
"version": "v1",
"created": "Thu, 13 Oct 2022 13:53:22 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 14:42:09 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Albora",
"Giambattista",
""
],
[
"Rossi-Mori",
"Lavinia",
""
],
[
"Zaccaria",
"Andrea",
""
]
] | TITLE: Sapling Similarity: a performing and interpretable memory-based tool for
recommendation
ABSTRACT: Many bipartite networks describe systems where an edge represents a relation
between a user and an item. Measuring the similarity between either users or
items is the basis of memory-based collaborative filtering, a widely used
method to build a recommender system with the purpose of proposing items to
users. When the edges of the network are unweighted, the popular common
neighbors-based approaches, allowing only positive similarity values, neglect
the possibility and the effect of two users (or two items) being very
dissimilar. Moreover, they underperform with respect to model-based (machine
learning) approaches, although providing higher interpretability. Inspired by
the functioning of Decision Trees, we propose a method to compute similarity
that allows also negative values, the Sapling Similarity. The key idea is to
look at how the information that a user is connected to an item influences our
prior estimation of the probability that another user is connected to the same
item: if it is reduced, then the similarity between the two users will be
negative, otherwise, it will be positive. We show that, when used to build
memory-based collaborative filtering, Sapling Similarity provides better
recommendations than existing similarity metrics. Then we compare the Sapling
Similarity Collaborative Filtering (SSCF, a hybrid of the item-based and the
user-based) with state-of-the-art models using standard datasets. Even if SSCF
depends on only one straightforward hyperparameter, it has comparable or higher
recommending accuracy, and outperforms all other models on the Amazon-Book
dataset, while retaining the high explainability of memory-based approaches.
| no_new_dataset | 0.709007 |
2210.07775 | Peihua Mai | Peihua Mai, Yan Pang | PrivMVMF: Privacy-Preserving Multi-View Matrix Factorization for
Recommender Systems | null | null | 10.1109/TAI.2023.3240700 | null | cs.IR cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With an increasing focus on data privacy, there have been pilot studies on
recommender systems in a federated learning (FL) framework, where multiple
parties collaboratively train a model without sharing their data. Most of these
studies assume that the conventional FL framework can fully protect user
privacy. However, there are serious privacy risks in matrix factorization in
federated recommender systems based on our study. This paper first provides a
rigorous theoretical analysis of the server reconstruction attack in four
scenarios in federated recommender systems, followed by comprehensive
experiments. The empirical results demonstrate that the FL server could infer
users' information with accuracy >80% based on the uploaded gradients from FL
nodes. The robustness analysis suggests that our reconstruction attack analysis
outperforms the random guess by >30% under Laplace noises with b no larger than
0.5 for all scenarios. Then, the paper proposes a new privacy-preserving
framework based on homomorphic encryption, Privacy-Preserving Multi-View Matrix
Factorization (PrivMVMF), to enhance user data privacy protection in federated
recommender systems. The proposed PrivMVMF is successfully implemented and
tested thoroughly with the MovieLens dataset.
| [
{
"version": "v1",
"created": "Thu, 29 Sep 2022 03:21:24 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Mai",
"Peihua",
""
],
[
"Pang",
"Yan",
""
]
] | TITLE: PrivMVMF: Privacy-Preserving Multi-View Matrix Factorization for
Recommender Systems
ABSTRACT: With an increasing focus on data privacy, there have been pilot studies on
recommender systems in a federated learning (FL) framework, where multiple
parties collaboratively train a model without sharing their data. Most of these
studies assume that the conventional FL framework can fully protect user
privacy. However, there are serious privacy risks in matrix factorization in
federated recommender systems based on our study. This paper first provides a
rigorous theoretical analysis of the server reconstruction attack in four
scenarios in federated recommender systems, followed by comprehensive
experiments. The empirical results demonstrate that the FL server could infer
users' information with accuracy >80% based on the uploaded gradients from FL
nodes. The robustness analysis suggests that our reconstruction attack analysis
outperforms the random guess by >30% under Laplace noises with b no larger than
0.5 for all scenarios. Then, the paper proposes a new privacy-preserving
framework based on homomorphic encryption, Privacy-Preserving Multi-View Matrix
Factorization (PrivMVMF), to enhance user data privacy protection in federated
recommender systems. The proposed PrivMVMF is successfully implemented and
tested thoroughly with the MovieLens dataset.
| no_new_dataset | 0.708975 |
2210.09245 | Haoming Li | Haoming Li, Xinzhuo Lin, Yang Zhou, Xiang Li, Yuchi Huo, Jiming Chen
and Qi Ye | Contact2Grasp: 3D Grasp Synthesis via Hand-Object Contact Constraint | Accepted at IJCAI 2023 | null | null | null | cs.RO cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | 3D grasp synthesis generates grasping poses given an input object. Existing
works tackle the problem by learning a direct mapping from objects to the
distributions of grasping poses. However, because the physical contact is
sensitive to small changes in pose, the high-nonlinear mapping between 3D
object representation to valid poses is considerably non-smooth, leading to
poor generation efficiency and restricted generality. To tackle the challenge,
we introduce an intermediate variable for grasp contact areas to constrain the
grasp generation; in other words, we factorize the mapping into two sequential
stages by assuming that grasping poses are fully constrained given contact
maps: 1) we first learn contact map distributions to generate the potential
contact maps for grasps; 2) then learn a mapping from the contact maps to the
grasping poses. Further, we propose a penetration-aware optimization with the
generated contacts as a consistency constraint for grasp refinement. Extensive
validations on two public datasets show that our method outperforms
state-of-the-art methods regarding grasp generation on various metrics.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2022 16:39:25 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Feb 2023 10:38:52 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 07:53:13 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Li",
"Haoming",
""
],
[
"Lin",
"Xinzhuo",
""
],
[
"Zhou",
"Yang",
""
],
[
"Li",
"Xiang",
""
],
[
"Huo",
"Yuchi",
""
],
[
"Chen",
"Jiming",
""
],
[
"Ye",
"Qi",
""
]
] | TITLE: Contact2Grasp: 3D Grasp Synthesis via Hand-Object Contact Constraint
ABSTRACT: 3D grasp synthesis generates grasping poses given an input object. Existing
works tackle the problem by learning a direct mapping from objects to the
distributions of grasping poses. However, because the physical contact is
sensitive to small changes in pose, the high-nonlinear mapping between 3D
object representation to valid poses is considerably non-smooth, leading to
poor generation efficiency and restricted generality. To tackle the challenge,
we introduce an intermediate variable for grasp contact areas to constrain the
grasp generation; in other words, we factorize the mapping into two sequential
stages by assuming that grasping poses are fully constrained given contact
maps: 1) we first learn contact map distributions to generate the potential
contact maps for grasps; 2) then learn a mapping from the contact maps to the
grasping poses. Further, we propose a penetration-aware optimization with the
generated contacts as a consistency constraint for grasp refinement. Extensive
validations on two public datasets show that our method outperforms
state-of-the-art methods regarding grasp generation on various metrics.
| no_new_dataset | 0.709422 |
2210.17166 | Hubert Etienne Dr. | Hubert Etienne and Onur \c{C}elebi | Listen to what they say: Better understand and detect online
misinformation with user feedback | null | null | 10.54501/jots.v1i5.106 | null | cs.SI cs.CY cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Social media users who report content are key allies in the management of
online misinformation, however, no research has been conducted yet to
understand their role and the different trends underlying their reporting
activity. We suggest an original approach to studying misinformation: examining
it from the reporting users perspective at the content-level and comparatively
across regions and platforms. We propose the first classification of reported
content pieces, resulting from a review of c. 9,000 items reported on Facebook
and Instagram in France, the UK, and the US in June 2020. This allows us to
observe meaningful distinctions regarding reporting content between countries
and platforms as it significantly varies in volume, type, topic, and
manipulation technique. Examining six of these techniques, we identify a novel
one that is specific to Instagram US and significantly more sophisticated than
others, potentially presenting a concrete challenge for algorithmic detection
and human moderation. We also identify four reporting behaviours, from which we
derive four types of noise capable of explaining half of the inaccuracy found
in content reported as misinformation. We finally show that breaking down the
user reporting signal into a plurality of behaviours allows to train a simple,
although competitive, classifier on a small dataset with a combination of basic
users-reports to classify the different types of reported content pieces.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2022 09:21:01 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Etienne",
"Hubert",
""
],
[
"Çelebi",
"Onur",
""
]
] | TITLE: Listen to what they say: Better understand and detect online
misinformation with user feedback
ABSTRACT: Social media users who report content are key allies in the management of
online misinformation, however, no research has been conducted yet to
understand their role and the different trends underlying their reporting
activity. We suggest an original approach to studying misinformation: examining
it from the reporting users perspective at the content-level and comparatively
across regions and platforms. We propose the first classification of reported
content pieces, resulting from a review of c. 9,000 items reported on Facebook
and Instagram in France, the UK, and the US in June 2020. This allows us to
observe meaningful distinctions regarding reporting content between countries
and platforms as it significantly varies in volume, type, topic, and
manipulation technique. Examining six of these techniques, we identify a novel
one that is specific to Instagram US and significantly more sophisticated than
others, potentially presenting a concrete challenge for algorithmic detection
and human moderation. We also identify four reporting behaviours, from which we
derive four types of noise capable of explaining half of the inaccuracy found
in content reported as misinformation. We finally show that breaking down the
user reporting signal into a plurality of behaviours allows to train a simple,
although competitive, classifier on a small dataset with a combination of basic
users-reports to classify the different types of reported content pieces.
| no_new_dataset | 0.70695 |
2211.00151 | Yangyi Chen | Yangyi Chen, Lifan Yuan, Ganqu Cui, Zhiyuan Liu, Heng Ji | A Close Look into the Calibration of Pre-trained Language Models | Accepted to ACL 2023 main conference. Code is available at:
https://github.com/lifan-yuan/PLMCalibration | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Pre-trained language models (PLMs) may fail in giving reliable estimates of
their predictive uncertainty. We take a close look into this problem, aiming to
answer two questions: (1) Do PLMs learn to become calibrated in the training
process? (2) How effective are existing calibration methods? For the first
question, we conduct fine-grained control experiments to study the dynamic
change in PLMs' calibration performance in training. We consider six factors as
control variables, including dataset difficulty, available training samples,
training steps, the number of tunable parameters, model scale, and pretraining.
We observe a consistent change in calibration performance across six factors.
We find that PLMs don't learn to become calibrated in training, evidenced by
the continual increase in confidence, no matter whether the predictions are
correct or not. We highlight that our finding somewhat contradicts two
established conclusions: (a) Larger PLMs are more calibrated; (b) Pretraining
improves model calibration. Next, we study the effectiveness of existing
calibration methods in mitigating the overconfidence issue. Besides unlearnable
calibration methods (e.g., label smoothing), we adapt and extend two recently
proposed learnable methods that directly collect data to train models to have
reasonable confidence estimations. Experimental results show that learnable
methods significantly reduce PLMs' confidence in wrong predictions. The code is
available at \url{https://github.com/lifan-yuan/PLMCalibration}.
| [
{
"version": "v1",
"created": "Mon, 31 Oct 2022 21:31:07 GMT"
},
{
"version": "v2",
"created": "Mon, 28 Nov 2022 21:15:34 GMT"
},
{
"version": "v3",
"created": "Mon, 8 May 2023 05:22:46 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Chen",
"Yangyi",
""
],
[
"Yuan",
"Lifan",
""
],
[
"Cui",
"Ganqu",
""
],
[
"Liu",
"Zhiyuan",
""
],
[
"Ji",
"Heng",
""
]
] | TITLE: A Close Look into the Calibration of Pre-trained Language Models
ABSTRACT: Pre-trained language models (PLMs) may fail in giving reliable estimates of
their predictive uncertainty. We take a close look into this problem, aiming to
answer two questions: (1) Do PLMs learn to become calibrated in the training
process? (2) How effective are existing calibration methods? For the first
question, we conduct fine-grained control experiments to study the dynamic
change in PLMs' calibration performance in training. We consider six factors as
control variables, including dataset difficulty, available training samples,
training steps, the number of tunable parameters, model scale, and pretraining.
We observe a consistent change in calibration performance across six factors.
We find that PLMs don't learn to become calibrated in training, evidenced by
the continual increase in confidence, no matter whether the predictions are
correct or not. We highlight that our finding somewhat contradicts two
established conclusions: (a) Larger PLMs are more calibrated; (b) Pretraining
improves model calibration. Next, we study the effectiveness of existing
calibration methods in mitigating the overconfidence issue. Besides unlearnable
calibration methods (e.g., label smoothing), we adapt and extend two recently
proposed learnable methods that directly collect data to train models to have
reasonable confidence estimations. Experimental results show that learnable
methods significantly reduce PLMs' confidence in wrong predictions. The code is
available at \url{https://github.com/lifan-yuan/PLMCalibration}.
| no_new_dataset | 0.707398 |
2211.00342 | Alexandra Vioni | Alexandra Vioni, Georgia Maniati, Nikolaos Ellinas, June Sig Sung,
Inchul Hwang, Aimilios Chalamandaris, Pirros Tsiakoulis | Investigating Content-Aware Neural Text-To-Speech MOS Prediction Using
Prosodic and Linguistic Features | Proceedings of ICASSP 2023 | null | 10.1109/ICASSP49357.2023.10096255 | null | cs.SD cs.CL cs.LG eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current state-of-the-art methods for automatic synthetic speech evaluation
are based on MOS prediction neural models. Such MOS prediction models include
MOSNet and LDNet that use spectral features as input, and SSL-MOS that relies
on a pretrained self-supervised learning model that directly uses the speech
signal as input. In modern high-quality neural TTS systems, prosodic
appropriateness with regard to the spoken content is a decisive factor for
speech naturalness. For this reason, we propose to include prosodic and
linguistic features as additional inputs in MOS prediction systems, and
evaluate their impact on the prediction outcome. We consider phoneme level F0
and duration features as prosodic inputs, as well as Tacotron encoder outputs,
POS tags and BERT embeddings as higher-level linguistic inputs. All MOS
prediction systems are trained on SOMOS, a neural TTS-only dataset with
crowdsourced naturalness MOS evaluations. Results show that the proposed
additional features are beneficial in the MOS prediction task, by improving the
predicted MOS scores' correlation with the ground truths, both at
utterance-level and system-level predictions.
| [
{
"version": "v1",
"created": "Tue, 1 Nov 2022 09:18:50 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 13:43:50 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Vioni",
"Alexandra",
""
],
[
"Maniati",
"Georgia",
""
],
[
"Ellinas",
"Nikolaos",
""
],
[
"Sung",
"June Sig",
""
],
[
"Hwang",
"Inchul",
""
],
[
"Chalamandaris",
"Aimilios",
""
],
[
"Tsiakoulis",
"Pirros",
""
]
] | TITLE: Investigating Content-Aware Neural Text-To-Speech MOS Prediction Using
Prosodic and Linguistic Features
ABSTRACT: Current state-of-the-art methods for automatic synthetic speech evaluation
are based on MOS prediction neural models. Such MOS prediction models include
MOSNet and LDNet that use spectral features as input, and SSL-MOS that relies
on a pretrained self-supervised learning model that directly uses the speech
signal as input. In modern high-quality neural TTS systems, prosodic
appropriateness with regard to the spoken content is a decisive factor for
speech naturalness. For this reason, we propose to include prosodic and
linguistic features as additional inputs in MOS prediction systems, and
evaluate their impact on the prediction outcome. We consider phoneme level F0
and duration features as prosodic inputs, as well as Tacotron encoder outputs,
POS tags and BERT embeddings as higher-level linguistic inputs. All MOS
prediction systems are trained on SOMOS, a neural TTS-only dataset with
crowdsourced naturalness MOS evaluations. Results show that the proposed
additional features are beneficial in the MOS prediction task, by improving the
predicted MOS scores' correlation with the ground truths, both at
utterance-level and system-level predictions.
| no_new_dataset | 0.710245 |
2211.02578 | Luis Oala | Luis Oala, Marco Aversa, Gabriel Nobis, Kurt Willis, Yoan
Neuenschwander, Mich\`ele Buck, Christian Matek, Jerome Extermann, Enrico
Pomarico, Wojciech Samek, Roderick Murray-Smith, Christoph Clausen, Bruno
Sanguinetti | Data Models for Dataset Drift Controls in Machine Learning With Optical
Images | Published as a journal paper in the Transactions on Machine Learning
Research 2023 (TMLR) available at https://openreview.net/forum?id=I4IkGmgFJz | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Camera images are ubiquitous in machine learning research. They also play a
central role in the delivery of important services spanning medicine and
environmental surveying. However, the application of machine learning models in
these domains has been limited because of robustness concerns. A primary
failure mode are performance drops due to differences between the training and
deployment data. While there are methods to prospectively validate the
robustness of machine learning models to such dataset drifts, existing
approaches do not account for explicit models of the primary object of
interest: the data. This limits our ability to study and understand the
relationship between data generation and downstream machine learning model
performance in a physically accurate manner. In this study, we demonstrate how
to overcome this limitation by pairing traditional machine learning with
physical optics to obtain explicit and differentiable data models. We
demonstrate how such data models can be constructed for image data and used to
control downstream machine learning model performance related to dataset drift.
The findings are distilled into three applications. First, drift synthesis
enables the controlled generation of physically faithful drift test cases to
power model selection and targeted generalization. Second, the gradient
connection between machine learning task model and data model allows advanced,
precise tolerancing of task model sensitivity to changes in the data
generation. These drift forensics can be used to precisely specify the
acceptable data environments in which a task model may be run. Third, drift
optimization opens up the possibility to create drifts that can help the task
model learn better faster, effectively optimizing the data generating process
itself. A guide to access the open code and datasets is available at
https://github.com/aiaudit-org/raw2logit.
| [
{
"version": "v1",
"created": "Fri, 4 Nov 2022 16:50:10 GMT"
},
{
"version": "v2",
"created": "Thu, 19 Jan 2023 18:59:45 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 05:58:05 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Oala",
"Luis",
""
],
[
"Aversa",
"Marco",
""
],
[
"Nobis",
"Gabriel",
""
],
[
"Willis",
"Kurt",
""
],
[
"Neuenschwander",
"Yoan",
""
],
[
"Buck",
"Michèle",
""
],
[
"Matek",
"Christian",
""
],
[
"Extermann",
"Jerome",
""
],
[
"Pomarico",
"Enrico",
""
],
[
"Samek",
"Wojciech",
""
],
[
"Murray-Smith",
"Roderick",
""
],
[
"Clausen",
"Christoph",
""
],
[
"Sanguinetti",
"Bruno",
""
]
] | TITLE: Data Models for Dataset Drift Controls in Machine Learning With Optical
Images
ABSTRACT: Camera images are ubiquitous in machine learning research. They also play a
central role in the delivery of important services spanning medicine and
environmental surveying. However, the application of machine learning models in
these domains has been limited because of robustness concerns. A primary
failure mode are performance drops due to differences between the training and
deployment data. While there are methods to prospectively validate the
robustness of machine learning models to such dataset drifts, existing
approaches do not account for explicit models of the primary object of
interest: the data. This limits our ability to study and understand the
relationship between data generation and downstream machine learning model
performance in a physically accurate manner. In this study, we demonstrate how
to overcome this limitation by pairing traditional machine learning with
physical optics to obtain explicit and differentiable data models. We
demonstrate how such data models can be constructed for image data and used to
control downstream machine learning model performance related to dataset drift.
The findings are distilled into three applications. First, drift synthesis
enables the controlled generation of physically faithful drift test cases to
power model selection and targeted generalization. Second, the gradient
connection between machine learning task model and data model allows advanced,
precise tolerancing of task model sensitivity to changes in the data
generation. These drift forensics can be used to precisely specify the
acceptable data environments in which a task model may be run. Third, drift
optimization opens up the possibility to create drifts that can help the task
model learn better faster, effectively optimizing the data generating process
itself. A guide to access the open code and datasets is available at
https://github.com/aiaudit-org/raw2logit.
| no_new_dataset | 0.708988 |
2211.08639 | Huaxiong Li | Haoxing Chen and Zhangxuan Gu and Yaohui Li and Jun Lan and Changhua
Meng and Weiqiang Wang and Huaxiong Li | Hierarchical Dynamic Image Harmonization | null | null | null | null | cs.CV cs.AI cs.MM | http://creativecommons.org/licenses/by/4.0/ | Image harmonization is a critical task in computer vision, which aims to
adjust the foreground to make it compatible with the background. Recent works
mainly focus on using global transformations (i.e., normalization and color
curve rendering) to achieve visual consistency. However, these models ignore
local visual consistency and their huge model sizes limit their harmonization
ability on edge devices. In this paper, we propose a hierarchical dynamic
network (HDNet) to adapt features from local to global view for better feature
transformation in efficient image harmonization. Inspired by the success of
various dynamic models, local dynamic (LD) module and mask-aware global dynamic
(MGD) module are proposed in this paper. Specifically, LD matches local
representations between the foreground and background regions based on semantic
similarities, then adaptively adjust every foreground local representation
according to the appearance of its $K$-nearest neighbor background regions. In
this way, LD can produce more realistic images at a more fine-grained level,
and simultaneously enjoy the characteristic of semantic alignment. The MGD
effectively applies distinct convolution to the foreground and background,
learning the representations of foreground and background regions as well as
their correlations to the global harmonization, facilitating local visual
consistency for the images much more efficiently. Experimental results
demonstrate that the proposed HDNet significantly reduces the total model
parameters by more than 80\% compared to previous methods, while still
attaining state-of-the-art performance on the popular iHarmony4 dataset.
Notably, the HDNet achieves a 4\% improvement in PSNR and a 19\% reduction in
MSE compared to the prior state-of-the-art methods.
| [
{
"version": "v1",
"created": "Wed, 16 Nov 2022 03:15:19 GMT"
},
{
"version": "v2",
"created": "Thu, 4 May 2023 08:36:09 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 10:04:06 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Chen",
"Haoxing",
""
],
[
"Gu",
"Zhangxuan",
""
],
[
"Li",
"Yaohui",
""
],
[
"Lan",
"Jun",
""
],
[
"Meng",
"Changhua",
""
],
[
"Wang",
"Weiqiang",
""
],
[
"Li",
"Huaxiong",
""
]
] | TITLE: Hierarchical Dynamic Image Harmonization
ABSTRACT: Image harmonization is a critical task in computer vision, which aims to
adjust the foreground to make it compatible with the background. Recent works
mainly focus on using global transformations (i.e., normalization and color
curve rendering) to achieve visual consistency. However, these models ignore
local visual consistency and their huge model sizes limit their harmonization
ability on edge devices. In this paper, we propose a hierarchical dynamic
network (HDNet) to adapt features from local to global view for better feature
transformation in efficient image harmonization. Inspired by the success of
various dynamic models, local dynamic (LD) module and mask-aware global dynamic
(MGD) module are proposed in this paper. Specifically, LD matches local
representations between the foreground and background regions based on semantic
similarities, then adaptively adjust every foreground local representation
according to the appearance of its $K$-nearest neighbor background regions. In
this way, LD can produce more realistic images at a more fine-grained level,
and simultaneously enjoy the characteristic of semantic alignment. The MGD
effectively applies distinct convolution to the foreground and background,
learning the representations of foreground and background regions as well as
their correlations to the global harmonization, facilitating local visual
consistency for the images much more efficiently. Experimental results
demonstrate that the proposed HDNet significantly reduces the total model
parameters by more than 80\% compared to previous methods, while still
attaining state-of-the-art performance on the popular iHarmony4 dataset.
Notably, the HDNet achieves a 4\% improvement in PSNR and a 19\% reduction in
MSE compared to the prior state-of-the-art methods.
| no_new_dataset | 0.711819 |
2211.08914 | Qi Guo | Qi Guo, Yong Qi, Saiyu Qi, Di Wu | Dual Class-Aware Contrastive Federated Semi-Supervised Learning | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated semi-supervised learning (FSSL), facilitates labeled clients and
unlabeled clients jointly training a global model without sharing private data.
Existing FSSL methods predominantly employ pseudo-labeling and consistency
regularization to exploit the knowledge of unlabeled data, achieving notable
success in raw data utilization. However, these training processes are hindered
by large deviations between uploaded local models of labeled and unlabeled
clients, as well as confirmation bias introduced by noisy pseudo-labels, both
of which negatively affect the global model's performance. In this paper, we
present a novel FSSL method called Dual Class-aware Contrastive Federated
Semi-Supervised Learning (DCCFSSL). This method accounts for both the local
class-aware distribution of each client's data and the global class-aware
distribution of all clients' data within the feature space. By implementing a
dual class-aware contrastive module, DCCFSSL establishes a unified training
objective for different clients to tackle large deviations and incorporates
contrastive information in the feature space to mitigate confirmation bias.
Moreover, DCCFSSL introduces an authentication-reweighted aggregation technique
to improve the server's aggregation robustness. Our comprehensive experiments
show that DCCFSSL outperforms current state-of-the-art methods on three
benchmark datasets and surpasses the FedAvg with relabeled unlabeled clients on
CIFAR-10, CIFAR-100, and STL-10 datasets. To our knowledge, we are the first to
present an FSSL method that utilizes only 10\% labeled clients, while still
achieving superior performance compared to standard federated supervised
learning, which uses all clients with labeled data.
| [
{
"version": "v1",
"created": "Wed, 16 Nov 2022 13:54:31 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 15:01:36 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Guo",
"Qi",
""
],
[
"Qi",
"Yong",
""
],
[
"Qi",
"Saiyu",
""
],
[
"Wu",
"Di",
""
]
] | TITLE: Dual Class-Aware Contrastive Federated Semi-Supervised Learning
ABSTRACT: Federated semi-supervised learning (FSSL), facilitates labeled clients and
unlabeled clients jointly training a global model without sharing private data.
Existing FSSL methods predominantly employ pseudo-labeling and consistency
regularization to exploit the knowledge of unlabeled data, achieving notable
success in raw data utilization. However, these training processes are hindered
by large deviations between uploaded local models of labeled and unlabeled
clients, as well as confirmation bias introduced by noisy pseudo-labels, both
of which negatively affect the global model's performance. In this paper, we
present a novel FSSL method called Dual Class-aware Contrastive Federated
Semi-Supervised Learning (DCCFSSL). This method accounts for both the local
class-aware distribution of each client's data and the global class-aware
distribution of all clients' data within the feature space. By implementing a
dual class-aware contrastive module, DCCFSSL establishes a unified training
objective for different clients to tackle large deviations and incorporates
contrastive information in the feature space to mitigate confirmation bias.
Moreover, DCCFSSL introduces an authentication-reweighted aggregation technique
to improve the server's aggregation robustness. Our comprehensive experiments
show that DCCFSSL outperforms current state-of-the-art methods on three
benchmark datasets and surpasses the FedAvg with relabeled unlabeled clients on
CIFAR-10, CIFAR-100, and STL-10 datasets. To our knowledge, we are the first to
present an FSSL method that utilizes only 10\% labeled clients, while still
achieving superior performance compared to standard federated supervised
learning, which uses all clients with labeled data.
| no_new_dataset | 0.710051 |
2211.13319 | Tanzila Rahman | Tanzila Rahman, Hsin-Ying Lee, Jian Ren, Sergey Tulyakov, Shweta
Mahajan, Leonid Sigal | Make-A-Story: Visual Memory Conditioned Consistent Story Generation | 11 pages | null | null | null | cs.CV | http://creativecommons.org/publicdomain/zero/1.0/ | There has been a recent explosion of impressive generative models that can
produce high quality images (or videos) conditioned on text descriptions.
However, all such approaches rely on conditional sentences that contain
unambiguous descriptions of scenes and main actors in them. Therefore employing
such models for more complex task of story visualization, where naturally
references and co-references exist, and one requires to reason about when to
maintain consistency of actors and backgrounds across frames/scenes, and when
not to, based on story progression, remains a challenge. In this work, we
address the aforementioned challenges and propose a novel autoregressive
diffusion-based framework with a visual memory module that implicitly captures
the actor and background context across the generated frames.
Sentence-conditioned soft attention over the memories enables effective
reference resolution and learns to maintain scene and actor consistency when
needed. To validate the effectiveness of our approach, we extend the MUGEN
dataset and introduce additional characters, backgrounds and referencing in
multi-sentence storylines. Our experiments for story generation on the MUGEN,
the PororoSV and the FlintstonesSV dataset show that our method not only
outperforms prior state-of-the-art in generating frames with high visual
quality, which are consistent with the story, but also models appropriate
correspondences between the characters and the background.
| [
{
"version": "v1",
"created": "Wed, 23 Nov 2022 21:38:51 GMT"
},
{
"version": "v2",
"created": "Sat, 31 Dec 2022 10:17:03 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 02:25:31 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Rahman",
"Tanzila",
""
],
[
"Lee",
"Hsin-Ying",
""
],
[
"Ren",
"Jian",
""
],
[
"Tulyakov",
"Sergey",
""
],
[
"Mahajan",
"Shweta",
""
],
[
"Sigal",
"Leonid",
""
]
] | TITLE: Make-A-Story: Visual Memory Conditioned Consistent Story Generation
ABSTRACT: There has been a recent explosion of impressive generative models that can
produce high quality images (or videos) conditioned on text descriptions.
However, all such approaches rely on conditional sentences that contain
unambiguous descriptions of scenes and main actors in them. Therefore employing
such models for more complex task of story visualization, where naturally
references and co-references exist, and one requires to reason about when to
maintain consistency of actors and backgrounds across frames/scenes, and when
not to, based on story progression, remains a challenge. In this work, we
address the aforementioned challenges and propose a novel autoregressive
diffusion-based framework with a visual memory module that implicitly captures
the actor and background context across the generated frames.
Sentence-conditioned soft attention over the memories enables effective
reference resolution and learns to maintain scene and actor consistency when
needed. To validate the effectiveness of our approach, we extend the MUGEN
dataset and introduce additional characters, backgrounds and referencing in
multi-sentence storylines. Our experiments for story generation on the MUGEN,
the PororoSV and the FlintstonesSV dataset show that our method not only
outperforms prior state-of-the-art in generating frames with high visual
quality, which are consistent with the story, but also models appropriate
correspondences between the characters and the background.
| no_new_dataset | 0.687223 |
2211.13490 | Wenbin Wang | Yuanyuan Liu, Wenbin Wang, Yibing Zhan, Shaoze Feng, Kejun Liu, Zhe
Chen | Pose-disentangled Contrastive Learning for Self-supervised Facial
Representation | Accepted by CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Self-supervised facial representation has recently attracted increasing
attention due to its ability to perform face understanding without relying on
large-scale annotated datasets heavily. However, analytically, current
contrastive-based self-supervised learning (SSL) still performs
unsatisfactorily for learning facial representation. More specifically,
existing contrastive learning (CL) tends to learn pose-invariant features that
cannot depict the pose details of faces, compromising the learning performance.
To conquer the above limitation of CL, we propose a novel Pose-disentangled
Contrastive Learning (PCL) method for general self-supervised facial
representation. Our PCL first devises a pose-disentangled decoder (PDD) with a
delicately designed orthogonalizing regulation, which disentangles the
pose-related features from the face-aware features; therefore, pose-related and
other pose-unrelated facial information could be performed in individual
subnetworks and do not affect each other's training. Furthermore, we introduce
a pose-related contrastive learning scheme that learns pose-related information
based on data augmentation of the same image, which would deliver more
effective face-aware representation for various downstream tasks. We conducted
linear evaluation on four challenging downstream facial understanding tasks,
ie, facial expression recognition, face recognition, AU detection and head pose
estimation. Experimental results demonstrate that our method significantly
outperforms state-of-the-art SSL methods. Code is available at
https://github.com/DreamMr/PCL}{https://github.com/DreamMr/PCL
| [
{
"version": "v1",
"created": "Thu, 24 Nov 2022 09:30:51 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 06:37:40 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Liu",
"Yuanyuan",
""
],
[
"Wang",
"Wenbin",
""
],
[
"Zhan",
"Yibing",
""
],
[
"Feng",
"Shaoze",
""
],
[
"Liu",
"Kejun",
""
],
[
"Chen",
"Zhe",
""
]
] | TITLE: Pose-disentangled Contrastive Learning for Self-supervised Facial
Representation
ABSTRACT: Self-supervised facial representation has recently attracted increasing
attention due to its ability to perform face understanding without relying on
large-scale annotated datasets heavily. However, analytically, current
contrastive-based self-supervised learning (SSL) still performs
unsatisfactorily for learning facial representation. More specifically,
existing contrastive learning (CL) tends to learn pose-invariant features that
cannot depict the pose details of faces, compromising the learning performance.
To conquer the above limitation of CL, we propose a novel Pose-disentangled
Contrastive Learning (PCL) method for general self-supervised facial
representation. Our PCL first devises a pose-disentangled decoder (PDD) with a
delicately designed orthogonalizing regulation, which disentangles the
pose-related features from the face-aware features; therefore, pose-related and
other pose-unrelated facial information could be performed in individual
subnetworks and do not affect each other's training. Furthermore, we introduce
a pose-related contrastive learning scheme that learns pose-related information
based on data augmentation of the same image, which would deliver more
effective face-aware representation for various downstream tasks. We conducted
linear evaluation on four challenging downstream facial understanding tasks,
ie, facial expression recognition, face recognition, AU detection and head pose
estimation. Experimental results demonstrate that our method significantly
outperforms state-of-the-art SSL methods. Code is available at
https://github.com/DreamMr/PCL}{https://github.com/DreamMr/PCL
| no_new_dataset | 0.708824 |
2212.05895 | Juhua Liu | Haibin He, Xinyuan Chen, Chaoyue Wang, Juhua Liu, Bo Du, Dacheng Tao,
Yu Qiao | Diff-Font: Diffusion Model for Robust One-Shot Font Generation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Font generation is a difficult and time-consuming task, especially in those
languages using ideograms that have complicated structures with a large number
of characters, such as Chinese. To solve this problem, few-shot font generation
and even one-shot font generation have attracted a lot of attention. However,
most existing font generation methods may still suffer from (i) large
cross-font gap challenge; (ii) subtle cross-font variation problem; and (iii)
incorrect generation of complicated characters. In this paper, we propose a
novel one-shot font generation method based on a diffusion model, named
Diff-Font, which can be stably trained on large datasets. The proposed model
aims to generate the entire font library by giving only one sample as the
reference. Specifically, a large stroke-wise dataset is constructed, and a
stroke-wise diffusion model is proposed to preserve the structure and the
completion of each generated character. To our best knowledge, the proposed
Diff-Font is the first work that developed diffusion models to handle the font
generation task. The well-trained Diff-Font is not only robust to font gap and
font variation, but also achieved promising performance on difficult character
generation. Compared to previous font generation methods, our model reaches
state-of-the-art performance both qualitatively and quantitatively.
| [
{
"version": "v1",
"created": "Mon, 12 Dec 2022 13:51:50 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Apr 2023 15:28:18 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 15:37:56 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"He",
"Haibin",
""
],
[
"Chen",
"Xinyuan",
""
],
[
"Wang",
"Chaoyue",
""
],
[
"Liu",
"Juhua",
""
],
[
"Du",
"Bo",
""
],
[
"Tao",
"Dacheng",
""
],
[
"Qiao",
"Yu",
""
]
] | TITLE: Diff-Font: Diffusion Model for Robust One-Shot Font Generation
ABSTRACT: Font generation is a difficult and time-consuming task, especially in those
languages using ideograms that have complicated structures with a large number
of characters, such as Chinese. To solve this problem, few-shot font generation
and even one-shot font generation have attracted a lot of attention. However,
most existing font generation methods may still suffer from (i) large
cross-font gap challenge; (ii) subtle cross-font variation problem; and (iii)
incorrect generation of complicated characters. In this paper, we propose a
novel one-shot font generation method based on a diffusion model, named
Diff-Font, which can be stably trained on large datasets. The proposed model
aims to generate the entire font library by giving only one sample as the
reference. Specifically, a large stroke-wise dataset is constructed, and a
stroke-wise diffusion model is proposed to preserve the structure and the
completion of each generated character. To our best knowledge, the proposed
Diff-Font is the first work that developed diffusion models to handle the font
generation task. The well-trained Diff-Font is not only robust to font gap and
font variation, but also achieved promising performance on difficult character
generation. Compared to previous font generation methods, our model reaches
state-of-the-art performance both qualitatively and quantitatively.
| no_new_dataset | 0.711055 |
2212.09420 | Daoguang Zan | Daoguang Zan, Bei Chen, Fengji Zhang, Dianjie Lu, Bingchao Wu, Bei
Guan, Yongji Wang, Jian-Guang Lou | Large Language Models Meet NL2Code: A Survey | Accepted to the main conference of ACL 2023 (long paper) | null | null | null | cs.SE cs.AI cs.CL cs.PL | http://creativecommons.org/licenses/by/4.0/ | The task of generating code from a natural language description, or NL2Code,
is considered a pressing and significant challenge in code intelligence. Thanks
to the rapid development of pre-training techniques, surging large language
models are being proposed for code, sparking the advances in NL2Code. To
facilitate further research and applications in this field, in this paper, we
present a comprehensive survey of 27 existing large language models for
NL2Code, and also review benchmarks and metrics. We provide an intuitive
comparison of all existing models on the HumanEval benchmark. Through in-depth
observation and analysis, we provide some insights and conclude that the key
factors contributing to the success of large language models for NL2Code are
"Large Size, Premium Data, Expert Tuning". In addition, we discuss challenges
and opportunities regarding the gap between models and humans. We also create a
website https://nl2code.github.io to track the latest progress through
crowd-sourcing. To the best of our knowledge, this is the first survey of large
language models for NL2Code, and we believe it will contribute to the ongoing
development of the field.
| [
{
"version": "v1",
"created": "Mon, 19 Dec 2022 12:55:32 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 10:25:41 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Zan",
"Daoguang",
""
],
[
"Chen",
"Bei",
""
],
[
"Zhang",
"Fengji",
""
],
[
"Lu",
"Dianjie",
""
],
[
"Wu",
"Bingchao",
""
],
[
"Guan",
"Bei",
""
],
[
"Wang",
"Yongji",
""
],
[
"Lou",
"Jian-Guang",
""
]
] | TITLE: Large Language Models Meet NL2Code: A Survey
ABSTRACT: The task of generating code from a natural language description, or NL2Code,
is considered a pressing and significant challenge in code intelligence. Thanks
to the rapid development of pre-training techniques, surging large language
models are being proposed for code, sparking the advances in NL2Code. To
facilitate further research and applications in this field, in this paper, we
present a comprehensive survey of 27 existing large language models for
NL2Code, and also review benchmarks and metrics. We provide an intuitive
comparison of all existing models on the HumanEval benchmark. Through in-depth
observation and analysis, we provide some insights and conclude that the key
factors contributing to the success of large language models for NL2Code are
"Large Size, Premium Data, Expert Tuning". In addition, we discuss challenges
and opportunities regarding the gap between models and humans. We also create a
website https://nl2code.github.io to track the latest progress through
crowd-sourcing. To the best of our knowledge, this is the first survey of large
language models for NL2Code, and we believe it will contribute to the ongoing
development of the field.
| no_new_dataset | 0.71027 |
2212.10846 | JiaXian Guo | Jiaxian Guo, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Boyang Li,
Dacheng Tao, Steven C.H. Hoi | From Images to Textual Prompts: Zero-shot VQA with Frozen Large Language
Models | CVPR 2023 Camera Ready Version | null | null | null | cs.CV cs.MM | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have demonstrated excellent zero-shot
generalization to new language tasks. However, effective utilization of LLMs
for zero-shot visual question-answering (VQA) remains challenging, primarily
due to the modality disconnection and task disconnection between LLM and VQA
task. End-to-end training on vision and language data may bridge the
disconnections, but is inflexible and computationally expensive. To address
this issue, we propose \emph{Img2Prompt}, a plug-and-play module that provides
the prompts that can bridge the aforementioned modality and task
disconnections, so that LLMs can perform zero-shot VQA tasks without end-to-end
training. In order to provide such prompts, we further employ LLM-agnostic
models to provide prompts that can describe image content and self-constructed
question-answer pairs, which can effectively guide LLM to perform zero-shot VQA
tasks. Img2Prompt offers the following benefits: 1) It can flexibly work with
various LLMs to perform VQA. 2)~Without the needing of end-to-end training, it
significantly reduces the cost of deploying LLM for zero-shot VQA tasks. 3) It
achieves comparable or better performance than methods relying on end-to-end
training. For example, we outperform Flamingo \cite{Deepmind:Flamingo2022} by
5.6\% on VQAv2. On the challenging A-OKVQA dataset, our method even outperforms
few-shot methods by as much as 20\%.
| [
{
"version": "v1",
"created": "Wed, 21 Dec 2022 08:39:36 GMT"
},
{
"version": "v2",
"created": "Sat, 4 Mar 2023 12:42:54 GMT"
},
{
"version": "v3",
"created": "Mon, 8 May 2023 06:04:04 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Guo",
"Jiaxian",
""
],
[
"Li",
"Junnan",
""
],
[
"Li",
"Dongxu",
""
],
[
"Tiong",
"Anthony Meng Huat",
""
],
[
"Li",
"Boyang",
""
],
[
"Tao",
"Dacheng",
""
],
[
"Hoi",
"Steven C. H.",
""
]
] | TITLE: From Images to Textual Prompts: Zero-shot VQA with Frozen Large Language
Models
ABSTRACT: Large language models (LLMs) have demonstrated excellent zero-shot
generalization to new language tasks. However, effective utilization of LLMs
for zero-shot visual question-answering (VQA) remains challenging, primarily
due to the modality disconnection and task disconnection between LLM and VQA
task. End-to-end training on vision and language data may bridge the
disconnections, but is inflexible and computationally expensive. To address
this issue, we propose \emph{Img2Prompt}, a plug-and-play module that provides
the prompts that can bridge the aforementioned modality and task
disconnections, so that LLMs can perform zero-shot VQA tasks without end-to-end
training. In order to provide such prompts, we further employ LLM-agnostic
models to provide prompts that can describe image content and self-constructed
question-answer pairs, which can effectively guide LLM to perform zero-shot VQA
tasks. Img2Prompt offers the following benefits: 1) It can flexibly work with
various LLMs to perform VQA. 2)~Without the needing of end-to-end training, it
significantly reduces the cost of deploying LLM for zero-shot VQA tasks. 3) It
achieves comparable or better performance than methods relying on end-to-end
training. For example, we outperform Flamingo \cite{Deepmind:Flamingo2022} by
5.6\% on VQAv2. On the challenging A-OKVQA dataset, our method even outperforms
few-shot methods by as much as 20\%.
| no_new_dataset | 0.708603 |
2301.01630 | Emiliano Staffoli | Emiliano Staffoli, Mattia Mancinelli, Paolo Bettotti, Lorenzo Pavesi | Equalization of a 10 Gbps IMDD signal by a small silicon photonics time
delayed neural network | 14 pages, 6 figures | Photon. Res. 11, 878-886 (2023) | 10.1364/PRJ.483356 | null | cs.ET physics.app-ph physics.optics | http://creativecommons.org/licenses/by-nc-nd/4.0/ | A small 4-channels time-delayed complex perceptron is used as a silicon
photonics neural network (NN) device to compensate for chromatic dispersion in
optical fiber links. The NN device is experimentally tested with
non-return-to-zero optical signals at 10 Gbps after propagation through up to
125 km optical fiber link. During the learning phase, a separation-loss
function is optimized in order to maximally separate the transmitted levels of
0s from the 1s, which implies an optimization of the bit-error-rate. Testing of
the NN device shows that the excess losses introduced by the NN device are
compensated by the gain in transmitted signal equalization for a link longer
than 100 km. The measured data are reproduced by a model which accounts for the
optical link and the neural network device. This allows simulating the network
performances for higher data rates, where the device shows improvement with
respect to the benchmark both in terms of performance as well as ease of use.
| [
{
"version": "v1",
"created": "Wed, 4 Jan 2023 14:06:12 GMT"
},
{
"version": "v2",
"created": "Fri, 10 Mar 2023 15:51:39 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Staffoli",
"Emiliano",
""
],
[
"Mancinelli",
"Mattia",
""
],
[
"Bettotti",
"Paolo",
""
],
[
"Pavesi",
"Lorenzo",
""
]
] | TITLE: Equalization of a 10 Gbps IMDD signal by a small silicon photonics time
delayed neural network
ABSTRACT: A small 4-channels time-delayed complex perceptron is used as a silicon
photonics neural network (NN) device to compensate for chromatic dispersion in
optical fiber links. The NN device is experimentally tested with
non-return-to-zero optical signals at 10 Gbps after propagation through up to
125 km optical fiber link. During the learning phase, a separation-loss
function is optimized in order to maximally separate the transmitted levels of
0s from the 1s, which implies an optimization of the bit-error-rate. Testing of
the NN device shows that the excess losses introduced by the NN device are
compensated by the gain in transmitted signal equalization for a link longer
than 100 km. The measured data are reproduced by a model which accounts for the
optical link and the neural network device. This allows simulating the network
performances for higher data rates, where the device shows improvement with
respect to the benchmark both in terms of performance as well as ease of use.
| no_new_dataset | 0.710998 |
2301.13364 | Wenzhuo Song | Wenzhuo Song, Shoujin Wang, Yan Wang, Kunpeng Liu, Xueyan Liu, Minghao
Yin | A Counterfactual Collaborative Session-based Recommender System | accepted by the ACM WebConf 2023 | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most session-based recommender systems (SBRSs) focus on extracting
information from the observed items in the current session of a user to predict
a next item, ignoring the causes outside the session (called outer-session
causes, OSCs) that influence the user's selection of items. However, these
causes widely exist in the real world, and few studies have investigated their
role in SBRSs. In this work, we analyze the causalities and correlations of the
OSCs in SBRSs from the perspective of causal inference. We find that the OSCs
are essentially the confounders in SBRSs, which leads to spurious correlations
in the data used to train SBRS models. To address this problem, we propose a
novel SBRS framework named COCO-SBRS (COunterfactual COllaborative
Session-Based Recommender Systems) to learn the causality between OSCs and
user-item interactions in SBRSs. COCO-SBRS first adopts a self-supervised
approach to pre-train a recommendation model by designing pseudo-labels of
causes for each user's selection of the item in data to guide the training
process. Next, COCO-SBRS adopts counterfactual inference to recommend items
based on the outputs of the pre-trained recommendation model considering the
causalities to alleviate the data sparsity problem. As a result, COCO-SBRS can
learn the causalities in data, preventing the model from learning spurious
correlations. The experimental results of our extensive experiments conducted
on three real-world datasets demonstrate the superiority of our proposed
framework over ten representative SBRSs.
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2023 02:08:34 GMT"
},
{
"version": "v2",
"created": "Tue, 7 Feb 2023 13:19:38 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 06:11:51 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Song",
"Wenzhuo",
""
],
[
"Wang",
"Shoujin",
""
],
[
"Wang",
"Yan",
""
],
[
"Liu",
"Kunpeng",
""
],
[
"Liu",
"Xueyan",
""
],
[
"Yin",
"Minghao",
""
]
] | TITLE: A Counterfactual Collaborative Session-based Recommender System
ABSTRACT: Most session-based recommender systems (SBRSs) focus on extracting
information from the observed items in the current session of a user to predict
a next item, ignoring the causes outside the session (called outer-session
causes, OSCs) that influence the user's selection of items. However, these
causes widely exist in the real world, and few studies have investigated their
role in SBRSs. In this work, we analyze the causalities and correlations of the
OSCs in SBRSs from the perspective of causal inference. We find that the OSCs
are essentially the confounders in SBRSs, which leads to spurious correlations
in the data used to train SBRS models. To address this problem, we propose a
novel SBRS framework named COCO-SBRS (COunterfactual COllaborative
Session-Based Recommender Systems) to learn the causality between OSCs and
user-item interactions in SBRSs. COCO-SBRS first adopts a self-supervised
approach to pre-train a recommendation model by designing pseudo-labels of
causes for each user's selection of the item in data to guide the training
process. Next, COCO-SBRS adopts counterfactual inference to recommend items
based on the outputs of the pre-trained recommendation model considering the
causalities to alleviate the data sparsity problem. As a result, COCO-SBRS can
learn the causalities in data, preventing the model from learning spurious
correlations. The experimental results of our extensive experiments conducted
on three real-world datasets demonstrate the superiority of our proposed
framework over ten representative SBRSs.
| no_new_dataset | 0.711794 |
2301.13384 | Ekkasit Pinyoanuntapong | Ekkasit Pinyoanuntapong, Ayman Ali, Kalvik Jakkala, Pu Wang, Minwoo
Lee, Qucheng Peng, Chen Chen, Zhi Sun | GaitSADA: Self-Aligned Domain Adaptation for mmWave Gait Recognition | Submitted to IEEE MASS 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | mmWave radar-based gait recognition is a novel user identification method
that captures human gait biometrics from mmWave radar return signals. This
technology offers privacy protection and is resilient to weather and lighting
conditions. However, its generalization performance is yet unknown and limits
its practical deployment. To address this problem, in this paper, a
non-synthetic dataset is collected and analyzed to reveal the presence of
spatial and temporal domain shifts in mmWave gait biometric data, which
significantly impacts identification accuracy. To mitigate this issue, a novel
self-aligned domain adaptation method called GaitSADA is proposed. GaitSADA
improves system generalization performance by using a two-stage semi-supervised
model training approach. The first stage employs semi-supervised contrastive
learning to learn a compact gait representation from both source and target
domain data, aligning source-target domain distributions implicitly. The second
stage uses semi-supervised consistency training with centroid alignment to
further close source-target domain gap by pseudo-labelling the target-domain
samples, clustering together the samples belonging to the same class but from
different domains, and pushing the class centroid close to the weight vector of
each class. Experiments show that GaitSADA outperforms representative domain
adaptation methods with an improvement ranging from 15.41\% to 26.32\% on
average accuracy in low data regimes. Code and dataset will be available at
https://exitudio.github.io/GaitSADA
| [
{
"version": "v1",
"created": "Tue, 31 Jan 2023 03:21:08 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Feb 2023 01:59:23 GMT"
},
{
"version": "v3",
"created": "Fri, 5 May 2023 18:01:18 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Pinyoanuntapong",
"Ekkasit",
""
],
[
"Ali",
"Ayman",
""
],
[
"Jakkala",
"Kalvik",
""
],
[
"Wang",
"Pu",
""
],
[
"Lee",
"Minwoo",
""
],
[
"Peng",
"Qucheng",
""
],
[
"Chen",
"Chen",
""
],
[
"Sun",
"Zhi",
""
]
] | TITLE: GaitSADA: Self-Aligned Domain Adaptation for mmWave Gait Recognition
ABSTRACT: mmWave radar-based gait recognition is a novel user identification method
that captures human gait biometrics from mmWave radar return signals. This
technology offers privacy protection and is resilient to weather and lighting
conditions. However, its generalization performance is yet unknown and limits
its practical deployment. To address this problem, in this paper, a
non-synthetic dataset is collected and analyzed to reveal the presence of
spatial and temporal domain shifts in mmWave gait biometric data, which
significantly impacts identification accuracy. To mitigate this issue, a novel
self-aligned domain adaptation method called GaitSADA is proposed. GaitSADA
improves system generalization performance by using a two-stage semi-supervised
model training approach. The first stage employs semi-supervised contrastive
learning to learn a compact gait representation from both source and target
domain data, aligning source-target domain distributions implicitly. The second
stage uses semi-supervised consistency training with centroid alignment to
further close source-target domain gap by pseudo-labelling the target-domain
samples, clustering together the samples belonging to the same class but from
different domains, and pushing the class centroid close to the weight vector of
each class. Experiments show that GaitSADA outperforms representative domain
adaptation methods with an improvement ranging from 15.41\% to 26.32\% on
average accuracy in low data regimes. Code and dataset will be available at
https://exitudio.github.io/GaitSADA
| no_new_dataset | 0.593079 |
2302.06761 | Yuan He | Yuan He, Jiaoyan Chen, Ernesto Jim\'enez-Ruiz, Hang Dong, Ian Horrocks | Language Model Analysis for Ontology Subsumption Inference | Accepted at Findings of ACL 2023; OntoLAMA Datasets are available at:
https://huggingface.co/datasets/krr-oxford/OntoLAMA (Huggingface) or
https://doi.org/10.5281/zenodo.6480540 (Zenodo) | null | null | null | cs.CL cs.AI cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Investigating whether pre-trained language models (LMs) can function as
knowledge bases (KBs) has raised wide research interests recently. However,
existing works focus on simple, triple-based, relational KBs, but omit more
sophisticated, logic-based, conceptualised KBs such as OWL ontologies. To
investigate an LM's knowledge of ontologies, we propose OntoLAMA, a set of
inference-based probing tasks and datasets from ontology subsumption axioms
involving both atomic and complex concepts. We conduct extensive experiments on
ontologies of different domains and scales, and our results demonstrate that
LMs encode relatively less background knowledge of Subsumption Inference (SI)
than traditional Natural Language Inference (NLI) but can improve on SI
significantly when a small number of samples are given. We will open-source our
code and datasets.
| [
{
"version": "v1",
"created": "Tue, 14 Feb 2023 00:21:56 GMT"
},
{
"version": "v2",
"created": "Tue, 2 May 2023 16:06:38 GMT"
},
{
"version": "v3",
"created": "Mon, 8 May 2023 17:21:40 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"He",
"Yuan",
""
],
[
"Chen",
"Jiaoyan",
""
],
[
"Jiménez-Ruiz",
"Ernesto",
""
],
[
"Dong",
"Hang",
""
],
[
"Horrocks",
"Ian",
""
]
] | TITLE: Language Model Analysis for Ontology Subsumption Inference
ABSTRACT: Investigating whether pre-trained language models (LMs) can function as
knowledge bases (KBs) has raised wide research interests recently. However,
existing works focus on simple, triple-based, relational KBs, but omit more
sophisticated, logic-based, conceptualised KBs such as OWL ontologies. To
investigate an LM's knowledge of ontologies, we propose OntoLAMA, a set of
inference-based probing tasks and datasets from ontology subsumption axioms
involving both atomic and complex concepts. We conduct extensive experiments on
ontologies of different domains and scales, and our results demonstrate that
LMs encode relatively less background knowledge of Subsumption Inference (SI)
than traditional Natural Language Inference (NLI) but can improve on SI
significantly when a small number of samples are given. We will open-source our
code and datasets.
| no_new_dataset | 0.702568 |
2302.08751 | Seunghyeon Seo | Seunghyeon Seo, Jaeyoung Yoo, Jihye Hwang, Nojun Kwak | MDPose: Real-Time Multi-Person Pose Estimation via Mixture Density Model | UAI 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | One of the major challenges in multi-person pose estimation is instance-aware
keypoint estimation. Previous methods address this problem by leveraging an
off-the-shelf detector, heuristic post-grouping process or explicit instance
identification process, hindering further improvements in the inference speed
which is an important factor for practical applications. From the statistical
point of view, those additional processes for identifying instances are
necessary to bypass learning the high-dimensional joint distribution of human
keypoints, which is a critical factor for another major challenge, the
occlusion scenario. In this work, we propose a novel framework of single-stage
instance-aware pose estimation by modeling the joint distribution of human
keypoints with a mixture density model, termed as MDPose. Our MDPose estimates
the distribution of human keypoints' coordinates using a mixture density model
with an instance-aware keypoint head consisting simply of 8 convolutional
layers. It is trained by minimizing the negative log-likelihood of the ground
truth keypoints. Also, we propose a simple yet effective training strategy,
Random Keypoint Grouping (RKG), which significantly alleviates the underflow
problem leading to successful learning of relations between keypoints. On
OCHuman dataset, which consists of images with highly occluded people, our
MDPose achieves state-of-the-art performance by successfully learning the
high-dimensional joint distribution of human keypoints. Furthermore, our MDPose
shows significant improvement in inference speed with a competitive accuracy on
MS COCO, a widely-used human keypoint dataset, thanks to the proposed much
simpler single-stage pipeline.
| [
{
"version": "v1",
"created": "Fri, 17 Feb 2023 08:29:33 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 12:22:30 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Seo",
"Seunghyeon",
""
],
[
"Yoo",
"Jaeyoung",
""
],
[
"Hwang",
"Jihye",
""
],
[
"Kwak",
"Nojun",
""
]
] | TITLE: MDPose: Real-Time Multi-Person Pose Estimation via Mixture Density Model
ABSTRACT: One of the major challenges in multi-person pose estimation is instance-aware
keypoint estimation. Previous methods address this problem by leveraging an
off-the-shelf detector, heuristic post-grouping process or explicit instance
identification process, hindering further improvements in the inference speed
which is an important factor for practical applications. From the statistical
point of view, those additional processes for identifying instances are
necessary to bypass learning the high-dimensional joint distribution of human
keypoints, which is a critical factor for another major challenge, the
occlusion scenario. In this work, we propose a novel framework of single-stage
instance-aware pose estimation by modeling the joint distribution of human
keypoints with a mixture density model, termed as MDPose. Our MDPose estimates
the distribution of human keypoints' coordinates using a mixture density model
with an instance-aware keypoint head consisting simply of 8 convolutional
layers. It is trained by minimizing the negative log-likelihood of the ground
truth keypoints. Also, we propose a simple yet effective training strategy,
Random Keypoint Grouping (RKG), which significantly alleviates the underflow
problem leading to successful learning of relations between keypoints. On
OCHuman dataset, which consists of images with highly occluded people, our
MDPose achieves state-of-the-art performance by successfully learning the
high-dimensional joint distribution of human keypoints. Furthermore, our MDPose
shows significant improvement in inference speed with a competitive accuracy on
MS COCO, a widely-used human keypoint dataset, thanks to the proposed much
simpler single-stage pipeline.
| no_new_dataset | 0.707569 |
2302.08888 | Qiying Yu | Qiying Yu, Yang Liu, Yimu Wang, Ke Xu, Jingjing Liu | Multimodal Federated Learning via Contrastive Representation Ensemble | ICLR 2023, update | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the increasing amount of multimedia data on modern mobile systems and
IoT infrastructures, harnessing these rich multimodal data without breaching
user privacy becomes a critical issue. Federated learning (FL) serves as a
privacy-conscious alternative to centralized machine learning. However,
existing FL methods extended to multimodal data all rely on model aggregation
on single modality level, which restrains the server and clients to have
identical model architecture for each modality. This limits the global model in
terms of both model complexity and data capacity, not to mention task
diversity. In this work, we propose Contrastive Representation Ensemble and
Aggregation for Multimodal FL (CreamFL), a multimodal federated learning
framework that enables training larger server models from clients with
heterogeneous model architectures and data modalities, while only communicating
knowledge on public dataset. To achieve better multimodal representation
fusion, we design a global-local cross-modal ensemble strategy to aggregate
client representations. To mitigate local model drift caused by two
unprecedented heterogeneous factors stemming from multimodal discrepancy
(modality gap and task gap), we further propose two inter-modal and intra-modal
contrasts to regularize local training, which complements information of the
absent modality for uni-modal clients and regularizes local clients to head
towards global consensus. Thorough evaluations and ablation studies on
image-text retrieval and visual question answering tasks showcase the
superiority of CreamFL over state-of-the-art FL methods and its practical
value.
| [
{
"version": "v1",
"created": "Fri, 17 Feb 2023 14:17:44 GMT"
},
{
"version": "v2",
"created": "Mon, 13 Mar 2023 12:14:55 GMT"
},
{
"version": "v3",
"created": "Sat, 6 May 2023 02:26:32 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Yu",
"Qiying",
""
],
[
"Liu",
"Yang",
""
],
[
"Wang",
"Yimu",
""
],
[
"Xu",
"Ke",
""
],
[
"Liu",
"Jingjing",
""
]
] | TITLE: Multimodal Federated Learning via Contrastive Representation Ensemble
ABSTRACT: With the increasing amount of multimedia data on modern mobile systems and
IoT infrastructures, harnessing these rich multimodal data without breaching
user privacy becomes a critical issue. Federated learning (FL) serves as a
privacy-conscious alternative to centralized machine learning. However,
existing FL methods extended to multimodal data all rely on model aggregation
on single modality level, which restrains the server and clients to have
identical model architecture for each modality. This limits the global model in
terms of both model complexity and data capacity, not to mention task
diversity. In this work, we propose Contrastive Representation Ensemble and
Aggregation for Multimodal FL (CreamFL), a multimodal federated learning
framework that enables training larger server models from clients with
heterogeneous model architectures and data modalities, while only communicating
knowledge on public dataset. To achieve better multimodal representation
fusion, we design a global-local cross-modal ensemble strategy to aggregate
client representations. To mitigate local model drift caused by two
unprecedented heterogeneous factors stemming from multimodal discrepancy
(modality gap and task gap), we further propose two inter-modal and intra-modal
contrasts to regularize local training, which complements information of the
absent modality for uni-modal clients and regularizes local clients to head
towards global consensus. Thorough evaluations and ablation studies on
image-text retrieval and visual question answering tasks showcase the
superiority of CreamFL over state-of-the-art FL methods and its practical
value.
| no_new_dataset | 0.709422 |
2302.09309 | Yuqian Fu | Yuqian Fu, Yu Xie, Yanwei Fu, Yu-Gang Jiang | StyleAdv: Meta Style Adversarial Training for Cross-Domain Few-Shot
Learning | accepted by CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cross-Domain Few-Shot Learning (CD-FSL) is a recently emerging task that
tackles few-shot learning across different domains. It aims at transferring
prior knowledge learned on the source dataset to novel target datasets. The
CD-FSL task is especially challenged by the huge domain gap between different
datasets. Critically, such a domain gap actually comes from the changes of
visual styles, and wave-SAN empirically shows that spanning the style
distribution of the source data helps alleviate this issue. However, wave-SAN
simply swaps styles of two images. Such a vanilla operation makes the generated
styles ``real'' and ``easy'', which still fall into the original set of the
source styles. Thus, inspired by vanilla adversarial learning, a novel
model-agnostic meta Style Adversarial training (StyleAdv) method together with
a novel style adversarial attack method is proposed for CD-FSL. Particularly,
our style attack method synthesizes both ``virtual'' and ``hard'' adversarial
styles for model training. This is achieved by perturbing the original style
with the signed style gradients. By continually attacking styles and forcing
the model to recognize these challenging adversarial styles, our model is
gradually robust to the visual styles, thus boosting the generalization ability
for novel target datasets. Besides the typical CNN-based backbone, we also
employ our StyleAdv method on large-scale pretrained vision transformer.
Extensive experiments conducted on eight various target datasets show the
effectiveness of our method. Whether built upon ResNet or ViT, we achieve the
new state of the art for CD-FSL. Code is available at
https://github.com/lovelyqian/StyleAdv-CDFSL.
| [
{
"version": "v1",
"created": "Sat, 18 Feb 2023 11:54:37 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 11:52:31 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Fu",
"Yuqian",
""
],
[
"Xie",
"Yu",
""
],
[
"Fu",
"Yanwei",
""
],
[
"Jiang",
"Yu-Gang",
""
]
] | TITLE: StyleAdv: Meta Style Adversarial Training for Cross-Domain Few-Shot
Learning
ABSTRACT: Cross-Domain Few-Shot Learning (CD-FSL) is a recently emerging task that
tackles few-shot learning across different domains. It aims at transferring
prior knowledge learned on the source dataset to novel target datasets. The
CD-FSL task is especially challenged by the huge domain gap between different
datasets. Critically, such a domain gap actually comes from the changes of
visual styles, and wave-SAN empirically shows that spanning the style
distribution of the source data helps alleviate this issue. However, wave-SAN
simply swaps styles of two images. Such a vanilla operation makes the generated
styles ``real'' and ``easy'', which still fall into the original set of the
source styles. Thus, inspired by vanilla adversarial learning, a novel
model-agnostic meta Style Adversarial training (StyleAdv) method together with
a novel style adversarial attack method is proposed for CD-FSL. Particularly,
our style attack method synthesizes both ``virtual'' and ``hard'' adversarial
styles for model training. This is achieved by perturbing the original style
with the signed style gradients. By continually attacking styles and forcing
the model to recognize these challenging adversarial styles, our model is
gradually robust to the visual styles, thus boosting the generalization ability
for novel target datasets. Besides the typical CNN-based backbone, we also
employ our StyleAdv method on large-scale pretrained vision transformer.
Extensive experiments conducted on eight various target datasets show the
effectiveness of our method. Whether built upon ResNet or ViT, we achieve the
new state of the art for CD-FSL. Code is available at
https://github.com/lovelyqian/StyleAdv-CDFSL.
| no_new_dataset | 0.709988 |
2303.01732 | Takato Yasuno | Takato Yasuno, Masahiro Okano, Junichiro Fujii | One-class Damage Detector Using Deeper Fully-Convolutional Data
Descriptions for Civil Application | 11 pages, 9 figures, 3 tables | Adv. Artif. Intell. Mach. Learn., 3 (2):996-1011 (Published On:
02-May-2023) | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Infrastructure managers must maintain high standards to ensure user
satisfaction during the lifecycle of infrastructures. Surveillance cameras and
visual inspections have enabled progress in automating the detection of
anomalous features and assessing the occurrence of deterioration. However,
collecting damage data is typically time consuming and requires repeated
inspections. The one-class damage detection approach has an advantage in that
normal images can be used to optimize model parameters. Additionally, visual
evaluation of heatmaps enables us to understand localized anomalous features.
The authors highlight damage vision applications utilized in the robust
property and localized damage explainability. First, we propose a civil-purpose
application for automating one-class damage detection reproducing a fully
convolutional data description (FCDD) as a baseline model. We have obtained
accurate and explainable results demonstrating experimental studies on concrete
damage and steel corrosion in civil engineering. Additionally, to develop a
more robust application, we applied our method to another outdoor domain that
contains complex and noisy backgrounds using natural disaster datasets
collected using various devices. Furthermore, we propose a valuable solution of
deeper FCDDs focusing on other powerful backbones to improve the performance of
damage detection and implement ablation studies on disaster datasets. The key
results indicate that the deeper FCDDs outperformed the baseline FCDD on
datasets representing natural disaster damage caused by hurricanes, typhoons,
earthquakes, and four-event disasters.
| [
{
"version": "v1",
"created": "Fri, 3 Mar 2023 06:27:15 GMT"
},
{
"version": "v2",
"created": "Sat, 18 Mar 2023 02:54:02 GMT"
},
{
"version": "v3",
"created": "Mon, 8 May 2023 16:12:44 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Yasuno",
"Takato",
""
],
[
"Okano",
"Masahiro",
""
],
[
"Fujii",
"Junichiro",
""
]
] | TITLE: One-class Damage Detector Using Deeper Fully-Convolutional Data
Descriptions for Civil Application
ABSTRACT: Infrastructure managers must maintain high standards to ensure user
satisfaction during the lifecycle of infrastructures. Surveillance cameras and
visual inspections have enabled progress in automating the detection of
anomalous features and assessing the occurrence of deterioration. However,
collecting damage data is typically time consuming and requires repeated
inspections. The one-class damage detection approach has an advantage in that
normal images can be used to optimize model parameters. Additionally, visual
evaluation of heatmaps enables us to understand localized anomalous features.
The authors highlight damage vision applications utilized in the robust
property and localized damage explainability. First, we propose a civil-purpose
application for automating one-class damage detection reproducing a fully
convolutional data description (FCDD) as a baseline model. We have obtained
accurate and explainable results demonstrating experimental studies on concrete
damage and steel corrosion in civil engineering. Additionally, to develop a
more robust application, we applied our method to another outdoor domain that
contains complex and noisy backgrounds using natural disaster datasets
collected using various devices. Furthermore, we propose a valuable solution of
deeper FCDDs focusing on other powerful backbones to improve the performance of
damage detection and implement ablation studies on disaster datasets. The key
results indicate that the deeper FCDDs outperformed the baseline FCDD on
datasets representing natural disaster damage caused by hurricanes, typhoons,
earthquakes, and four-event disasters.
| no_new_dataset | 0.711387 |
2303.01740 | Juan Hu | Juan Hu, Xin Liao, Difei Gao, Satoshi Tsutsui, Qian Wang, Zheng Qin,
Mike Zheng Shou | Mover: Mask and Recovery based Facial Part Consistency Aware Method for
Deepfake Video Detection | null | null | null | null | cs.CV cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deepfake techniques have been widely used for malicious purposes, prompting
extensive research interest in developing Deepfake detection methods. Deepfake
manipulations typically involve tampering with facial parts, which can result
in inconsistencies across different parts of the face. For instance, Deepfake
techniques may change smiling lips to an upset lip, while the eyes remain
smiling. Existing detection methods depend on specific indicators of forgery,
which tend to disappear as the forgery patterns are improved. To address the
limitation, we propose Mover, a new Deepfake detection model that exploits
unspecific facial part inconsistencies, which are inevitable weaknesses of
Deepfake videos. Mover randomly masks regions of interest (ROIs) and recovers
faces to learn unspecific features, which makes it difficult for fake faces to
be recovered, while real faces can be easily recovered. Specifically, given a
real face image, we first pretrain a masked autoencoder to learn facial part
consistency by dividing faces into three parts and randomly masking ROIs, which
are then recovered based on the unmasked facial parts. Furthermore, to maximize
the discrepancy between real and fake videos, we propose a novel model with
dual networks that utilize the pretrained encoder and masked autoencoder,
respectively. 1) The pretrained encoder is finetuned for capturing the encoding
of inconsistent information in the given video. 2) The pretrained masked
autoencoder is utilized for mapping faces and distinguishing real and fake
videos. Our extensive experiments on standard benchmarks demonstrate that Mover
is highly effective.
| [
{
"version": "v1",
"created": "Fri, 3 Mar 2023 06:57:22 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 02:23:25 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Hu",
"Juan",
""
],
[
"Liao",
"Xin",
""
],
[
"Gao",
"Difei",
""
],
[
"Tsutsui",
"Satoshi",
""
],
[
"Wang",
"Qian",
""
],
[
"Qin",
"Zheng",
""
],
[
"Shou",
"Mike Zheng",
""
]
] | TITLE: Mover: Mask and Recovery based Facial Part Consistency Aware Method for
Deepfake Video Detection
ABSTRACT: Deepfake techniques have been widely used for malicious purposes, prompting
extensive research interest in developing Deepfake detection methods. Deepfake
manipulations typically involve tampering with facial parts, which can result
in inconsistencies across different parts of the face. For instance, Deepfake
techniques may change smiling lips to an upset lip, while the eyes remain
smiling. Existing detection methods depend on specific indicators of forgery,
which tend to disappear as the forgery patterns are improved. To address the
limitation, we propose Mover, a new Deepfake detection model that exploits
unspecific facial part inconsistencies, which are inevitable weaknesses of
Deepfake videos. Mover randomly masks regions of interest (ROIs) and recovers
faces to learn unspecific features, which makes it difficult for fake faces to
be recovered, while real faces can be easily recovered. Specifically, given a
real face image, we first pretrain a masked autoencoder to learn facial part
consistency by dividing faces into three parts and randomly masking ROIs, which
are then recovered based on the unmasked facial parts. Furthermore, to maximize
the discrepancy between real and fake videos, we propose a novel model with
dual networks that utilize the pretrained encoder and masked autoencoder,
respectively. 1) The pretrained encoder is finetuned for capturing the encoding
of inconsistent information in the given video. 2) The pretrained masked
autoencoder is utilized for mapping faces and distinguishing real and fake
videos. Our extensive experiments on standard benchmarks demonstrate that Mover
is highly effective.
| no_new_dataset | 0.711186 |
2303.02105 | Joyanta Jyoti Mondal | Jannatun Noor, Rizwanul Haque Ratul, Mir Rownak Ali Uday, Joyanta
Jyoti Mondal, Md. Sadiqul Islam Sakif, A. B. M. Alim Al Islam | Sherlock in OSS: A Novel Approach of Content-Based Searching in Object
Storage System | 13 Pages, 9 Figures, Submitted to IEEE Transactions on Parallel and
Distributed Systems for possible publication. arXiv admin note: substantial
text overlap with arXiv:1306.3075 by other authors; substantial text overlap
with arXiv:1910.05786 by other authors without attribution | null | null | null | cs.DC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Object Storage Systems (OSS) inside a cloud promise scalability, durability,
availability, and concurrency. However, open-source OSS does not have a
specific approach to letting users and administrators search based on the data,
which is contained inside the object storage, without involving the entire
cloud infrastructure. Therefore, in this paper, we propose Sherlock, a novel
Content-Based Searching (CoBS) architecture to extract additional information
from images and documents. Here, we store the additional information in an
Elasticsearch-enabled database, which helps us to search for our desired data
based on its contents. This approach works in two sequential stages. First, the
data will be uploaded to a classifier that will determine the data type and
send it to the specific model for the data. Here, the images that are being
uploaded are sent to our trained model for object detection, and the documents
are sent for keyword extraction. Next, the extracted information is sent to
Elasticsearch, which enables searching based on the contents. Because the
precision of the models is so fundamental to the search's correctness, we train
our models with comprehensive datasets (Microsoft COCO Dataset for multimedia
data and SemEval2017 Dataset for document data). Furthermore, we put our
designed architecture to the test with a real-world implementation of an
open-source OSS called OpenStack Swift. We upload images into the dataset of
our implementation in various segments to find out the efficacy of our proposed
model in real-life Swift object storage.
| [
{
"version": "v1",
"created": "Tue, 24 Jan 2023 07:11:14 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 14:44:33 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Noor",
"Jannatun",
""
],
[
"Ratul",
"Rizwanul Haque",
""
],
[
"Uday",
"Mir Rownak Ali",
""
],
[
"Mondal",
"Joyanta Jyoti",
""
],
[
"Sakif",
"Md. Sadiqul Islam",
""
],
[
"Islam",
"A. B. M. Alim Al",
""
]
] | TITLE: Sherlock in OSS: A Novel Approach of Content-Based Searching in Object
Storage System
ABSTRACT: Object Storage Systems (OSS) inside a cloud promise scalability, durability,
availability, and concurrency. However, open-source OSS does not have a
specific approach to letting users and administrators search based on the data,
which is contained inside the object storage, without involving the entire
cloud infrastructure. Therefore, in this paper, we propose Sherlock, a novel
Content-Based Searching (CoBS) architecture to extract additional information
from images and documents. Here, we store the additional information in an
Elasticsearch-enabled database, which helps us to search for our desired data
based on its contents. This approach works in two sequential stages. First, the
data will be uploaded to a classifier that will determine the data type and
send it to the specific model for the data. Here, the images that are being
uploaded are sent to our trained model for object detection, and the documents
are sent for keyword extraction. Next, the extracted information is sent to
Elasticsearch, which enables searching based on the contents. Because the
precision of the models is so fundamental to the search's correctness, we train
our models with comprehensive datasets (Microsoft COCO Dataset for multimedia
data and SemEval2017 Dataset for document data). Furthermore, we put our
designed architecture to the test with a real-world implementation of an
open-source OSS called OpenStack Swift. We upload images into the dataset of
our implementation in various segments to find out the efficacy of our proposed
model in real-life Swift object storage.
| no_new_dataset | 0.710201 |
2303.04201 | Shantanu Ghosh | Shantanu Ghosh, Zheng Feng, Jiang Bian, Kevin Butler, Mattia Prosperi | DR-VIDAL -- Doubly Robust Variational Information-theoretic Deep
Adversarial Learning for Counterfactual Prediction and Treatment Effect
Estimation on Real World Data | AMIA Annual Symposium, 2022
(https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10148269/) | Journal="AMIA Annu Symp Proc", Year="2022",
Volume="2022",Pages="485--494" | null | null | cs.LG stat.ME | http://creativecommons.org/licenses/by/4.0/ | Determining causal effects of interventions onto outcomes from real-world,
observational (non-randomized) data, e.g., treatment repurposing using
electronic health records, is challenging due to underlying bias. Causal deep
learning has improved over traditional techniques for estimating individualized
treatment effects (ITE). We present the Doubly Robust Variational
Information-theoretic Deep Adversarial Learning (DR-VIDAL), a novel generative
framework that combines two joint models of treatment and outcome, ensuring an
unbiased ITE estimation even when one of the two is misspecified. DR-VIDAL
integrates: (i) a variational autoencoder (VAE) to factorize confounders into
latent variables according to causal assumptions; (ii) an information-theoretic
generative adversarial network (Info-GAN) to generate counterfactuals; (iii) a
doubly robust block incorporating treatment propensities for outcome
predictions. On synthetic and real-world datasets (Infant Health and
Development Program, Twin Birth Registry, and National Supported Work Program),
DR-VIDAL achieves better performance than other non-generative and generative
methods. In conclusion, DR-VIDAL uniquely fuses causal assumptions, VAE,
Info-GAN, and doubly robustness into a comprehensive, performant framework.
Code is available at: https://github.com/Shantanu48114860/DR-VIDAL-AMIA-22
under MIT license.
| [
{
"version": "v1",
"created": "Tue, 7 Mar 2023 19:44:58 GMT"
},
{
"version": "v2",
"created": "Thu, 4 May 2023 14:44:26 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 17:36:04 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Ghosh",
"Shantanu",
""
],
[
"Feng",
"Zheng",
""
],
[
"Bian",
"Jiang",
""
],
[
"Butler",
"Kevin",
""
],
[
"Prosperi",
"Mattia",
""
]
] | TITLE: DR-VIDAL -- Doubly Robust Variational Information-theoretic Deep
Adversarial Learning for Counterfactual Prediction and Treatment Effect
Estimation on Real World Data
ABSTRACT: Determining causal effects of interventions onto outcomes from real-world,
observational (non-randomized) data, e.g., treatment repurposing using
electronic health records, is challenging due to underlying bias. Causal deep
learning has improved over traditional techniques for estimating individualized
treatment effects (ITE). We present the Doubly Robust Variational
Information-theoretic Deep Adversarial Learning (DR-VIDAL), a novel generative
framework that combines two joint models of treatment and outcome, ensuring an
unbiased ITE estimation even when one of the two is misspecified. DR-VIDAL
integrates: (i) a variational autoencoder (VAE) to factorize confounders into
latent variables according to causal assumptions; (ii) an information-theoretic
generative adversarial network (Info-GAN) to generate counterfactuals; (iii) a
doubly robust block incorporating treatment propensities for outcome
predictions. On synthetic and real-world datasets (Infant Health and
Development Program, Twin Birth Registry, and National Supported Work Program),
DR-VIDAL achieves better performance than other non-generative and generative
methods. In conclusion, DR-VIDAL uniquely fuses causal assumptions, VAE,
Info-GAN, and doubly robustness into a comprehensive, performant framework.
Code is available at: https://github.com/Shantanu48114860/DR-VIDAL-AMIA-22
under MIT license.
| no_new_dataset | 0.710635 |
2303.04222 | Hannah Rose Kirk Miss | Hannah Rose Kirk, Wenjie Yin, Bertie Vidgen, Paul R\"ottger | SemEval-2023 Task 10: Explainable Detection of Online Sexism | SemEval-2023 Task 10 (ACL 2023) | null | null | null | cs.CL cs.CY | http://creativecommons.org/licenses/by/4.0/ | Online sexism is a widespread and harmful phenomenon. Automated tools can
assist the detection of sexism at scale. Binary detection, however, disregards
the diversity of sexist content, and fails to provide clear explanations for
why something is sexist. To address this issue, we introduce SemEval Task 10 on
the Explainable Detection of Online Sexism (EDOS). We make three main
contributions: i) a novel hierarchical taxonomy of sexist content, which
includes granular vectors of sexism to aid explainability; ii) a new dataset of
20,000 social media comments with fine-grained labels, along with larger
unlabelled datasets for model adaptation; and iii) baseline models as well as
an analysis of the methods, results and errors for participant submissions to
our task.
| [
{
"version": "v1",
"created": "Tue, 7 Mar 2023 20:28:39 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 14:34:49 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Kirk",
"Hannah Rose",
""
],
[
"Yin",
"Wenjie",
""
],
[
"Vidgen",
"Bertie",
""
],
[
"Röttger",
"Paul",
""
]
] | TITLE: SemEval-2023 Task 10: Explainable Detection of Online Sexism
ABSTRACT: Online sexism is a widespread and harmful phenomenon. Automated tools can
assist the detection of sexism at scale. Binary detection, however, disregards
the diversity of sexist content, and fails to provide clear explanations for
why something is sexist. To address this issue, we introduce SemEval Task 10 on
the Explainable Detection of Online Sexism (EDOS). We make three main
contributions: i) a novel hierarchical taxonomy of sexist content, which
includes granular vectors of sexism to aid explainability; ii) a new dataset of
20,000 social media comments with fine-grained labels, along with larger
unlabelled datasets for model adaptation; and iii) baseline models as well as
an analysis of the methods, results and errors for participant submissions to
our task.
| new_dataset | 0.714609 |
2303.04611 | Furong Ye | Furong Ye and Frank Neumann and Jacob de Nobel and Aneta Neumann and
Thomas B\"ack | Towards Self-adaptive Mutation in Evolutionary Multi-Objective
Algorithms | submitted to FOGA 2023 | null | null | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Parameter control has succeeded in accelerating the convergence process of
evolutionary algorithms. While empirical and theoretical studies have shed
light on the behavior of algorithms for single-objective optimization, little
is known about how self-adaptation influences multi-objective evolutionary
algorithms. In this work, we contribute (1) extensive experimental analysis of
the Global Simple Evolutionary Multi-objective Algorithm (GSEMO) variants on
classic problems, such as OneMinMax, LOTZ, COCZ, and (2) a novel version of
GSEMO with self-adaptive mutation.
To enable self-adaptation in GSEMO, we explore three self-adaptive mutation
techniques from single-objective optimization and use various performance
metrics, such as hypervolume and inverted generational distance, to guide the
adaptation. Our experiments show that adapting the mutation rate based on
single-objective optimization and hypervolume can speed up the convergence of
GSEMO. Moreover, we propose a GSEMO with self-adaptive mutation, which
considers optimizing for single objectives and adjusts the mutation rate for
each solution individually. Our results demonstrate that the proposed method
outperforms the GSEMO with static mutation rates across all the tested
problems.
This work provides a comprehensive benchmarking study for MOEAs and
complements existing theoretical runtime analysis. Our proposed algorithm
addresses interesting issues for designing MOEAs for future practical
applications.
| [
{
"version": "v1",
"created": "Wed, 8 Mar 2023 14:26:46 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 12:26:27 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Ye",
"Furong",
""
],
[
"Neumann",
"Frank",
""
],
[
"de Nobel",
"Jacob",
""
],
[
"Neumann",
"Aneta",
""
],
[
"Bäck",
"Thomas",
""
]
] | TITLE: Towards Self-adaptive Mutation in Evolutionary Multi-Objective
Algorithms
ABSTRACT: Parameter control has succeeded in accelerating the convergence process of
evolutionary algorithms. While empirical and theoretical studies have shed
light on the behavior of algorithms for single-objective optimization, little
is known about how self-adaptation influences multi-objective evolutionary
algorithms. In this work, we contribute (1) extensive experimental analysis of
the Global Simple Evolutionary Multi-objective Algorithm (GSEMO) variants on
classic problems, such as OneMinMax, LOTZ, COCZ, and (2) a novel version of
GSEMO with self-adaptive mutation.
To enable self-adaptation in GSEMO, we explore three self-adaptive mutation
techniques from single-objective optimization and use various performance
metrics, such as hypervolume and inverted generational distance, to guide the
adaptation. Our experiments show that adapting the mutation rate based on
single-objective optimization and hypervolume can speed up the convergence of
GSEMO. Moreover, we propose a GSEMO with self-adaptive mutation, which
considers optimizing for single objectives and adjusts the mutation rate for
each solution individually. Our results demonstrate that the proposed method
outperforms the GSEMO with static mutation rates across all the tested
problems.
This work provides a comprehensive benchmarking study for MOEAs and
complements existing theoretical runtime analysis. Our proposed algorithm
addresses interesting issues for designing MOEAs for future practical
applications.
| no_new_dataset | 0.707196 |
2303.06791 | Arun Tejasvi Chaganty | Arun Tejasvi Chaganty, Megan Leszczynski, Shu Zhang, Ravi Ganti,
Krisztian Balog, Filip Radlinski | Beyond Single Items: Exploring User Preferences in Item Sets with the
Conversational Playlist Curation Dataset | Appearing in Proceedings of the 46th International ACM SIGIR
Conference on Research and Development in Information Retrieval | null | 10.1145/3539618.3591881 | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Users in consumption domains, like music, are often able to more efficiently
provide preferences over a set of items (e.g. a playlist or radio) than over
single items (e.g. songs). Unfortunately, this is an underexplored area of
research, with most existing recommendation systems limited to understanding
preferences over single items. Curating an item set exponentiates the search
space that recommender systems must consider (all subsets of items!): this
motivates conversational approaches-where users explicitly state or refine
their preferences and systems elicit preferences in natural language-as an
efficient way to understand user needs. We call this task conversational item
set curation and present a novel data collection methodology that efficiently
collects realistic preferences about item sets in a conversational setting by
observing both item-level and set-level feedback. We apply this methodology to
music recommendation to build the Conversational Playlist Curation Dataset
(CPCD), where we show that it leads raters to express preferences that would
not be otherwise expressed. Finally, we propose a wide range of conversational
retrieval models as baselines for this task and evaluate them on the dataset.
| [
{
"version": "v1",
"created": "Mon, 13 Mar 2023 00:39:04 GMT"
},
{
"version": "v2",
"created": "Fri, 5 May 2023 20:31:45 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Chaganty",
"Arun Tejasvi",
""
],
[
"Leszczynski",
"Megan",
""
],
[
"Zhang",
"Shu",
""
],
[
"Ganti",
"Ravi",
""
],
[
"Balog",
"Krisztian",
""
],
[
"Radlinski",
"Filip",
""
]
] | TITLE: Beyond Single Items: Exploring User Preferences in Item Sets with the
Conversational Playlist Curation Dataset
ABSTRACT: Users in consumption domains, like music, are often able to more efficiently
provide preferences over a set of items (e.g. a playlist or radio) than over
single items (e.g. songs). Unfortunately, this is an underexplored area of
research, with most existing recommendation systems limited to understanding
preferences over single items. Curating an item set exponentiates the search
space that recommender systems must consider (all subsets of items!): this
motivates conversational approaches-where users explicitly state or refine
their preferences and systems elicit preferences in natural language-as an
efficient way to understand user needs. We call this task conversational item
set curation and present a novel data collection methodology that efficiently
collects realistic preferences about item sets in a conversational setting by
observing both item-level and set-level feedback. We apply this methodology to
music recommendation to build the Conversational Playlist Curation Dataset
(CPCD), where we show that it leads raters to express preferences that would
not be otherwise expressed. Finally, we propose a wide range of conversational
retrieval models as baselines for this task and evaluate them on the dataset.
| new_dataset | 0.543833 |
2303.08475 | Runyang Feng | Runyang Feng, Yixing Gao, Xueqing Ma, Tze Ho Elden Tse, Hyung Jin
Chang | Mutual Information-Based Temporal Difference Learning for Human Pose
Estimation in Video | This paper is accepted to CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Temporal modeling is crucial for multi-frame human pose estimation. Most
existing methods directly employ optical flow or deformable convolution to
predict full-spectrum motion fields, which might incur numerous irrelevant
cues, such as a nearby person or background. Without further efforts to
excavate meaningful motion priors, their results are suboptimal, especially in
complicated spatiotemporal interactions. On the other hand, the temporal
difference has the ability to encode representative motion information which
can potentially be valuable for pose estimation but has not been fully
exploited. In this paper, we present a novel multi-frame human pose estimation
framework, which employs temporal differences across frames to model dynamic
contexts and engages mutual information objectively to facilitate useful motion
information disentanglement. To be specific, we design a multi-stage Temporal
Difference Encoder that performs incremental cascaded learning conditioned on
multi-stage feature difference sequences to derive informative motion
representation. We further propose a Representation Disentanglement module from
the mutual information perspective, which can grasp discriminative
task-relevant motion signals by explicitly defining useful and noisy
constituents of the raw motion features and minimizing their mutual
information. These place us to rank No.1 in the Crowd Pose Estimation in
Complex Events Challenge on benchmark dataset HiEve, and achieve
state-of-the-art performance on three benchmarks PoseTrack2017, PoseTrack2018,
and PoseTrack21.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 09:29:03 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 13:43:41 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Feng",
"Runyang",
""
],
[
"Gao",
"Yixing",
""
],
[
"Ma",
"Xueqing",
""
],
[
"Tse",
"Tze Ho Elden",
""
],
[
"Chang",
"Hyung Jin",
""
]
] | TITLE: Mutual Information-Based Temporal Difference Learning for Human Pose
Estimation in Video
ABSTRACT: Temporal modeling is crucial for multi-frame human pose estimation. Most
existing methods directly employ optical flow or deformable convolution to
predict full-spectrum motion fields, which might incur numerous irrelevant
cues, such as a nearby person or background. Without further efforts to
excavate meaningful motion priors, their results are suboptimal, especially in
complicated spatiotemporal interactions. On the other hand, the temporal
difference has the ability to encode representative motion information which
can potentially be valuable for pose estimation but has not been fully
exploited. In this paper, we present a novel multi-frame human pose estimation
framework, which employs temporal differences across frames to model dynamic
contexts and engages mutual information objectively to facilitate useful motion
information disentanglement. To be specific, we design a multi-stage Temporal
Difference Encoder that performs incremental cascaded learning conditioned on
multi-stage feature difference sequences to derive informative motion
representation. We further propose a Representation Disentanglement module from
the mutual information perspective, which can grasp discriminative
task-relevant motion signals by explicitly defining useful and noisy
constituents of the raw motion features and minimizing their mutual
information. These place us to rank No.1 in the Crowd Pose Estimation in
Complex Events Challenge on benchmark dataset HiEve, and achieve
state-of-the-art performance on three benchmarks PoseTrack2017, PoseTrack2018,
and PoseTrack21.
| no_new_dataset | 0.707859 |
2303.08610 | Sungho Lee | Sungho Lee, Jaehyun Park, Seungryeol Paik, and Kyogu Lee | Blind Estimation of Audio Processing Graph | Accepted to ICASSP 2023 | null | null | null | cs.SD eess.AS | http://creativecommons.org/licenses/by-sa/4.0/ | Musicians and audio engineers sculpt and transform their sounds by connecting
multiple processors, forming an audio processing graph. However, most
deep-learning methods overlook this real-world practice and assume fixed graph
settings. To bridge this gap, we develop a system that reconstructs the entire
graph from a given reference audio. We first generate a realistic
graph-reference pair dataset and train a simple blind estimation system
composed of a convolutional reference encoder and a transformer-based graph
decoder. We apply our model to singing voice effects and drum mixing estimation
tasks. Evaluation results show that our method can reconstruct complex signal
routings, including multi-band processing and sidechaining.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 13:31:56 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 00:48:32 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Lee",
"Sungho",
""
],
[
"Park",
"Jaehyun",
""
],
[
"Paik",
"Seungryeol",
""
],
[
"Lee",
"Kyogu",
""
]
] | TITLE: Blind Estimation of Audio Processing Graph
ABSTRACT: Musicians and audio engineers sculpt and transform their sounds by connecting
multiple processors, forming an audio processing graph. However, most
deep-learning methods overlook this real-world practice and assume fixed graph
settings. To bridge this gap, we develop a system that reconstructs the entire
graph from a given reference audio. We first generate a realistic
graph-reference pair dataset and train a simple blind estimation system
composed of a convolutional reference encoder and a transformer-based graph
decoder. We apply our model to singing voice effects and drum mixing estimation
tasks. Evaluation results show that our method can reconstruct complex signal
routings, including multi-band processing and sidechaining.
| new_dataset | 0.71287 |
2303.09266 | Yun Zhu | Boren Hu, Yun Zhu, Jiacheng Li, Siliang Tang | SmartBERT: A Promotion of Dynamic Early Exiting Mechanism for
Accelerating BERT Inference | Accepted by IJCAI2023 | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dynamic early exiting has been proven to improve the inference speed of the
pre-trained language model like BERT. However, all samples must go through all
consecutive layers before early exiting and more complex samples usually go
through more layers, which still exists redundant computation. In this paper,
we propose a novel dynamic early exiting combined with layer skipping for BERT
inference named SmartBERT, which adds a skipping gate and an exiting operator
into each layer of BERT. SmartBERT can adaptively skip some layers and
adaptively choose whether to exit. Besides, we propose cross-layer contrastive
learning and combine it into our training phases to boost the intermediate
layers and classifiers which would be beneficial for early exiting. To keep the
consistent usage of skipping gates between training and inference phases, we
propose a hard weight mechanism during training phase. We conduct experiments
on eight classification datasets of the GLUE benchmark. Experimental results
show that SmartBERT achieves 2-3x computation reduction with minimal accuracy
drops compared with BERT and our method outperforms previous methods in both
efficiency and accuracy. Moreover, in some complex datasets like RTE and WNLI,
we prove that the early exiting based on entropy hardly works, and the skipping
mechanism is essential for reducing computation.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 12:44:16 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 13:05:00 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Hu",
"Boren",
""
],
[
"Zhu",
"Yun",
""
],
[
"Li",
"Jiacheng",
""
],
[
"Tang",
"Siliang",
""
]
] | TITLE: SmartBERT: A Promotion of Dynamic Early Exiting Mechanism for
Accelerating BERT Inference
ABSTRACT: Dynamic early exiting has been proven to improve the inference speed of the
pre-trained language model like BERT. However, all samples must go through all
consecutive layers before early exiting and more complex samples usually go
through more layers, which still exists redundant computation. In this paper,
we propose a novel dynamic early exiting combined with layer skipping for BERT
inference named SmartBERT, which adds a skipping gate and an exiting operator
into each layer of BERT. SmartBERT can adaptively skip some layers and
adaptively choose whether to exit. Besides, we propose cross-layer contrastive
learning and combine it into our training phases to boost the intermediate
layers and classifiers which would be beneficial for early exiting. To keep the
consistent usage of skipping gates between training and inference phases, we
propose a hard weight mechanism during training phase. We conduct experiments
on eight classification datasets of the GLUE benchmark. Experimental results
show that SmartBERT achieves 2-3x computation reduction with minimal accuracy
drops compared with BERT and our method outperforms previous methods in both
efficiency and accuracy. Moreover, in some complex datasets like RTE and WNLI,
we prove that the early exiting based on entropy hardly works, and the skipping
mechanism is essential for reducing computation.
| no_new_dataset | 0.711437 |
2303.09859 | David Samuel | David Samuel, Andrey Kutuzov, Lilja {\O}vrelid and Erik Velldal | Trained on 100 million words and still in shape: BERT meets British
National Corpus | Accepted to EACL 2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While modern masked language models (LMs) are trained on ever larger corpora,
we here explore the effects of down-scaling training to a modestly-sized but
representative, well-balanced, and publicly available English text source --
the British National Corpus. We show that pre-training on this carefully
curated corpus can reach better performance than the original BERT model. We
argue that this type of corpora has great potential as a language modeling
benchmark. To showcase this potential, we present fair, reproducible and
data-efficient comparative studies of LMs, in which we evaluate several
training objectives and model architectures and replicate previous empirical
results in a systematic way. We propose an optimized LM architecture called
LTG-BERT.
| [
{
"version": "v1",
"created": "Fri, 17 Mar 2023 09:53:33 GMT"
},
{
"version": "v2",
"created": "Wed, 29 Mar 2023 09:00:21 GMT"
},
{
"version": "v3",
"created": "Fri, 5 May 2023 23:59:35 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Samuel",
"David",
""
],
[
"Kutuzov",
"Andrey",
""
],
[
"Øvrelid",
"Lilja",
""
],
[
"Velldal",
"Erik",
""
]
] | TITLE: Trained on 100 million words and still in shape: BERT meets British
National Corpus
ABSTRACT: While modern masked language models (LMs) are trained on ever larger corpora,
we here explore the effects of down-scaling training to a modestly-sized but
representative, well-balanced, and publicly available English text source --
the British National Corpus. We show that pre-training on this carefully
curated corpus can reach better performance than the original BERT model. We
argue that this type of corpora has great potential as a language modeling
benchmark. To showcase this potential, we present fair, reproducible and
data-efficient comparative studies of LMs, in which we evaluate several
training objectives and model architectures and replicate previous empirical
results in a systematic way. We propose an optimized LM architecture called
LTG-BERT.
| no_new_dataset | 0.703371 |
2303.11637 | Yang Shen | Yang Shen and Xuhao Sun and Xiu-Shen Wei | Equiangular Basis Vectors | CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose Equiangular Basis Vectors (EBVs) for classification tasks. In deep
neural networks, models usually end with a k-way fully connected layer with
softmax to handle different classification tasks. The learning objective of
these methods can be summarized as mapping the learned feature representations
to the samples' label space. While in metric learning approaches, the main
objective is to learn a transformation function that maps training data points
from the original space to a new space where similar points are closer while
dissimilar points become farther apart. Different from previous methods, our
EBVs generate normalized vector embeddings as "predefined classifiers" which
are required to not only be with the equal status between each other, but also
be as orthogonal as possible. By minimizing the spherical distance of the
embedding of an input between its categorical EBV in training, the predictions
can be obtained by identifying the categorical EBV with the smallest distance
during inference. Various experiments on the ImageNet-1K dataset and other
downstream tasks demonstrate that our method outperforms the general fully
connected classifier while it does not introduce huge additional computation
compared with classical metric learning methods. Our EBVs won the first place
in the 2022 DIGIX Global AI Challenge, and our code is open-source and
available at https://github.com/NJUST-VIPGroup/Equiangular-Basis-Vectors.
| [
{
"version": "v1",
"created": "Tue, 21 Mar 2023 07:08:51 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 06:25:15 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Shen",
"Yang",
""
],
[
"Sun",
"Xuhao",
""
],
[
"Wei",
"Xiu-Shen",
""
]
] | TITLE: Equiangular Basis Vectors
ABSTRACT: We propose Equiangular Basis Vectors (EBVs) for classification tasks. In deep
neural networks, models usually end with a k-way fully connected layer with
softmax to handle different classification tasks. The learning objective of
these methods can be summarized as mapping the learned feature representations
to the samples' label space. While in metric learning approaches, the main
objective is to learn a transformation function that maps training data points
from the original space to a new space where similar points are closer while
dissimilar points become farther apart. Different from previous methods, our
EBVs generate normalized vector embeddings as "predefined classifiers" which
are required to not only be with the equal status between each other, but also
be as orthogonal as possible. By minimizing the spherical distance of the
embedding of an input between its categorical EBV in training, the predictions
can be obtained by identifying the categorical EBV with the smallest distance
during inference. Various experiments on the ImageNet-1K dataset and other
downstream tasks demonstrate that our method outperforms the general fully
connected classifier while it does not introduce huge additional computation
compared with classical metric learning methods. Our EBVs won the first place
in the 2022 DIGIX Global AI Challenge, and our code is open-source and
available at https://github.com/NJUST-VIPGroup/Equiangular-Basis-Vectors.
| no_new_dataset | 0.709636 |
2303.15708 | Hanjia Lyu | Jinsheng Pan, Weihong Qi, Zichen Wang, Hanjia Lyu, Jiebo Luo | Bias or Diversity? Unraveling Fine-Grained Thematic Discrepancy in U.S.
News Headlines | Accepted for publication in Proceedings of the Workshop on News Media
and Computational Journalism (MEDIATE), AAAI International Conference on Web
and Social Media (ICWSM), 2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There is a broad consensus that news media outlets incorporate ideological
biases in their news articles. However, prior studies on measuring the
discrepancies among media outlets and further dissecting the origins of
thematic differences suffer from small sample sizes and limited scope and
granularity. In this study, we use a large dataset of 1.8 million news
headlines from major U.S. media outlets spanning from 2014 to 2022 to
thoroughly track and dissect the fine-grained thematic discrepancy in U.S. news
media. We employ multiple correspondence analysis (MCA) to quantify the
fine-grained thematic discrepancy related to four prominent topics - domestic
politics, economic issues, social issues, and foreign affairs in order to
derive a more holistic analysis. Additionally, we compare the most frequent
$n$-grams in media headlines to provide further qualitative insights into our
analysis. Our findings indicate that on domestic politics and social issues,
the discrepancy can be attributed to a certain degree of media bias. Meanwhile,
the discrepancy in reporting foreign affairs is largely attributed to the
diversity in individual journalistic styles. Finally, U.S. media outlets show
consistency and high similarity in their coverage of economic issues.
| [
{
"version": "v1",
"created": "Tue, 28 Mar 2023 03:31:37 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 03:57:30 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Pan",
"Jinsheng",
""
],
[
"Qi",
"Weihong",
""
],
[
"Wang",
"Zichen",
""
],
[
"Lyu",
"Hanjia",
""
],
[
"Luo",
"Jiebo",
""
]
] | TITLE: Bias or Diversity? Unraveling Fine-Grained Thematic Discrepancy in U.S.
News Headlines
ABSTRACT: There is a broad consensus that news media outlets incorporate ideological
biases in their news articles. However, prior studies on measuring the
discrepancies among media outlets and further dissecting the origins of
thematic differences suffer from small sample sizes and limited scope and
granularity. In this study, we use a large dataset of 1.8 million news
headlines from major U.S. media outlets spanning from 2014 to 2022 to
thoroughly track and dissect the fine-grained thematic discrepancy in U.S. news
media. We employ multiple correspondence analysis (MCA) to quantify the
fine-grained thematic discrepancy related to four prominent topics - domestic
politics, economic issues, social issues, and foreign affairs in order to
derive a more holistic analysis. Additionally, we compare the most frequent
$n$-grams in media headlines to provide further qualitative insights into our
analysis. Our findings indicate that on domestic politics and social issues,
the discrepancy can be attributed to a certain degree of media bias. Meanwhile,
the discrepancy in reporting foreign affairs is largely attributed to the
diversity in individual journalistic styles. Finally, U.S. media outlets show
consistency and high similarity in their coverage of economic issues.
| no_new_dataset | 0.694303 |
2304.01238 | Maxime Labonne | Maxime Labonne and Sean Moran | Spam-T5: Benchmarking Large Language Models for Few-Shot Email Spam
Detection | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This paper investigates the effectiveness of large language models (LLMs) in
email spam detection by comparing prominent models from three distinct
families: BERT-like, Sentence Transformers, and Seq2Seq. Additionally, we
examine well-established machine learning techniques for spam detection, such
as Na\"ive Bayes and LightGBM, as baseline methods. We assess the performance
of these models across four public datasets, utilizing different numbers of
training samples (full training set and few-shot settings). Our findings reveal
that, in the majority of cases, LLMs surpass the performance of the popular
baseline techniques, particularly in few-shot scenarios. This adaptability
renders LLMs uniquely suited to spam detection tasks, where labeled samples are
limited in number and models require frequent updates. Additionally, we
introduce Spam-T5, a Flan-T5 model that has been specifically adapted and
fine-tuned for the purpose of detecting email spam. Our results demonstrate
that Spam-T5 surpasses baseline models and other LLMs in the majority of
scenarios, particularly when there are a limited number of training samples
available. Our code is publicly available at
https://github.com/jpmorganchase/emailspamdetection.
| [
{
"version": "v1",
"created": "Mon, 3 Apr 2023 10:27:53 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Apr 2023 13:38:54 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 10:57:51 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Labonne",
"Maxime",
""
],
[
"Moran",
"Sean",
""
]
] | TITLE: Spam-T5: Benchmarking Large Language Models for Few-Shot Email Spam
Detection
ABSTRACT: This paper investigates the effectiveness of large language models (LLMs) in
email spam detection by comparing prominent models from three distinct
families: BERT-like, Sentence Transformers, and Seq2Seq. Additionally, we
examine well-established machine learning techniques for spam detection, such
as Na\"ive Bayes and LightGBM, as baseline methods. We assess the performance
of these models across four public datasets, utilizing different numbers of
training samples (full training set and few-shot settings). Our findings reveal
that, in the majority of cases, LLMs surpass the performance of the popular
baseline techniques, particularly in few-shot scenarios. This adaptability
renders LLMs uniquely suited to spam detection tasks, where labeled samples are
limited in number and models require frequent updates. Additionally, we
introduce Spam-T5, a Flan-T5 model that has been specifically adapted and
fine-tuned for the purpose of detecting email spam. Our results demonstrate
that Spam-T5 surpasses baseline models and other LLMs in the majority of
scenarios, particularly when there are a limited number of training samples
available. Our code is publicly available at
https://github.com/jpmorganchase/emailspamdetection.
| no_new_dataset | 0.710051 |
2304.03144 | Jun Wu | Jun Wu, Xuesong Ye, and Yanyuet Man | BotTriNet: A Unified and Efficient Embedding for Social Bots Detection
via Metric Learning | null | null | null | null | cs.AI cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid and accurate identification of bot accounts in online social
networks is an ongoing challenge. In this paper, we propose BOTTRINET, a
unified embedding framework that leverages the textual content posted by
accounts to detect bots. Our approach is based on the premise that account
personalities and habits can be revealed through their contextual content. To
achieve this, we designed a triplet network that refines raw embeddings using
metric learning techniques. The BOTTRINET framework produces word, sentence,
and account embeddings, which we evaluate on a real-world dataset, CRESCI2017,
consisting of three bot account categories and five bot sample sets. Our
approach achieves state-of-the-art performance on two content-intensive bot
sets, with an average accuracy of 98.34% and f1score of 97.99%. Moreover, our
method makes a significant breakthrough on four content-less bot sets, with an
average accuracy improvement of 11.52% and an average f1score increase of
16.70%. Our contribution is twofold: First, we propose a unified and effective
framework that combines various embeddings for bot detection. Second, we
demonstrate that metric learning techniques can be applied in this context to
refine raw embeddings and improve classification performance. Our approach
outperforms prior works and sets a new standard for bot detection in social
networks.
| [
{
"version": "v1",
"created": "Thu, 6 Apr 2023 15:28:58 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Apr 2023 04:24:41 GMT"
},
{
"version": "v3",
"created": "Tue, 18 Apr 2023 16:47:05 GMT"
},
{
"version": "v4",
"created": "Sat, 6 May 2023 15:09:36 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Wu",
"Jun",
""
],
[
"Ye",
"Xuesong",
""
],
[
"Man",
"Yanyuet",
""
]
] | TITLE: BotTriNet: A Unified and Efficient Embedding for Social Bots Detection
via Metric Learning
ABSTRACT: The rapid and accurate identification of bot accounts in online social
networks is an ongoing challenge. In this paper, we propose BOTTRINET, a
unified embedding framework that leverages the textual content posted by
accounts to detect bots. Our approach is based on the premise that account
personalities and habits can be revealed through their contextual content. To
achieve this, we designed a triplet network that refines raw embeddings using
metric learning techniques. The BOTTRINET framework produces word, sentence,
and account embeddings, which we evaluate on a real-world dataset, CRESCI2017,
consisting of three bot account categories and five bot sample sets. Our
approach achieves state-of-the-art performance on two content-intensive bot
sets, with an average accuracy of 98.34% and f1score of 97.99%. Moreover, our
method makes a significant breakthrough on four content-less bot sets, with an
average accuracy improvement of 11.52% and an average f1score increase of
16.70%. Our contribution is twofold: First, we propose a unified and effective
framework that combines various embeddings for bot detection. Second, we
demonstrate that metric learning techniques can be applied in this context to
refine raw embeddings and improve classification performance. Our approach
outperforms prior works and sets a new standard for bot detection in social
networks.
| no_new_dataset | 0.710879 |
2304.04507 | Raktim Kumar Mondol | Raktim Kumar Mondol, Ewan K.A. Millar, Peter H Graham, Lois Browne,
Arcot Sowmya, Erik Meijering | hist2RNA: An efficient deep learning architecture to predict gene
expression from breast cancer histopathology images | 16 pages, 10 figures, 2 tables | https://www.mdpi.com/2072-6694/15/9/2569 | 10.3390/cancers15092569 | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Gene expression can be used to subtype breast cancer with improved prediction
of risk of recurrence and treatment responsiveness over that obtained using
routine immunohistochemistry (IHC). However, in the clinic, molecular profiling
is primarily used for ER+ breast cancer, which is costly, tissue destructive,
requires specialized platforms and takes several weeks to obtain a result. Deep
learning algorithms can effectively extract morphological patterns in digital
histopathology images to predict molecular phenotypes quickly and
cost-effectively. We propose a new, computationally efficient approach called
hist2RNA inspired by bulk RNA-sequencing techniques to predict the expression
of 138 genes (incorporated from six commercially available molecular profiling
tests), including luminal PAM50 subtype, from hematoxylin and eosin (H&E)
stained whole slide images (WSIs). The training phase involves the aggregation
of extracted features for each patient from a pretrained model to predict gene
expression at the patient level using annotated H&E images from The Cancer
Genome Atlas (TCGA, n=335). We demonstrate successful gene prediction on a
held-out test set (n = 160, corr = 0.82 across patients, corr = 0.29 across
genes) and perform exploratory analysis on an external tissue microarray (TMA)
dataset (n = 498) with known IHC and survival information. Our model is able to
predict gene expression and luminal PAM50 subtype (Luminal A versus Luminal B)
on the TMA dataset with prognostic significance for overall survival in
univariate analysis (c-index = 0.56, hazard ratio = 2.16 (95% CI 1.12-3.06), p
< 5 x 10-3), and independent significance in multivariate analysis
incorporating standard clinicopathological variables (c-index = 0.65, hazard
ratio = 1.85 (95% CI 1.30-2.68), p < 5 x 10-3).
| [
{
"version": "v1",
"created": "Mon, 10 Apr 2023 10:54:32 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 15:35:48 GMT"
},
{
"version": "v3",
"created": "Tue, 2 May 2023 16:33:35 GMT"
},
{
"version": "v4",
"created": "Sun, 7 May 2023 06:26:25 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Mondol",
"Raktim Kumar",
""
],
[
"Millar",
"Ewan K. A.",
""
],
[
"Graham",
"Peter H",
""
],
[
"Browne",
"Lois",
""
],
[
"Sowmya",
"Arcot",
""
],
[
"Meijering",
"Erik",
""
]
] | TITLE: hist2RNA: An efficient deep learning architecture to predict gene
expression from breast cancer histopathology images
ABSTRACT: Gene expression can be used to subtype breast cancer with improved prediction
of risk of recurrence and treatment responsiveness over that obtained using
routine immunohistochemistry (IHC). However, in the clinic, molecular profiling
is primarily used for ER+ breast cancer, which is costly, tissue destructive,
requires specialized platforms and takes several weeks to obtain a result. Deep
learning algorithms can effectively extract morphological patterns in digital
histopathology images to predict molecular phenotypes quickly and
cost-effectively. We propose a new, computationally efficient approach called
hist2RNA inspired by bulk RNA-sequencing techniques to predict the expression
of 138 genes (incorporated from six commercially available molecular profiling
tests), including luminal PAM50 subtype, from hematoxylin and eosin (H&E)
stained whole slide images (WSIs). The training phase involves the aggregation
of extracted features for each patient from a pretrained model to predict gene
expression at the patient level using annotated H&E images from The Cancer
Genome Atlas (TCGA, n=335). We demonstrate successful gene prediction on a
held-out test set (n = 160, corr = 0.82 across patients, corr = 0.29 across
genes) and perform exploratory analysis on an external tissue microarray (TMA)
dataset (n = 498) with known IHC and survival information. Our model is able to
predict gene expression and luminal PAM50 subtype (Luminal A versus Luminal B)
on the TMA dataset with prognostic significance for overall survival in
univariate analysis (c-index = 0.56, hazard ratio = 2.16 (95% CI 1.12-3.06), p
< 5 x 10-3), and independent significance in multivariate analysis
incorporating standard clinicopathological variables (c-index = 0.65, hazard
ratio = 1.85 (95% CI 1.30-2.68), p < 5 x 10-3).
| no_new_dataset | 0.714061 |
2304.04811 | Yida Mu | Yida Mu, Ye Jiang, Freddy Heppell, Iknoor Singh, Carolina Scarton,
Kalina Bontcheva, Xingyi Song | A Large-Scale Comparative Study of Accurate COVID-19 Information versus
Misinformation | ICWSM TrueHealth 2023 | null | null | null | cs.CL | http://creativecommons.org/publicdomain/zero/1.0/ | The COVID-19 pandemic led to an infodemic where an overwhelming amount of
COVID-19 related content was being disseminated at high velocity through social
media. This made it challenging for citizens to differentiate between accurate
and inaccurate information about COVID-19. This motivated us to carry out a
comparative study of the characteristics of COVID-19 misinformation versus
those of accurate COVID-19 information through a large-scale computational
analysis of over 242 million tweets. The study makes comparisons alongside four
key aspects: 1) the distribution of topics, 2) the live status of tweets, 3)
language analysis and 4) the spreading power over time. An added contribution
of this study is the creation of a COVID-19 misinformation classification
dataset. Finally, we demonstrate that this new dataset helps improve
misinformation classification by more than 9\% based on average F1 measure.
| [
{
"version": "v1",
"created": "Mon, 10 Apr 2023 18:44:41 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 13:55:02 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Mu",
"Yida",
""
],
[
"Jiang",
"Ye",
""
],
[
"Heppell",
"Freddy",
""
],
[
"Singh",
"Iknoor",
""
],
[
"Scarton",
"Carolina",
""
],
[
"Bontcheva",
"Kalina",
""
],
[
"Song",
"Xingyi",
""
]
] | TITLE: A Large-Scale Comparative Study of Accurate COVID-19 Information versus
Misinformation
ABSTRACT: The COVID-19 pandemic led to an infodemic where an overwhelming amount of
COVID-19 related content was being disseminated at high velocity through social
media. This made it challenging for citizens to differentiate between accurate
and inaccurate information about COVID-19. This motivated us to carry out a
comparative study of the characteristics of COVID-19 misinformation versus
those of accurate COVID-19 information through a large-scale computational
analysis of over 242 million tweets. The study makes comparisons alongside four
key aspects: 1) the distribution of topics, 2) the live status of tweets, 3)
language analysis and 4) the spreading power over time. An added contribution
of this study is the creation of a COVID-19 misinformation classification
dataset. Finally, we demonstrate that this new dataset helps improve
misinformation classification by more than 9\% based on average F1 measure.
| new_dataset | 0.713432 |
2304.08682 | Aisha Urooj | Aisha Urooj Khan, Hilde Kuehne, Bo Wu, Kim Chheu, Walid Bousselham,
Chuang Gan, Niels Lobo, Mubarak Shah | Learning Situation Hyper-Graphs for Video Question Answering | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Answering questions about complex situations in videos requires not only
capturing the presence of actors, objects, and their relations but also the
evolution of these relationships over time. A situation hyper-graph is a
representation that describes situations as scene sub-graphs for video frames
and hyper-edges for connected sub-graphs and has been proposed to capture all
such information in a compact structured form. In this work, we propose an
architecture for Video Question Answering (VQA) that enables answering
questions related to video content by predicting situation hyper-graphs, coined
Situation Hyper-Graph based Video Question Answering (SHG-VQA). To this end, we
train a situation hyper-graph decoder to implicitly identify graph
representations with actions and object/human-object relationships from the
input video clip. and to use cross-attention between the predicted situation
hyper-graphs and the question embedding to predict the correct answer. The
proposed method is trained in an end-to-end manner and optimized by a VQA loss
with the cross-entropy function and a Hungarian matching loss for the situation
graph prediction. The effectiveness of the proposed architecture is extensively
evaluated on two challenging benchmarks: AGQA and STAR. Our results show that
learning the underlying situation hyper-graphs helps the system to
significantly improve its performance for novel challenges of video
question-answering tasks.
| [
{
"version": "v1",
"created": "Tue, 18 Apr 2023 01:23:11 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 06:44:56 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Khan",
"Aisha Urooj",
""
],
[
"Kuehne",
"Hilde",
""
],
[
"Wu",
"Bo",
""
],
[
"Chheu",
"Kim",
""
],
[
"Bousselham",
"Walid",
""
],
[
"Gan",
"Chuang",
""
],
[
"Lobo",
"Niels",
""
],
[
"Shah",
"Mubarak",
""
]
] | TITLE: Learning Situation Hyper-Graphs for Video Question Answering
ABSTRACT: Answering questions about complex situations in videos requires not only
capturing the presence of actors, objects, and their relations but also the
evolution of these relationships over time. A situation hyper-graph is a
representation that describes situations as scene sub-graphs for video frames
and hyper-edges for connected sub-graphs and has been proposed to capture all
such information in a compact structured form. In this work, we propose an
architecture for Video Question Answering (VQA) that enables answering
questions related to video content by predicting situation hyper-graphs, coined
Situation Hyper-Graph based Video Question Answering (SHG-VQA). To this end, we
train a situation hyper-graph decoder to implicitly identify graph
representations with actions and object/human-object relationships from the
input video clip. and to use cross-attention between the predicted situation
hyper-graphs and the question embedding to predict the correct answer. The
proposed method is trained in an end-to-end manner and optimized by a VQA loss
with the cross-entropy function and a Hungarian matching loss for the situation
graph prediction. The effectiveness of the proposed architecture is extensively
evaluated on two challenging benchmarks: AGQA and STAR. Our results show that
learning the underlying situation hyper-graphs helps the system to
significantly improve its performance for novel challenges of video
question-answering tasks.
| no_new_dataset | 0.710025 |
2304.09324 | Sheng He | Sheng He, Rina Bao, Jingpeng Li, Jeffrey Stout, Atle Bjornerud, P.
Ellen Grant, Yangming Ou | Computer-Vision Benchmark Segment-Anything Model (SAM) in Medical
Images: Accuracy in 12 Datasets | Technical Report | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Background: The segment-anything model (SAM), introduced in April 2023, shows
promise as a benchmark model and a universal solution to segment various
natural images. It comes without previously-required re-training or fine-tuning
specific to each new dataset.
Purpose: To test SAM's accuracy in various medical image segmentation tasks
and investigate potential factors that may affect its accuracy in medical
images.
Methods: SAM was tested on 12 public medical image segmentation datasets
involving 7,451 subjects. The accuracy was measured by the Dice overlap between
the algorithm-segmented and ground-truth masks. SAM was compared with five
state-of-the-art algorithms specifically designed for medical image
segmentation tasks. Associations of SAM's accuracy with six factors were
computed, independently and jointly, including segmentation difficulties as
measured by segmentation ability score and by Dice overlap in U-Net, image
dimension, size of the target region, image modality, and contrast.
Results: The Dice overlaps from SAM were significantly lower than the five
medical-image-based algorithms in all 12 medical image segmentation datasets,
by a margin of 0.1-0.5 and even 0.6-0.7 Dice. SAM-Semantic was significantly
associated with medical image segmentation difficulty and the image modality,
and SAM-Point and SAM-Box were significantly associated with image segmentation
difficulty, image dimension, target region size, and target-vs-background
contrast. All these 3 variations of SAM were more accurate in 2D medical
images, larger target region sizes, easier cases with a higher Segmentation
Ability score and higher U-Net Dice, and higher foreground-background contrast.
| [
{
"version": "v1",
"created": "Tue, 18 Apr 2023 22:16:49 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Apr 2023 22:41:34 GMT"
},
{
"version": "v3",
"created": "Fri, 5 May 2023 18:58:52 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"He",
"Sheng",
""
],
[
"Bao",
"Rina",
""
],
[
"Li",
"Jingpeng",
""
],
[
"Stout",
"Jeffrey",
""
],
[
"Bjornerud",
"Atle",
""
],
[
"Grant",
"P. Ellen",
""
],
[
"Ou",
"Yangming",
""
]
] | TITLE: Computer-Vision Benchmark Segment-Anything Model (SAM) in Medical
Images: Accuracy in 12 Datasets
ABSTRACT: Background: The segment-anything model (SAM), introduced in April 2023, shows
promise as a benchmark model and a universal solution to segment various
natural images. It comes without previously-required re-training or fine-tuning
specific to each new dataset.
Purpose: To test SAM's accuracy in various medical image segmentation tasks
and investigate potential factors that may affect its accuracy in medical
images.
Methods: SAM was tested on 12 public medical image segmentation datasets
involving 7,451 subjects. The accuracy was measured by the Dice overlap between
the algorithm-segmented and ground-truth masks. SAM was compared with five
state-of-the-art algorithms specifically designed for medical image
segmentation tasks. Associations of SAM's accuracy with six factors were
computed, independently and jointly, including segmentation difficulties as
measured by segmentation ability score and by Dice overlap in U-Net, image
dimension, size of the target region, image modality, and contrast.
Results: The Dice overlaps from SAM were significantly lower than the five
medical-image-based algorithms in all 12 medical image segmentation datasets,
by a margin of 0.1-0.5 and even 0.6-0.7 Dice. SAM-Semantic was significantly
associated with medical image segmentation difficulty and the image modality,
and SAM-Point and SAM-Box were significantly associated with image segmentation
difficulty, image dimension, target region size, and target-vs-background
contrast. All these 3 variations of SAM were more accurate in 2D medical
images, larger target region sizes, easier cases with a higher Segmentation
Ability score and higher U-Net Dice, and higher foreground-background contrast.
| no_new_dataset | 0.71103 |
2304.11702 | Jiawei Xiong | Jiawei Xiong, Olga Fink, Jian Zhou, Yizhong Ma | Controlled physics-informed data generation for deep learning-based
remaining useful life prediction under unseen operation conditions | 22 pages,12 figures | Mechanical Systems and Signal Processing, 197, 110359 (2023) | 10.1016/j.ymssp.2023.110359 | null | cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Limited availability of representative time-to-failure (TTF) trajectories
either limits the performance of deep learning (DL)-based approaches on
remaining useful life (RUL) prediction in practice or even precludes their
application. Generating synthetic data that is physically plausible is a
promising way to tackle this challenge. In this study, a novel hybrid framework
combining the controlled physics-informed data generation approach with a deep
learning-based prediction model for prognostics is proposed. In the proposed
framework, a new controlled physics-informed generative adversarial network
(CPI-GAN) is developed to generate synthetic degradation trajectories that are
physically interpretable and diverse. Five basic physics constraints are
proposed as the controllable settings in the generator. A physics-informed loss
function with penalty is designed as the regularization term, which ensures
that the changing trend of system health state recorded in the synthetic data
is consistent with the underlying physical laws. Then, the generated synthetic
data is used as input of the DL-based prediction model to obtain the RUL
estimations. The proposed framework is evaluated based on new Commercial
Modular Aero-Propulsion System Simulation (N-CMAPSS), a turbofan engine
prognostics dataset where a limited avail-ability of TTF trajectories is
assumed. The experimental results demonstrate that the proposed framework is
able to generate synthetic TTF trajectories that are consistent with underlying
degradation trends. The generated trajectories enable to significantly improve
the accuracy of RUL predictions.
| [
{
"version": "v1",
"created": "Sun, 23 Apr 2023 17:34:26 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 09:22:46 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Xiong",
"Jiawei",
""
],
[
"Fink",
"Olga",
""
],
[
"Zhou",
"Jian",
""
],
[
"Ma",
"Yizhong",
""
]
] | TITLE: Controlled physics-informed data generation for deep learning-based
remaining useful life prediction under unseen operation conditions
ABSTRACT: Limited availability of representative time-to-failure (TTF) trajectories
either limits the performance of deep learning (DL)-based approaches on
remaining useful life (RUL) prediction in practice or even precludes their
application. Generating synthetic data that is physically plausible is a
promising way to tackle this challenge. In this study, a novel hybrid framework
combining the controlled physics-informed data generation approach with a deep
learning-based prediction model for prognostics is proposed. In the proposed
framework, a new controlled physics-informed generative adversarial network
(CPI-GAN) is developed to generate synthetic degradation trajectories that are
physically interpretable and diverse. Five basic physics constraints are
proposed as the controllable settings in the generator. A physics-informed loss
function with penalty is designed as the regularization term, which ensures
that the changing trend of system health state recorded in the synthetic data
is consistent with the underlying physical laws. Then, the generated synthetic
data is used as input of the DL-based prediction model to obtain the RUL
estimations. The proposed framework is evaluated based on new Commercial
Modular Aero-Propulsion System Simulation (N-CMAPSS), a turbofan engine
prognostics dataset where a limited avail-ability of TTF trajectories is
assumed. The experimental results demonstrate that the proposed framework is
able to generate synthetic TTF trajectories that are consistent with underlying
degradation trends. The generated trajectories enable to significantly improve
the accuracy of RUL predictions.
| no_new_dataset | 0.71022 |
2304.13833 | Yuji Saikai | Yuji Saikai and Khue-Dung Dang | Mixtures of Gaussian process experts based on kernel stick-breaking
processes | null | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | Mixtures of Gaussian process experts is a class of models that can
simultaneously address two of the key limitations inherent in standard Gaussian
processes: scalability and predictive performance. In particular, models that
use Dirichlet processes as gating functions permit straightforward
interpretation and automatic selection of the number of experts in a mixture.
While the existing models are intuitive and capable of capturing
non-stationarity, multi-modality and heteroskedasticity, the simplicity of
their gating functions may limit the predictive performance when applied to
complex data-generating processes. Capitalising on the recent advancement in
the dependent Dirichlet processes literature, we propose a new mixture model of
Gaussian process experts based on kernel stick-breaking processes. Our model
maintains the intuitive appeal yet improve the performance of the existing
models. To make it practical, we design a sampler for posterior computation
based on the slice sampling. The model behaviour and improved predictive
performance are demonstrated in experiments using six datasets.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 21:23:01 GMT"
},
{
"version": "v2",
"created": "Fri, 5 May 2023 21:00:46 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Saikai",
"Yuji",
""
],
[
"Dang",
"Khue-Dung",
""
]
] | TITLE: Mixtures of Gaussian process experts based on kernel stick-breaking
processes
ABSTRACT: Mixtures of Gaussian process experts is a class of models that can
simultaneously address two of the key limitations inherent in standard Gaussian
processes: scalability and predictive performance. In particular, models that
use Dirichlet processes as gating functions permit straightforward
interpretation and automatic selection of the number of experts in a mixture.
While the existing models are intuitive and capable of capturing
non-stationarity, multi-modality and heteroskedasticity, the simplicity of
their gating functions may limit the predictive performance when applied to
complex data-generating processes. Capitalising on the recent advancement in
the dependent Dirichlet processes literature, we propose a new mixture model of
Gaussian process experts based on kernel stick-breaking processes. Our model
maintains the intuitive appeal yet improve the performance of the existing
models. To make it practical, we design a sampler for posterior computation
based on the slice sampling. The model behaviour and improved predictive
performance are demonstrated in experiments using six datasets.
| no_new_dataset | 0.710641 |
2304.13874 | Paul Owoicho | Paul Owoicho, Ivan Sekuli\'c, Mohammad Aliannejadi, Jeffrey Dalton,
and Fabio Crestani | Exploiting Simulated User Feedback for Conversational Search: Ranking,
Rewriting, and Beyond | 11 pages, 2 figures, to be published in SIGIR 2023 | null | 10.1145/3539618.3591683 | null | cs.IR cs.HC | http://creativecommons.org/licenses/by-sa/4.0/ | This research aims to explore various methods for assessing user feedback in
mixed-initiative conversational search (CS) systems. While CS systems enjoy
profuse advancements across multiple aspects, recent research fails to
successfully incorporate feedback from the users. One of the main reasons for
that is the lack of system-user conversational interaction data. To this end,
we propose a user simulator-based framework for multi-turn interactions with a
variety of mixed-initiative CS systems. Specifically, we develop a user
simulator, dubbed ConvSim, that, once initialized with an information need
description, is capable of providing feedback to a system's responses, as well
as answering potential clarifying questions. Our experiments on a wide variety
of state-of-the-art passage retrieval and neural re-ranking models show that
effective utilization of user feedback can lead to 16% retrieval performance
increase in terms of nDCG@3. Moreover, we observe consistent improvements as
the number of feedback rounds increases (35% relative improvement in terms of
nDCG@3 after three rounds). This points to a research gap in the development of
specific feedback processing modules and opens a potential for significant
advancements in CS. To support further research in the topic, we release over
30,000 transcripts of system-simulator interactions based on well-established
CS datasets.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 23:40:55 GMT"
},
{
"version": "v2",
"created": "Tue, 2 May 2023 21:58:33 GMT"
},
{
"version": "v3",
"created": "Sun, 7 May 2023 09:33:10 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Owoicho",
"Paul",
""
],
[
"Sekulić",
"Ivan",
""
],
[
"Aliannejadi",
"Mohammad",
""
],
[
"Dalton",
"Jeffrey",
""
],
[
"Crestani",
"Fabio",
""
]
] | TITLE: Exploiting Simulated User Feedback for Conversational Search: Ranking,
Rewriting, and Beyond
ABSTRACT: This research aims to explore various methods for assessing user feedback in
mixed-initiative conversational search (CS) systems. While CS systems enjoy
profuse advancements across multiple aspects, recent research fails to
successfully incorporate feedback from the users. One of the main reasons for
that is the lack of system-user conversational interaction data. To this end,
we propose a user simulator-based framework for multi-turn interactions with a
variety of mixed-initiative CS systems. Specifically, we develop a user
simulator, dubbed ConvSim, that, once initialized with an information need
description, is capable of providing feedback to a system's responses, as well
as answering potential clarifying questions. Our experiments on a wide variety
of state-of-the-art passage retrieval and neural re-ranking models show that
effective utilization of user feedback can lead to 16% retrieval performance
increase in terms of nDCG@3. Moreover, we observe consistent improvements as
the number of feedback rounds increases (35% relative improvement in terms of
nDCG@3 after three rounds). This points to a research gap in the development of
specific feedback processing modules and opens a potential for significant
advancements in CS. To support further research in the topic, we release over
30,000 transcripts of system-simulator interactions based on well-established
CS datasets.
| no_new_dataset | 0.707859 |
2305.00082 | Jun Kataoka | Jun Kataoka and Hyunsoo Yoon | AVATAR: Adversarial self-superVised domain Adaptation network for TARget
domain | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This paper presents an unsupervised domain adaptation (UDA) method for
predicting unlabeled target domain data, specific to complex UDA tasks where
the domain gap is significant. Mainstream UDA models aim to learn from both
domains and improve target discrimination by utilizing labeled source domain
data. However, the performance boost may be limited when the discrepancy
between the source and target domains is large or the target domain contains
outliers. To explicitly address this issue, we propose the Adversarial
self-superVised domain Adaptation network for the TARget domain (AVATAR)
algorithm. It outperforms state-of-the-art UDA models by concurrently reducing
domain discrepancy while enhancing discrimination through domain adversarial
learning, self-supervised learning, and sample selection strategy for the
target domain, all guided by deep clustering. Our proposed model significantly
outperforms state-of-the-art methods on three UDA benchmarks, and extensive
ablation studies and experiments demonstrate the effectiveness of our approach
for addressing complex UDA tasks.
| [
{
"version": "v1",
"created": "Fri, 28 Apr 2023 20:31:56 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 03:35:14 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Kataoka",
"Jun",
""
],
[
"Yoon",
"Hyunsoo",
""
]
] | TITLE: AVATAR: Adversarial self-superVised domain Adaptation network for TARget
domain
ABSTRACT: This paper presents an unsupervised domain adaptation (UDA) method for
predicting unlabeled target domain data, specific to complex UDA tasks where
the domain gap is significant. Mainstream UDA models aim to learn from both
domains and improve target discrimination by utilizing labeled source domain
data. However, the performance boost may be limited when the discrepancy
between the source and target domains is large or the target domain contains
outliers. To explicitly address this issue, we propose the Adversarial
self-superVised domain Adaptation network for the TARget domain (AVATAR)
algorithm. It outperforms state-of-the-art UDA models by concurrently reducing
domain discrepancy while enhancing discrimination through domain adversarial
learning, self-supervised learning, and sample selection strategy for the
target domain, all guided by deep clustering. Our proposed model significantly
outperforms state-of-the-art methods on three UDA benchmarks, and extensive
ablation studies and experiments demonstrate the effectiveness of our approach
for addressing complex UDA tasks.
| no_new_dataset | 0.709837 |
2305.00109 | Christian Mattjie | Christian Mattjie and Luis Vinicius de Moura and Rafaela Cappelari
Ravazio and Lucas Silveira Kupssinsk\"u and Ot\'avio Parraga and Marcelo
Mussi Delucis and Rodrigo Coelho Barros | Zero-shot performance of the Segment Anything Model (SAM) in 2D medical
imaging: A comprehensive evaluation and practical guidelines | 18 pages, 3 Tables, 10 Figures with additional supplementary material
with 1 Table | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Segmentation in medical imaging is a critical component for the diagnosis,
monitoring, and treatment of various diseases and medical conditions.
Presently, the medical segmentation landscape is dominated by numerous
specialized deep learning models, each fine-tuned for specific segmentation
tasks and image modalities. The recently-introduced Segment Anything Model
(SAM) employs the ViT neural architecture and harnesses a massive training
dataset to segment nearly any object; however, its suitability to the medical
domain has not yet been investigated. In this study, we explore the zero-shot
performance of SAM in medical imaging by implementing eight distinct prompt
strategies across six datasets from four imaging modalities, including X-ray,
ultrasound, dermatoscopy, and colonoscopy. Our findings reveal that SAM's
zero-shot performance is not only comparable to, but in certain cases,
surpasses the current state-of-the-art. Based on these results, we propose
practical guidelines that require minimal interaction while consistently
yielding robust outcomes across all assessed contexts. The source code, along
with a demonstration of the recommended guidelines, can be accessed at
https://github.com/Malta-Lab/SAM-zero-shot-in-Medical-Imaging.
| [
{
"version": "v1",
"created": "Fri, 28 Apr 2023 22:07:24 GMT"
},
{
"version": "v2",
"created": "Fri, 5 May 2023 19:02:03 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Mattjie",
"Christian",
""
],
[
"de Moura",
"Luis Vinicius",
""
],
[
"Ravazio",
"Rafaela Cappelari",
""
],
[
"Kupssinskü",
"Lucas Silveira",
""
],
[
"Parraga",
"Otávio",
""
],
[
"Delucis",
"Marcelo Mussi",
""
],
[
"Barros",
"Rodrigo Coelho",
""
]
] | TITLE: Zero-shot performance of the Segment Anything Model (SAM) in 2D medical
imaging: A comprehensive evaluation and practical guidelines
ABSTRACT: Segmentation in medical imaging is a critical component for the diagnosis,
monitoring, and treatment of various diseases and medical conditions.
Presently, the medical segmentation landscape is dominated by numerous
specialized deep learning models, each fine-tuned for specific segmentation
tasks and image modalities. The recently-introduced Segment Anything Model
(SAM) employs the ViT neural architecture and harnesses a massive training
dataset to segment nearly any object; however, its suitability to the medical
domain has not yet been investigated. In this study, we explore the zero-shot
performance of SAM in medical imaging by implementing eight distinct prompt
strategies across six datasets from four imaging modalities, including X-ray,
ultrasound, dermatoscopy, and colonoscopy. Our findings reveal that SAM's
zero-shot performance is not only comparable to, but in certain cases,
surpasses the current state-of-the-art. Based on these results, we propose
practical guidelines that require minimal interaction while consistently
yielding robust outcomes across all assessed contexts. The source code, along
with a demonstration of the recommended guidelines, can be accessed at
https://github.com/Malta-Lab/SAM-zero-shot-in-Medical-Imaging.
| no_new_dataset | 0.707234 |
2305.00514 | Li Long | Long Li, Junwei Han, Ni Zhang, Nian Liu, Salman Khan, Hisham
Cholakkal, Rao Muhammad Anwer, and Fahad Shahbaz Khan | Discriminative Co-Saliency and Background Mining Transformer for
Co-Salient Object Detection | Accepted by CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Most previous co-salient object detection works mainly focus on extracting
co-salient cues via mining the consistency relations across images while
ignoring explicit exploration of background regions. In this paper, we propose
a Discriminative co-saliency and background Mining Transformer framework (DMT)
based on several economical multi-grained correlation modules to explicitly
mine both co-saliency and background information and effectively model their
discrimination. Specifically, we first propose a region-to-region correlation
module for introducing inter-image relations to pixel-wise segmentation
features while maintaining computational efficiency. Then, we use two types of
pre-defined tokens to mine co-saliency and background information via our
proposed contrast-induced pixel-to-token correlation and co-saliency
token-to-token correlation modules. We also design a token-guided feature
refinement module to enhance the discriminability of the segmentation features
under the guidance of the learned tokens. We perform iterative mutual promotion
for the segmentation feature extraction and token construction. Experimental
results on three benchmark datasets demonstrate the effectiveness of our
proposed method. The source code is available at:
https://github.com/dragonlee258079/DMT.
| [
{
"version": "v1",
"created": "Sun, 30 Apr 2023 15:56:47 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 01:59:52 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Li",
"Long",
""
],
[
"Han",
"Junwei",
""
],
[
"Zhang",
"Ni",
""
],
[
"Liu",
"Nian",
""
],
[
"Khan",
"Salman",
""
],
[
"Cholakkal",
"Hisham",
""
],
[
"Anwer",
"Rao Muhammad",
""
],
[
"Khan",
"Fahad Shahbaz",
""
]
] | TITLE: Discriminative Co-Saliency and Background Mining Transformer for
Co-Salient Object Detection
ABSTRACT: Most previous co-salient object detection works mainly focus on extracting
co-salient cues via mining the consistency relations across images while
ignoring explicit exploration of background regions. In this paper, we propose
a Discriminative co-saliency and background Mining Transformer framework (DMT)
based on several economical multi-grained correlation modules to explicitly
mine both co-saliency and background information and effectively model their
discrimination. Specifically, we first propose a region-to-region correlation
module for introducing inter-image relations to pixel-wise segmentation
features while maintaining computational efficiency. Then, we use two types of
pre-defined tokens to mine co-saliency and background information via our
proposed contrast-induced pixel-to-token correlation and co-saliency
token-to-token correlation modules. We also design a token-guided feature
refinement module to enhance the discriminability of the segmentation features
under the guidance of the learned tokens. We perform iterative mutual promotion
for the segmentation feature extraction and token construction. Experimental
results on three benchmark datasets demonstrate the effectiveness of our
proposed method. The source code is available at:
https://github.com/dragonlee258079/DMT.
| no_new_dataset | 0.709617 |
2305.01044 | Md Mahamudul Hasan | Md Mahamudul Hasan, Moi Hoon Yap, Md Kamrul Hasan | Venn Diagram Multi-label Class Interpretation of Diabetic Foot Ulcer
with Color and Sharpness Enhancement | The Paper is not complete, more modifications are needed | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | DFU is a severe complication of diabetes that can lead to amputation of the
lower limb if not treated properly. Inspired by the 2021 Diabetic Foot Ulcer
Grand Challenge, researchers designed automated multi-class classification of
DFU, including infection, ischaemia, both of these conditions, and none of
these conditions. However, it remains a challenge as classification accuracy is
still not satisfactory. This paper proposes a Venn Diagram interpretation of
multi-label CNN-based method, utilizing different image enhancement strategies,
to improve the multi-class DFU classification. We propose to reduce the four
classes into two since both class wounds can be interpreted as the simultaneous
occurrence of infection and ischaemia and none class wounds as the absence of
infection and ischaemia. We introduce a novel Venn Diagram representation block
in the classifier to interpret all four classes from these two classes. To make
our model more resilient, we propose enhancing the perceptual quality of DFU
images, particularly blurry or inconsistently lit DFU images, by performing
color and sharpness enhancements on them. We also employ a fine-tuned
optimization technique, adaptive sharpness aware minimization, to improve the
CNN model generalization performance. The proposed method is evaluated on the
test dataset of DFUC2021, containing 5,734 images and the results are compared
with the top-3 winning entries of DFUC2021. Our proposed approach outperforms
these existing approaches and achieves Macro-Average F1, Recall and Precision
scores of 0.6592, 0.6593, and 0.6652, respectively.Additionally, We perform
ablation studies and image quality measurements to further interpret our
proposed method. This proposed method will benefit patients with DFUs since it
tackles the inconsistencies in captured images and can be employed for a more
robust remote DFU wound classification.
| [
{
"version": "v1",
"created": "Mon, 1 May 2023 19:06:28 GMT"
},
{
"version": "v2",
"created": "Fri, 5 May 2023 19:25:01 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Hasan",
"Md Mahamudul",
""
],
[
"Yap",
"Moi Hoon",
""
],
[
"Hasan",
"Md Kamrul",
""
]
] | TITLE: Venn Diagram Multi-label Class Interpretation of Diabetic Foot Ulcer
with Color and Sharpness Enhancement
ABSTRACT: DFU is a severe complication of diabetes that can lead to amputation of the
lower limb if not treated properly. Inspired by the 2021 Diabetic Foot Ulcer
Grand Challenge, researchers designed automated multi-class classification of
DFU, including infection, ischaemia, both of these conditions, and none of
these conditions. However, it remains a challenge as classification accuracy is
still not satisfactory. This paper proposes a Venn Diagram interpretation of
multi-label CNN-based method, utilizing different image enhancement strategies,
to improve the multi-class DFU classification. We propose to reduce the four
classes into two since both class wounds can be interpreted as the simultaneous
occurrence of infection and ischaemia and none class wounds as the absence of
infection and ischaemia. We introduce a novel Venn Diagram representation block
in the classifier to interpret all four classes from these two classes. To make
our model more resilient, we propose enhancing the perceptual quality of DFU
images, particularly blurry or inconsistently lit DFU images, by performing
color and sharpness enhancements on them. We also employ a fine-tuned
optimization technique, adaptive sharpness aware minimization, to improve the
CNN model generalization performance. The proposed method is evaluated on the
test dataset of DFUC2021, containing 5,734 images and the results are compared
with the top-3 winning entries of DFUC2021. Our proposed approach outperforms
these existing approaches and achieves Macro-Average F1, Recall and Precision
scores of 0.6592, 0.6593, and 0.6652, respectively.Additionally, We perform
ablation studies and image quality measurements to further interpret our
proposed method. This proposed method will benefit patients with DFUs since it
tackles the inconsistencies in captured images and can be employed for a more
robust remote DFU wound classification.
| no_new_dataset | 0.708584 |
2305.01241 | Hendric Vo{\ss} | Hendric Vo{\ss} and Stefan Kopp | AQ-GT: a Temporally Aligned and Quantized GRU-Transformer for Co-Speech
Gesture Synthesis | null | null | null | null | cs.HC cs.GR cs.LG cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The generation of realistic and contextually relevant co-speech gestures is a
challenging yet increasingly important task in the creation of multimodal
artificial agents. Prior methods focused on learning a direct correspondence
between co-speech gesture representations and produced motions, which created
seemingly natural but often unconvincing gestures during human assessment. We
present an approach to pre-train partial gesture sequences using a generative
adversarial network with a quantization pipeline. The resulting codebook
vectors serve as both input and output in our framework, forming the basis for
the generation and reconstruction of gestures. By learning the mapping of a
latent space representation as opposed to directly mapping it to a vector
representation, this framework facilitates the generation of highly realistic
and expressive gestures that closely replicate human movement and behavior,
while simultaneously avoiding artifacts in the generation process. We evaluate
our approach by comparing it with established methods for generating co-speech
gestures as well as with existing datasets of human behavior. We also perform
an ablation study to assess our findings. The results show that our approach
outperforms the current state of the art by a clear margin and is partially
indistinguishable from human gesturing. We make our data pipeline and the
generation framework publicly available.
| [
{
"version": "v1",
"created": "Tue, 2 May 2023 07:59:38 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 11:59:12 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Voß",
"Hendric",
""
],
[
"Kopp",
"Stefan",
""
]
] | TITLE: AQ-GT: a Temporally Aligned and Quantized GRU-Transformer for Co-Speech
Gesture Synthesis
ABSTRACT: The generation of realistic and contextually relevant co-speech gestures is a
challenging yet increasingly important task in the creation of multimodal
artificial agents. Prior methods focused on learning a direct correspondence
between co-speech gesture representations and produced motions, which created
seemingly natural but often unconvincing gestures during human assessment. We
present an approach to pre-train partial gesture sequences using a generative
adversarial network with a quantization pipeline. The resulting codebook
vectors serve as both input and output in our framework, forming the basis for
the generation and reconstruction of gestures. By learning the mapping of a
latent space representation as opposed to directly mapping it to a vector
representation, this framework facilitates the generation of highly realistic
and expressive gestures that closely replicate human movement and behavior,
while simultaneously avoiding artifacts in the generation process. We evaluate
our approach by comparing it with established methods for generating co-speech
gestures as well as with existing datasets of human behavior. We also perform
an ablation study to assess our findings. The results show that our approach
outperforms the current state of the art by a clear margin and is partially
indistinguishable from human gesturing. We make our data pipeline and the
generation framework publicly available.
| no_new_dataset | 0.71022 |
2305.01620 | Siddhant Arora | Siddhant Arora, Hayato Futami, Shih-Lun Wu, Jessica Huynh, Yifan Peng,
Yosuke Kashiwagi, Emiru Tsunoo, Brian Yan, Shinji Watanabe | A Study on the Integration of Pipeline and E2E SLU systems for Spoken
Semantic Parsing toward STOP Quality Challenge | First Place in Track 1 of STOP Challenge, which is part of ICASSP
Signal Processing Grand Challenge 2023 | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently there have been efforts to introduce new benchmark tasks for spoken
language understanding (SLU), like semantic parsing. In this paper, we describe
our proposed spoken semantic parsing system for the quality track (Track 1) in
Spoken Language Understanding Grand Challenge which is part of ICASSP Signal
Processing Grand Challenge 2023. We experiment with both end-to-end and
pipeline systems for this task. Strong automatic speech recognition (ASR)
models like Whisper and pretrained Language models (LM) like BART are utilized
inside our SLU framework to boost performance. We also investigate the output
level combination of various models to get an exact match accuracy of 80.8,
which won the 1st place at the challenge.
| [
{
"version": "v1",
"created": "Tue, 2 May 2023 17:25:19 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 16:35:31 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Arora",
"Siddhant",
""
],
[
"Futami",
"Hayato",
""
],
[
"Wu",
"Shih-Lun",
""
],
[
"Huynh",
"Jessica",
""
],
[
"Peng",
"Yifan",
""
],
[
"Kashiwagi",
"Yosuke",
""
],
[
"Tsunoo",
"Emiru",
""
],
[
"Yan",
"Brian",
""
],
[
"Watanabe",
"Shinji",
""
]
] | TITLE: A Study on the Integration of Pipeline and E2E SLU systems for Spoken
Semantic Parsing toward STOP Quality Challenge
ABSTRACT: Recently there have been efforts to introduce new benchmark tasks for spoken
language understanding (SLU), like semantic parsing. In this paper, we describe
our proposed spoken semantic parsing system for the quality track (Track 1) in
Spoken Language Understanding Grand Challenge which is part of ICASSP Signal
Processing Grand Challenge 2023. We experiment with both end-to-end and
pipeline systems for this task. Strong automatic speech recognition (ASR)
models like Whisper and pretrained Language models (LM) like BART are utilized
inside our SLU framework to boost performance. We also investigate the output
level combination of various models to get an exact match accuracy of 80.8,
which won the 1st place at the challenge.
| no_new_dataset | 0.678114 |
2305.02200 | Chen Ling | Chen Ling, Junji Jiang, Junxiang Wang, My Thai, Lukas Xue, James Song,
Meikang Qiu, Liang Zhao | Deep Graph Representation Learning and Optimization for Influence
Maximization | In Proceedings of the 40th International Conference on Machine
Learning (ICML 2023), Honolulu, Hawaii, USA. PMLR 202, 2023 | null | null | null | cs.SI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Influence maximization (IM) is formulated as selecting a set of initial users
from a social network to maximize the expected number of influenced users.
Researchers have made great progress in designing various traditional methods,
and their theoretical design and performance gain are close to a limit. In the
past few years, learning-based IM methods have emerged to achieve stronger
generalization ability to unknown graphs than traditional ones. However, the
development of learning-based IM methods is still limited by fundamental
obstacles, including 1) the difficulty of effectively solving the objective
function; 2) the difficulty of characterizing the diversified underlying
diffusion patterns; and 3) the difficulty of adapting the solution under
various node-centrality-constrained IM variants. To cope with the above
challenges, we design a novel framework DeepIM to generatively characterize the
latent representation of seed sets, and we propose to learn the diversified
information diffusion pattern in a data-driven and end-to-end manner. Finally,
we design a novel objective function to infer optimal seed sets under flexible
node-centrality-based budget constraints. Extensive analyses are conducted over
both synthetic and real-world datasets to demonstrate the overall performance
of DeepIM. The code and data are available at:
https://github.com/triplej0079/DeepIM.
| [
{
"version": "v1",
"created": "Mon, 1 May 2023 15:45:01 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 15:02:48 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Ling",
"Chen",
""
],
[
"Jiang",
"Junji",
""
],
[
"Wang",
"Junxiang",
""
],
[
"Thai",
"My",
""
],
[
"Xue",
"Lukas",
""
],
[
"Song",
"James",
""
],
[
"Qiu",
"Meikang",
""
],
[
"Zhao",
"Liang",
""
]
] | TITLE: Deep Graph Representation Learning and Optimization for Influence
Maximization
ABSTRACT: Influence maximization (IM) is formulated as selecting a set of initial users
from a social network to maximize the expected number of influenced users.
Researchers have made great progress in designing various traditional methods,
and their theoretical design and performance gain are close to a limit. In the
past few years, learning-based IM methods have emerged to achieve stronger
generalization ability to unknown graphs than traditional ones. However, the
development of learning-based IM methods is still limited by fundamental
obstacles, including 1) the difficulty of effectively solving the objective
function; 2) the difficulty of characterizing the diversified underlying
diffusion patterns; and 3) the difficulty of adapting the solution under
various node-centrality-constrained IM variants. To cope with the above
challenges, we design a novel framework DeepIM to generatively characterize the
latent representation of seed sets, and we propose to learn the diversified
information diffusion pattern in a data-driven and end-to-end manner. Finally,
we design a novel objective function to infer optimal seed sets under flexible
node-centrality-based budget constraints. Extensive analyses are conducted over
both synthetic and real-world datasets to demonstrate the overall performance
of DeepIM. The code and data are available at:
https://github.com/triplej0079/DeepIM.
| no_new_dataset | 0.70878 |
2305.02350 | Fabian Ziegner | Fabian Ziegner, Janos Borst, Andreas Niekler, Martin Potthast | Using Language Models on Low-end Hardware | 5+4 pages, 6 tables; fixed affiliation | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper evaluates the viability of using fixed language models for
training text classification networks on low-end hardware. We combine language
models with a CNN architecture and put together a comprehensive benchmark with
8 datasets covering single-label and multi-label classification of topic,
sentiment, and genre. Our observations are distilled into a list of trade-offs,
concluding that there are scenarios, where not fine-tuning a language model
yields competitive effectiveness at faster training, requiring only a quarter
of the memory compared to fine-tuning.
| [
{
"version": "v1",
"created": "Wed, 3 May 2023 18:00:03 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 15:43:52 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Ziegner",
"Fabian",
""
],
[
"Borst",
"Janos",
""
],
[
"Niekler",
"Andreas",
""
],
[
"Potthast",
"Martin",
""
]
] | TITLE: Using Language Models on Low-end Hardware
ABSTRACT: This paper evaluates the viability of using fixed language models for
training text classification networks on low-end hardware. We combine language
models with a CNN architecture and put together a comprehensive benchmark with
8 datasets covering single-label and multi-label classification of topic,
sentiment, and genre. Our observations are distilled into a list of trade-offs,
concluding that there are scenarios, where not fine-tuning a language model
yields competitive effectiveness at faster training, requiring only a quarter
of the memory compared to fine-tuning.
| no_new_dataset | 0.661212 |
2305.02412 | Yue Wu | Yue Wu, So Yeon Min, Yonatan Bisk, Ruslan Salakhutdinov, Amos Azaria,
Yuanzhi Li, Tom Mitchell, Shrimai Prabhumoye | Plan, Eliminate, and Track -- Language Models are Good Teachers for
Embodied Agents | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Pre-trained large language models (LLMs) capture procedural knowledge about
the world. Recent work has leveraged LLM's ability to generate abstract plans
to simplify challenging control tasks, either by action scoring, or action
modeling (fine-tuning). However, the transformer architecture inherits several
constraints that make it difficult for the LLM to directly serve as the agent:
e.g. limited input lengths, fine-tuning inefficiency, bias from pre-training,
and incompatibility with non-text environments. To maintain compatibility with
a low-level trainable actor, we propose to instead use the knowledge in LLMs to
simplify the control problem, rather than solving it. We propose the Plan,
Eliminate, and Track (PET) framework. The Plan module translates a task
description into a list of high-level sub-tasks. The Eliminate module masks out
irrelevant objects and receptacles from the observation for the current
sub-task. Finally, the Track module determines whether the agent has
accomplished each sub-task. On the AlfWorld instruction following benchmark,
the PET framework leads to a significant 15% improvement over SOTA for
generalization to human goal specifications.
| [
{
"version": "v1",
"created": "Wed, 3 May 2023 20:11:22 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 05:33:10 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Wu",
"Yue",
""
],
[
"Min",
"So Yeon",
""
],
[
"Bisk",
"Yonatan",
""
],
[
"Salakhutdinov",
"Ruslan",
""
],
[
"Azaria",
"Amos",
""
],
[
"Li",
"Yuanzhi",
""
],
[
"Mitchell",
"Tom",
""
],
[
"Prabhumoye",
"Shrimai",
""
]
] | TITLE: Plan, Eliminate, and Track -- Language Models are Good Teachers for
Embodied Agents
ABSTRACT: Pre-trained large language models (LLMs) capture procedural knowledge about
the world. Recent work has leveraged LLM's ability to generate abstract plans
to simplify challenging control tasks, either by action scoring, or action
modeling (fine-tuning). However, the transformer architecture inherits several
constraints that make it difficult for the LLM to directly serve as the agent:
e.g. limited input lengths, fine-tuning inefficiency, bias from pre-training,
and incompatibility with non-text environments. To maintain compatibility with
a low-level trainable actor, we propose to instead use the knowledge in LLMs to
simplify the control problem, rather than solving it. We propose the Plan,
Eliminate, and Track (PET) framework. The Plan module translates a task
description into a list of high-level sub-tasks. The Eliminate module masks out
irrelevant objects and receptacles from the observation for the current
sub-task. Finally, the Track module determines whether the agent has
accomplished each sub-task. On the AlfWorld instruction following benchmark,
the PET framework leads to a significant 15% improvement over SOTA for
generalization to human goal specifications.
| no_new_dataset | 0.708427 |
2305.02743 | Shun-Cheng Wu | Shun-Cheng Wu, Keisuke Tateno, Nassir Navab, Federico Tombari | Incremental 3D Semantic Scene Graph Prediction from RGB Sequences | The paper has been accepted in CVPR23 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | 3D semantic scene graphs are a powerful holistic representation as they
describe the individual objects and depict the relation between them. They are
compact high-level graphs that enable many tasks requiring scene reasoning. In
real-world settings, existing 3D estimation methods produce robust predictions
that mostly rely on dense inputs. In this work, we propose a real-time
framework that incrementally builds a consistent 3D semantic scene graph of a
scene given an RGB image sequence. Our method consists of a novel incremental
entity estimation pipeline and a scene graph prediction network. The proposed
pipeline simultaneously reconstructs a sparse point map and fuses entity
estimation from the input images. The proposed network estimates 3D semantic
scene graphs with iterative message passing using multi-view and geometric
features extracted from the scene entities. Extensive experiments on the 3RScan
dataset show the effectiveness of the proposed method in this challenging task,
outperforming state-of-the-art approaches.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 11:32:16 GMT"
},
{
"version": "v2",
"created": "Sat, 6 May 2023 20:15:08 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Wu",
"Shun-Cheng",
""
],
[
"Tateno",
"Keisuke",
""
],
[
"Navab",
"Nassir",
""
],
[
"Tombari",
"Federico",
""
]
] | TITLE: Incremental 3D Semantic Scene Graph Prediction from RGB Sequences
ABSTRACT: 3D semantic scene graphs are a powerful holistic representation as they
describe the individual objects and depict the relation between them. They are
compact high-level graphs that enable many tasks requiring scene reasoning. In
real-world settings, existing 3D estimation methods produce robust predictions
that mostly rely on dense inputs. In this work, we propose a real-time
framework that incrementally builds a consistent 3D semantic scene graph of a
scene given an RGB image sequence. Our method consists of a novel incremental
entity estimation pipeline and a scene graph prediction network. The proposed
pipeline simultaneously reconstructs a sparse point map and fuses entity
estimation from the input images. The proposed network estimates 3D semantic
scene graphs with iterative message passing using multi-view and geometric
features extracted from the scene entities. Extensive experiments on the 3RScan
dataset show the effectiveness of the proposed method in this challenging task,
outperforming state-of-the-art approaches.
| no_new_dataset | 0.70978 |
2305.02769 | Tahira Shehzadi | Tahira Shehzadi, Khurram Azeem Hashmi, Didier Stricker, Marcus Liwicki
and Muhammad Zeshan Afzal | Towards End-to-End Semi-Supervised Table Detection with Deformable
Transformer | ICDAR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Table detection is the task of classifying and localizing table objects
within document images. With the recent development in deep learning methods,
we observe remarkable success in table detection. However, a significant amount
of labeled data is required to train these models effectively. Many
semi-supervised approaches are introduced to mitigate the need for a
substantial amount of label data. These approaches use CNN-based detectors that
rely on anchor proposals and post-processing stages such as NMS. To tackle
these limitations, this paper presents a novel end-to-end semi-supervised table
detection method that employs the deformable transformer for detecting table
objects. We evaluate our semi-supervised method on PubLayNet, DocBank, ICADR-19
and TableBank datasets, and it achieves superior performance compared to
previous methods. It outperforms the fully supervised method (Deformable
transformer) by +3.4 points on 10\% labels of TableBank-both dataset and the
previous CNN-based semi-supervised approach (Soft Teacher) by +1.8 points on
10\% labels of PubLayNet dataset. We hope this work opens new possibilities
towards semi-supervised and unsupervised table detection methods.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 12:15:15 GMT"
},
{
"version": "v2",
"created": "Sun, 7 May 2023 20:06:18 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Shehzadi",
"Tahira",
""
],
[
"Hashmi",
"Khurram Azeem",
""
],
[
"Stricker",
"Didier",
""
],
[
"Liwicki",
"Marcus",
""
],
[
"Afzal",
"Muhammad Zeshan",
""
]
] | TITLE: Towards End-to-End Semi-Supervised Table Detection with Deformable
Transformer
ABSTRACT: Table detection is the task of classifying and localizing table objects
within document images. With the recent development in deep learning methods,
we observe remarkable success in table detection. However, a significant amount
of labeled data is required to train these models effectively. Many
semi-supervised approaches are introduced to mitigate the need for a
substantial amount of label data. These approaches use CNN-based detectors that
rely on anchor proposals and post-processing stages such as NMS. To tackle
these limitations, this paper presents a novel end-to-end semi-supervised table
detection method that employs the deformable transformer for detecting table
objects. We evaluate our semi-supervised method on PubLayNet, DocBank, ICADR-19
and TableBank datasets, and it achieves superior performance compared to
previous methods. It outperforms the fully supervised method (Deformable
transformer) by +3.4 points on 10\% labels of TableBank-both dataset and the
previous CNN-based semi-supervised approach (Soft Teacher) by +1.8 points on
10\% labels of PubLayNet dataset. We hope this work opens new possibilities
towards semi-supervised and unsupervised table detection methods.
| no_new_dataset | 0.711638 |
2305.02893 | Quan Liu | Quan Liu, Yunsong Zhou, Hongzi Zhu, Shan Chang, Minyi Guo | APR: Online Distant Point Cloud Registration Through Aggregated Point
Cloud Reconstruction | in Proceedings of the International Joint Conference on Artificial
Intelligence (IJCAI), 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For many driving safety applications, it is of great importance to accurately
register LiDAR point clouds generated on distant moving vehicles. However, such
point clouds have extremely different point density and sensor perspective on
the same object, making registration on such point clouds very hard. In this
paper, we propose a novel feature extraction framework, called APR, for online
distant point cloud registration. Specifically, APR leverages an autoencoder
design, where the autoencoder reconstructs a denser aggregated point cloud with
several frames instead of the original single input point cloud. Our design
forces the encoder to extract features with rich local geometry information
based on one single input point cloud. Such features are then used for online
distant point cloud registration. We conduct extensive experiments against
state-of-the-art (SOTA) feature extractors on KITTI and nuScenes datasets.
Results show that APR outperforms all other extractors by a large margin,
increasing average registration recall of SOTA extractors by 7.1% on LoKITTI
and 4.6% on LoNuScenes. Code is available at https://github.com/liuQuan98/APR.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 15:01:56 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 11:44:10 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Liu",
"Quan",
""
],
[
"Zhou",
"Yunsong",
""
],
[
"Zhu",
"Hongzi",
""
],
[
"Chang",
"Shan",
""
],
[
"Guo",
"Minyi",
""
]
] | TITLE: APR: Online Distant Point Cloud Registration Through Aggregated Point
Cloud Reconstruction
ABSTRACT: For many driving safety applications, it is of great importance to accurately
register LiDAR point clouds generated on distant moving vehicles. However, such
point clouds have extremely different point density and sensor perspective on
the same object, making registration on such point clouds very hard. In this
paper, we propose a novel feature extraction framework, called APR, for online
distant point cloud registration. Specifically, APR leverages an autoencoder
design, where the autoencoder reconstructs a denser aggregated point cloud with
several frames instead of the original single input point cloud. Our design
forces the encoder to extract features with rich local geometry information
based on one single input point cloud. Such features are then used for online
distant point cloud registration. We conduct extensive experiments against
state-of-the-art (SOTA) feature extractors on KITTI and nuScenes datasets.
Results show that APR outperforms all other extractors by a large margin,
increasing average registration recall of SOTA extractors by 7.1% on LoKITTI
and 4.6% on LoNuScenes. Code is available at https://github.com/liuQuan98/APR.
| no_new_dataset | 0.713644 |
2305.02931 | Zhao Kang | Erlin Pan, Zhao Kang | Beyond Homophily: Reconstructing Structure for Graph-agnostic Clustering | Accepted by ICML 2023 | null | null | null | cs.SI cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Graph neural networks (GNNs) based methods have achieved impressive
performance on node clustering task. However, they are designed on the
homophilic assumption of graph and clustering on heterophilic graph is
overlooked. Due to the lack of labels, it is impossible to first identify a
graph as homophilic or heterophilic before a suitable GNN model can be found.
Hence, clustering on real-world graph with various levels of homophily poses a
new challenge to the graph research community. To fill this gap, we propose a
novel graph clustering method, which contains three key components: graph
reconstruction, a mixed filter, and dual graph clustering network. To be
graph-agnostic, we empirically construct two graphs which are high homophily
and heterophily from each data. The mixed filter based on the new graphs
extracts both low-frequency and high-frequency information. To reduce the
adverse coupling between node attribute and topological structure, we
separately map them into two subspaces in dual graph clustering network.
Extensive experiments on 11 benchmark graphs demonstrate our promising
performance. In particular, our method dominates others on heterophilic graphs.
| [
{
"version": "v1",
"created": "Wed, 3 May 2023 01:49:01 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Pan",
"Erlin",
""
],
[
"Kang",
"Zhao",
""
]
] | TITLE: Beyond Homophily: Reconstructing Structure for Graph-agnostic Clustering
ABSTRACT: Graph neural networks (GNNs) based methods have achieved impressive
performance on node clustering task. However, they are designed on the
homophilic assumption of graph and clustering on heterophilic graph is
overlooked. Due to the lack of labels, it is impossible to first identify a
graph as homophilic or heterophilic before a suitable GNN model can be found.
Hence, clustering on real-world graph with various levels of homophily poses a
new challenge to the graph research community. To fill this gap, we propose a
novel graph clustering method, which contains three key components: graph
reconstruction, a mixed filter, and dual graph clustering network. To be
graph-agnostic, we empirically construct two graphs which are high homophily
and heterophily from each data. The mixed filter based on the new graphs
extracts both low-frequency and high-frequency information. To reduce the
adverse coupling between node attribute and topological structure, we
separately map them into two subspaces in dual graph clustering network.
Extensive experiments on 11 benchmark graphs demonstrate our promising
performance. In particular, our method dominates others on heterophilic graphs.
| no_new_dataset | 0.711425 |
2305.03045 | Peng-Shuai Wang | Peng-Shuai Wang | OctFormer: Octree-based Transformers for 3D Point Clouds | SIGGRAPH 2023, Journal Track | ACM Transactions on Graphics (SIGGRAPH), 42, 4 (August 2023), 11
pages | 10.1145/3592131 | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose octree-based transformers, named OctFormer, for 3D point cloud
learning. OctFormer can not only serve as a general and effective backbone for
3D point cloud segmentation and object detection but also have linear
complexity and is scalable for large-scale point clouds. The key challenge in
applying transformers to point clouds is reducing the quadratic, thus
overwhelming, computation complexity of attentions. To combat this issue,
several works divide point clouds into non-overlapping windows and constrain
attentions in each local window. However, the point number in each window
varies greatly, impeding the efficient execution on GPU. Observing that
attentions are robust to the shapes of local windows, we propose a novel octree
attention, which leverages sorted shuffled keys of octrees to partition point
clouds into local windows containing a fixed number of points while permitting
shapes of windows to change freely. And we also introduce dilated octree
attention to expand the receptive field further. Our octree attention can be
implemented in 10 lines of code with open-sourced libraries and runs 17 times
faster than other point cloud attentions when the point number exceeds 200k.
Built upon the octree attention, OctFormer can be easily scaled up and achieves
state-of-the-art performances on a series of 3D segmentation and detection
benchmarks, surpassing previous sparse-voxel-based CNNs and point cloud
transformers in terms of both efficiency and effectiveness. Notably, on the
challenging ScanNet200 dataset, OctFormer outperforms sparse-voxel-based CNNs
by 7.3 in mIoU. Our code and trained models are available at
https://wang-ps.github.io/octformer.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 17:59:05 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 00:31:54 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Wang",
"Peng-Shuai",
""
]
] | TITLE: OctFormer: Octree-based Transformers for 3D Point Clouds
ABSTRACT: We propose octree-based transformers, named OctFormer, for 3D point cloud
learning. OctFormer can not only serve as a general and effective backbone for
3D point cloud segmentation and object detection but also have linear
complexity and is scalable for large-scale point clouds. The key challenge in
applying transformers to point clouds is reducing the quadratic, thus
overwhelming, computation complexity of attentions. To combat this issue,
several works divide point clouds into non-overlapping windows and constrain
attentions in each local window. However, the point number in each window
varies greatly, impeding the efficient execution on GPU. Observing that
attentions are robust to the shapes of local windows, we propose a novel octree
attention, which leverages sorted shuffled keys of octrees to partition point
clouds into local windows containing a fixed number of points while permitting
shapes of windows to change freely. And we also introduce dilated octree
attention to expand the receptive field further. Our octree attention can be
implemented in 10 lines of code with open-sourced libraries and runs 17 times
faster than other point cloud attentions when the point number exceeds 200k.
Built upon the octree attention, OctFormer can be easily scaled up and achieves
state-of-the-art performances on a series of 3D segmentation and detection
benchmarks, surpassing previous sparse-voxel-based CNNs and point cloud
transformers in terms of both efficiency and effectiveness. Notably, on the
challenging ScanNet200 dataset, OctFormer outperforms sparse-voxel-based CNNs
by 7.3 in mIoU. Our code and trained models are available at
https://wang-ps.github.io/octformer.
| no_new_dataset | 0.710791 |
2305.03259 | Xingyu Zhu | Xingyu Zhu, Xin Wang, Jonathan Freer, Hyung Jin Chang, Yixing Gao | Clothes Grasping and Unfolding Based on RGB-D Semantic Segmentation | This paper is accepted to ICRA 2023 | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Clothes grasping and unfolding is a core step in robotic-assisted dressing.
Most existing works leverage depth images of clothes to train a deep
learning-based model to recognize suitable grasping points. These methods often
utilize physics engines to synthesize depth images to reduce the cost of real
labeled data collection. However, the natural domain gap between synthetic and
real images often leads to poor performance of these methods on real data.
Furthermore, these approaches often struggle in scenarios where grasping points
are occluded by the clothing item itself. To address the above challenges, we
propose a novel Bi-directional Fractal Cross Fusion Network (BiFCNet) for
semantic segmentation, enabling recognition of graspable regions in order to
provide more possibilities for grasping. Instead of using depth images only, we
also utilize RGB images with rich color features as input to our network in
which the Fractal Cross Fusion (FCF) module fuses RGB and depth data by
considering global complex features based on fractal geometry. To reduce the
cost of real data collection, we further propose a data augmentation method
based on an adversarial strategy, in which the color and geometric
transformations simultaneously process RGB and depth data while maintaining the
label correspondence. Finally, we present a pipeline for clothes grasping and
unfolding from the perspective of semantic segmentation, through the addition
of a strategy for grasp point selection from segmentation regions based on
clothing flatness measures, while taking into account the grasping direction.
We evaluate our BiFCNet on the public dataset NYUDv2 and obtained comparable
performance to current state-of-the-art models. We also deploy our model on a
Baxter robot, running extensive grasping and unfolding experiments as part of
our ablation studies, achieving an 84% success rate.
| [
{
"version": "v1",
"created": "Fri, 5 May 2023 03:21:55 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 12:44:30 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Zhu",
"Xingyu",
""
],
[
"Wang",
"Xin",
""
],
[
"Freer",
"Jonathan",
""
],
[
"Chang",
"Hyung Jin",
""
],
[
"Gao",
"Yixing",
""
]
] | TITLE: Clothes Grasping and Unfolding Based on RGB-D Semantic Segmentation
ABSTRACT: Clothes grasping and unfolding is a core step in robotic-assisted dressing.
Most existing works leverage depth images of clothes to train a deep
learning-based model to recognize suitable grasping points. These methods often
utilize physics engines to synthesize depth images to reduce the cost of real
labeled data collection. However, the natural domain gap between synthetic and
real images often leads to poor performance of these methods on real data.
Furthermore, these approaches often struggle in scenarios where grasping points
are occluded by the clothing item itself. To address the above challenges, we
propose a novel Bi-directional Fractal Cross Fusion Network (BiFCNet) for
semantic segmentation, enabling recognition of graspable regions in order to
provide more possibilities for grasping. Instead of using depth images only, we
also utilize RGB images with rich color features as input to our network in
which the Fractal Cross Fusion (FCF) module fuses RGB and depth data by
considering global complex features based on fractal geometry. To reduce the
cost of real data collection, we further propose a data augmentation method
based on an adversarial strategy, in which the color and geometric
transformations simultaneously process RGB and depth data while maintaining the
label correspondence. Finally, we present a pipeline for clothes grasping and
unfolding from the perspective of semantic segmentation, through the addition
of a strategy for grasp point selection from segmentation regions based on
clothing flatness measures, while taking into account the grasping direction.
We evaluate our BiFCNet on the public dataset NYUDv2 and obtained comparable
performance to current state-of-the-art models. We also deploy our model on a
Baxter robot, running extensive grasping and unfolding experiments as part of
our ablation studies, achieving an 84% success rate.
| no_new_dataset | 0.712826 |
2305.03655 | Yufei Li | Yufei Li, Zexin Li, Yingfan Gao, Cong Liu | White-Box Multi-Objective Adversarial Attack on Dialogue Generation | ACL 2023 main conference long paper | null | null | null | cs.CL cs.CR cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Pre-trained transformers are popular in state-of-the-art dialogue generation
(DG) systems. Such language models are, however, vulnerable to various
adversarial samples as studied in traditional tasks such as text
classification, which inspires our curiosity about their robustness in DG
systems. One main challenge of attacking DG models is that perturbations on the
current sentence can hardly degrade the response accuracy because the unchanged
chat histories are also considered for decision-making. Instead of merely
pursuing pitfalls of performance metrics such as BLEU, ROUGE, we observe that
crafting adversarial samples to force longer generation outputs benefits attack
effectiveness -- the generated responses are typically irrelevant, lengthy, and
repetitive. To this end, we propose a white-box multi-objective attack method
called DGSlow. Specifically, DGSlow balances two objectives -- generation
accuracy and length, via a gradient-based multi-objective optimizer and applies
an adaptive searching mechanism to iteratively craft adversarial samples with
only a few modifications. Comprehensive experiments on four benchmark datasets
demonstrate that DGSlow could significantly degrade state-of-the-art DG models
with a higher success rate than traditional accuracy-based methods. Besides,
our crafted sentences also exhibit strong transferability in attacking other
models.
| [
{
"version": "v1",
"created": "Fri, 5 May 2023 16:21:24 GMT"
},
{
"version": "v2",
"created": "Mon, 8 May 2023 15:16:05 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Li",
"Yufei",
""
],
[
"Li",
"Zexin",
""
],
[
"Gao",
"Yingfan",
""
],
[
"Liu",
"Cong",
""
]
] | TITLE: White-Box Multi-Objective Adversarial Attack on Dialogue Generation
ABSTRACT: Pre-trained transformers are popular in state-of-the-art dialogue generation
(DG) systems. Such language models are, however, vulnerable to various
adversarial samples as studied in traditional tasks such as text
classification, which inspires our curiosity about their robustness in DG
systems. One main challenge of attacking DG models is that perturbations on the
current sentence can hardly degrade the response accuracy because the unchanged
chat histories are also considered for decision-making. Instead of merely
pursuing pitfalls of performance metrics such as BLEU, ROUGE, we observe that
crafting adversarial samples to force longer generation outputs benefits attack
effectiveness -- the generated responses are typically irrelevant, lengthy, and
repetitive. To this end, we propose a white-box multi-objective attack method
called DGSlow. Specifically, DGSlow balances two objectives -- generation
accuracy and length, via a gradient-based multi-objective optimizer and applies
an adaptive searching mechanism to iteratively craft adversarial samples with
only a few modifications. Comprehensive experiments on four benchmark datasets
demonstrate that DGSlow could significantly degrade state-of-the-art DG models
with a higher success rate than traditional accuracy-based methods. Besides,
our crafted sentences also exhibit strong transferability in attacking other
models.
| no_new_dataset | 0.7076 |
2305.03737 | Jawid Ahmad Baktash | Jawid Ahmad Baktash, Mursal Dawodi, Mohammad Zarif Joya, Nematullah
Hassanzada | Tuning Traditional Language Processing Approaches for Pashto Text
Classification | arXiv admin note: substantial text overlap with arXiv:2305.03201 | International Journal on Cybernetics & Informatics (IJCI) Vol. 12,
No.2, April 2023 | 10.5121/ijci.2023.120222 | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Today text classification becomes critical task for concerned individuals for
numerous purposes. Hence, several researches have been conducted to develop
automatic text classification for national and international languages.
However, the need for an automatic text categorization system for local
languages is felt. The main aim of this study is to establish a Pashto
automatic text classification system. In order to pursue this work, we built a
Pashto corpus which is a collection of Pashto documents due to the
unavailability of public datasets of Pashto text documents. Besides, this study
compares several models containing both statistical and neural network machine
learning techniques including Multilayer Perceptron (MLP), Support Vector
Machine (SVM), K Nearest Neighbor (KNN), decision tree, gaussian na\"ive Bayes,
multinomial na\"ive Bayes, random forest, and logistic regression to discover
the most effective approach. Moreover, this investigation evaluates two
different feature extraction methods including unigram, and Time Frequency
Inverse Document Frequency (IFIDF). Subsequently, this research obtained
average testing accuracy rate 94% using MLP classification algorithm and TFIDF
feature extraction method in this context.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 22:57:45 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Baktash",
"Jawid Ahmad",
""
],
[
"Dawodi",
"Mursal",
""
],
[
"Joya",
"Mohammad Zarif",
""
],
[
"Hassanzada",
"Nematullah",
""
]
] | TITLE: Tuning Traditional Language Processing Approaches for Pashto Text
Classification
ABSTRACT: Today text classification becomes critical task for concerned individuals for
numerous purposes. Hence, several researches have been conducted to develop
automatic text classification for national and international languages.
However, the need for an automatic text categorization system for local
languages is felt. The main aim of this study is to establish a Pashto
automatic text classification system. In order to pursue this work, we built a
Pashto corpus which is a collection of Pashto documents due to the
unavailability of public datasets of Pashto text documents. Besides, this study
compares several models containing both statistical and neural network machine
learning techniques including Multilayer Perceptron (MLP), Support Vector
Machine (SVM), K Nearest Neighbor (KNN), decision tree, gaussian na\"ive Bayes,
multinomial na\"ive Bayes, random forest, and logistic regression to discover
the most effective approach. Moreover, this investigation evaluates two
different feature extraction methods including unigram, and Time Frequency
Inverse Document Frequency (IFIDF). Subsequently, this research obtained
average testing accuracy rate 94% using MLP classification algorithm and TFIDF
feature extraction method in this context.
| new_dataset | 0.701125 |
2305.03741 | Mengran Li | Xiaochuan Zhang, Mengran Li, Ye Wang, Haojun Fei | AmGCL: Feature Imputation of Attribute Missing Graph via Self-supervised
Contrastive Learning | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Attribute graphs are ubiquitous in multimedia applications, and graph
representation learning (GRL) has been successful in analyzing attribute graph
data. However, incomplete graph data and missing node attributes can have a
negative impact on media knowledge discovery. Existing methods for handling
attribute missing graph have limited assumptions or fail to capture complex
attribute-graph dependencies. To address these challenges, we propose Attribute
missing Graph Contrastive Learning (AmGCL), a framework for handling missing
node attributes in attribute graph data. AmGCL leverages Dirichlet energy
minimization-based feature precoding to encode in missing attributes and a
self-supervised Graph Augmentation Contrastive Learning Structure (GACLS) to
learn latent variables from the encoded-in data. Specifically, AmGCL utilizies
feature reconstruction based on structure-attribute energy minimization while
maximizes the lower bound of evidence for latent representation mutual
information. Our experimental results on multiple real-world datasets
demonstrate that AmGCL outperforms state-of-the-art methods in both feature
imputation and node classification tasks, indicating the effectiveness of our
proposed method in real-world attribute graph analysis tasks.
| [
{
"version": "v1",
"created": "Fri, 5 May 2023 07:03:24 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Zhang",
"Xiaochuan",
""
],
[
"Li",
"Mengran",
""
],
[
"Wang",
"Ye",
""
],
[
"Fei",
"Haojun",
""
]
] | TITLE: AmGCL: Feature Imputation of Attribute Missing Graph via Self-supervised
Contrastive Learning
ABSTRACT: Attribute graphs are ubiquitous in multimedia applications, and graph
representation learning (GRL) has been successful in analyzing attribute graph
data. However, incomplete graph data and missing node attributes can have a
negative impact on media knowledge discovery. Existing methods for handling
attribute missing graph have limited assumptions or fail to capture complex
attribute-graph dependencies. To address these challenges, we propose Attribute
missing Graph Contrastive Learning (AmGCL), a framework for handling missing
node attributes in attribute graph data. AmGCL leverages Dirichlet energy
minimization-based feature precoding to encode in missing attributes and a
self-supervised Graph Augmentation Contrastive Learning Structure (GACLS) to
learn latent variables from the encoded-in data. Specifically, AmGCL utilizies
feature reconstruction based on structure-attribute energy minimization while
maximizes the lower bound of evidence for latent representation mutual
information. Our experimental results on multiple real-world datasets
demonstrate that AmGCL outperforms state-of-the-art methods in both feature
imputation and node classification tasks, indicating the effectiveness of our
proposed method in real-world attribute graph analysis tasks.
| no_new_dataset | 0.709585 |
2305.03742 | Hanlin Zhang | Hanlin Zhang, Jiani Huang, Ziyang Li, Mayur Naik, Eric Xing | Improved Logical Reasoning of Language Models via Differentiable
Symbolic Programming | ACL 2023 Findings | null | null | null | cs.AI cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pre-trained large language models (LMs) struggle to perform logical reasoning
reliably despite advances in scale and compositionality. In this work, we
tackle this challenge through the lens of symbolic programming. We propose
DSR-LM, a Differentiable Symbolic Reasoning framework where pre-trained LMs
govern the perception of factual knowledge, and a symbolic module performs
deductive reasoning. In contrast to works that rely on hand-crafted logic
rules, our differentiable symbolic reasoning framework efficiently learns
weighted rules and applies semantic loss to further improve LMs. DSR-LM is
scalable, interpretable, and allows easy integration of prior knowledge,
thereby supporting extensive symbolic programming to robustly derive a logical
conclusion. The results of our experiments suggest that DSR-LM improves the
logical reasoning abilities of pre-trained language models, resulting in a
significant increase in accuracy of over 20% on deductive reasoning benchmarks.
Furthermore, DSR-LM outperforms a variety of competitive baselines when faced
with systematic changes in sequence length.
| [
{
"version": "v1",
"created": "Fri, 5 May 2023 07:24:46 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Zhang",
"Hanlin",
""
],
[
"Huang",
"Jiani",
""
],
[
"Li",
"Ziyang",
""
],
[
"Naik",
"Mayur",
""
],
[
"Xing",
"Eric",
""
]
] | TITLE: Improved Logical Reasoning of Language Models via Differentiable
Symbolic Programming
ABSTRACT: Pre-trained large language models (LMs) struggle to perform logical reasoning
reliably despite advances in scale and compositionality. In this work, we
tackle this challenge through the lens of symbolic programming. We propose
DSR-LM, a Differentiable Symbolic Reasoning framework where pre-trained LMs
govern the perception of factual knowledge, and a symbolic module performs
deductive reasoning. In contrast to works that rely on hand-crafted logic
rules, our differentiable symbolic reasoning framework efficiently learns
weighted rules and applies semantic loss to further improve LMs. DSR-LM is
scalable, interpretable, and allows easy integration of prior knowledge,
thereby supporting extensive symbolic programming to robustly derive a logical
conclusion. The results of our experiments suggest that DSR-LM improves the
logical reasoning abilities of pre-trained language models, resulting in a
significant increase in accuracy of over 20% on deductive reasoning benchmarks.
Furthermore, DSR-LM outperforms a variety of competitive baselines when faced
with systematic changes in sequence length.
| no_new_dataset | 0.707992 |
2305.03743 | Anthony Frion | Anthony Frion, Lucas Drumetz, Guillaume Tochon, Mauro Dalla Mura,
Abdeldjalil A\"issa El Bey | Learning Sentinel-2 reflectance dynamics for data-driven assimilation
and forecasting | null | null | null | null | eess.IV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Over the last few years, massive amounts of satellite multispectral and
hyperspectral images covering the Earth's surface have been made publicly
available for scientific purpose, for example through the European Copernicus
project. Simultaneously, the development of self-supervised learning (SSL)
methods has sparked great interest in the remote sensing community, enabling to
learn latent representations from unlabeled data to help treating downstream
tasks for which there is few annotated examples, such as interpolation,
forecasting or unmixing. Following this line, we train a deep learning model
inspired from the Koopman operator theory to model long-term reflectance
dynamics in an unsupervised way. We show that this trained model, being
differentiable, can be used as a prior for data assimilation in a
straightforward way. Our datasets, which are composed of Sentinel-2
multispectral image time series, are publicly released with several levels of
treatment.
| [
{
"version": "v1",
"created": "Fri, 5 May 2023 10:04:03 GMT"
}
] | 2023-05-09T00:00:00 | [
[
"Frion",
"Anthony",
""
],
[
"Drumetz",
"Lucas",
""
],
[
"Tochon",
"Guillaume",
""
],
[
"Mura",
"Mauro Dalla",
""
],
[
"Bey",
"Abdeldjalil Aïssa El",
""
]
] | TITLE: Learning Sentinel-2 reflectance dynamics for data-driven assimilation
and forecasting
ABSTRACT: Over the last few years, massive amounts of satellite multispectral and
hyperspectral images covering the Earth's surface have been made publicly
available for scientific purpose, for example through the European Copernicus
project. Simultaneously, the development of self-supervised learning (SSL)
methods has sparked great interest in the remote sensing community, enabling to
learn latent representations from unlabeled data to help treating downstream
tasks for which there is few annotated examples, such as interpolation,
forecasting or unmixing. Following this line, we train a deep learning model
inspired from the Koopman operator theory to model long-term reflectance
dynamics in an unsupervised way. We show that this trained model, being
differentiable, can be used as a prior for data assimilation in a
straightforward way. Our datasets, which are composed of Sentinel-2
multispectral image time series, are publicly released with several levels of
treatment.
| new_dataset | 0.693512 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.