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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2304.13495 | Pol Jan\'e-Soneira | Pol Jan\'e-Soneira, Ionela Prodan, Albertus Johannes Malan, S\"oren
Hohmann | On MPC-based Strategies for Optimal Voltage References in DC Microgrids | null | null | null | null | eess.SY cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern power systems are characterized by low inertia and fast voltage
dynamics due to the increase of sources connecting via power electronics and
the removal of large traditional thermal generators. Power electronics are
commonly equipped with fast controllers that are able to reach a desired
voltage setpoint within seconds. In this paper, we propose and compare two
approaches using Model Predictive Control (MPC) to compute optimal voltage
references for the power electronic devices in order to minimize the losses in
a DC microgrid: i) a traditional setpoint-tracking MPC which receives a
previously computed optimal setpoint; ii) an economic MPC which does not
require a priori computed setpoints. We show that the economic MPC outperforms
the setpoint-tracking MPC in simulations with the CIGRE benchmark system when
multiple load disturbances occur. Some insights and discussions related to the
stability of the closed-loop system using its dissipativity properties are
highlighted for both approaches.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 12:39:46 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Jané-Soneira",
"Pol",
""
],
[
"Prodan",
"Ionela",
""
],
[
"Malan",
"Albertus Johannes",
""
],
[
"Hohmann",
"Sören",
""
]
] | TITLE: On MPC-based Strategies for Optimal Voltage References in DC Microgrids
ABSTRACT: Modern power systems are characterized by low inertia and fast voltage
dynamics due to the increase of sources connecting via power electronics and
the removal of large traditional thermal generators. Power electronics are
commonly equipped with fast controllers that are able to reach a desired
voltage setpoint within seconds. In this paper, we propose and compare two
approaches using Model Predictive Control (MPC) to compute optimal voltage
references for the power electronic devices in order to minimize the losses in
a DC microgrid: i) a traditional setpoint-tracking MPC which receives a
previously computed optimal setpoint; ii) an economic MPC which does not
require a priori computed setpoints. We show that the economic MPC outperforms
the setpoint-tracking MPC in simulations with the CIGRE benchmark system when
multiple load disturbances occur. Some insights and discussions related to the
stability of the closed-loop system using its dissipativity properties are
highlighted for both approaches.
|
2304.13506 | Ruida Zeng | Ruida Zeng, Aadarsh Jha, Ashwin Kumar, Terry Luo | Cloud-Based Deep Learning: End-To-End Full-Stack Handwritten Digit
Recognition | null | null | null | null | cs.DC cs.CV | http://creativecommons.org/licenses/by/4.0/ | Herein, we present Stratus, an end-to-end full-stack deep learning
application deployed on the cloud. The rise of productionized deep learning
necessitates infrastructure in the cloud that can provide such service (IaaS).
In this paper, we explore the use of modern cloud infrastructure and
micro-services to deliver accurate and high-speed predictions to an end-user,
using a Deep Neural Network (DNN) to predict handwritten digit input,
interfaced via a full-stack application. We survey tooling from Spark ML,
Apache Kafka, Chameleon Cloud, Ansible, Vagrant, Python Flask, Docker, and
Kubernetes in order to realize this machine learning pipeline. Through our
cloud-based approach, we are able to demonstrate benchmark performance on the
MNIST dataset with a deep learning model.
| [
{
"version": "v1",
"created": "Wed, 1 Feb 2023 05:40:28 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Zeng",
"Ruida",
""
],
[
"Jha",
"Aadarsh",
""
],
[
"Kumar",
"Ashwin",
""
],
[
"Luo",
"Terry",
""
]
] | TITLE: Cloud-Based Deep Learning: End-To-End Full-Stack Handwritten Digit
Recognition
ABSTRACT: Herein, we present Stratus, an end-to-end full-stack deep learning
application deployed on the cloud. The rise of productionized deep learning
necessitates infrastructure in the cloud that can provide such service (IaaS).
In this paper, we explore the use of modern cloud infrastructure and
micro-services to deliver accurate and high-speed predictions to an end-user,
using a Deep Neural Network (DNN) to predict handwritten digit input,
interfaced via a full-stack application. We survey tooling from Spark ML,
Apache Kafka, Chameleon Cloud, Ansible, Vagrant, Python Flask, Docker, and
Kubernetes in order to realize this machine learning pipeline. Through our
cloud-based approach, we are able to demonstrate benchmark performance on the
MNIST dataset with a deep learning model.
|
2304.13513 | Xiaoqing Liu | Xiaoqing Liu, Kengo Araki, Shota Harada, Akihiko Yoshizawa, Kazuhiro
Terada, Mariyo Kurata, Naoki Nakajima, Hiroyuki Abe, Tetsuo Ushiku, Ryoma
Bise | Cluster Entropy: Active Domain Adaptation in Pathological Image
Segmentation | Accepted by IEEE ISBI'23 | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The domain shift in pathological segmentation is an important problem, where
a network trained by a source domain (collected at a specific hospital) does
not work well in the target domain (from different hospitals) due to the
different image features. Due to the problems of class imbalance and different
class prior of pathology, typical unsupervised domain adaptation methods do not
work well by aligning the distribution of source domain and target domain. In
this paper, we propose a cluster entropy for selecting an effective whole slide
image (WSI) that is used for semi-supervised domain adaptation. This approach
can measure how the image features of the WSI cover the entire distribution of
the target domain by calculating the entropy of each cluster and can
significantly improve the performance of domain adaptation. Our approach
achieved competitive results against the prior arts on datasets collected from
two hospitals.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 12:53:41 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Liu",
"Xiaoqing",
""
],
[
"Araki",
"Kengo",
""
],
[
"Harada",
"Shota",
""
],
[
"Yoshizawa",
"Akihiko",
""
],
[
"Terada",
"Kazuhiro",
""
],
[
"Kurata",
"Mariyo",
""
],
[
"Nakajima",
"Naoki",
""
],
[
"Abe",
"Hiroyuki",
""
],
[
"Ushiku",
"Tetsuo",
""
],
[
"Bise",
"Ryoma",
""
]
] | TITLE: Cluster Entropy: Active Domain Adaptation in Pathological Image
Segmentation
ABSTRACT: The domain shift in pathological segmentation is an important problem, where
a network trained by a source domain (collected at a specific hospital) does
not work well in the target domain (from different hospitals) due to the
different image features. Due to the problems of class imbalance and different
class prior of pathology, typical unsupervised domain adaptation methods do not
work well by aligning the distribution of source domain and target domain. In
this paper, we propose a cluster entropy for selecting an effective whole slide
image (WSI) that is used for semi-supervised domain adaptation. This approach
can measure how the image features of the WSI cover the entire distribution of
the target domain by calculating the entropy of each cluster and can
significantly improve the performance of domain adaptation. Our approach
achieved competitive results against the prior arts on datasets collected from
two hospitals.
|
2304.13518 | Yuqi Han | Yuqi Han and Tao Yu and Xiaohang Yu and Yuwang Wang and Qionghai Dai | Super-NeRF: View-consistent Detail Generation for NeRF super-resolution | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The neural radiance field (NeRF) achieved remarkable success in modeling 3D
scenes and synthesizing high-fidelity novel views. However, existing NeRF-based
methods focus more on the make full use of the image resolution to generate
novel views, but less considering the generation of details under the limited
input resolution. In analogy to the extensive usage of image super-resolution,
NeRF super-resolution is an effective way to generate the high-resolution
implicit representation of 3D scenes and holds great potential applications. Up
to now, such an important topic is still under-explored. In this paper, we
propose a NeRF super-resolution method, named Super-NeRF, to generate
high-resolution NeRF from only low-resolution inputs. Given multi-view
low-resolution images, Super-NeRF constructs a consistency-controlling
super-resolution module to generate view-consistent high-resolution details for
NeRF. Specifically, an optimizable latent code is introduced for each
low-resolution input image to control the 2D super-resolution images to
converge to the view-consistent output. The latent codes of each low-resolution
image are optimized synergistically with the target Super-NeRF representation
to fully utilize the view consistency constraint inherent in NeRF construction.
We verify the effectiveness of Super-NeRF on synthetic, real-world, and
AI-generated NeRF datasets. Super-NeRF achieves state-of-the-art NeRF
super-resolution performance on high-resolution detail generation and
cross-view consistency.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 12:54:40 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Han",
"Yuqi",
""
],
[
"Yu",
"Tao",
""
],
[
"Yu",
"Xiaohang",
""
],
[
"Wang",
"Yuwang",
""
],
[
"Dai",
"Qionghai",
""
]
] | TITLE: Super-NeRF: View-consistent Detail Generation for NeRF super-resolution
ABSTRACT: The neural radiance field (NeRF) achieved remarkable success in modeling 3D
scenes and synthesizing high-fidelity novel views. However, existing NeRF-based
methods focus more on the make full use of the image resolution to generate
novel views, but less considering the generation of details under the limited
input resolution. In analogy to the extensive usage of image super-resolution,
NeRF super-resolution is an effective way to generate the high-resolution
implicit representation of 3D scenes and holds great potential applications. Up
to now, such an important topic is still under-explored. In this paper, we
propose a NeRF super-resolution method, named Super-NeRF, to generate
high-resolution NeRF from only low-resolution inputs. Given multi-view
low-resolution images, Super-NeRF constructs a consistency-controlling
super-resolution module to generate view-consistent high-resolution details for
NeRF. Specifically, an optimizable latent code is introduced for each
low-resolution input image to control the 2D super-resolution images to
converge to the view-consistent output. The latent codes of each low-resolution
image are optimized synergistically with the target Super-NeRF representation
to fully utilize the view consistency constraint inherent in NeRF construction.
We verify the effectiveness of Super-NeRF on synthetic, real-world, and
AI-generated NeRF datasets. Super-NeRF achieves state-of-the-art NeRF
super-resolution performance on high-resolution detail generation and
cross-view consistency.
|
2304.13530 | Sol\`ene Tarride | Sol\`ene Tarride and M\'elodie Boillet and Christopher Kermorvant | Key-value information extraction from full handwritten pages | null | null | null | null | cs.CV cs.AI cs.IR | http://creativecommons.org/licenses/by/4.0/ | We propose a Transformer-based approach for information extraction from
digitized handwritten documents. Our approach combines, in a single model, the
different steps that were so far performed by separate models: feature
extraction, handwriting recognition and named entity recognition. We compare
this integrated approach with traditional two-stage methods that perform
handwriting recognition before named entity recognition, and present results at
different levels: line, paragraph, and page. Our experiments show that
attention-based models are especially interesting when applied on full pages,
as they do not require any prior segmentation step. Finally, we show that they
are able to learn from key-value annotations: a list of important words with
their corresponding named entities. We compare our models to state-of-the-art
methods on three public databases (IAM, ESPOSALLES, and POPP) and outperform
previous performances on all three datasets.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 13:06:55 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Tarride",
"Solène",
""
],
[
"Boillet",
"Mélodie",
""
],
[
"Kermorvant",
"Christopher",
""
]
] | TITLE: Key-value information extraction from full handwritten pages
ABSTRACT: We propose a Transformer-based approach for information extraction from
digitized handwritten documents. Our approach combines, in a single model, the
different steps that were so far performed by separate models: feature
extraction, handwriting recognition and named entity recognition. We compare
this integrated approach with traditional two-stage methods that perform
handwriting recognition before named entity recognition, and present results at
different levels: line, paragraph, and page. Our experiments show that
attention-based models are especially interesting when applied on full pages,
as they do not require any prior segmentation step. Finally, we show that they
are able to learn from key-value annotations: a list of important words with
their corresponding named entities. We compare our models to state-of-the-art
methods on three public databases (IAM, ESPOSALLES, and POPP) and outperform
previous performances on all three datasets.
|
2304.13536 | Daniel G. Krakowczyk | Daniel G. Krakowczyk, Paul Prasse, David R. Reich, Sebastian
Lapuschkin, Tobias Scheffer, Lena A. J\"ager | Bridging the Gap: Gaze Events as Interpretable Concepts to Explain Deep
Neural Sequence Models | Preprint for ETRA '23: 2023 Symposium on Eye Tracking Research and
Applications | null | 10.1145/3588015.3588412 | null | cs.LG cs.HC | http://creativecommons.org/licenses/by/4.0/ | Recent work in XAI for eye tracking data has evaluated the suitability of
feature attribution methods to explain the output of deep neural sequence
models for the task of oculomotric biometric identification. These methods
provide saliency maps to highlight important input features of a specific eye
gaze sequence. However, to date, its localization analysis has been lacking a
quantitative approach across entire datasets. In this work, we employ
established gaze event detection algorithms for fixations and saccades and
quantitatively evaluate the impact of these events by determining their concept
influence. Input features that belong to saccades are shown to be substantially
more important than features that belong to fixations. By dissecting saccade
events into sub-events, we are able to show that gaze samples that are close to
the saccadic peak velocity are most influential. We further investigate the
effect of event properties like saccadic amplitude or fixational dispersion on
the resulting concept influence.
| [
{
"version": "v1",
"created": "Wed, 12 Apr 2023 10:15:31 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Krakowczyk",
"Daniel G.",
""
],
[
"Prasse",
"Paul",
""
],
[
"Reich",
"David R.",
""
],
[
"Lapuschkin",
"Sebastian",
""
],
[
"Scheffer",
"Tobias",
""
],
[
"Jäger",
"Lena A.",
""
]
] | TITLE: Bridging the Gap: Gaze Events as Interpretable Concepts to Explain Deep
Neural Sequence Models
ABSTRACT: Recent work in XAI for eye tracking data has evaluated the suitability of
feature attribution methods to explain the output of deep neural sequence
models for the task of oculomotric biometric identification. These methods
provide saliency maps to highlight important input features of a specific eye
gaze sequence. However, to date, its localization analysis has been lacking a
quantitative approach across entire datasets. In this work, we employ
established gaze event detection algorithms for fixations and saccades and
quantitatively evaluate the impact of these events by determining their concept
influence. Input features that belong to saccades are shown to be substantially
more important than features that belong to fixations. By dissecting saccade
events into sub-events, we are able to show that gaze samples that are close to
the saccadic peak velocity are most influential. We further investigate the
effect of event properties like saccadic amplitude or fixational dispersion on
the resulting concept influence.
|
2304.13549 | Abdelaziz Salama | Abdelaziz Salama, Syed Ali Zaidi, Des McLernon, Mohammed M. H. Qazzaz | FLCC: Efficient Distributed Federated Learning on IoMT over CSMA/CA | null | null | null | null | cs.DC cs.LG cs.SY eess.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Federated Learning (FL) has emerged as a promising approach for privacy
preservation, allowing sharing of the model parameters between users and the
cloud server rather than the raw local data. FL approaches have been adopted as
a cornerstone of distributed machine learning (ML) to solve several complex use
cases. FL presents an interesting interplay between communication and ML
performance when implemented over distributed wireless nodes. Both the dynamics
of networking and learning play an important role. In this article, we
investigate the performance of FL on an application that might be used to
improve a remote healthcare system over ad hoc networks which employ CSMA/CA to
schedule its transmissions. Our FL over CSMA/CA (FLCC) model is designed to
eliminate untrusted devices and harness frequency reuse and spatial clustering
techniques to improve the throughput required for coordinating a distributed
implementation of FL in the wireless network.
In our proposed model, frequency allocation is performed on the basis of
spatial clustering performed using virtual cells. Each cell assigns a FL server
and dedicated carrier frequencies to exchange the updated model's parameters
within the cell. We present two metrics to evaluate the network performance: 1)
probability of successful transmission while minimizing the interference, and
2) performance of distributed FL model in terms of accuracy and loss while
considering the networking dynamics.
We benchmark the proposed approach using a well-known MNIST dataset for
performance evaluation. We demonstrate that the proposed approach outperforms
the baseline FL algorithms in terms of explicitly defining the chosen users'
criteria and achieving high accuracy in a robust network.
| [
{
"version": "v1",
"created": "Wed, 29 Mar 2023 16:36:42 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Salama",
"Abdelaziz",
""
],
[
"Zaidi",
"Syed Ali",
""
],
[
"McLernon",
"Des",
""
],
[
"Qazzaz",
"Mohammed M. H.",
""
]
] | TITLE: FLCC: Efficient Distributed Federated Learning on IoMT over CSMA/CA
ABSTRACT: Federated Learning (FL) has emerged as a promising approach for privacy
preservation, allowing sharing of the model parameters between users and the
cloud server rather than the raw local data. FL approaches have been adopted as
a cornerstone of distributed machine learning (ML) to solve several complex use
cases. FL presents an interesting interplay between communication and ML
performance when implemented over distributed wireless nodes. Both the dynamics
of networking and learning play an important role. In this article, we
investigate the performance of FL on an application that might be used to
improve a remote healthcare system over ad hoc networks which employ CSMA/CA to
schedule its transmissions. Our FL over CSMA/CA (FLCC) model is designed to
eliminate untrusted devices and harness frequency reuse and spatial clustering
techniques to improve the throughput required for coordinating a distributed
implementation of FL in the wireless network.
In our proposed model, frequency allocation is performed on the basis of
spatial clustering performed using virtual cells. Each cell assigns a FL server
and dedicated carrier frequencies to exchange the updated model's parameters
within the cell. We present two metrics to evaluate the network performance: 1)
probability of successful transmission while minimizing the interference, and
2) performance of distributed FL model in terms of accuracy and loss while
considering the networking dynamics.
We benchmark the proposed approach using a well-known MNIST dataset for
performance evaluation. We demonstrate that the proposed approach outperforms
the baseline FL algorithms in terms of explicitly defining the chosen users'
criteria and achieving high accuracy in a robust network.
|
2304.13574 | Debayan Bhattacharya | Debayan Bhattacharya, Sarah Latus, Finn Behrendt, Florin Thimm, Dennis
Eggert, Christian Betz, Alexander Schlaefer | Tissue Classification During Needle Insertion Using Self-Supervised
Contrastive Learning and Optical Coherence Tomography | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | Needle positioning is essential for various medical applications such as
epidural anaesthesia. Physicians rely on their instincts while navigating the
needle in epidural spaces. Thereby, identifying the tissue structures may be
helpful to the physician as they can provide additional feedback in the needle
insertion process. To this end, we propose a deep neural network that
classifies the tissues from the phase and intensity data of complex OCT signals
acquired at the needle tip. We investigate the performance of the deep neural
network in a limited labelled dataset scenario and propose a novel contrastive
pretraining strategy that learns invariant representation for phase and
intensity data. We show that with 10% of the training set, our proposed
pretraining strategy helps the model achieve an F1 score of 0.84 whereas the
model achieves an F1 score of 0.60 without it. Further, we analyse the
importance of phase and intensity individually towards tissue classification.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 14:11:04 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Bhattacharya",
"Debayan",
""
],
[
"Latus",
"Sarah",
""
],
[
"Behrendt",
"Finn",
""
],
[
"Thimm",
"Florin",
""
],
[
"Eggert",
"Dennis",
""
],
[
"Betz",
"Christian",
""
],
[
"Schlaefer",
"Alexander",
""
]
] | TITLE: Tissue Classification During Needle Insertion Using Self-Supervised
Contrastive Learning and Optical Coherence Tomography
ABSTRACT: Needle positioning is essential for various medical applications such as
epidural anaesthesia. Physicians rely on their instincts while navigating the
needle in epidural spaces. Thereby, identifying the tissue structures may be
helpful to the physician as they can provide additional feedback in the needle
insertion process. To this end, we propose a deep neural network that
classifies the tissues from the phase and intensity data of complex OCT signals
acquired at the needle tip. We investigate the performance of the deep neural
network in a limited labelled dataset scenario and propose a novel contrastive
pretraining strategy that learns invariant representation for phase and
intensity data. We show that with 10% of the training set, our proposed
pretraining strategy helps the model achieve an F1 score of 0.84 whereas the
model achieves an F1 score of 0.60 without it. Further, we analyse the
importance of phase and intensity individually towards tissue classification.
|
2304.13594 | Andre Vauvelle | Andre Vauvelle, Benjamin Wild, Aylin Cakiroglu, Roland Eils, Spiros
Denaxas | Diffsurv: Differentiable sorting for censored time-to-event data | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Survival analysis is a crucial semi-supervised task in machine learning with
numerous real-world applications, particularly in healthcare. Currently, the
most common approach to survival analysis is based on Cox's partial likelihood,
which can be interpreted as a ranking model optimized on a lower bound of the
concordance index. This relation between ranking models and Cox's partial
likelihood considers only pairwise comparisons. Recent work has developed
differentiable sorting methods which relax this pairwise independence
assumption, enabling the ranking of sets of samples. However, current
differentiable sorting methods cannot account for censoring, a key factor in
many real-world datasets. To address this limitation, we propose a novel method
called Diffsurv. We extend differentiable sorting methods to handle censored
tasks by predicting matrices of possible permutations that take into account
the label uncertainty introduced by censored samples. We contrast this approach
with methods derived from partial likelihood and ranking losses. Our
experiments show that Diffsurv outperforms established baselines in various
simulated and real-world risk prediction scenarios. Additionally, we
demonstrate the benefits of the algorithmic supervision enabled by Diffsurv by
presenting a novel method for top-k risk prediction that outperforms current
methods.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 14:42:31 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Vauvelle",
"Andre",
""
],
[
"Wild",
"Benjamin",
""
],
[
"Cakiroglu",
"Aylin",
""
],
[
"Eils",
"Roland",
""
],
[
"Denaxas",
"Spiros",
""
]
] | TITLE: Diffsurv: Differentiable sorting for censored time-to-event data
ABSTRACT: Survival analysis is a crucial semi-supervised task in machine learning with
numerous real-world applications, particularly in healthcare. Currently, the
most common approach to survival analysis is based on Cox's partial likelihood,
which can be interpreted as a ranking model optimized on a lower bound of the
concordance index. This relation between ranking models and Cox's partial
likelihood considers only pairwise comparisons. Recent work has developed
differentiable sorting methods which relax this pairwise independence
assumption, enabling the ranking of sets of samples. However, current
differentiable sorting methods cannot account for censoring, a key factor in
many real-world datasets. To address this limitation, we propose a novel method
called Diffsurv. We extend differentiable sorting methods to handle censored
tasks by predicting matrices of possible permutations that take into account
the label uncertainty introduced by censored samples. We contrast this approach
with methods derived from partial likelihood and ranking losses. Our
experiments show that Diffsurv outperforms established baselines in various
simulated and real-world risk prediction scenarios. Additionally, we
demonstrate the benefits of the algorithmic supervision enabled by Diffsurv by
presenting a novel method for top-k risk prediction that outperforms current
methods.
|
2304.13606 | Sol\`ene Tarride | Sol\`ene Tarride and M\'elodie Boillet and Jean-Fran\c{c}ois Moufflet
and Christopher Kermorvant | SIMARA: a database for key-value information extraction from full pages | null | null | null | null | cs.CV cs.DB | http://creativecommons.org/licenses/by/4.0/ | We propose a new database for information extraction from historical
handwritten documents. The corpus includes 5,393 finding aids from six
different series, dating from the 18th-20th centuries. Finding aids are
handwritten documents that contain metadata describing older archives. They are
stored in the National Archives of France and are used by archivists to
identify and find archival documents. Each document is annotated at page-level,
and contains seven fields to retrieve. The localization of each field is not
available in such a way that this dataset encourages research on
segmentation-free systems for information extraction. We propose a model based
on the Transformer architecture trained for end-to-end information extraction
and provide three sets for training, validation and testing, to ensure fair
comparison with future works. The database is freely accessible at
https://zenodo.org/record/7868059.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 15:00:04 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Tarride",
"Solène",
""
],
[
"Boillet",
"Mélodie",
""
],
[
"Moufflet",
"Jean-François",
""
],
[
"Kermorvant",
"Christopher",
""
]
] | TITLE: SIMARA: a database for key-value information extraction from full pages
ABSTRACT: We propose a new database for information extraction from historical
handwritten documents. The corpus includes 5,393 finding aids from six
different series, dating from the 18th-20th centuries. Finding aids are
handwritten documents that contain metadata describing older archives. They are
stored in the National Archives of France and are used by archivists to
identify and find archival documents. Each document is annotated at page-level,
and contains seven fields to retrieve. The localization of each field is not
available in such a way that this dataset encourages research on
segmentation-free systems for information extraction. We propose a model based
on the Transformer architecture trained for end-to-end information extraction
and provide three sets for training, validation and testing, to ensure fair
comparison with future works. The database is freely accessible at
https://zenodo.org/record/7868059.
|
2304.13618 | Peng Liu | Peng Liu, Jonas Golde, Joseph Morgenstern, Sebastian Bodenstedt,
Chenpan Li, Yujia Hu, Zhaoyu Chen, Edmund Koch, Marcus Neudert, Stefanie
Speidel | Non-rigid Point Cloud Registration for Middle Ear Diagnostics with
Endoscopic Optical Coherence Tomography | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Purpose: Middle ear infection is the most prevalent inflammatory disease,
especially among the pediatric population. Current diagnostic methods are
subjective and depend on visual cues from an otoscope, which is limited for
otologists to identify pathology. To address this shortcoming, endoscopic
optical coherence tomography (OCT) provides both morphological and functional
in-vivo measurements of the middle ear. However, due to the shadow of prior
structures, interpretation of OCT images is challenging and time-consuming. To
facilitate fast diagnosis and measurement, improvement in the readability of
OCT data is achieved by merging morphological knowledge from ex-vivo middle ear
models with OCT volumetric data, so that OCT applications can be further
promoted in daily clinical settings. Methods: We propose C2P-Net: a two-staged
non-rigid registration pipeline for complete to partial point clouds, which are
sampled from ex-vivo and in-vivo OCT models, respectively. To overcome the lack
of labeled training data, a fast and effective generation pipeline in Blender3D
is designed to simulate middle ear shapes and extract in-vivo noisy and partial
point clouds. Results: We evaluate the performance of C2P-Net through
experiments on both synthetic and real OCT datasets. The results demonstrate
that C2P-Net is generalized to unseen middle ear point clouds and capable of
handling realistic noise and incompleteness in synthetic and real OCT data.
Conclusion: In this work, we aim to enable diagnosis of middle ear structures
with the assistance of OCT images. We propose C2P-Net: a two-staged non-rigid
registration pipeline for point clouds to support the interpretation of in-vivo
noisy and partial OCT images for the first time. Code is available at:
https://gitlab.com/nct\_tso\_public/c2p-net.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 15:20:48 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Liu",
"Peng",
""
],
[
"Golde",
"Jonas",
""
],
[
"Morgenstern",
"Joseph",
""
],
[
"Bodenstedt",
"Sebastian",
""
],
[
"Li",
"Chenpan",
""
],
[
"Hu",
"Yujia",
""
],
[
"Chen",
"Zhaoyu",
""
],
[
"Koch",
"Edmund",
""
],
[
"Neudert",
"Marcus",
""
],
[
"Speidel",
"Stefanie",
""
]
] | TITLE: Non-rigid Point Cloud Registration for Middle Ear Diagnostics with
Endoscopic Optical Coherence Tomography
ABSTRACT: Purpose: Middle ear infection is the most prevalent inflammatory disease,
especially among the pediatric population. Current diagnostic methods are
subjective and depend on visual cues from an otoscope, which is limited for
otologists to identify pathology. To address this shortcoming, endoscopic
optical coherence tomography (OCT) provides both morphological and functional
in-vivo measurements of the middle ear. However, due to the shadow of prior
structures, interpretation of OCT images is challenging and time-consuming. To
facilitate fast diagnosis and measurement, improvement in the readability of
OCT data is achieved by merging morphological knowledge from ex-vivo middle ear
models with OCT volumetric data, so that OCT applications can be further
promoted in daily clinical settings. Methods: We propose C2P-Net: a two-staged
non-rigid registration pipeline for complete to partial point clouds, which are
sampled from ex-vivo and in-vivo OCT models, respectively. To overcome the lack
of labeled training data, a fast and effective generation pipeline in Blender3D
is designed to simulate middle ear shapes and extract in-vivo noisy and partial
point clouds. Results: We evaluate the performance of C2P-Net through
experiments on both synthetic and real OCT datasets. The results demonstrate
that C2P-Net is generalized to unseen middle ear point clouds and capable of
handling realistic noise and incompleteness in synthetic and real OCT data.
Conclusion: In this work, we aim to enable diagnosis of middle ear structures
with the assistance of OCT images. We propose C2P-Net: a two-staged non-rigid
registration pipeline for point clouds to support the interpretation of in-vivo
noisy and partial OCT images for the first time. Code is available at:
https://gitlab.com/nct\_tso\_public/c2p-net.
|
2304.13634 | Saheed Abdullahi Salahudeen | Saheed Abdullahi Salahudeen, Falalu Ibrahim Lawan, Ahmad Mustapha
Wali, Amina Abubakar Imam, Aliyu Rabiu Shuaibu, Aliyu Yusuf, Nur Bala Rabiu,
Musa Bello, Shamsuddeen Umaru Adamu, Saminu Mohammad Aliyu, Murja Sani
Gadanya, Sanah Abdullahi Muaz, Mahmoud Said Ahmad, Abdulkadir Abdullahi,
Abdulmalik Yusuf Jamoh | HausaNLP at SemEval-2023 Task 12: Leveraging African Low Resource
TweetData for Sentiment Analysis | null | null | null | null | cs.CL | http://creativecommons.org/publicdomain/zero/1.0/ | We present the findings of SemEval-2023 Task 12, a shared task on sentiment
analysis for low-resource African languages using Twitter dataset. The task
featured three subtasks; subtask A is monolingual sentiment classification with
12 tracks which are all monolingual languages, subtask B is multilingual
sentiment classification using the tracks in subtask A and subtask C is a
zero-shot sentiment classification. We present the results and findings of
subtask A, subtask B and subtask C. We also release the code on github. Our
goal is to leverage low-resource tweet data using pre-trained Afro-xlmr-large,
AfriBERTa-Large, Bert-base-arabic-camelbert-da-sentiment (Arabic-camelbert),
Multilingual-BERT (mBERT) and BERT models for sentiment analysis of 14 African
languages. The datasets for these subtasks consists of a gold standard
multi-class labeled Twitter datasets from these languages. Our results
demonstrate that Afro-xlmr-large model performed better compared to the other
models in most of the languages datasets. Similarly, Nigerian languages: Hausa,
Igbo, and Yoruba achieved better performance compared to other languages and
this can be attributed to the higher volume of data present in the languages.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 15:47:50 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Salahudeen",
"Saheed Abdullahi",
""
],
[
"Lawan",
"Falalu Ibrahim",
""
],
[
"Wali",
"Ahmad Mustapha",
""
],
[
"Imam",
"Amina Abubakar",
""
],
[
"Shuaibu",
"Aliyu Rabiu",
""
],
[
"Yusuf",
"Aliyu",
""
],
[
"Rabiu",
"Nur Bala",
""
],
[
"Bello",
"Musa",
""
],
[
"Adamu",
"Shamsuddeen Umaru",
""
],
[
"Aliyu",
"Saminu Mohammad",
""
],
[
"Gadanya",
"Murja Sani",
""
],
[
"Muaz",
"Sanah Abdullahi",
""
],
[
"Ahmad",
"Mahmoud Said",
""
],
[
"Abdullahi",
"Abdulkadir",
""
],
[
"Jamoh",
"Abdulmalik Yusuf",
""
]
] | TITLE: HausaNLP at SemEval-2023 Task 12: Leveraging African Low Resource
TweetData for Sentiment Analysis
ABSTRACT: We present the findings of SemEval-2023 Task 12, a shared task on sentiment
analysis for low-resource African languages using Twitter dataset. The task
featured three subtasks; subtask A is monolingual sentiment classification with
12 tracks which are all monolingual languages, subtask B is multilingual
sentiment classification using the tracks in subtask A and subtask C is a
zero-shot sentiment classification. We present the results and findings of
subtask A, subtask B and subtask C. We also release the code on github. Our
goal is to leverage low-resource tweet data using pre-trained Afro-xlmr-large,
AfriBERTa-Large, Bert-base-arabic-camelbert-da-sentiment (Arabic-camelbert),
Multilingual-BERT (mBERT) and BERT models for sentiment analysis of 14 African
languages. The datasets for these subtasks consists of a gold standard
multi-class labeled Twitter datasets from these languages. Our results
demonstrate that Afro-xlmr-large model performed better compared to the other
models in most of the languages datasets. Similarly, Nigerian languages: Hausa,
Igbo, and Yoruba achieved better performance compared to other languages and
this can be attributed to the higher volume of data present in the languages.
|
2304.13638 | Rahul Kumar Gupta | Rahul Gupta, Mario Paolone | Experimental Validation of Model-less Robust Voltage Control using
Measurement-based Estimated Voltage Sensitivity Coefficients | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Increasing adoption of smart meters and phasor measurement units (PMUs) in
power distribution networks are enabling the adoption of data-driven/model-less
control schemes to mitigate grid issues such as over/under voltages and
power-flow congestions. However, such a scheme can lead to
infeasible/inaccurate control decisions due to measurement inaccuracies. In
this context, the authors' previous work proposed a robust measurement-based
control scheme accounting for the uncertainties of the estimated models. In
this scheme, a recursive least squares (RLS)-based method estimates the grid
model (in the form of voltage magnitude sensitivity coefficients). Then, a
robust control problem optimizes power set-points of distributed energy
resources (DERs) such that the nodal voltage limits are satisfied. The
estimated voltage sensitivity coefficients are used to model the nodal
voltages, and the control robustness is achieved by accounting for their
uncertainties. This work presents the first experimental validation of such a
robust model-less control scheme on a real power distribution grid. The scheme
is applied for voltage control by regulating two photovoltaic (PV) inverters
connected in a real microgrid which is a replica of the CIGRE benchmark
microgrid network at the EPFL Distributed Electrical Systems Laboratory.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 15:54:39 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Gupta",
"Rahul",
""
],
[
"Paolone",
"Mario",
""
]
] | TITLE: Experimental Validation of Model-less Robust Voltage Control using
Measurement-based Estimated Voltage Sensitivity Coefficients
ABSTRACT: Increasing adoption of smart meters and phasor measurement units (PMUs) in
power distribution networks are enabling the adoption of data-driven/model-less
control schemes to mitigate grid issues such as over/under voltages and
power-flow congestions. However, such a scheme can lead to
infeasible/inaccurate control decisions due to measurement inaccuracies. In
this context, the authors' previous work proposed a robust measurement-based
control scheme accounting for the uncertainties of the estimated models. In
this scheme, a recursive least squares (RLS)-based method estimates the grid
model (in the form of voltage magnitude sensitivity coefficients). Then, a
robust control problem optimizes power set-points of distributed energy
resources (DERs) such that the nodal voltage limits are satisfied. The
estimated voltage sensitivity coefficients are used to model the nodal
voltages, and the control robustness is achieved by accounting for their
uncertainties. This work presents the first experimental validation of such a
robust model-less control scheme on a real power distribution grid. The scheme
is applied for voltage control by regulating two photovoltaic (PV) inverters
connected in a real microgrid which is a replica of the CIGRE benchmark
microgrid network at the EPFL Distributed Electrical Systems Laboratory.
|
2304.13639 | Qingyong Hu | Zhao Song, Ke Yang, Naiyang Guan, Junjie Zhu, Peng Qiao, Qingyong Hu | PVP: Pre-trained Visual Parameter-Efficient Tuning | null | null | null | null | cs.CV cs.AI cs.LG eess.IV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Large-scale pre-trained transformers have demonstrated remarkable success in
various computer vision tasks. However, it is still highly challenging to fully
fine-tune these models for downstream tasks due to their high computational and
storage costs. Recently, Parameter-Efficient Tuning (PETuning) techniques,
e.g., Visual Prompt Tuning (VPT) and Low-Rank Adaptation (LoRA), have
significantly reduced the computation and storage cost by inserting lightweight
prompt modules into the pre-trained models and tuning these prompt modules with
a small number of trainable parameters, while keeping the transformer backbone
frozen. Although only a few parameters need to be adjusted, most PETuning
methods still require a significant amount of downstream task training data to
achieve good results. The performance is inadequate on low-data regimes,
especially when there are only one or two examples per class. To this end, we
first empirically identify the poor performance is mainly due to the
inappropriate way of initializing prompt modules, which has also been verified
in the pre-trained language models. Next, we propose a Pre-trained Visual
Parameter-efficient (PVP) Tuning framework, which pre-trains the
parameter-efficient tuning modules first and then leverages the pre-trained
modules along with the pre-trained transformer backbone to perform
parameter-efficient tuning on downstream tasks. Experiment results on five
Fine-Grained Visual Classification (FGVC) and VTAB-1k datasets demonstrate that
our proposed method significantly outperforms state-of-the-art PETuning
methods.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 15:55:29 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Song",
"Zhao",
""
],
[
"Yang",
"Ke",
""
],
[
"Guan",
"Naiyang",
""
],
[
"Zhu",
"Junjie",
""
],
[
"Qiao",
"Peng",
""
],
[
"Hu",
"Qingyong",
""
]
] | TITLE: PVP: Pre-trained Visual Parameter-Efficient Tuning
ABSTRACT: Large-scale pre-trained transformers have demonstrated remarkable success in
various computer vision tasks. However, it is still highly challenging to fully
fine-tune these models for downstream tasks due to their high computational and
storage costs. Recently, Parameter-Efficient Tuning (PETuning) techniques,
e.g., Visual Prompt Tuning (VPT) and Low-Rank Adaptation (LoRA), have
significantly reduced the computation and storage cost by inserting lightweight
prompt modules into the pre-trained models and tuning these prompt modules with
a small number of trainable parameters, while keeping the transformer backbone
frozen. Although only a few parameters need to be adjusted, most PETuning
methods still require a significant amount of downstream task training data to
achieve good results. The performance is inadequate on low-data regimes,
especially when there are only one or two examples per class. To this end, we
first empirically identify the poor performance is mainly due to the
inappropriate way of initializing prompt modules, which has also been verified
in the pre-trained language models. Next, we propose a Pre-trained Visual
Parameter-efficient (PVP) Tuning framework, which pre-trains the
parameter-efficient tuning modules first and then leverages the pre-trained
modules along with the pre-trained transformer backbone to perform
parameter-efficient tuning on downstream tasks. Experiment results on five
Fine-Grained Visual Classification (FGVC) and VTAB-1k datasets demonstrate that
our proposed method significantly outperforms state-of-the-art PETuning
methods.
|
2304.13649 | Alireza Salemi | Alireza Salemi, Juan Altmayer Pizzorno, Hamed Zamani | A Symmetric Dual Encoding Dense Retrieval Framework for
Knowledge-Intensive Visual Question Answering | null | null | null | null | cs.CV cs.CL cs.IR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Knowledge-Intensive Visual Question Answering (KI-VQA) refers to answering a
question about an image whose answer does not lie in the image. This paper
presents a new pipeline for KI-VQA tasks, consisting of a retriever and a
reader. First, we introduce DEDR, a symmetric dual encoding dense retrieval
framework in which documents and queries are encoded into a shared embedding
space using uni-modal (textual) and multi-modal encoders. We introduce an
iterative knowledge distillation approach that bridges the gap between the
representation spaces in these two encoders. Extensive evaluation on two
well-established KI-VQA datasets, i.e., OK-VQA and FVQA, suggests that DEDR
outperforms state-of-the-art baselines by 11.6% and 30.9% on OK-VQA and FVQA,
respectively. Utilizing the passages retrieved by DEDR, we further introduce
MM-FiD, an encoder-decoder multi-modal fusion-in-decoder model, for generating
a textual answer for KI-VQA tasks. MM-FiD encodes the question, the image, and
each retrieved passage separately and uses all passages jointly in its decoder.
Compared to competitive baselines in the literature, this approach leads to
5.5% and 8.5% improvements in terms of question answering accuracy on OK-VQA
and FVQA, respectively.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 16:14:39 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Salemi",
"Alireza",
""
],
[
"Pizzorno",
"Juan Altmayer",
""
],
[
"Zamani",
"Hamed",
""
]
] | TITLE: A Symmetric Dual Encoding Dense Retrieval Framework for
Knowledge-Intensive Visual Question Answering
ABSTRACT: Knowledge-Intensive Visual Question Answering (KI-VQA) refers to answering a
question about an image whose answer does not lie in the image. This paper
presents a new pipeline for KI-VQA tasks, consisting of a retriever and a
reader. First, we introduce DEDR, a symmetric dual encoding dense retrieval
framework in which documents and queries are encoded into a shared embedding
space using uni-modal (textual) and multi-modal encoders. We introduce an
iterative knowledge distillation approach that bridges the gap between the
representation spaces in these two encoders. Extensive evaluation on two
well-established KI-VQA datasets, i.e., OK-VQA and FVQA, suggests that DEDR
outperforms state-of-the-art baselines by 11.6% and 30.9% on OK-VQA and FVQA,
respectively. Utilizing the passages retrieved by DEDR, we further introduce
MM-FiD, an encoder-decoder multi-modal fusion-in-decoder model, for generating
a textual answer for KI-VQA tasks. MM-FiD encodes the question, the image, and
each retrieved passage separately and uses all passages jointly in its decoder.
Compared to competitive baselines in the literature, this approach leads to
5.5% and 8.5% improvements in terms of question answering accuracy on OK-VQA
and FVQA, respectively.
|
2304.13651 | Zitian Tang | Zitian Tang, Wenjie Ye, Wei-Chiu Ma, Hang Zhao | What Happened 3 Seconds Ago? Inferring the Past with Thermal Imaging | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Inferring past human motion from RGB images is challenging due to the
inherent uncertainty of the prediction problem. Thermal images, on the other
hand, encode traces of past human-object interactions left in the environment
via thermal radiation measurement. Based on this observation, we collect the
first RGB-Thermal dataset for human motion analysis, dubbed Thermal-IM. Then we
develop a three-stage neural network model for accurate past human pose
estimation. Comprehensive experiments show that thermal cues significantly
reduce the ambiguities of this task, and the proposed model achieves remarkable
performance. The dataset is available at
https://github.com/ZitianTang/Thermal-IM.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 16:23:10 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Tang",
"Zitian",
""
],
[
"Ye",
"Wenjie",
""
],
[
"Ma",
"Wei-Chiu",
""
],
[
"Zhao",
"Hang",
""
]
] | TITLE: What Happened 3 Seconds Ago? Inferring the Past with Thermal Imaging
ABSTRACT: Inferring past human motion from RGB images is challenging due to the
inherent uncertainty of the prediction problem. Thermal images, on the other
hand, encode traces of past human-object interactions left in the environment
via thermal radiation measurement. Based on this observation, we collect the
first RGB-Thermal dataset for human motion analysis, dubbed Thermal-IM. Then we
develop a three-stage neural network model for accurate past human pose
estimation. Comprehensive experiments show that thermal cues significantly
reduce the ambiguities of this task, and the proposed model achieves remarkable
performance. The dataset is available at
https://github.com/ZitianTang/Thermal-IM.
|
2304.13677 | Ping Li | Xinyi Zheng, Weijie Zhao, Xiaoyun Li, Ping Li | Building K-Anonymous User Cohorts with\\ Consecutive Consistent Weighted
Sampling (CCWS) | null | null | null | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | To retrieve personalized campaigns and creatives while protecting user
privacy, digital advertising is shifting from member-based identity to
cohort-based identity. Under such identity regime, an accurate and efficient
cohort building algorithm is desired to group users with similar
characteristics. In this paper, we propose a scalable $K$-anonymous cohort
building algorithm called {\em consecutive consistent weighted sampling}
(CCWS). The proposed method combines the spirit of the ($p$-powered) consistent
weighted sampling and hierarchical clustering, so that the $K$-anonymity is
ensured by enforcing a lower bound on the size of cohorts. Evaluations on a
LinkedIn dataset consisting of $>70$M users and ads campaigns demonstrate that
CCWS achieves substantial improvements over several hashing-based methods
including sign random projections (SignRP), minwise hashing (MinHash), as well
as the vanilla CWS.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 16:46:01 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Zheng",
"Xinyi",
""
],
[
"Zhao",
"Weijie",
""
],
[
"Li",
"Xiaoyun",
""
],
[
"Li",
"Ping",
""
]
] | TITLE: Building K-Anonymous User Cohorts with\\ Consecutive Consistent Weighted
Sampling (CCWS)
ABSTRACT: To retrieve personalized campaigns and creatives while protecting user
privacy, digital advertising is shifting from member-based identity to
cohort-based identity. Under such identity regime, an accurate and efficient
cohort building algorithm is desired to group users with similar
characteristics. In this paper, we propose a scalable $K$-anonymous cohort
building algorithm called {\em consecutive consistent weighted sampling}
(CCWS). The proposed method combines the spirit of the ($p$-powered) consistent
weighted sampling and hierarchical clustering, so that the $K$-anonymity is
ensured by enforcing a lower bound on the size of cohorts. Evaluations on a
LinkedIn dataset consisting of $>70$M users and ads campaigns demonstrate that
CCWS achieves substantial improvements over several hashing-based methods
including sign random projections (SignRP), minwise hashing (MinHash), as well
as the vanilla CWS.
|
2304.13717 | Theophile Berteloot | Th\'eophile Berteloot, Richard Khoury, Audrey Durand | Association Rules Mining with Auto-Encoders | null | null | null | null | cs.LG cs.DB | http://creativecommons.org/licenses/by/4.0/ | Association rule mining is one of the most studied research fields of data
mining, with applications ranging from grocery basket problems to explainable
classification systems. Classical association rule mining algorithms have
several limitations, especially with regards to their high execution times and
number of rules produced. Over the past decade, neural network solutions have
been used to solve various optimization problems, such as classification,
regression or clustering. However there are still no efficient way association
rules using neural networks. In this paper, we present an auto-encoder solution
to mine association rule called ARM-AE. We compare our algorithm to FP-Growth
and NSGAII on three categorical datasets, and show that our algorithm discovers
high support and confidence rule set and has a better execution time than
classical methods while preserving the quality of the rule set produced.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 17:55:44 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Berteloot",
"Théophile",
""
],
[
"Khoury",
"Richard",
""
],
[
"Durand",
"Audrey",
""
]
] | TITLE: Association Rules Mining with Auto-Encoders
ABSTRACT: Association rule mining is one of the most studied research fields of data
mining, with applications ranging from grocery basket problems to explainable
classification systems. Classical association rule mining algorithms have
several limitations, especially with regards to their high execution times and
number of rules produced. Over the past decade, neural network solutions have
been used to solve various optimization problems, such as classification,
regression or clustering. However there are still no efficient way association
rules using neural networks. In this paper, we present an auto-encoder solution
to mine association rule called ARM-AE. We compare our algorithm to FP-Growth
and NSGAII on three categorical datasets, and show that our algorithm discovers
high support and confidence rule set and has a better execution time than
classical methods while preserving the quality of the rule set produced.
|
2304.13722 | Arantxa Casanova | Arantxa Casanova, Marl\`ene Careil, Adriana Romero-Soriano,
Christopher J. Pal, Jakob Verbeek, Michal Drozdzal | Controllable Image Generation via Collage Representations | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in conditional generative image models have enabled
impressive results. On the one hand, text-based conditional models have
achieved remarkable generation quality, by leveraging large-scale datasets of
image-text pairs. To enable fine-grained controllability, however, text-based
models require long prompts, whose details may be ignored by the model. On the
other hand, layout-based conditional models have also witnessed significant
advances. These models rely on bounding boxes or segmentation maps for precise
spatial conditioning in combination with coarse semantic labels. The semantic
labels, however, cannot be used to express detailed appearance characteristics.
In this paper, we approach fine-grained scene controllability through image
collages which allow a rich visual description of the desired scene as well as
the appearance and location of the objects therein, without the need of class
nor attribute labels. We introduce "mixing and matching scenes" (M&Ms), an
approach that consists of an adversarially trained generative image model which
is conditioned on appearance features and spatial positions of the different
elements in a collage, and integrates these into a coherent image. We train our
model on the OpenImages (OI) dataset and evaluate it on collages derived from
OI and MS-COCO datasets. Our experiments on the OI dataset show that M&Ms
outperforms baselines in terms of fine-grained scene controllability while
being very competitive in terms of image quality and sample diversity. On the
MS-COCO dataset, we highlight the generalization ability of our model by
outperforming DALL-E in terms of the zero-shot FID metric, despite using two
magnitudes fewer parameters and data. Collage based generative models have the
potential to advance content creation in an efficient and effective way as they
are intuitive to use and yield high quality generations.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 17:58:39 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Casanova",
"Arantxa",
""
],
[
"Careil",
"Marlène",
""
],
[
"Romero-Soriano",
"Adriana",
""
],
[
"Pal",
"Christopher J.",
""
],
[
"Verbeek",
"Jakob",
""
],
[
"Drozdzal",
"Michal",
""
]
] | TITLE: Controllable Image Generation via Collage Representations
ABSTRACT: Recent advances in conditional generative image models have enabled
impressive results. On the one hand, text-based conditional models have
achieved remarkable generation quality, by leveraging large-scale datasets of
image-text pairs. To enable fine-grained controllability, however, text-based
models require long prompts, whose details may be ignored by the model. On the
other hand, layout-based conditional models have also witnessed significant
advances. These models rely on bounding boxes or segmentation maps for precise
spatial conditioning in combination with coarse semantic labels. The semantic
labels, however, cannot be used to express detailed appearance characteristics.
In this paper, we approach fine-grained scene controllability through image
collages which allow a rich visual description of the desired scene as well as
the appearance and location of the objects therein, without the need of class
nor attribute labels. We introduce "mixing and matching scenes" (M&Ms), an
approach that consists of an adversarially trained generative image model which
is conditioned on appearance features and spatial positions of the different
elements in a collage, and integrates these into a coherent image. We train our
model on the OpenImages (OI) dataset and evaluate it on collages derived from
OI and MS-COCO datasets. Our experiments on the OI dataset show that M&Ms
outperforms baselines in terms of fine-grained scene controllability while
being very competitive in terms of image quality and sample diversity. On the
MS-COCO dataset, we highlight the generalization ability of our model by
outperforming DALL-E in terms of the zero-shot FID metric, despite using two
magnitudes fewer parameters and data. Collage based generative models have the
potential to advance content creation in an efficient and effective way as they
are intuitive to use and yield high quality generations.
|
2304.13723 | Stephen Tian | Stephen Tian, Chelsea Finn, Jiajun Wu | A Control-Centric Benchmark for Video Prediction | ICLR 2023 | null | null | null | cs.CV cs.AI cs.LG cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video is a promising source of knowledge for embodied agents to learn models
of the world's dynamics. Large deep networks have become increasingly effective
at modeling complex video data in a self-supervised manner, as evaluated by
metrics based on human perceptual similarity or pixel-wise comparison. However,
it remains unclear whether current metrics are accurate indicators of
performance on downstream tasks. We find empirically that for planning robotic
manipulation, existing metrics can be unreliable at predicting execution
success. To address this, we propose a benchmark for action-conditioned video
prediction in the form of a control benchmark that evaluates a given model for
simulated robotic manipulation through sampling-based planning. Our benchmark,
Video Prediction for Visual Planning ($VP^2$), includes simulated environments
with 11 task categories and 310 task instance definitions, a full planning
implementation, and training datasets containing scripted interaction
trajectories for each task category. A central design goal of our benchmark is
to expose a simple interface -- a single forward prediction call -- so it is
straightforward to evaluate almost any action-conditioned video prediction
model. We then leverage our benchmark to study the effects of scaling model
size, quantity of training data, and model ensembling by analyzing five
highly-performant video prediction models, finding that while scale can improve
perceptual quality when modeling visually diverse settings, other attributes
such as uncertainty awareness can also aid planning performance.
| [
{
"version": "v1",
"created": "Wed, 26 Apr 2023 17:59:45 GMT"
}
] | 2023-04-27T00:00:00 | [
[
"Tian",
"Stephen",
""
],
[
"Finn",
"Chelsea",
""
],
[
"Wu",
"Jiajun",
""
]
] | TITLE: A Control-Centric Benchmark for Video Prediction
ABSTRACT: Video is a promising source of knowledge for embodied agents to learn models
of the world's dynamics. Large deep networks have become increasingly effective
at modeling complex video data in a self-supervised manner, as evaluated by
metrics based on human perceptual similarity or pixel-wise comparison. However,
it remains unclear whether current metrics are accurate indicators of
performance on downstream tasks. We find empirically that for planning robotic
manipulation, existing metrics can be unreliable at predicting execution
success. To address this, we propose a benchmark for action-conditioned video
prediction in the form of a control benchmark that evaluates a given model for
simulated robotic manipulation through sampling-based planning. Our benchmark,
Video Prediction for Visual Planning ($VP^2$), includes simulated environments
with 11 task categories and 310 task instance definitions, a full planning
implementation, and training datasets containing scripted interaction
trajectories for each task category. A central design goal of our benchmark is
to expose a simple interface -- a single forward prediction call -- so it is
straightforward to evaluate almost any action-conditioned video prediction
model. We then leverage our benchmark to study the effects of scaling model
size, quantity of training data, and model ensembling by analyzing five
highly-performant video prediction models, finding that while scale can improve
perceptual quality when modeling visually diverse settings, other attributes
such as uncertainty awareness can also aid planning performance.
|
1903.11406 | Hung Nghiep Tran | Hung Nghiep Tran, Atsuhiro Takasu | Analyzing Knowledge Graph Embedding Methods from a Multi-Embedding
Interaction Perspective | DSI4 at EDBT/ICDT 2019; add source code at
https://github.com/tranhungnghiep/AnalyzeKGE | Data Science for Industry 4.0 at EDBT/ICDT 2019 | null | null | cs.LG cs.AI stat.ML | http://creativecommons.org/licenses/by/4.0/ | Knowledge graph is a popular format for representing knowledge, with many
applications to semantic search engines, question-answering systems, and
recommender systems. Real-world knowledge graphs are usually incomplete, so
knowledge graph embedding methods, such as Canonical decomposition/Parallel
factorization (CP), DistMult, and ComplEx, have been proposed to address this
issue. These methods represent entities and relations as embedding vectors in
semantic space and predict the links between them. The embedding vectors
themselves contain rich semantic information and can be used in other
applications such as data analysis. However, mechanisms in these models and the
embedding vectors themselves vary greatly, making it difficult to understand
and compare them. Given this lack of understanding, we risk using them
ineffectively or incorrectly, particularly for complicated models, such as CP,
with two role-based embedding vectors, or the state-of-the-art ComplEx model,
with complex-valued embedding vectors. In this paper, we propose a
multi-embedding interaction mechanism as a new approach to uniting and
generalizing these models. We derive them theoretically via this mechanism and
provide empirical analyses and comparisons between them. We also propose a new
multi-embedding model based on quaternion algebra and show that it achieves
promising results using popular benchmarks. Source code is available on GitHub
at https://github.com/tranhungnghiep/AnalyzeKGE.
| [
{
"version": "v1",
"created": "Wed, 27 Mar 2019 13:09:16 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Oct 2019 04:34:16 GMT"
},
{
"version": "v3",
"created": "Thu, 14 May 2020 19:58:51 GMT"
},
{
"version": "v4",
"created": "Tue, 25 Apr 2023 07:53:02 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Tran",
"Hung Nghiep",
""
],
[
"Takasu",
"Atsuhiro",
""
]
] | TITLE: Analyzing Knowledge Graph Embedding Methods from a Multi-Embedding
Interaction Perspective
ABSTRACT: Knowledge graph is a popular format for representing knowledge, with many
applications to semantic search engines, question-answering systems, and
recommender systems. Real-world knowledge graphs are usually incomplete, so
knowledge graph embedding methods, such as Canonical decomposition/Parallel
factorization (CP), DistMult, and ComplEx, have been proposed to address this
issue. These methods represent entities and relations as embedding vectors in
semantic space and predict the links between them. The embedding vectors
themselves contain rich semantic information and can be used in other
applications such as data analysis. However, mechanisms in these models and the
embedding vectors themselves vary greatly, making it difficult to understand
and compare them. Given this lack of understanding, we risk using them
ineffectively or incorrectly, particularly for complicated models, such as CP,
with two role-based embedding vectors, or the state-of-the-art ComplEx model,
with complex-valued embedding vectors. In this paper, we propose a
multi-embedding interaction mechanism as a new approach to uniting and
generalizing these models. We derive them theoretically via this mechanism and
provide empirical analyses and comparisons between them. We also propose a new
multi-embedding model based on quaternion algebra and show that it achieves
promising results using popular benchmarks. Source code is available on GitHub
at https://github.com/tranhungnghiep/AnalyzeKGE.
|
2012.13873 | Hao Zhang | Fuzhao Xue, Aixin Sun, Hao Zhang, Jinjie Ni, Eng Siong Chng | An Embarrassingly Simple Model for Dialogue Relation Extraction | Accepted by ICASSP 2022 | null | 10.1109/ICASSP43922.2022.9747486 | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dialogue relation extraction (RE) is to predict the relation type of two
entities mentioned in a dialogue. In this paper, we propose a simple yet
effective model named SimpleRE for the RE task. SimpleRE captures the
interrelations among multiple relations in a dialogue through a novel input
format named BERT Relation Token Sequence (BRS). In BRS, multiple [CLS] tokens
are used to capture possible relations between different pairs of entities
mentioned in the dialogue. A Relation Refinement Gate (RRG) is then designed to
extract relation-specific semantic representation in an adaptive manner.
Experiments on the DialogRE dataset show that SimpleRE achieves the best
performance, with much shorter training time. Further, SimpleRE outperforms all
direct baselines on sentence-level RE without using external resources.
| [
{
"version": "v1",
"created": "Sun, 27 Dec 2020 06:22:23 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Jan 2022 04:17:02 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Xue",
"Fuzhao",
""
],
[
"Sun",
"Aixin",
""
],
[
"Zhang",
"Hao",
""
],
[
"Ni",
"Jinjie",
""
],
[
"Chng",
"Eng Siong",
""
]
] | TITLE: An Embarrassingly Simple Model for Dialogue Relation Extraction
ABSTRACT: Dialogue relation extraction (RE) is to predict the relation type of two
entities mentioned in a dialogue. In this paper, we propose a simple yet
effective model named SimpleRE for the RE task. SimpleRE captures the
interrelations among multiple relations in a dialogue through a novel input
format named BERT Relation Token Sequence (BRS). In BRS, multiple [CLS] tokens
are used to capture possible relations between different pairs of entities
mentioned in the dialogue. A Relation Refinement Gate (RRG) is then designed to
extract relation-specific semantic representation in an adaptive manner.
Experiments on the DialogRE dataset show that SimpleRE achieves the best
performance, with much shorter training time. Further, SimpleRE outperforms all
direct baselines on sentence-level RE without using external resources.
|
2103.07302 | Jingchao Peng | Peng Jingchao, Zhao Haitao, Hu Zhengwei, Zhuang Yi, Wang Bofan | Siamese Infrared and Visible Light Fusion Network for RGB-T Tracking | null | null | 10.1007/s13042-023-01833-6 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Due to the different photosensitive properties of infrared and visible light,
the registered RGB-T image pairs shot in the same scene exhibit quite different
characteristics. This paper proposes a siamese infrared and visible light
fusion Network (SiamIVFN) for RBG-T image-based tracking. SiamIVFN contains two
main subnetworks: a complementary-feature-fusion network (CFFN) and a
contribution-aggregation network (CAN). CFFN utilizes a two-stream multilayer
convolutional structure whose filters for each layer are partially coupled to
fuse the features extracted from infrared images and visible light images. CFFN
is a feature-level fusion network, which can cope with the misalignment of the
RGB-T image pairs. Through adaptively calculating the contributions of infrared
and visible light features obtained from CFFN, CAN makes the tracker robust
under various light conditions. Experiments on two RGB-T tracking benchmark
datasets demonstrate that the proposed SiamIVFN has achieved state-of-the-art
performance. The tracking speed of SiamIVFN is 147.6FPS, the current fastest
RGB-T fusion tracker.
| [
{
"version": "v1",
"created": "Fri, 12 Mar 2021 14:20:41 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Jingchao",
"Peng",
""
],
[
"Haitao",
"Zhao",
""
],
[
"Zhengwei",
"Hu",
""
],
[
"Yi",
"Zhuang",
""
],
[
"Bofan",
"Wang",
""
]
] | TITLE: Siamese Infrared and Visible Light Fusion Network for RGB-T Tracking
ABSTRACT: Due to the different photosensitive properties of infrared and visible light,
the registered RGB-T image pairs shot in the same scene exhibit quite different
characteristics. This paper proposes a siamese infrared and visible light
fusion Network (SiamIVFN) for RBG-T image-based tracking. SiamIVFN contains two
main subnetworks: a complementary-feature-fusion network (CFFN) and a
contribution-aggregation network (CAN). CFFN utilizes a two-stream multilayer
convolutional structure whose filters for each layer are partially coupled to
fuse the features extracted from infrared images and visible light images. CFFN
is a feature-level fusion network, which can cope with the misalignment of the
RGB-T image pairs. Through adaptively calculating the contributions of infrared
and visible light features obtained from CFFN, CAN makes the tracker robust
under various light conditions. Experiments on two RGB-T tracking benchmark
datasets demonstrate that the proposed SiamIVFN has achieved state-of-the-art
performance. The tracking speed of SiamIVFN is 147.6FPS, the current fastest
RGB-T fusion tracker.
|
2105.08481 | Hao Zhang | Hao Zhang, Aixin Sun, Wei Jing, Liangli Zhen, Joey Tianyi Zhou, Rick
Siow Mong Goh | Parallel Attention Network with Sequence Matching for Video Grounding | 15 pages, 10 figures, 7 tables, Findings at ACL 2021 | null | 10.18653/v1/2021.findings-acl.69 | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a video, video grounding aims to retrieve a temporal moment that
semantically corresponds to a language query. In this work, we propose a
Parallel Attention Network with Sequence matching (SeqPAN) to address the
challenges in this task: multi-modal representation learning, and target moment
boundary prediction. We design a self-guided parallel attention module to
effectively capture self-modal contexts and cross-modal attentive information
between video and text. Inspired by sequence labeling tasks in natural language
processing, we split the ground truth moment into begin, inside, and end
regions. We then propose a sequence matching strategy to guide start/end
boundary predictions using region labels. Experimental results on three
datasets show that SeqPAN is superior to state-of-the-art methods. Furthermore,
the effectiveness of the self-guided parallel attention module and the sequence
matching module is verified.
| [
{
"version": "v1",
"created": "Tue, 18 May 2021 12:43:20 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Zhang",
"Hao",
""
],
[
"Sun",
"Aixin",
""
],
[
"Jing",
"Wei",
""
],
[
"Zhen",
"Liangli",
""
],
[
"Zhou",
"Joey Tianyi",
""
],
[
"Goh",
"Rick Siow Mong",
""
]
] | TITLE: Parallel Attention Network with Sequence Matching for Video Grounding
ABSTRACT: Given a video, video grounding aims to retrieve a temporal moment that
semantically corresponds to a language query. In this work, we propose a
Parallel Attention Network with Sequence matching (SeqPAN) to address the
challenges in this task: multi-modal representation learning, and target moment
boundary prediction. We design a self-guided parallel attention module to
effectively capture self-modal contexts and cross-modal attentive information
between video and text. Inspired by sequence labeling tasks in natural language
processing, we split the ground truth moment into begin, inside, and end
regions. We then propose a sequence matching strategy to guide start/end
boundary predictions using region labels. Experimental results on three
datasets show that SeqPAN is superior to state-of-the-art methods. Furthermore,
the effectiveness of the self-guided parallel attention module and the sequence
matching module is verified.
|
2108.02859 | Markus Dreyer | Markus Dreyer, Mengwen Liu, Feng Nan, Sandeep Atluri, Sujith Ravi | Evaluating the Tradeoff Between Abstractiveness and Factuality in
Abstractive Summarization | Accepted at EACL 2023 (Findings) | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural models for abstractive summarization tend to generate output that is
fluent and well-formed but lacks semantic faithfulness, or factuality, with
respect to the input documents. In this paper, we analyze the tradeoff between
abstractiveness and factuality of generated summaries across multiple datasets
and models, using extensive human evaluations of factuality. In our analysis,
we visualize the rates of change in factuality as we gradually increase
abstractiveness using a decoding constraint, and we observe that, while
increased abstractiveness generally leads to a drop in factuality, the rate of
factuality decay depends on factors such as the data that the system was
trained on. We introduce two datasets with human factuality judgements; one
containing 10.2k generated summaries with systematically varied degrees of
abstractiveness; the other containing 4.2k summaries from five different
summarization models. We propose new factuality metrics that adjust for the
degree of abstractiveness, and we use them to compare the
abstractiveness-adjusted factuality of previous summarization works, providing
baselines for future work.
| [
{
"version": "v1",
"created": "Thu, 5 Aug 2021 21:28:20 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 20:48:32 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Dreyer",
"Markus",
""
],
[
"Liu",
"Mengwen",
""
],
[
"Nan",
"Feng",
""
],
[
"Atluri",
"Sandeep",
""
],
[
"Ravi",
"Sujith",
""
]
] | TITLE: Evaluating the Tradeoff Between Abstractiveness and Factuality in
Abstractive Summarization
ABSTRACT: Neural models for abstractive summarization tend to generate output that is
fluent and well-formed but lacks semantic faithfulness, or factuality, with
respect to the input documents. In this paper, we analyze the tradeoff between
abstractiveness and factuality of generated summaries across multiple datasets
and models, using extensive human evaluations of factuality. In our analysis,
we visualize the rates of change in factuality as we gradually increase
abstractiveness using a decoding constraint, and we observe that, while
increased abstractiveness generally leads to a drop in factuality, the rate of
factuality decay depends on factors such as the data that the system was
trained on. We introduce two datasets with human factuality judgements; one
containing 10.2k generated summaries with systematically varied degrees of
abstractiveness; the other containing 4.2k summaries from five different
summarization models. We propose new factuality metrics that adjust for the
degree of abstractiveness, and we use them to compare the
abstractiveness-adjusted factuality of previous summarization works, providing
baselines for future work.
|
2111.14451 | Xin Huang | Xin Huang, Qi Zhang, Ying Feng, Hongdong Li, Xuan Wang, Qing Wang | HDR-NeRF: High Dynamic Range Neural Radiance Fields | Accepted to CVPR 2022. Project page:
https://xhuangcv.github.io/hdr-nerf/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present High Dynamic Range Neural Radiance Fields (HDR-NeRF) to recover an
HDR radiance field from a set of low dynamic range (LDR) views with different
exposures. Using the HDR-NeRF, we are able to generate both novel HDR views and
novel LDR views under different exposures. The key to our method is to model
the physical imaging process, which dictates that the radiance of a scene point
transforms to a pixel value in the LDR image with two implicit functions: a
radiance field and a tone mapper. The radiance field encodes the scene radiance
(values vary from 0 to +infty), which outputs the density and radiance of a ray
by giving corresponding ray origin and ray direction. The tone mapper models
the mapping process that a ray hitting on the camera sensor becomes a pixel
value. The color of the ray is predicted by feeding the radiance and the
corresponding exposure time into the tone mapper. We use the classic volume
rendering technique to project the output radiance, colors, and densities into
HDR and LDR images, while only the input LDR images are used as the
supervision. We collect a new forward-facing HDR dataset to evaluate the
proposed method. Experimental results on synthetic and real-world scenes
validate that our method can not only accurately control the exposures of
synthesized views but also render views with a high dynamic range.
| [
{
"version": "v1",
"created": "Mon, 29 Nov 2021 11:06:39 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Dec 2021 08:40:26 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Mar 2022 05:18:40 GMT"
},
{
"version": "v4",
"created": "Tue, 25 Apr 2023 11:49:08 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Huang",
"Xin",
""
],
[
"Zhang",
"Qi",
""
],
[
"Feng",
"Ying",
""
],
[
"Li",
"Hongdong",
""
],
[
"Wang",
"Xuan",
""
],
[
"Wang",
"Qing",
""
]
] | TITLE: HDR-NeRF: High Dynamic Range Neural Radiance Fields
ABSTRACT: We present High Dynamic Range Neural Radiance Fields (HDR-NeRF) to recover an
HDR radiance field from a set of low dynamic range (LDR) views with different
exposures. Using the HDR-NeRF, we are able to generate both novel HDR views and
novel LDR views under different exposures. The key to our method is to model
the physical imaging process, which dictates that the radiance of a scene point
transforms to a pixel value in the LDR image with two implicit functions: a
radiance field and a tone mapper. The radiance field encodes the scene radiance
(values vary from 0 to +infty), which outputs the density and radiance of a ray
by giving corresponding ray origin and ray direction. The tone mapper models
the mapping process that a ray hitting on the camera sensor becomes a pixel
value. The color of the ray is predicted by feeding the radiance and the
corresponding exposure time into the tone mapper. We use the classic volume
rendering technique to project the output radiance, colors, and densities into
HDR and LDR images, while only the input LDR images are used as the
supervision. We collect a new forward-facing HDR dataset to evaluate the
proposed method. Experimental results on synthetic and real-world scenes
validate that our method can not only accurately control the exposures of
synthesized views but also render views with a high dynamic range.
|
2112.12275 | Felipe S. Abrah\~ao | Felipe S. Abrah\~ao, Hector Zenil, Fabio Porto, Michael Winter, Klaus
Wehmuth, Itala M. L. D'Ottaviano | A Simplicity Bubble Problem in Formal-Theoretic Learning Systems | null | null | null | null | cs.IT cs.AI cs.LG cs.LO math.IT math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When mining large datasets in order to predict new data, limitations of the
principles behind statistical machine learning pose a serious challenge not
only to the Big Data deluge, but also to the traditional assumptions that data
generating processes are biased toward low algorithmic complexity. Even when
one assumes an underlying algorithmic-informational bias toward simplicity in
finite dataset generators, we show that current approaches to machine learning
(including deep learning, or any formal-theoretic hybrid mix of top-down AI and
statistical machine learning approaches), can always be deceived, naturally or
artificially, by sufficiently large datasets. In particular, we demonstrate
that, for every learning algorithm (with or without access to a formal theory),
there is a sufficiently large dataset size above which the algorithmic
probability of an unpredictable deceiver is an upper bound (up to a
multiplicative constant that only depends on the learning algorithm) for the
algorithmic probability of any other larger dataset. In other words, very large
and complex datasets can deceive learning algorithms into a ``simplicity
bubble'' as likely as any other particular non-deceiving dataset. These
deceiving datasets guarantee that any prediction effected by the learning
algorithm will unpredictably diverge from the high-algorithmic-complexity
globally optimal solution while converging toward the
low-algorithmic-complexity locally optimal solution, although the latter is
deemed a global one by the learning algorithm. We discuss the framework and
additional empirical conditions to be met in order to circumvent this deceptive
phenomenon, moving away from statistical machine learning towards a stronger
type of machine learning based on, and motivated by, the intrinsic power of
algorithmic information theory and computability theory.
| [
{
"version": "v1",
"created": "Wed, 22 Dec 2021 23:44:47 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 15:14:07 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Abrahão",
"Felipe S.",
""
],
[
"Zenil",
"Hector",
""
],
[
"Porto",
"Fabio",
""
],
[
"Winter",
"Michael",
""
],
[
"Wehmuth",
"Klaus",
""
],
[
"D'Ottaviano",
"Itala M. L.",
""
]
] | TITLE: A Simplicity Bubble Problem in Formal-Theoretic Learning Systems
ABSTRACT: When mining large datasets in order to predict new data, limitations of the
principles behind statistical machine learning pose a serious challenge not
only to the Big Data deluge, but also to the traditional assumptions that data
generating processes are biased toward low algorithmic complexity. Even when
one assumes an underlying algorithmic-informational bias toward simplicity in
finite dataset generators, we show that current approaches to machine learning
(including deep learning, or any formal-theoretic hybrid mix of top-down AI and
statistical machine learning approaches), can always be deceived, naturally or
artificially, by sufficiently large datasets. In particular, we demonstrate
that, for every learning algorithm (with or without access to a formal theory),
there is a sufficiently large dataset size above which the algorithmic
probability of an unpredictable deceiver is an upper bound (up to a
multiplicative constant that only depends on the learning algorithm) for the
algorithmic probability of any other larger dataset. In other words, very large
and complex datasets can deceive learning algorithms into a ``simplicity
bubble'' as likely as any other particular non-deceiving dataset. These
deceiving datasets guarantee that any prediction effected by the learning
algorithm will unpredictably diverge from the high-algorithmic-complexity
globally optimal solution while converging toward the
low-algorithmic-complexity locally optimal solution, although the latter is
deemed a global one by the learning algorithm. We discuss the framework and
additional empirical conditions to be met in order to circumvent this deceptive
phenomenon, moving away from statistical machine learning towards a stronger
type of machine learning based on, and motivated by, the intrinsic power of
algorithmic information theory and computability theory.
|
2203.08459 | Antoine Nzeyimana | Antoine Nzeyimana, Andre Niyongabo Rubungo | KinyaBERT: a Morphology-aware Kinyarwanda Language Model | ACL 2022 | null | 10.18653/v1/2022.acl-long.367 | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Pre-trained language models such as BERT have been successful at tackling
many natural language processing tasks. However, the unsupervised sub-word
tokenization methods commonly used in these models (e.g., byte-pair encoding -
BPE) are sub-optimal at handling morphologically rich languages. Even given a
morphological analyzer, naive sequencing of morphemes into a standard BERT
architecture is inefficient at capturing morphological compositionality and
expressing word-relative syntactic regularities. We address these challenges by
proposing a simple yet effective two-tier BERT architecture that leverages a
morphological analyzer and explicitly represents morphological
compositionality. Despite the success of BERT, most of its evaluations have
been conducted on high-resource languages, obscuring its applicability on
low-resource languages. We evaluate our proposed method on the low-resource
morphologically rich Kinyarwanda language, naming the proposed model
architecture KinyaBERT. A robust set of experimental results reveal that
KinyaBERT outperforms solid baselines by 2% in F1 score on a named entity
recognition task and by 4.3% in average score of a machine-translated GLUE
benchmark. KinyaBERT fine-tuning has better convergence and achieves more
robust results on multiple tasks even in the presence of translation noise.
| [
{
"version": "v1",
"created": "Wed, 16 Mar 2022 08:36:14 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Mar 2022 12:35:21 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Nzeyimana",
"Antoine",
""
],
[
"Rubungo",
"Andre Niyongabo",
""
]
] | TITLE: KinyaBERT: a Morphology-aware Kinyarwanda Language Model
ABSTRACT: Pre-trained language models such as BERT have been successful at tackling
many natural language processing tasks. However, the unsupervised sub-word
tokenization methods commonly used in these models (e.g., byte-pair encoding -
BPE) are sub-optimal at handling morphologically rich languages. Even given a
morphological analyzer, naive sequencing of morphemes into a standard BERT
architecture is inefficient at capturing morphological compositionality and
expressing word-relative syntactic regularities. We address these challenges by
proposing a simple yet effective two-tier BERT architecture that leverages a
morphological analyzer and explicitly represents morphological
compositionality. Despite the success of BERT, most of its evaluations have
been conducted on high-resource languages, obscuring its applicability on
low-resource languages. We evaluate our proposed method on the low-resource
morphologically rich Kinyarwanda language, naming the proposed model
architecture KinyaBERT. A robust set of experimental results reveal that
KinyaBERT outperforms solid baselines by 2% in F1 score on a named entity
recognition task and by 4.3% in average score of a machine-translated GLUE
benchmark. KinyaBERT fine-tuning has better convergence and achieves more
robust results on multiple tasks even in the presence of translation noise.
|
2206.07706 | Jiahao Xie | Jiahao Xie, Wei Li, Xiaohang Zhan, Ziwei Liu, Yew Soon Ong, Chen
Change Loy | Masked Frequency Modeling for Self-Supervised Visual Pre-Training | ICLR 2023. Project page:
https://www.mmlab-ntu.com/project/mfm/index.html Code:
https://github.com/Jiahao000/MFM | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present Masked Frequency Modeling (MFM), a unified frequency-domain-based
approach for self-supervised pre-training of visual models. Instead of randomly
inserting mask tokens to the input embeddings in the spatial domain, in this
paper, we shift the perspective to the frequency domain. Specifically, MFM
first masks out a portion of frequency components of the input image and then
predicts the missing frequencies on the frequency spectrum. Our key insight is
that predicting masked components in the frequency domain is more ideal to
reveal underlying image patterns rather than predicting masked patches in the
spatial domain, due to the heavy spatial redundancy. Our findings suggest that
with the right configuration of mask-and-predict strategy, both the structural
information within high-frequency components and the low-level statistics among
low-frequency counterparts are useful in learning good representations. For the
first time, MFM demonstrates that, for both ViT and CNN, a simple non-Siamese
framework can learn meaningful representations even using none of the
following: (i) extra data, (ii) extra model, (iii) mask token. Experimental
results on image classification and semantic segmentation, as well as several
robustness benchmarks show the competitive performance and advanced robustness
of MFM compared with recent masked image modeling approaches. Furthermore, we
also comprehensively investigate the effectiveness of classical image
restoration tasks for representation learning from a unified frequency
perspective and reveal their intriguing relations with our MFM approach.
| [
{
"version": "v1",
"created": "Wed, 15 Jun 2022 17:58:30 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 17:29:15 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Xie",
"Jiahao",
""
],
[
"Li",
"Wei",
""
],
[
"Zhan",
"Xiaohang",
""
],
[
"Liu",
"Ziwei",
""
],
[
"Ong",
"Yew Soon",
""
],
[
"Loy",
"Chen Change",
""
]
] | TITLE: Masked Frequency Modeling for Self-Supervised Visual Pre-Training
ABSTRACT: We present Masked Frequency Modeling (MFM), a unified frequency-domain-based
approach for self-supervised pre-training of visual models. Instead of randomly
inserting mask tokens to the input embeddings in the spatial domain, in this
paper, we shift the perspective to the frequency domain. Specifically, MFM
first masks out a portion of frequency components of the input image and then
predicts the missing frequencies on the frequency spectrum. Our key insight is
that predicting masked components in the frequency domain is more ideal to
reveal underlying image patterns rather than predicting masked patches in the
spatial domain, due to the heavy spatial redundancy. Our findings suggest that
with the right configuration of mask-and-predict strategy, both the structural
information within high-frequency components and the low-level statistics among
low-frequency counterparts are useful in learning good representations. For the
first time, MFM demonstrates that, for both ViT and CNN, a simple non-Siamese
framework can learn meaningful representations even using none of the
following: (i) extra data, (ii) extra model, (iii) mask token. Experimental
results on image classification and semantic segmentation, as well as several
robustness benchmarks show the competitive performance and advanced robustness
of MFM compared with recent masked image modeling approaches. Furthermore, we
also comprehensively investigate the effectiveness of classical image
restoration tasks for representation learning from a unified frequency
perspective and reveal their intriguing relations with our MFM approach.
|
2206.10747 | Marcell Tam\'as Kurbucz | Marcell Stippinger, D\'avid Han\'ak, Marcell T. Kurbucz, Gergely
Hancz\'ar, Oliv\'er M. T\"orteli, Zolt\'an Somogyv\'ari | BiometricBlender: Ultra-high dimensional, multi-class synthetic data
generator to imitate biometric feature space | DISCLAIMER: This is a preprint article. A final peer reviewed article
has now been published in SoftwareX. DOI:
https://doi.org/10.1016/j.softx.2023.101366 | SoftwareX 22 (2023) 101366 | 10.1016/j.softx.2023.101366 | null | cs.LG cs.AI cs.DB stat.CO | http://creativecommons.org/licenses/by/4.0/ | The lack of freely available (real-life or synthetic) high or ultra-high
dimensional, multi-class datasets may hamper the rapidly growing research on
feature screening, especially in the field of biometrics, where the usage of
such datasets is common. This paper reports a Python package called
BiometricBlender, which is an ultra-high dimensional, multi-class synthetic
data generator to benchmark a wide range of feature screening methods. During
the data generation process, the overall usefulness and the intercorrelations
of blended features can be controlled by the user, thus the synthetic feature
space is able to imitate the key properties of a real biometric dataset.
| [
{
"version": "v1",
"created": "Tue, 21 Jun 2022 21:42:27 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 17:01:22 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Stippinger",
"Marcell",
""
],
[
"Hanák",
"Dávid",
""
],
[
"Kurbucz",
"Marcell T.",
""
],
[
"Hanczár",
"Gergely",
""
],
[
"Törteli",
"Olivér M.",
""
],
[
"Somogyvári",
"Zoltán",
""
]
] | TITLE: BiometricBlender: Ultra-high dimensional, multi-class synthetic data
generator to imitate biometric feature space
ABSTRACT: The lack of freely available (real-life or synthetic) high or ultra-high
dimensional, multi-class datasets may hamper the rapidly growing research on
feature screening, especially in the field of biometrics, where the usage of
such datasets is common. This paper reports a Python package called
BiometricBlender, which is an ultra-high dimensional, multi-class synthetic
data generator to benchmark a wide range of feature screening methods. During
the data generation process, the overall usefulness and the intercorrelations
of blended features can be controlled by the user, thus the synthetic feature
space is able to imitate the key properties of a real biometric dataset.
|
2207.12194 | Sen Pei | Sen Pei, Jiaxi Sun, Richard Yi Da Xu, Shiming Xiang, and Gaofeng Meng | Domain Decorrelation with Potential Energy Ranking | 2022 ECCV jury award, accepted by AAAI 2023 | AAAI 2023 Oral | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning systems, especially the methods based on deep learning,
enjoy great success in modern computer vision tasks under experimental
settings. Generally, these classic deep learning methods are built on the
\emph{i.i.d.} assumption, supposing the training and test data are drawn from a
similar distribution independently and identically. However, the aforementioned
\emph{i.i.d.} assumption is in general unavailable in the real-world scenario,
and as a result, leads to sharp performance decay of deep learning algorithms.
Behind this, domain shift is one of the primary factors to be blamed. In order
to tackle this problem, we propose using \textbf{Po}tential \textbf{E}nergy
\textbf{R}anking (PoER) to decouple the object feature and the domain feature
(\emph{i.e.,} appearance feature) in given images, promoting the learning of
label-discriminative features while filtering out the irrelevant correlations
between the objects and the background. PoER helps the neural networks to
capture label-related features which contain the domain information first in
shallow layers and then distills the label-discriminative representations out
progressively, enforcing the neural networks to be aware of the characteristic
of objects and background which is vital to the generation of domain-invariant
features. PoER reports superior performance on domain generalization
benchmarks, improving the average top-1 accuracy by at least 1.20\% compared to
the existing methods. Moreover, we use PoER in the ECCV 2022 NICO
Challenge\footnote{https://nicochallenge.com}, achieving top place with only a
vanilla ResNet-18. The code has been made available at
https://github.com/ForeverPs/PoER.
| [
{
"version": "v1",
"created": "Mon, 25 Jul 2022 13:33:53 GMT"
},
{
"version": "v2",
"created": "Tue, 26 Jul 2022 01:27:35 GMT"
},
{
"version": "v3",
"created": "Fri, 16 Dec 2022 06:41:05 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Pei",
"Sen",
""
],
[
"Sun",
"Jiaxi",
""
],
[
"Da Xu",
"Richard Yi",
""
],
[
"Xiang",
"Shiming",
""
],
[
"Meng",
"Gaofeng",
""
]
] | TITLE: Domain Decorrelation with Potential Energy Ranking
ABSTRACT: Machine learning systems, especially the methods based on deep learning,
enjoy great success in modern computer vision tasks under experimental
settings. Generally, these classic deep learning methods are built on the
\emph{i.i.d.} assumption, supposing the training and test data are drawn from a
similar distribution independently and identically. However, the aforementioned
\emph{i.i.d.} assumption is in general unavailable in the real-world scenario,
and as a result, leads to sharp performance decay of deep learning algorithms.
Behind this, domain shift is one of the primary factors to be blamed. In order
to tackle this problem, we propose using \textbf{Po}tential \textbf{E}nergy
\textbf{R}anking (PoER) to decouple the object feature and the domain feature
(\emph{i.e.,} appearance feature) in given images, promoting the learning of
label-discriminative features while filtering out the irrelevant correlations
between the objects and the background. PoER helps the neural networks to
capture label-related features which contain the domain information first in
shallow layers and then distills the label-discriminative representations out
progressively, enforcing the neural networks to be aware of the characteristic
of objects and background which is vital to the generation of domain-invariant
features. PoER reports superior performance on domain generalization
benchmarks, improving the average top-1 accuracy by at least 1.20\% compared to
the existing methods. Moreover, we use PoER in the ECCV 2022 NICO
Challenge\footnote{https://nicochallenge.com}, achieving top place with only a
vanilla ResNet-18. The code has been made available at
https://github.com/ForeverPs/PoER.
|
2208.01485 | Ashish Kumar | Ashish Kumar, R.K. Agrawal, Leve Joseph | IterMiUnet: A lightweight architecture for automatic blood vessel
segmentation | null | null | 10.1007/s11042-023-15433-7 | null | eess.IV cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | The automatic segmentation of blood vessels in fundus images can help analyze
the condition of retinal vasculature, which is crucial for identifying various
systemic diseases like hypertension, diabetes, etc. Despite the success of Deep
Learning-based models in this segmentation task, most of them are heavily
parametrized and thus have limited use in practical applications. This paper
proposes IterMiUnet, a new lightweight convolution-based segmentation model
that requires significantly fewer parameters and yet delivers performance
similar to existing models. The model makes use of the excellent segmentation
capabilities of Iternet architecture but overcomes its heavily parametrized
nature by incorporating the encoder-decoder structure of MiUnet model within
it. Thus, the new model reduces parameters without any compromise with the
network's depth, which is necessary to learn abstract hierarchical concepts in
deep models. This lightweight segmentation model speeds up training and
inference time and is potentially helpful in the medical domain where data is
scarce and, therefore, heavily parametrized models tend to overfit. The
proposed model was evaluated on three publicly available datasets: DRIVE,
STARE, and CHASE-DB1. Further cross-training and inter-rater variability
evaluations have also been performed. The proposed model has a lot of potential
to be utilized as a tool for the early diagnosis of many diseases.
| [
{
"version": "v1",
"created": "Tue, 2 Aug 2022 14:33:14 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Kumar",
"Ashish",
""
],
[
"Agrawal",
"R. K.",
""
],
[
"Joseph",
"Leve",
""
]
] | TITLE: IterMiUnet: A lightweight architecture for automatic blood vessel
segmentation
ABSTRACT: The automatic segmentation of blood vessels in fundus images can help analyze
the condition of retinal vasculature, which is crucial for identifying various
systemic diseases like hypertension, diabetes, etc. Despite the success of Deep
Learning-based models in this segmentation task, most of them are heavily
parametrized and thus have limited use in practical applications. This paper
proposes IterMiUnet, a new lightweight convolution-based segmentation model
that requires significantly fewer parameters and yet delivers performance
similar to existing models. The model makes use of the excellent segmentation
capabilities of Iternet architecture but overcomes its heavily parametrized
nature by incorporating the encoder-decoder structure of MiUnet model within
it. Thus, the new model reduces parameters without any compromise with the
network's depth, which is necessary to learn abstract hierarchical concepts in
deep models. This lightweight segmentation model speeds up training and
inference time and is potentially helpful in the medical domain where data is
scarce and, therefore, heavily parametrized models tend to overfit. The
proposed model was evaluated on three publicly available datasets: DRIVE,
STARE, and CHASE-DB1. Further cross-training and inter-rater variability
evaluations have also been performed. The proposed model has a lot of potential
to be utilized as a tool for the early diagnosis of many diseases.
|
2208.12044 | Jingcai Guo | Jingcai Guo, Song Guo, Jie Zhang, Ziming Liu | Fed-FSNet: Mitigating Non-I.I.D. Federated Learning via Fuzzy
Synthesizing Network | null | null | null | null | cs.CR cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated learning (FL) has emerged as a promising privacy-preserving
distributed machine learning framework recently. It aims at collaboratively
learning a shared global model by performing distributed training locally on
edge devices and aggregating local models into a global one without centralized
raw data sharing in the cloud server. However, due to the large local data
heterogeneities (Non-I.I.D. data) across edge devices, the FL may easily obtain
a global model that can produce more shifted gradients on local datasets,
thereby degrading the model performance or even suffering from the
non-convergence during training. In this paper, we propose a novel FL training
framework, dubbed Fed-FSNet, using a properly designed Fuzzy Synthesizing
Network (FSNet) to mitigate the Non-I.I.D. FL at-the-source. Concretely, we
maintain an edge-agnostic hidden model in the cloud server to estimate a
less-accurate while direction-aware inversion of the global model. The hidden
model can then fuzzily synthesize several mimic I.I.D. data samples (sample
features) conditioned on only the global model, which can be shared by edge
devices to facilitate the FL training towards faster and better convergence.
Moreover, since the synthesizing process involves neither access to the
parameters/updates of local models nor analyzing individual local model
outputs, our framework can still ensure the privacy of FL. Experimental results
on several FL benchmarks demonstrate that our method can significantly mitigate
the Non-I.I.D. issue and obtain better performance against other representative
methods.
| [
{
"version": "v1",
"created": "Sun, 21 Aug 2022 18:40:51 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 08:45:05 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Guo",
"Jingcai",
""
],
[
"Guo",
"Song",
""
],
[
"Zhang",
"Jie",
""
],
[
"Liu",
"Ziming",
""
]
] | TITLE: Fed-FSNet: Mitigating Non-I.I.D. Federated Learning via Fuzzy
Synthesizing Network
ABSTRACT: Federated learning (FL) has emerged as a promising privacy-preserving
distributed machine learning framework recently. It aims at collaboratively
learning a shared global model by performing distributed training locally on
edge devices and aggregating local models into a global one without centralized
raw data sharing in the cloud server. However, due to the large local data
heterogeneities (Non-I.I.D. data) across edge devices, the FL may easily obtain
a global model that can produce more shifted gradients on local datasets,
thereby degrading the model performance or even suffering from the
non-convergence during training. In this paper, we propose a novel FL training
framework, dubbed Fed-FSNet, using a properly designed Fuzzy Synthesizing
Network (FSNet) to mitigate the Non-I.I.D. FL at-the-source. Concretely, we
maintain an edge-agnostic hidden model in the cloud server to estimate a
less-accurate while direction-aware inversion of the global model. The hidden
model can then fuzzily synthesize several mimic I.I.D. data samples (sample
features) conditioned on only the global model, which can be shared by edge
devices to facilitate the FL training towards faster and better convergence.
Moreover, since the synthesizing process involves neither access to the
parameters/updates of local models nor analyzing individual local model
outputs, our framework can still ensure the privacy of FL. Experimental results
on several FL benchmarks demonstrate that our method can significantly mitigate
the Non-I.I.D. issue and obtain better performance against other representative
methods.
|
2210.04977 | Rima Arnaout | Chinmayee Athalye, Rima Arnaout | Domain-guided data augmentation for deep learning on medical imaging | 18 pages, 6 Tables, 3 Figures | null | 10.1371/journal.pone.0282532 | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | While domain-specific data augmentation can be useful in training neural
networks for medical imaging tasks, such techniques have not been widely used
to date. Here, we test whether domain-specific data augmentation is useful for
medical imaging using a well-benchmarked task: view classification on fetal
ultrasound FETAL-125 and OB-125 datasets. We found that using a
context-preserving cut-paste strategy, we could create valid training data as
measured by performance of the resulting trained model on the benchmark test
dataset. When used in an online fashion, models trained on this data performed
similarly to those trained using traditional data augmentation (FETAL-125
F-score 85.33+/-0.24 vs 86.89+/-0.60, p-value 0.0139; OB-125 F-score
74.60+/-0.11 vs 72.43+/-0.62, p-value 0.0039). Furthermore, the ability to
perform augmentations during training time, as well as the ability to apply
chosen augmentations equally across data classes, are important considerations
in designing a bespoke data augmentation. Finally, we provide open-source code
to facilitate running bespoke data augmentations in an online fashion. Taken
together, this work expands the ability to design and apply domain-guided data
augmentations for medical imaging tasks.
| [
{
"version": "v1",
"created": "Mon, 10 Oct 2022 19:21:03 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Athalye",
"Chinmayee",
""
],
[
"Arnaout",
"Rima",
""
]
] | TITLE: Domain-guided data augmentation for deep learning on medical imaging
ABSTRACT: While domain-specific data augmentation can be useful in training neural
networks for medical imaging tasks, such techniques have not been widely used
to date. Here, we test whether domain-specific data augmentation is useful for
medical imaging using a well-benchmarked task: view classification on fetal
ultrasound FETAL-125 and OB-125 datasets. We found that using a
context-preserving cut-paste strategy, we could create valid training data as
measured by performance of the resulting trained model on the benchmark test
dataset. When used in an online fashion, models trained on this data performed
similarly to those trained using traditional data augmentation (FETAL-125
F-score 85.33+/-0.24 vs 86.89+/-0.60, p-value 0.0139; OB-125 F-score
74.60+/-0.11 vs 72.43+/-0.62, p-value 0.0039). Furthermore, the ability to
perform augmentations during training time, as well as the ability to apply
chosen augmentations equally across data classes, are important considerations
in designing a bespoke data augmentation. Finally, we provide open-source code
to facilitate running bespoke data augmentations in an online fashion. Taken
together, this work expands the ability to design and apply domain-guided data
augmentations for medical imaging tasks.
|
2210.08332 | Yiqiao Jin | Yiqiao Jin, Yunsheng Bai, Yanqiao Zhu, Yizhou Sun, Wei Wang | Code Recommendation for Open Source Software Developers | In Proceedings of The WebConf 2023 (WWW 2023) | null | 10.1145/3543507.3583503 | null | cs.SE cs.IR cs.LG cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Open Source Software (OSS) is forming the spines of technology
infrastructures, attracting millions of talents to contribute. Notably, it is
challenging and critical to consider both the developers' interests and the
semantic features of the project code to recommend appropriate development
tasks to OSS developers. In this paper, we formulate the novel problem of code
recommendation, whose purpose is to predict the future contribution behaviors
of developers given their interaction history, the semantic features of source
code, and the hierarchical file structures of projects. Considering the complex
interactions among multiple parties within the system, we propose CODER, a
novel graph-based code recommendation framework for open source software
developers. CODER jointly models microscopic user-code interactions and
macroscopic user-project interactions via a heterogeneous graph and further
bridges the two levels of information through aggregation on file-structure
graphs that reflect the project hierarchy. Moreover, due to the lack of
reliable benchmarks, we construct three large-scale datasets to facilitate
future research in this direction. Extensive experiments show that our CODER
framework achieves superior performance under various experimental settings,
including intra-project, cross-project, and cold-start recommendation. We will
release all the datasets, code, and utilities for data retrieval upon the
acceptance of this work.
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2022 16:40:36 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Oct 2022 13:46:18 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Apr 2023 11:53:04 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Jin",
"Yiqiao",
""
],
[
"Bai",
"Yunsheng",
""
],
[
"Zhu",
"Yanqiao",
""
],
[
"Sun",
"Yizhou",
""
],
[
"Wang",
"Wei",
""
]
] | TITLE: Code Recommendation for Open Source Software Developers
ABSTRACT: Open Source Software (OSS) is forming the spines of technology
infrastructures, attracting millions of talents to contribute. Notably, it is
challenging and critical to consider both the developers' interests and the
semantic features of the project code to recommend appropriate development
tasks to OSS developers. In this paper, we formulate the novel problem of code
recommendation, whose purpose is to predict the future contribution behaviors
of developers given their interaction history, the semantic features of source
code, and the hierarchical file structures of projects. Considering the complex
interactions among multiple parties within the system, we propose CODER, a
novel graph-based code recommendation framework for open source software
developers. CODER jointly models microscopic user-code interactions and
macroscopic user-project interactions via a heterogeneous graph and further
bridges the two levels of information through aggregation on file-structure
graphs that reflect the project hierarchy. Moreover, due to the lack of
reliable benchmarks, we construct three large-scale datasets to facilitate
future research in this direction. Extensive experiments show that our CODER
framework achieves superior performance under various experimental settings,
including intra-project, cross-project, and cold-start recommendation. We will
release all the datasets, code, and utilities for data retrieval upon the
acceptance of this work.
|
2210.11231 | Hongmin Xiao | Yucong Lin, Hongming Xiao, Jiani Liu, Zichao Lin, Keming Lu, Feifei
Wang, Wei Wei | Knowledge-Enhanced Relation Extraction Dataset | 20 pages, 6 figures, submitted to Neural Computing and Applications | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recently, knowledge-enhanced methods leveraging auxiliary knowledge graphs
have emerged in relation extraction, surpassing traditional text-based
approaches. However, to our best knowledge, there is currently no public
dataset available that encompasses both evidence sentences and knowledge graphs
for knowledge-enhanced relation extraction. To address this gap, we introduce
the Knowledge-Enhanced Relation Extraction Dataset (KERED). KERED annotates
each sentence with a relational fact, and it provides knowledge context for
entities through entity linking. Using our curated dataset, We compared
contemporary relation extraction methods under two prevalent task settings:
sentence-level and bag-level. The experimental result shows the knowledge
graphs provided by KERED can support knowledge-enhanced relation extraction
methods. We believe that KERED offers high-quality relation extraction datasets
with corresponding knowledge graphs for evaluating the performance of
knowledge-enhanced relation extraction methods. Our dataset is available at:
\url{https://figshare.com/projects/KERED/134459}
| [
{
"version": "v1",
"created": "Wed, 19 Oct 2022 13:23:10 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Nov 2022 02:24:23 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Apr 2023 08:46:30 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Lin",
"Yucong",
""
],
[
"Xiao",
"Hongming",
""
],
[
"Liu",
"Jiani",
""
],
[
"Lin",
"Zichao",
""
],
[
"Lu",
"Keming",
""
],
[
"Wang",
"Feifei",
""
],
[
"Wei",
"Wei",
""
]
] | TITLE: Knowledge-Enhanced Relation Extraction Dataset
ABSTRACT: Recently, knowledge-enhanced methods leveraging auxiliary knowledge graphs
have emerged in relation extraction, surpassing traditional text-based
approaches. However, to our best knowledge, there is currently no public
dataset available that encompasses both evidence sentences and knowledge graphs
for knowledge-enhanced relation extraction. To address this gap, we introduce
the Knowledge-Enhanced Relation Extraction Dataset (KERED). KERED annotates
each sentence with a relational fact, and it provides knowledge context for
entities through entity linking. Using our curated dataset, We compared
contemporary relation extraction methods under two prevalent task settings:
sentence-level and bag-level. The experimental result shows the knowledge
graphs provided by KERED can support knowledge-enhanced relation extraction
methods. We believe that KERED offers high-quality relation extraction datasets
with corresponding knowledge graphs for evaluating the performance of
knowledge-enhanced relation extraction methods. Our dataset is available at:
\url{https://figshare.com/projects/KERED/134459}
|
2211.07907 | Namrata Deka | Namrata Deka and Danica J. Sutherland | MMD-B-Fair: Learning Fair Representations with Statistical Testing | null | Proceedings of The 26th International Conference on Artificial
Intelligence and Statistics, 2023, PMLR 206:9564-9576 | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | We introduce a method, MMD-B-Fair, to learn fair representations of data via
kernel two-sample testing. We find neural features of our data where a maximum
mean discrepancy (MMD) test cannot distinguish between representations of
different sensitive groups, while preserving information about the target
attributes. Minimizing the power of an MMD test is more difficult than
maximizing it (as done in previous work), because the test threshold's complex
behavior cannot be simply ignored. Our method exploits the simple asymptotics
of block testing schemes to efficiently find fair representations without
requiring complex adversarial optimization or generative modelling schemes
widely used by existing work on fair representation learning. We evaluate our
approach on various datasets, showing its ability to ``hide'' information about
sensitive attributes, and its effectiveness in downstream transfer tasks.
| [
{
"version": "v1",
"created": "Tue, 15 Nov 2022 05:25:38 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 13:17:19 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Apr 2023 11:56:09 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Deka",
"Namrata",
""
],
[
"Sutherland",
"Danica J.",
""
]
] | TITLE: MMD-B-Fair: Learning Fair Representations with Statistical Testing
ABSTRACT: We introduce a method, MMD-B-Fair, to learn fair representations of data via
kernel two-sample testing. We find neural features of our data where a maximum
mean discrepancy (MMD) test cannot distinguish between representations of
different sensitive groups, while preserving information about the target
attributes. Minimizing the power of an MMD test is more difficult than
maximizing it (as done in previous work), because the test threshold's complex
behavior cannot be simply ignored. Our method exploits the simple asymptotics
of block testing schemes to efficiently find fair representations without
requiring complex adversarial optimization or generative modelling schemes
widely used by existing work on fair representation learning. We evaluate our
approach on various datasets, showing its ability to ``hide'' information about
sensitive attributes, and its effectiveness in downstream transfer tasks.
|
2212.00975 | Jinyoung Park | Jinyoung Park, Hyeong Kyu Choi, Juyeon Ko, Hyeonjin Park, Ji-Hoon Kim,
Jisu Jeong, Kyungmin Kim, Hyunwoo J. Kim | Relation-Aware Language-Graph Transformer for Question Answering | AAAI2023 (accepted) | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Question Answering (QA) is a task that entails reasoning over natural
language contexts, and many relevant works augment language models (LMs) with
graph neural networks (GNNs) to encode the Knowledge Graph (KG) information.
However, most existing GNN-based modules for QA do not take advantage of rich
relational information of KGs and depend on limited information interaction
between the LM and the KG. To address these issues, we propose Question
Answering Transformer (QAT), which is designed to jointly reason over language
and graphs with respect to entity relations in a unified manner. Specifically,
QAT constructs Meta-Path tokens, which learn relation-centric embeddings based
on diverse structural and semantic relations. Then, our Relation-Aware
Self-Attention module comprehensively integrates different modalities via the
Cross-Modal Relative Position Bias, which guides information exchange between
relevant entites of different modalities. We validate the effectiveness of QAT
on commonsense question answering datasets like CommonsenseQA and OpenBookQA,
and on a medical question answering dataset, MedQA-USMLE. On all the datasets,
our method achieves state-of-the-art performance. Our code is available at
http://github.com/mlvlab/QAT.
| [
{
"version": "v1",
"created": "Fri, 2 Dec 2022 05:10:10 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 09:02:15 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Park",
"Jinyoung",
""
],
[
"Choi",
"Hyeong Kyu",
""
],
[
"Ko",
"Juyeon",
""
],
[
"Park",
"Hyeonjin",
""
],
[
"Kim",
"Ji-Hoon",
""
],
[
"Jeong",
"Jisu",
""
],
[
"Kim",
"Kyungmin",
""
],
[
"Kim",
"Hyunwoo J.",
""
]
] | TITLE: Relation-Aware Language-Graph Transformer for Question Answering
ABSTRACT: Question Answering (QA) is a task that entails reasoning over natural
language contexts, and many relevant works augment language models (LMs) with
graph neural networks (GNNs) to encode the Knowledge Graph (KG) information.
However, most existing GNN-based modules for QA do not take advantage of rich
relational information of KGs and depend on limited information interaction
between the LM and the KG. To address these issues, we propose Question
Answering Transformer (QAT), which is designed to jointly reason over language
and graphs with respect to entity relations in a unified manner. Specifically,
QAT constructs Meta-Path tokens, which learn relation-centric embeddings based
on diverse structural and semantic relations. Then, our Relation-Aware
Self-Attention module comprehensively integrates different modalities via the
Cross-Modal Relative Position Bias, which guides information exchange between
relevant entites of different modalities. We validate the effectiveness of QAT
on commonsense question answering datasets like CommonsenseQA and OpenBookQA,
and on a medical question answering dataset, MedQA-USMLE. On all the datasets,
our method achieves state-of-the-art performance. Our code is available at
http://github.com/mlvlab/QAT.
|
2212.06065 | Lianrui Zuo | Lianrui Zuo, Yihao Liu, Yuan Xue, Blake E. Dewey, Samuel W. Remedios,
Savannah P. Hays, Murat Bilgel, Ellen M. Mowry, Scott D. Newsome, Peter A.
Calabresi, Susan M. Resnick, Jerry L. Prince, Aaron Carass | HACA3: A Unified Approach for Multi-site MR Image Harmonization | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The lack of standardization is a prominent issue in magnetic resonance (MR)
imaging. This often causes undesired contrast variations in the acquired images
due to differences in hardware and acquisition parameters. In recent years,
image synthesis-based MR harmonization with disentanglement has been proposed
to compensate for the undesired contrast variations. Despite the success of
existing methods, we argue that three major improvements can be made. First,
most existing methods are built upon the assumption that multi-contrast MR
images of the same subject share the same anatomy. This assumption is
questionable, since different MR contrasts are specialized to highlight
different anatomical features. Second, these methods often require a fixed set
of MR contrasts for training (e.g., both T1-weighted and T2-weighted images),
limiting their applicability. Lastly, existing methods are generally sensitive
to imaging artifacts. In this paper, we present Harmonization with
Attention-based Contrast, Anatomy, and Artifact Awareness (HACA3), a novel
approach to address these three issues. HACA3 incorporates an anatomy fusion
module that accounts for the inherent anatomical differences between MR
contrasts. Furthermore, HACA3 is also robust to imaging artifacts and can be
trained and applied to any set of MR contrasts. HACA3 is developed and
evaluated on diverse MR datasets acquired from 21 sites with varying field
strengths, scanner platforms, and acquisition protocols. Experiments show that
HACA3 achieves state-of-the-art performance under multiple image quality
metrics. We also demonstrate the applicability and versatility of HACA3 on
downstream tasks including white matter lesion segmentation and longitudinal
volumetric analyses.
| [
{
"version": "v1",
"created": "Mon, 12 Dec 2022 17:33:34 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 15:34:20 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Zuo",
"Lianrui",
""
],
[
"Liu",
"Yihao",
""
],
[
"Xue",
"Yuan",
""
],
[
"Dewey",
"Blake E.",
""
],
[
"Remedios",
"Samuel W.",
""
],
[
"Hays",
"Savannah P.",
""
],
[
"Bilgel",
"Murat",
""
],
[
"Mowry",
"Ellen M.",
""
],
[
"Newsome",
"Scott D.",
""
],
[
"Calabresi",
"Peter A.",
""
],
[
"Resnick",
"Susan M.",
""
],
[
"Prince",
"Jerry L.",
""
],
[
"Carass",
"Aaron",
""
]
] | TITLE: HACA3: A Unified Approach for Multi-site MR Image Harmonization
ABSTRACT: The lack of standardization is a prominent issue in magnetic resonance (MR)
imaging. This often causes undesired contrast variations in the acquired images
due to differences in hardware and acquisition parameters. In recent years,
image synthesis-based MR harmonization with disentanglement has been proposed
to compensate for the undesired contrast variations. Despite the success of
existing methods, we argue that three major improvements can be made. First,
most existing methods are built upon the assumption that multi-contrast MR
images of the same subject share the same anatomy. This assumption is
questionable, since different MR contrasts are specialized to highlight
different anatomical features. Second, these methods often require a fixed set
of MR contrasts for training (e.g., both T1-weighted and T2-weighted images),
limiting their applicability. Lastly, existing methods are generally sensitive
to imaging artifacts. In this paper, we present Harmonization with
Attention-based Contrast, Anatomy, and Artifact Awareness (HACA3), a novel
approach to address these three issues. HACA3 incorporates an anatomy fusion
module that accounts for the inherent anatomical differences between MR
contrasts. Furthermore, HACA3 is also robust to imaging artifacts and can be
trained and applied to any set of MR contrasts. HACA3 is developed and
evaluated on diverse MR datasets acquired from 21 sites with varying field
strengths, scanner platforms, and acquisition protocols. Experiments show that
HACA3 achieves state-of-the-art performance under multiple image quality
metrics. We also demonstrate the applicability and versatility of HACA3 on
downstream tasks including white matter lesion segmentation and longitudinal
volumetric analyses.
|
2301.02931 | Can Chen | Can Chen, Yingxue Zhang, Xue Liu, Mark Coates | Bidirectional Learning for Offline Model-based Biological Sequence
Design | Accepted by ICML2023, AI for Science, 13 pages,3 figures | null | null | null | cs.CE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Offline model-based optimization aims to maximize a black-box objective
function with a static dataset of designs and their scores. In this paper, we
focus on biological sequence design to maximize some sequence score. A recent
approach employs bidirectional learning, combining a forward mapping for
exploitation and a backward mapping for constraint, and it relies on the neural
tangent kernel (NTK) of an infinitely wide network to build a proxy model.
Though effective, the NTK cannot learn features because of its parametrization,
and its use prevents the incorporation of powerful pre-trained Language Models
(LMs) that can capture the rich biophysical information in millions of
biological sequences. We adopt an alternative proxy model, adding a linear head
to a pre-trained LM, and propose a linearization scheme. This yields a
closed-form loss and also takes into account the biophysical information in the
pre-trained LM. In addition, the forward mapping and the backward mapping play
different roles and thus deserve different weights during sequence
optimization. To achieve this, we train an auxiliary model and leverage its
weak supervision signal via a bi-level optimization framework to effectively
learn how to balance the two mappings. Further, by extending the framework, we
develop the first learning rate adaptation module \textit{Adaptive}-$\eta$,
which is compatible with all gradient-based algorithms for offline model-based
optimization. Experimental results on DNA/protein sequence design tasks verify
the effectiveness of our algorithm. Our code is
available~\href{https://anonymous.4open.science/r/BIB-ICLR2023-Submission/README.md}{here.}
| [
{
"version": "v1",
"created": "Sat, 7 Jan 2023 20:27:34 GMT"
},
{
"version": "v2",
"created": "Mon, 13 Feb 2023 01:18:11 GMT"
},
{
"version": "v3",
"created": "Mon, 24 Apr 2023 22:18:50 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Chen",
"Can",
""
],
[
"Zhang",
"Yingxue",
""
],
[
"Liu",
"Xue",
""
],
[
"Coates",
"Mark",
""
]
] | TITLE: Bidirectional Learning for Offline Model-based Biological Sequence
Design
ABSTRACT: Offline model-based optimization aims to maximize a black-box objective
function with a static dataset of designs and their scores. In this paper, we
focus on biological sequence design to maximize some sequence score. A recent
approach employs bidirectional learning, combining a forward mapping for
exploitation and a backward mapping for constraint, and it relies on the neural
tangent kernel (NTK) of an infinitely wide network to build a proxy model.
Though effective, the NTK cannot learn features because of its parametrization,
and its use prevents the incorporation of powerful pre-trained Language Models
(LMs) that can capture the rich biophysical information in millions of
biological sequences. We adopt an alternative proxy model, adding a linear head
to a pre-trained LM, and propose a linearization scheme. This yields a
closed-form loss and also takes into account the biophysical information in the
pre-trained LM. In addition, the forward mapping and the backward mapping play
different roles and thus deserve different weights during sequence
optimization. To achieve this, we train an auxiliary model and leverage its
weak supervision signal via a bi-level optimization framework to effectively
learn how to balance the two mappings. Further, by extending the framework, we
develop the first learning rate adaptation module \textit{Adaptive}-$\eta$,
which is compatible with all gradient-based algorithms for offline model-based
optimization. Experimental results on DNA/protein sequence design tasks verify
the effectiveness of our algorithm. Our code is
available~\href{https://anonymous.4open.science/r/BIB-ICLR2023-Submission/README.md}{here.}
|
2301.10222 | Angelika Ando | Angelika Ando, Spyros Gidaris, Andrei Bursuc, Gilles Puy, Alexandre
Boulch, Renaud Marlet | RangeViT: Towards Vision Transformers for 3D Semantic Segmentation in
Autonomous Driving | CVPR 2023. Code at https://github.com/valeoai/rangevit | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Casting semantic segmentation of outdoor LiDAR point clouds as a 2D problem,
e.g., via range projection, is an effective and popular approach. These
projection-based methods usually benefit from fast computations and, when
combined with techniques which use other point cloud representations, achieve
state-of-the-art results. Today, projection-based methods leverage 2D CNNs but
recent advances in computer vision show that vision transformers (ViTs) have
achieved state-of-the-art results in many image-based benchmarks. In this work,
we question if projection-based methods for 3D semantic segmentation can
benefit from these latest improvements on ViTs. We answer positively but only
after combining them with three key ingredients: (a) ViTs are notoriously hard
to train and require a lot of training data to learn powerful representations.
By preserving the same backbone architecture as for RGB images, we can exploit
the knowledge from long training on large image collections that are much
cheaper to acquire and annotate than point clouds. We reach our best results
with pre-trained ViTs on large image datasets. (b) We compensate ViTs' lack of
inductive bias by substituting a tailored convolutional stem for the classical
linear embedding layer. (c) We refine pixel-wise predictions with a
convolutional decoder and a skip connection from the convolutional stem to
combine low-level but fine-grained features of the the convolutional stem with
the high-level but coarse predictions of the ViT encoder. With these
ingredients, we show that our method, called RangeViT, outperforms existing
projection-based methods on nuScenes and SemanticKITTI. The code is available
at https://github.com/valeoai/rangevit.
| [
{
"version": "v1",
"created": "Tue, 24 Jan 2023 18:50:48 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 13:11:42 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Ando",
"Angelika",
""
],
[
"Gidaris",
"Spyros",
""
],
[
"Bursuc",
"Andrei",
""
],
[
"Puy",
"Gilles",
""
],
[
"Boulch",
"Alexandre",
""
],
[
"Marlet",
"Renaud",
""
]
] | TITLE: RangeViT: Towards Vision Transformers for 3D Semantic Segmentation in
Autonomous Driving
ABSTRACT: Casting semantic segmentation of outdoor LiDAR point clouds as a 2D problem,
e.g., via range projection, is an effective and popular approach. These
projection-based methods usually benefit from fast computations and, when
combined with techniques which use other point cloud representations, achieve
state-of-the-art results. Today, projection-based methods leverage 2D CNNs but
recent advances in computer vision show that vision transformers (ViTs) have
achieved state-of-the-art results in many image-based benchmarks. In this work,
we question if projection-based methods for 3D semantic segmentation can
benefit from these latest improvements on ViTs. We answer positively but only
after combining them with three key ingredients: (a) ViTs are notoriously hard
to train and require a lot of training data to learn powerful representations.
By preserving the same backbone architecture as for RGB images, we can exploit
the knowledge from long training on large image collections that are much
cheaper to acquire and annotate than point clouds. We reach our best results
with pre-trained ViTs on large image datasets. (b) We compensate ViTs' lack of
inductive bias by substituting a tailored convolutional stem for the classical
linear embedding layer. (c) We refine pixel-wise predictions with a
convolutional decoder and a skip connection from the convolutional stem to
combine low-level but fine-grained features of the the convolutional stem with
the high-level but coarse predictions of the ViT encoder. With these
ingredients, we show that our method, called RangeViT, outperforms existing
projection-based methods on nuScenes and SemanticKITTI. The code is available
at https://github.com/valeoai/rangevit.
|
2302.00839 | Shubhendu Trivedi | Zhen Lin, Shubhendu Trivedi, Cao Xiao, Jimeng Sun | Fast Online Value-Maximizing Prediction Sets with Conformal Cost Control | Proceedings of the 40th International Conference on Machine Learning,
Honolulu, Hawaii, USA. PMLR 202, 2023. 11 pages (main paper, including
references) + 10 pages (supplementary material) | null | null | null | cs.LG stat.ME stat.ML | http://creativecommons.org/licenses/by/4.0/ | Many real-world multi-label prediction problems involve set-valued
predictions that must satisfy specific requirements dictated by downstream
usage. We focus on a typical scenario where such requirements, separately
encoding $\textit{value}$ and $\textit{cost}$, compete with each other. For
instance, a hospital might expect a smart diagnosis system to capture as many
severe, often co-morbid, diseases as possible (the value), while maintaining
strict control over incorrect predictions (the cost). We present a general
pipeline, dubbed as FavMac, to maximize the value while controlling the cost in
such scenarios. FavMac can be combined with almost any multi-label classifier,
affording distribution-free theoretical guarantees on cost control. Moreover,
unlike prior works, it can handle real-world large-scale applications via a
carefully designed online update mechanism, which is of independent interest.
Our methodological and theoretical contributions are supported by experiments
on several healthcare tasks and synthetic datasets - FavMac furnishes higher
value compared with several variants and baselines while maintaining strict
cost control. Our code is available at https://github.com/zlin7/FavMac
| [
{
"version": "v1",
"created": "Thu, 2 Feb 2023 02:53:12 GMT"
},
{
"version": "v2",
"created": "Sat, 18 Feb 2023 05:53:51 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Apr 2023 17:48:32 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Lin",
"Zhen",
""
],
[
"Trivedi",
"Shubhendu",
""
],
[
"Xiao",
"Cao",
""
],
[
"Sun",
"Jimeng",
""
]
] | TITLE: Fast Online Value-Maximizing Prediction Sets with Conformal Cost Control
ABSTRACT: Many real-world multi-label prediction problems involve set-valued
predictions that must satisfy specific requirements dictated by downstream
usage. We focus on a typical scenario where such requirements, separately
encoding $\textit{value}$ and $\textit{cost}$, compete with each other. For
instance, a hospital might expect a smart diagnosis system to capture as many
severe, often co-morbid, diseases as possible (the value), while maintaining
strict control over incorrect predictions (the cost). We present a general
pipeline, dubbed as FavMac, to maximize the value while controlling the cost in
such scenarios. FavMac can be combined with almost any multi-label classifier,
affording distribution-free theoretical guarantees on cost control. Moreover,
unlike prior works, it can handle real-world large-scale applications via a
carefully designed online update mechanism, which is of independent interest.
Our methodological and theoretical contributions are supported by experiments
on several healthcare tasks and synthetic datasets - FavMac furnishes higher
value compared with several variants and baselines while maintaining strict
cost control. Our code is available at https://github.com/zlin7/FavMac
|
2302.02125 | Qihua Dong | Hao Du, Qihua Dong, Yan Xu, Jing Liao | Weakly-Supervised 3D Medical Image Segmentation using Geometric Prior
and Contrastive Similarity | Weakly-supervised Segmentation, Medical Image Segmentation,
Contrastive Similarity, Geometric Prior, Point Cloud | IEEE Trans. Med. Imaging, Early Access, pp. 1-1, April 24, 2023 | 10.1109/TMI.2023.3269523 | null | eess.IV cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Medical image segmentation is almost the most important pre-processing
procedure in computer-aided diagnosis but is also a very challenging task due
to the complex shapes of segments and various artifacts caused by medical
imaging, (i.e., low-contrast tissues, and non-homogenous textures). In this
paper, we propose a simple yet effective segmentation framework that
incorporates the geometric prior and contrastive similarity into the
weakly-supervised segmentation framework in a loss-based fashion. The proposed
geometric prior built on point cloud provides meticulous geometry to the
weakly-supervised segmentation proposal, which serves as better supervision
than the inherent property of the bounding-box annotation (i.e., height and
width). Furthermore, we propose contrastive similarity to encourage organ
pixels to gather around in the contrastive embedding space, which helps better
distinguish low-contrast tissues. The proposed contrastive embedding space can
make up for the poor representation of the conventionally-used gray space.
Extensive experiments are conducted to verify the effectiveness and the
robustness of the proposed weakly-supervised segmentation framework. The
proposed framework is superior to state-of-the-art weakly-supervised methods on
the following publicly accessible datasets: LiTS 2017 Challenge, KiTS 2021
Challenge, and LPBA40. We also dissect our method and evaluate the performance
of each component.
| [
{
"version": "v1",
"created": "Sat, 4 Feb 2023 07:55:30 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Du",
"Hao",
""
],
[
"Dong",
"Qihua",
""
],
[
"Xu",
"Yan",
""
],
[
"Liao",
"Jing",
""
]
] | TITLE: Weakly-Supervised 3D Medical Image Segmentation using Geometric Prior
and Contrastive Similarity
ABSTRACT: Medical image segmentation is almost the most important pre-processing
procedure in computer-aided diagnosis but is also a very challenging task due
to the complex shapes of segments and various artifacts caused by medical
imaging, (i.e., low-contrast tissues, and non-homogenous textures). In this
paper, we propose a simple yet effective segmentation framework that
incorporates the geometric prior and contrastive similarity into the
weakly-supervised segmentation framework in a loss-based fashion. The proposed
geometric prior built on point cloud provides meticulous geometry to the
weakly-supervised segmentation proposal, which serves as better supervision
than the inherent property of the bounding-box annotation (i.e., height and
width). Furthermore, we propose contrastive similarity to encourage organ
pixels to gather around in the contrastive embedding space, which helps better
distinguish low-contrast tissues. The proposed contrastive embedding space can
make up for the poor representation of the conventionally-used gray space.
Extensive experiments are conducted to verify the effectiveness and the
robustness of the proposed weakly-supervised segmentation framework. The
proposed framework is superior to state-of-the-art weakly-supervised methods on
the following publicly accessible datasets: LiTS 2017 Challenge, KiTS 2021
Challenge, and LPBA40. We also dissect our method and evaluate the performance
of each component.
|
2302.05505 | Hyunju Kim | Hyunju Kim, Jihoon Ko, Fanchen Bu, Kijung Shin | Characterization of Simplicial Complexes by Counting Simplets Beyond
Four Nodes | Accepted to WWW 2023 - The Web Conference 2023. Simplet 0 and Simplet
1 of size 4 have been swapped in Figure 2 | null | null | null | cs.SI cs.DS | http://creativecommons.org/licenses/by/4.0/ | Simplicial complexes are higher-order combinatorial structures which have
been used to represent real-world complex systems. In this paper, we
concentrate on the local patterns in simplicial complexes called simplets, a
generalization of graphlets. We formulate the problem of counting simplets of a
given size in a given simplicial complex. For this problem, we extend a
sampling algorithm based on color coding from graphs to simplicial complexes,
with essential technical novelty. We theoretically analyze our proposed
algorithm named SC3, showing its correctness, unbiasedness, convergence, and
time/space complexity. Through the extensive experiments on sixteen real-world
datasets, we show the superiority of SC3 in terms of accuracy, speed, and
scalability, compared to the baseline methods. Finally, we use the counts given
by SC3 for simplicial complex analysis, especially for characterization, which
is further used for simplicial complex clustering, where SC3 shows a strong
ability of characterization with domain-based similarity.
| [
{
"version": "v1",
"created": "Fri, 10 Feb 2023 20:34:49 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 14:31:23 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Kim",
"Hyunju",
""
],
[
"Ko",
"Jihoon",
""
],
[
"Bu",
"Fanchen",
""
],
[
"Shin",
"Kijung",
""
]
] | TITLE: Characterization of Simplicial Complexes by Counting Simplets Beyond
Four Nodes
ABSTRACT: Simplicial complexes are higher-order combinatorial structures which have
been used to represent real-world complex systems. In this paper, we
concentrate on the local patterns in simplicial complexes called simplets, a
generalization of graphlets. We formulate the problem of counting simplets of a
given size in a given simplicial complex. For this problem, we extend a
sampling algorithm based on color coding from graphs to simplicial complexes,
with essential technical novelty. We theoretically analyze our proposed
algorithm named SC3, showing its correctness, unbiasedness, convergence, and
time/space complexity. Through the extensive experiments on sixteen real-world
datasets, we show the superiority of SC3 in terms of accuracy, speed, and
scalability, compared to the baseline methods. Finally, we use the counts given
by SC3 for simplicial complex analysis, especially for characterization, which
is further used for simplicial complex clustering, where SC3 shows a strong
ability of characterization with domain-based similarity.
|
2302.13145 | Sankha Guria | Sankha Narayan Guria, Jeffrey S. Foster, David Van Horn | Absynthe: Abstract Interpretation-Guided Synthesis | 22 pages, 6 figures, this is a preprint of a paper conditionally
accepted at Programming Language Design and Implementation (PLDI) 2023 | null | 10.1145/3591285 | null | cs.PL | http://creativecommons.org/licenses/by/4.0/ | Synthesis tools have seen significant success in recent times. However, past
approaches often require a complete and accurate embedding of the source
language in the logic of the underlying solver, an approach difficult for
industrial-grade languages. Other approaches couple the semantics of the source
language with purpose-built synthesizers, necessarily tying the synthesis
engine to a particular language model. In this paper, we propose Absynthe, an
alternative approach based on user-defined abstract semantics that aims to be
both lightweight and language agnostic, yet effective in guiding the search for
programs. A synthesis goal in Absynthe is specified as an abstract
specification in a lightweight user-defined abstract domain and concrete test
cases. The synthesis engine is parameterized by the abstract semantics and
independent of the source language. Absynthe validates candidate programs
against test cases using the actual concrete language implementation to ensure
correctness. We formalize the synthesis rules for Absynthe and describe how the
key ideas are scaled-up in our implementation in Ruby. We evaluated Absynthe on
SyGuS strings benchmark and found it competitive with other enumerative search
solvers. Moreover, Absynthe's ability to combine abstract domains allows the
user to move along a cost spectrum, i.e., expressive domains prune more
programs but require more time. Finally, to verify Absynthe can act as a
general purpose synthesis tool, we use Absynthe to synthesize Pandas data frame
manipulating programs in Python using simple abstractions like types and column
labels of a data frame. Absynthe reaches parity with AutoPandas, a deep
learning based tool for the same benchmark suite. In summary, our results
demonstrate Absynthe is a promising step forward towards a general-purpose
approach to synthesis that may broaden the applicability of synthesis to more
$\ldots$
| [
{
"version": "v1",
"created": "Sat, 25 Feb 2023 19:27:56 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 21:09:55 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Guria",
"Sankha Narayan",
""
],
[
"Foster",
"Jeffrey S.",
""
],
[
"Van Horn",
"David",
""
]
] | TITLE: Absynthe: Abstract Interpretation-Guided Synthesis
ABSTRACT: Synthesis tools have seen significant success in recent times. However, past
approaches often require a complete and accurate embedding of the source
language in the logic of the underlying solver, an approach difficult for
industrial-grade languages. Other approaches couple the semantics of the source
language with purpose-built synthesizers, necessarily tying the synthesis
engine to a particular language model. In this paper, we propose Absynthe, an
alternative approach based on user-defined abstract semantics that aims to be
both lightweight and language agnostic, yet effective in guiding the search for
programs. A synthesis goal in Absynthe is specified as an abstract
specification in a lightweight user-defined abstract domain and concrete test
cases. The synthesis engine is parameterized by the abstract semantics and
independent of the source language. Absynthe validates candidate programs
against test cases using the actual concrete language implementation to ensure
correctness. We formalize the synthesis rules for Absynthe and describe how the
key ideas are scaled-up in our implementation in Ruby. We evaluated Absynthe on
SyGuS strings benchmark and found it competitive with other enumerative search
solvers. Moreover, Absynthe's ability to combine abstract domains allows the
user to move along a cost spectrum, i.e., expressive domains prune more
programs but require more time. Finally, to verify Absynthe can act as a
general purpose synthesis tool, we use Absynthe to synthesize Pandas data frame
manipulating programs in Python using simple abstractions like types and column
labels of a data frame. Absynthe reaches parity with AutoPandas, a deep
learning based tool for the same benchmark suite. In summary, our results
demonstrate Absynthe is a promising step forward towards a general-purpose
approach to synthesis that may broaden the applicability of synthesis to more
$\ldots$
|
2302.14293 | Shizuka Tsumita | Shizuka Tsumita, Shinpei Hayashi, Sousuke Amasaki | Large-Scale Evaluation of Method-Level Bug Localization with
FinerBench4BL | 10 pages, SANER 2023 | Proceedings of the 30th IEEE International Conference on Software
Analysis, Evolution and Reengineering, 815-824, 2023 | 10.1109/SANER56733.2023.00094 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bug localization is an important aspect of software maintenance because it
can locate modules that need to be changed to fix a specific bug. Although
method-level bug localization is helpful for developers, there are only a few
tools and techniques for this task; moreover, there is no large-scale framework
for their evaluation. In this paper, we present FinerBench4BL, an evaluation
framework for method-level information retrieval-based bug localization
techniques, and a comparative study using this framework. This framework was
semi-automatically constructed from Bench4BL, a file-level bug localization
evaluation framework, using a repository transformation approach. We converted
the original file-level version repositories provided by Bench4BL into
method-level repositories by repository transformation. Method-level data
components such as oracle methods can also be automatically derived by applying
the oracle generation approach via bug-commit linking in Bench4BL to the
generated method repositories. Furthermore, we tailored existing file-level bug
localization technique implementations at the method level. We created a
framework for method-level evaluation by merging the generated dataset and
implementations. The comparison results show that the method-level techniques
decreased accuracy whereas improved debugging efficiency compared to file-level
techniques.
| [
{
"version": "v1",
"created": "Tue, 28 Feb 2023 03:57:21 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Tsumita",
"Shizuka",
""
],
[
"Hayashi",
"Shinpei",
""
],
[
"Amasaki",
"Sousuke",
""
]
] | TITLE: Large-Scale Evaluation of Method-Level Bug Localization with
FinerBench4BL
ABSTRACT: Bug localization is an important aspect of software maintenance because it
can locate modules that need to be changed to fix a specific bug. Although
method-level bug localization is helpful for developers, there are only a few
tools and techniques for this task; moreover, there is no large-scale framework
for their evaluation. In this paper, we present FinerBench4BL, an evaluation
framework for method-level information retrieval-based bug localization
techniques, and a comparative study using this framework. This framework was
semi-automatically constructed from Bench4BL, a file-level bug localization
evaluation framework, using a repository transformation approach. We converted
the original file-level version repositories provided by Bench4BL into
method-level repositories by repository transformation. Method-level data
components such as oracle methods can also be automatically derived by applying
the oracle generation approach via bug-commit linking in Bench4BL to the
generated method repositories. Furthermore, we tailored existing file-level bug
localization technique implementations at the method level. We created a
framework for method-level evaluation by merging the generated dataset and
implementations. The comparison results show that the method-level techniques
decreased accuracy whereas improved debugging efficiency compared to file-level
techniques.
|
2303.01794 | Yuta Koreeda | Yuta Koreeda, Ken-ichi Yokote, Hiroaki Ozaki, Atsuki Yamaguchi, Masaya
Tsunokake, Yasuhiro Sogawa | Hitachi at SemEval-2023 Task 3: Exploring Cross-lingual Multi-task
Strategies for Genre and Framing Detection in Online News | Accepted at SemEval-2023 Task 3 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | This paper explains the participation of team Hitachi to SemEval-2023 Task 3
"Detecting the genre, the framing, and the persuasion techniques in online news
in a multi-lingual setup.'' Based on the multilingual, multi-task nature of the
task and the low-resource setting, we investigated different cross-lingual and
multi-task strategies for training the pretrained language models. Through
extensive experiments, we found that (a) cross-lingual/multi-task training, and
(b) collecting an external balanced dataset, can benefit the genre and framing
detection. We constructed ensemble models from the results and achieved the
highest macro-averaged F1 scores in Italian and Russian genre categorization
subtasks.
| [
{
"version": "v1",
"created": "Fri, 3 Mar 2023 09:12:55 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 05:44:49 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Koreeda",
"Yuta",
""
],
[
"Yokote",
"Ken-ichi",
""
],
[
"Ozaki",
"Hiroaki",
""
],
[
"Yamaguchi",
"Atsuki",
""
],
[
"Tsunokake",
"Masaya",
""
],
[
"Sogawa",
"Yasuhiro",
""
]
] | TITLE: Hitachi at SemEval-2023 Task 3: Exploring Cross-lingual Multi-task
Strategies for Genre and Framing Detection in Online News
ABSTRACT: This paper explains the participation of team Hitachi to SemEval-2023 Task 3
"Detecting the genre, the framing, and the persuasion techniques in online news
in a multi-lingual setup.'' Based on the multilingual, multi-task nature of the
task and the low-resource setting, we investigated different cross-lingual and
multi-task strategies for training the pretrained language models. Through
extensive experiments, we found that (a) cross-lingual/multi-task training, and
(b) collecting an external balanced dataset, can benefit the genre and framing
detection. We constructed ensemble models from the results and achieved the
highest macro-averaged F1 scores in Italian and Russian genre categorization
subtasks.
|
2303.06442 | PoYung Chou | Po-Yung Chou, Yu-Yung Kao, Cheng-Hung Lin | Fine-grained Visual Classification with High-temperature Refinement and
Background Suppression | Details of the previous experiments can be found in the technical
report: arXiv:2202.03822 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Fine-grained visual classification is a challenging task due to the high
similarity between categories and distinct differences among data within one
single category. To address the challenges, previous strategies have focused on
localizing subtle discrepancies between categories and enhencing the
discriminative features in them. However, the background also provides
important information that can tell the model which features are unnecessary or
even harmful for classification, and models that rely too heavily on subtle
features may overlook global features and contextual information. In this
paper, we propose a novel network called ``High-temperaturE Refinement and
Background Suppression'' (HERBS), which consists of two modules, namely, the
high-temperature refinement module and the background suppression module, for
extracting discriminative features and suppressing background noise,
respectively. The high-temperature refinement module allows the model to learn
the appropriate feature scales by refining the features map at different scales
and improving the learning of diverse features. And, the background suppression
module first splits the features map into foreground and background using
classification confidence scores and suppresses feature values in
low-confidence areas while enhancing discriminative features. The experimental
results show that the proposed HERBS effectively fuses features of varying
scales, suppresses background noise, discriminative features at appropriate
scales for fine-grained visual classification.The proposed method achieves
state-of-the-art performance on the CUB-200-2011 and NABirds benchmarks,
surpassing 93% accuracy on both datasets. Thus, HERBS presents a promising
solution for improving the performance of fine-grained visual classification
tasks. code: https://github.com/chou141253/FGVC-HERBS
| [
{
"version": "v1",
"created": "Sat, 11 Mar 2023 16:43:59 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 00:51:55 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Chou",
"Po-Yung",
""
],
[
"Kao",
"Yu-Yung",
""
],
[
"Lin",
"Cheng-Hung",
""
]
] | TITLE: Fine-grained Visual Classification with High-temperature Refinement and
Background Suppression
ABSTRACT: Fine-grained visual classification is a challenging task due to the high
similarity between categories and distinct differences among data within one
single category. To address the challenges, previous strategies have focused on
localizing subtle discrepancies between categories and enhencing the
discriminative features in them. However, the background also provides
important information that can tell the model which features are unnecessary or
even harmful for classification, and models that rely too heavily on subtle
features may overlook global features and contextual information. In this
paper, we propose a novel network called ``High-temperaturE Refinement and
Background Suppression'' (HERBS), which consists of two modules, namely, the
high-temperature refinement module and the background suppression module, for
extracting discriminative features and suppressing background noise,
respectively. The high-temperature refinement module allows the model to learn
the appropriate feature scales by refining the features map at different scales
and improving the learning of diverse features. And, the background suppression
module first splits the features map into foreground and background using
classification confidence scores and suppresses feature values in
low-confidence areas while enhancing discriminative features. The experimental
results show that the proposed HERBS effectively fuses features of varying
scales, suppresses background noise, discriminative features at appropriate
scales for fine-grained visual classification.The proposed method achieves
state-of-the-art performance on the CUB-200-2011 and NABirds benchmarks,
surpassing 93% accuracy on both datasets. Thus, HERBS presents a promising
solution for improving the performance of fine-grained visual classification
tasks. code: https://github.com/chou141253/FGVC-HERBS
|
2303.08119 | Jiuhai Chen | Jiuhai Chen, Lichang Chen, Chen Zhu, Tianyi Zhou | How Many Demonstrations Do You Need for In-context Learning? | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) are capable to perform complex reasoning by
in-context learning (ICL) when provided with a few input-output demonstrations
(demos) and more powerful when intermediate reasoning steps ("chain of thoughts
(CoT)") of the demos are given. Is it necessary to use multi-demo in ICL? In
this paper, we study ICL using fewer demos for each test query on the tasks
in~\cite{wei2022chain}. Surprisingly, we do not observe significant degradation
when using only one randomly chosen demo. To study this phenomenon, for each
test query, we categorize demos into "correct demos" leading to the correct
answer, and "wrong demos" resulting in wrong answers. Our analysis reveals an
inherent bias in those widely studied datasets: most demos are correct for a
majority of test queries, which explains the good performance of using one
random demo. Moreover, ICL (with and w/o CoT) using only one correct demo
significantly outperforms all-demo ICL adopted by most previous works,
indicating the weakness of LLMs in finding correct demo(s) for input queries,
which is difficult to evaluate on the biased datasets. Furthermore, we observe
a counterintuitive behavior of ICL using multi-demo, i.e., its accuracy
degrades(improves) when given more correct(wrong) demos. This implies that ICL
can be easily misguided by interference among demos and their spurious
correlations. Our analyses highlight several fundamental challenges that need
to be addressed in LLMs training, ICL, and benchmark design.
| [
{
"version": "v1",
"created": "Tue, 14 Mar 2023 17:50:45 GMT"
},
{
"version": "v2",
"created": "Mon, 3 Apr 2023 14:38:04 GMT"
},
{
"version": "v3",
"created": "Mon, 24 Apr 2023 22:28:54 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Chen",
"Jiuhai",
""
],
[
"Chen",
"Lichang",
""
],
[
"Zhu",
"Chen",
""
],
[
"Zhou",
"Tianyi",
""
]
] | TITLE: How Many Demonstrations Do You Need for In-context Learning?
ABSTRACT: Large language models (LLMs) are capable to perform complex reasoning by
in-context learning (ICL) when provided with a few input-output demonstrations
(demos) and more powerful when intermediate reasoning steps ("chain of thoughts
(CoT)") of the demos are given. Is it necessary to use multi-demo in ICL? In
this paper, we study ICL using fewer demos for each test query on the tasks
in~\cite{wei2022chain}. Surprisingly, we do not observe significant degradation
when using only one randomly chosen demo. To study this phenomenon, for each
test query, we categorize demos into "correct demos" leading to the correct
answer, and "wrong demos" resulting in wrong answers. Our analysis reveals an
inherent bias in those widely studied datasets: most demos are correct for a
majority of test queries, which explains the good performance of using one
random demo. Moreover, ICL (with and w/o CoT) using only one correct demo
significantly outperforms all-demo ICL adopted by most previous works,
indicating the weakness of LLMs in finding correct demo(s) for input queries,
which is difficult to evaluate on the biased datasets. Furthermore, we observe
a counterintuitive behavior of ICL using multi-demo, i.e., its accuracy
degrades(improves) when given more correct(wrong) demos. This implies that ICL
can be easily misguided by interference among demos and their spurious
correlations. Our analyses highlight several fundamental challenges that need
to be addressed in LLMs training, ICL, and benchmark design.
|
2303.14157 | Thuan Nguyen Hoang | Thuan Hoang Nguyen, Thanh Van Le, Anh Tran | Efficient Scale-Invariant Generator with Column-Row Entangled Pixel
Synthesis | Accepted to CVPR 2023; Project Page:
https://thuanz123.github.io/creps/ | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Any-scale image synthesis offers an efficient and scalable solution to
synthesize photo-realistic images at any scale, even going beyond 2K
resolution. However, existing GAN-based solutions depend excessively on
convolutions and a hierarchical architecture, which introduce inconsistency and
the $``$texture sticking$"$ issue when scaling the output resolution. From
another perspective, INR-based generators are scale-equivariant by design, but
their huge memory footprint and slow inference hinder these networks from being
adopted in large-scale or real-time systems. In this work, we propose
$\textbf{C}$olumn-$\textbf{R}$ow $\textbf{E}$ntangled $\textbf{P}$ixel
$\textbf{S}$ynthesis ($\textbf{CREPS}$), a new generative model that is both
efficient and scale-equivariant without using any spatial convolutions or
coarse-to-fine design. To save memory footprint and make the system scalable,
we employ a novel bi-line representation that decomposes layer-wise feature
maps into separate $``$thick$"$ column and row encodings. Experiments on
various datasets, including FFHQ, LSUN-Church, MetFaces, and Flickr-Scenery,
confirm CREPS' ability to synthesize scale-consistent and alias-free images at
any arbitrary resolution with proper training and inference speed. Code is
available at https://github.com/VinAIResearch/CREPS.
| [
{
"version": "v1",
"created": "Fri, 24 Mar 2023 17:12:38 GMT"
},
{
"version": "v2",
"created": "Wed, 29 Mar 2023 01:54:49 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Apr 2023 08:49:08 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Nguyen",
"Thuan Hoang",
""
],
[
"Van Le",
"Thanh",
""
],
[
"Tran",
"Anh",
""
]
] | TITLE: Efficient Scale-Invariant Generator with Column-Row Entangled Pixel
Synthesis
ABSTRACT: Any-scale image synthesis offers an efficient and scalable solution to
synthesize photo-realistic images at any scale, even going beyond 2K
resolution. However, existing GAN-based solutions depend excessively on
convolutions and a hierarchical architecture, which introduce inconsistency and
the $``$texture sticking$"$ issue when scaling the output resolution. From
another perspective, INR-based generators are scale-equivariant by design, but
their huge memory footprint and slow inference hinder these networks from being
adopted in large-scale or real-time systems. In this work, we propose
$\textbf{C}$olumn-$\textbf{R}$ow $\textbf{E}$ntangled $\textbf{P}$ixel
$\textbf{S}$ynthesis ($\textbf{CREPS}$), a new generative model that is both
efficient and scale-equivariant without using any spatial convolutions or
coarse-to-fine design. To save memory footprint and make the system scalable,
we employ a novel bi-line representation that decomposes layer-wise feature
maps into separate $``$thick$"$ column and row encodings. Experiments on
various datasets, including FFHQ, LSUN-Church, MetFaces, and Flickr-Scenery,
confirm CREPS' ability to synthesize scale-consistent and alias-free images at
any arbitrary resolution with proper training and inference speed. Code is
available at https://github.com/VinAIResearch/CREPS.
|
2303.15839 | Suncheng Xiang | Dan You, Pengcheng Xia, Qiuzhu Chen, Minghui Wu, Suncheng Xiang, Jun
Wang | AutoKary2022: A Large-Scale Densely Annotated Dataset for Chromosome
Instance Segmentation | Accepted by ICME 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automated chromosome instance segmentation from metaphase cell microscopic
images is critical for the diagnosis of chromosomal disorders (i.e., karyotype
analysis). However, it is still a challenging task due to lacking of densely
annotated datasets and the complicated morphologies of chromosomes, e.g., dense
distribution, arbitrary orientations, and wide range of lengths. To facilitate
the development of this area, we take a big step forward and manually construct
a large-scale densely annotated dataset named AutoKary2022, which contains over
27,000 chromosome instances in 612 microscopic images from 50 patients.
Specifically, each instance is annotated with a polygonal mask and a class
label to assist in precise chromosome detection and segmentation. On top of it,
we systematically investigate representative methods on this dataset and obtain
a number of interesting findings, which helps us have a deeper understanding of
the fundamental problems in chromosome instance segmentation. We hope this
dataset could advance research towards medical understanding. The dataset can
be available at:
https://github.com/wangjuncongyu/chromosome-instance-segmentation-dataset.
| [
{
"version": "v1",
"created": "Tue, 28 Mar 2023 09:15:55 GMT"
},
{
"version": "v2",
"created": "Fri, 31 Mar 2023 12:16:03 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Apr 2023 11:47:31 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"You",
"Dan",
""
],
[
"Xia",
"Pengcheng",
""
],
[
"Chen",
"Qiuzhu",
""
],
[
"Wu",
"Minghui",
""
],
[
"Xiang",
"Suncheng",
""
],
[
"Wang",
"Jun",
""
]
] | TITLE: AutoKary2022: A Large-Scale Densely Annotated Dataset for Chromosome
Instance Segmentation
ABSTRACT: Automated chromosome instance segmentation from metaphase cell microscopic
images is critical for the diagnosis of chromosomal disorders (i.e., karyotype
analysis). However, it is still a challenging task due to lacking of densely
annotated datasets and the complicated morphologies of chromosomes, e.g., dense
distribution, arbitrary orientations, and wide range of lengths. To facilitate
the development of this area, we take a big step forward and manually construct
a large-scale densely annotated dataset named AutoKary2022, which contains over
27,000 chromosome instances in 612 microscopic images from 50 patients.
Specifically, each instance is annotated with a polygonal mask and a class
label to assist in precise chromosome detection and segmentation. On top of it,
we systematically investigate representative methods on this dataset and obtain
a number of interesting findings, which helps us have a deeper understanding of
the fundamental problems in chromosome instance segmentation. We hope this
dataset could advance research towards medical understanding. The dataset can
be available at:
https://github.com/wangjuncongyu/chromosome-instance-segmentation-dataset.
|
2304.02787 | Pavlos Fragkogiannis | Pavlos Fragkogiannis, Martina Forster, Grace E. Lee, Dell Zhang | Context-Aware Classification of Legal Document Pages | 5 pages. To be published in the Proceedings of the 46th International
ACM SIGIR Conference on Research and Development in Information Retrieval
(SIGIR '23) | null | 10.1145/3539618.3591839 | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For many business applications that require the processing, indexing, and
retrieval of professional documents such as legal briefs (in PDF format etc.),
it is often essential to classify the pages of any given document into their
corresponding types beforehand. Most existing studies in the field of document
image classification either focus on single-page documents or treat multiple
pages in a document independently. Although in recent years a few techniques
have been proposed to exploit the context information from neighboring pages to
enhance document page classification, they typically cannot be utilized with
large pre-trained language models due to the constraint on input length. In
this paper, we present a simple but effective approach that overcomes the above
limitation. Specifically, we enhance the input with extra tokens carrying
sequential information about previous pages - introducing recurrence - which
enables the usage of pre-trained Transformer models like BERT for context-aware
page classification. Our experiments conducted on two legal datasets in English
and Portuguese respectively show that the proposed approach can significantly
improve the performance of document page classification compared to the
non-recurrent setup as well as the other context-aware baselines.
| [
{
"version": "v1",
"created": "Wed, 5 Apr 2023 23:14:58 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 14:59:24 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Fragkogiannis",
"Pavlos",
""
],
[
"Forster",
"Martina",
""
],
[
"Lee",
"Grace E.",
""
],
[
"Zhang",
"Dell",
""
]
] | TITLE: Context-Aware Classification of Legal Document Pages
ABSTRACT: For many business applications that require the processing, indexing, and
retrieval of professional documents such as legal briefs (in PDF format etc.),
it is often essential to classify the pages of any given document into their
corresponding types beforehand. Most existing studies in the field of document
image classification either focus on single-page documents or treat multiple
pages in a document independently. Although in recent years a few techniques
have been proposed to exploit the context information from neighboring pages to
enhance document page classification, they typically cannot be utilized with
large pre-trained language models due to the constraint on input length. In
this paper, we present a simple but effective approach that overcomes the above
limitation. Specifically, we enhance the input with extra tokens carrying
sequential information about previous pages - introducing recurrence - which
enables the usage of pre-trained Transformer models like BERT for context-aware
page classification. Our experiments conducted on two legal datasets in English
and Portuguese respectively show that the proposed approach can significantly
improve the performance of document page classification compared to the
non-recurrent setup as well as the other context-aware baselines.
|
2304.03973 | Alberto Marchisio | Alberto Marchisio and Antonio De Marco and Alessio Colucci and
Maurizio Martina and Muhammad Shafique | RobCaps: Evaluating the Robustness of Capsule Networks against Affine
Transformations and Adversarial Attacks | To appear at the 2023 International Joint Conference on Neural
Networks (IJCNN), Queensland, Australia, June 2023 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Capsule Networks (CapsNets) are able to hierarchically preserve the pose
relationships between multiple objects for image classification tasks. Other
than achieving high accuracy, another relevant factor in deploying CapsNets in
safety-critical applications is the robustness against input transformations
and malicious adversarial attacks.
In this paper, we systematically analyze and evaluate different factors
affecting the robustness of CapsNets, compared to traditional Convolutional
Neural Networks (CNNs). Towards a comprehensive comparison, we test two CapsNet
models and two CNN models on the MNIST, GTSRB, and CIFAR10 datasets, as well as
on the affine-transformed versions of such datasets. With a thorough analysis,
we show which properties of these architectures better contribute to increasing
the robustness and their limitations. Overall, CapsNets achieve better
robustness against adversarial examples and affine transformations, compared to
a traditional CNN with a similar number of parameters. Similar conclusions have
been derived for deeper versions of CapsNets and CNNs. Moreover, our results
unleash a key finding that the dynamic routing does not contribute much to
improving the CapsNets' robustness. Indeed, the main generalization
contribution is due to the hierarchical feature learning through capsules.
| [
{
"version": "v1",
"created": "Sat, 8 Apr 2023 09:58:35 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 10:35:37 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Marchisio",
"Alberto",
""
],
[
"De Marco",
"Antonio",
""
],
[
"Colucci",
"Alessio",
""
],
[
"Martina",
"Maurizio",
""
],
[
"Shafique",
"Muhammad",
""
]
] | TITLE: RobCaps: Evaluating the Robustness of Capsule Networks against Affine
Transformations and Adversarial Attacks
ABSTRACT: Capsule Networks (CapsNets) are able to hierarchically preserve the pose
relationships between multiple objects for image classification tasks. Other
than achieving high accuracy, another relevant factor in deploying CapsNets in
safety-critical applications is the robustness against input transformations
and malicious adversarial attacks.
In this paper, we systematically analyze and evaluate different factors
affecting the robustness of CapsNets, compared to traditional Convolutional
Neural Networks (CNNs). Towards a comprehensive comparison, we test two CapsNet
models and two CNN models on the MNIST, GTSRB, and CIFAR10 datasets, as well as
on the affine-transformed versions of such datasets. With a thorough analysis,
we show which properties of these architectures better contribute to increasing
the robustness and their limitations. Overall, CapsNets achieve better
robustness against adversarial examples and affine transformations, compared to
a traditional CNN with a similar number of parameters. Similar conclusions have
been derived for deeper versions of CapsNets and CNNs. Moreover, our results
unleash a key finding that the dynamic routing does not contribute much to
improving the CapsNets' robustness. Indeed, the main generalization
contribution is due to the hierarchical feature learning through capsules.
|
2304.06280 | Yuhan Liu | Yuhan Liu, Zhaoxuan Tan, Heng Wang, Shangbin Feng, Qinghua Zheng, and
Minnan Luo | BotMoE: Twitter Bot Detection with Community-Aware Mixtures of
Modal-Specific Experts | Accepted at SIGIR 2023 | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Twitter bot detection has become a crucial task in efforts to combat online
misinformation, mitigate election interference, and curb malicious propaganda.
However, advanced Twitter bots often attempt to mimic the characteristics of
genuine users through feature manipulation and disguise themselves to fit in
diverse user communities, posing challenges for existing Twitter bot detection
models. To this end, we propose BotMoE, a Twitter bot detection framework that
jointly utilizes multiple user information modalities (metadata, textual
content, network structure) to improve the detection of deceptive bots.
Furthermore, BotMoE incorporates a community-aware Mixture-of-Experts (MoE)
layer to improve domain generalization and adapt to different Twitter
communities. Specifically, BotMoE constructs modal-specific encoders for
metadata features, textual content, and graphical structure, which jointly
model Twitter users from three modal-specific perspectives. We then employ a
community-aware MoE layer to automatically assign users to different
communities and leverage the corresponding expert networks. Finally, user
representations from metadata, text, and graph perspectives are fused with an
expert fusion layer, combining all three modalities while measuring the
consistency of user information. Extensive experiments demonstrate that BotMoE
significantly advances the state-of-the-art on three Twitter bot detection
benchmarks. Studies also confirm that BotMoE captures advanced and evasive
bots, alleviates the reliance on training data, and better generalizes to new
and previously unseen user communities.
| [
{
"version": "v1",
"created": "Thu, 13 Apr 2023 05:55:10 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 08:04:46 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Liu",
"Yuhan",
""
],
[
"Tan",
"Zhaoxuan",
""
],
[
"Wang",
"Heng",
""
],
[
"Feng",
"Shangbin",
""
],
[
"Zheng",
"Qinghua",
""
],
[
"Luo",
"Minnan",
""
]
] | TITLE: BotMoE: Twitter Bot Detection with Community-Aware Mixtures of
Modal-Specific Experts
ABSTRACT: Twitter bot detection has become a crucial task in efforts to combat online
misinformation, mitigate election interference, and curb malicious propaganda.
However, advanced Twitter bots often attempt to mimic the characteristics of
genuine users through feature manipulation and disguise themselves to fit in
diverse user communities, posing challenges for existing Twitter bot detection
models. To this end, we propose BotMoE, a Twitter bot detection framework that
jointly utilizes multiple user information modalities (metadata, textual
content, network structure) to improve the detection of deceptive bots.
Furthermore, BotMoE incorporates a community-aware Mixture-of-Experts (MoE)
layer to improve domain generalization and adapt to different Twitter
communities. Specifically, BotMoE constructs modal-specific encoders for
metadata features, textual content, and graphical structure, which jointly
model Twitter users from three modal-specific perspectives. We then employ a
community-aware MoE layer to automatically assign users to different
communities and leverage the corresponding expert networks. Finally, user
representations from metadata, text, and graph perspectives are fused with an
expert fusion layer, combining all three modalities while measuring the
consistency of user information. Extensive experiments demonstrate that BotMoE
significantly advances the state-of-the-art on three Twitter bot detection
benchmarks. Studies also confirm that BotMoE captures advanced and evasive
bots, alleviates the reliance on training data, and better generalizes to new
and previously unseen user communities.
|
2304.07987 | Ruibin Yuan | Ge Zhang, Yemin Shi, Ruibo Liu, Ruibin Yuan, Yizhi Li, Siwei Dong, Yu
Shu, Zhaoqun Li, Zekun Wang, Chenghua Lin, Wenhao Huang, Jie Fu | Chinese Open Instruction Generalist: A Preliminary Release | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Instruction tuning is widely recognized as a key technique for building
generalist language models, which has attracted the attention of researchers
and the public with the release of InstructGPT~\citep{ouyang2022training} and
ChatGPT\footnote{\url{https://chat.openai.com/}}. Despite impressive progress
in English-oriented large-scale language models (LLMs), it is still
under-explored whether English-based foundation LLMs can perform similarly on
multilingual tasks compared to English tasks with well-designed instruction
tuning and how we can construct the corpora needed for the tuning. To remedy
this gap, we propose the project as an attempt to create a Chinese instruction
dataset by various methods adapted to the intrinsic characteristics of 4
sub-tasks. We collect around 200k Chinese instruction tuning samples, which
have been manually checked to guarantee high quality. We also summarize the
existing English and Chinese instruction corpora and briefly describe some
potential applications of the newly constructed Chinese instruction corpora.
The resulting \textbf{C}hinese \textbf{O}pen \textbf{I}nstruction
\textbf{G}eneralist (\textbf{COIG}) corpora are available in
Huggingface\footnote{\url{https://huggingface.co/datasets/BAAI/COIG}} and
Github\footnote{\url{https://github.com/BAAI-Zlab/COIG}}, and will be
continuously updated.
| [
{
"version": "v1",
"created": "Mon, 17 Apr 2023 04:45:06 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Apr 2023 04:46:57 GMT"
},
{
"version": "v3",
"created": "Fri, 21 Apr 2023 03:16:13 GMT"
},
{
"version": "v4",
"created": "Tue, 25 Apr 2023 01:50:19 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Zhang",
"Ge",
""
],
[
"Shi",
"Yemin",
""
],
[
"Liu",
"Ruibo",
""
],
[
"Yuan",
"Ruibin",
""
],
[
"Li",
"Yizhi",
""
],
[
"Dong",
"Siwei",
""
],
[
"Shu",
"Yu",
""
],
[
"Li",
"Zhaoqun",
""
],
[
"Wang",
"Zekun",
""
],
[
"Lin",
"Chenghua",
""
],
[
"Huang",
"Wenhao",
""
],
[
"Fu",
"Jie",
""
]
] | TITLE: Chinese Open Instruction Generalist: A Preliminary Release
ABSTRACT: Instruction tuning is widely recognized as a key technique for building
generalist language models, which has attracted the attention of researchers
and the public with the release of InstructGPT~\citep{ouyang2022training} and
ChatGPT\footnote{\url{https://chat.openai.com/}}. Despite impressive progress
in English-oriented large-scale language models (LLMs), it is still
under-explored whether English-based foundation LLMs can perform similarly on
multilingual tasks compared to English tasks with well-designed instruction
tuning and how we can construct the corpora needed for the tuning. To remedy
this gap, we propose the project as an attempt to create a Chinese instruction
dataset by various methods adapted to the intrinsic characteristics of 4
sub-tasks. We collect around 200k Chinese instruction tuning samples, which
have been manually checked to guarantee high quality. We also summarize the
existing English and Chinese instruction corpora and briefly describe some
potential applications of the newly constructed Chinese instruction corpora.
The resulting \textbf{C}hinese \textbf{O}pen \textbf{I}nstruction
\textbf{G}eneralist (\textbf{COIG}) corpora are available in
Huggingface\footnote{\url{https://huggingface.co/datasets/BAAI/COIG}} and
Github\footnote{\url{https://github.com/BAAI-Zlab/COIG}}, and will be
continuously updated.
|
2304.08953 | Adarsh Kumar | Adarsh Kumar and Pedro Sarmento | From Words to Music: A Study of Subword Tokenization Techniques in
Symbolic Music Generation | null | null | null | null | cs.SD cs.LG eess.AS | http://creativecommons.org/licenses/by/4.0/ | Subword tokenization has been widely successful in text-based natural
language processing (NLP) tasks with Transformer-based models. As Transformer
models become increasingly popular in symbolic music-related studies, it is
imperative to investigate the efficacy of subword tokenization in the symbolic
music domain. In this paper, we explore subword tokenization techniques, such
as byte-pair encoding (BPE), in symbolic music generation and its impact on the
overall structure of generated songs. Our experiments are based on three types
of MIDI datasets: single track-melody only, multi-track with a single
instrument, and multi-track and multi-instrument. We apply subword tokenization
on post-musical tokenization schemes and find that it enables the generation of
longer songs at the same time and improves the overall structure of the
generated music in terms of objective metrics like structure indicator (SI),
Pitch Class Entropy, etc. We also compare two subword tokenization methods, BPE
and Unigram, and observe that both methods lead to consistent improvements. Our
study suggests that subword tokenization is a promising technique for symbolic
music generation and may have broader implications for music composition,
particularly in cases involving complex data such as multi-track songs.
| [
{
"version": "v1",
"created": "Tue, 18 Apr 2023 12:46:12 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 11:16:53 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Kumar",
"Adarsh",
""
],
[
"Sarmento",
"Pedro",
""
]
] | TITLE: From Words to Music: A Study of Subword Tokenization Techniques in
Symbolic Music Generation
ABSTRACT: Subword tokenization has been widely successful in text-based natural
language processing (NLP) tasks with Transformer-based models. As Transformer
models become increasingly popular in symbolic music-related studies, it is
imperative to investigate the efficacy of subword tokenization in the symbolic
music domain. In this paper, we explore subword tokenization techniques, such
as byte-pair encoding (BPE), in symbolic music generation and its impact on the
overall structure of generated songs. Our experiments are based on three types
of MIDI datasets: single track-melody only, multi-track with a single
instrument, and multi-track and multi-instrument. We apply subword tokenization
on post-musical tokenization schemes and find that it enables the generation of
longer songs at the same time and improves the overall structure of the
generated music in terms of objective metrics like structure indicator (SI),
Pitch Class Entropy, etc. We also compare two subword tokenization methods, BPE
and Unigram, and observe that both methods lead to consistent improvements. Our
study suggests that subword tokenization is a promising technique for symbolic
music generation and may have broader implications for music composition,
particularly in cases involving complex data such as multi-track songs.
|
2304.09607 | Yaoxun Xu | Yaoxun Xu and Baiji Liu and Qiaochu Huang and, Xingchen Song and
Zhiyong Wu and Shiyin Kang and Helen Meng | CB-Conformer: Contextual biasing Conformer for biased word recognition | null | null | null | null | cs.SD cs.CL eess.AS | http://creativecommons.org/licenses/by/4.0/ | Due to the mismatch between the source and target domains, how to better
utilize the biased word information to improve the performance of the automatic
speech recognition model in the target domain becomes a hot research topic.
Previous approaches either decode with a fixed external language model or
introduce a sizeable biasing module, which leads to poor adaptability and slow
inference. In this work, we propose CB-Conformer to improve biased word
recognition by introducing the Contextual Biasing Module and the Self-Adaptive
Language Model to vanilla Conformer. The Contextual Biasing Module combines
audio fragments and contextual information, with only 0.2% model parameters of
the original Conformer. The Self-Adaptive Language Model modifies the internal
weights of biased words based on their recall and precision, resulting in a
greater focus on biased words and more successful integration with the
automatic speech recognition model than the standard fixed language model. In
addition, we construct and release an open-source Mandarin biased-word dataset
based on WenetSpeech. Experiments indicate that our proposed method brings a
15.34% character error rate reduction, a 14.13% biased word recall increase,
and a 6.80% biased word F1-score increase compared with the base Conformer.
| [
{
"version": "v1",
"created": "Wed, 19 Apr 2023 12:26:04 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 07:45:40 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Xu",
"Yaoxun",
""
],
[
"Liu",
"Baiji",
""
],
[
"and",
"Qiaochu Huang",
""
],
[
"Song",
"Xingchen",
""
],
[
"Wu",
"Zhiyong",
""
],
[
"Kang",
"Shiyin",
""
],
[
"Meng",
"Helen",
""
]
] | TITLE: CB-Conformer: Contextual biasing Conformer for biased word recognition
ABSTRACT: Due to the mismatch between the source and target domains, how to better
utilize the biased word information to improve the performance of the automatic
speech recognition model in the target domain becomes a hot research topic.
Previous approaches either decode with a fixed external language model or
introduce a sizeable biasing module, which leads to poor adaptability and slow
inference. In this work, we propose CB-Conformer to improve biased word
recognition by introducing the Contextual Biasing Module and the Self-Adaptive
Language Model to vanilla Conformer. The Contextual Biasing Module combines
audio fragments and contextual information, with only 0.2% model parameters of
the original Conformer. The Self-Adaptive Language Model modifies the internal
weights of biased words based on their recall and precision, resulting in a
greater focus on biased words and more successful integration with the
automatic speech recognition model than the standard fixed language model. In
addition, we construct and release an open-source Mandarin biased-word dataset
based on WenetSpeech. Experiments indicate that our proposed method brings a
15.34% character error rate reduction, a 14.13% biased word recall increase,
and a 6.80% biased word F1-score increase compared with the base Conformer.
|
2304.10089 | Qianhui Sun | Qianhui Sun, Qingyu Yang, Chongyi Li, Shangchen Zhou, Ruicheng Feng,
Yuekun Dai, Wenxiu Sun, Qingpeng Zhu, Chen Change Loy, Jinwei Gu | MIPI 2023 Challenge on RGBW Fusion: Methods and Results | CVPR 2023 Mobile Intelligent Photography and Imaging (MIPI)
Workshop--RGBW Sensor Fusion Challenge Report. Website:
https://mipi-challenge.org/MIPI2023/. arXiv admin note: substantial text
overlap with arXiv:2209.07530, arXiv:2209.08471, arXiv:2209.07060 | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Developing and integrating advanced image sensors with novel algorithms in
camera systems are prevalent with the increasing demand for computational
photography and imaging on mobile platforms. However, the lack of high-quality
data for research and the rare opportunity for an in-depth exchange of views
from industry and academia constrain the development of mobile intelligent
photography and imaging (MIPI). With the success of the 1st MIPI Workshop@ECCV
2022, we introduce the second MIPI challenge, including four tracks focusing on
novel image sensors and imaging algorithms. This paper summarizes and reviews
the RGBW Joint Fusion and Denoise track on MIPI 2023. In total, 69 participants
were successfully registered, and 4 teams submitted results in the final
testing phase. The final results are evaluated using objective metrics,
including PSNR, SSIM, LPIPS, and KLD. A detailed description of the top three
models developed in this challenge is provided in this paper. More details of
this challenge and the link to the dataset can be found at
https://mipi-challenge.org/MIPI2023/.
| [
{
"version": "v1",
"created": "Thu, 20 Apr 2023 05:02:03 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 21:13:04 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Sun",
"Qianhui",
""
],
[
"Yang",
"Qingyu",
""
],
[
"Li",
"Chongyi",
""
],
[
"Zhou",
"Shangchen",
""
],
[
"Feng",
"Ruicheng",
""
],
[
"Dai",
"Yuekun",
""
],
[
"Sun",
"Wenxiu",
""
],
[
"Zhu",
"Qingpeng",
""
],
[
"Loy",
"Chen Change",
""
],
[
"Gu",
"Jinwei",
""
]
] | TITLE: MIPI 2023 Challenge on RGBW Fusion: Methods and Results
ABSTRACT: Developing and integrating advanced image sensors with novel algorithms in
camera systems are prevalent with the increasing demand for computational
photography and imaging on mobile platforms. However, the lack of high-quality
data for research and the rare opportunity for an in-depth exchange of views
from industry and academia constrain the development of mobile intelligent
photography and imaging (MIPI). With the success of the 1st MIPI Workshop@ECCV
2022, we introduce the second MIPI challenge, including four tracks focusing on
novel image sensors and imaging algorithms. This paper summarizes and reviews
the RGBW Joint Fusion and Denoise track on MIPI 2023. In total, 69 participants
were successfully registered, and 4 teams submitted results in the final
testing phase. The final results are evaluated using objective metrics,
including PSNR, SSIM, LPIPS, and KLD. A detailed description of the top three
models developed in this challenge is provided in this paper. More details of
this challenge and the link to the dataset can be found at
https://mipi-challenge.org/MIPI2023/.
|
2304.10859 | Karthick Prasad Gunasekaran | Karthick Prasad Gunasekaran, B Chase Babrich, Saurabh Shirodkar, Hee
Hwang | Text2Time: Transformer-based Article Time Period Prediction | 8 Pages | null | 10.13140/RG.2.2.29195.36641 | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The task of predicting the publication period of text documents, such as news
articles, is an important but less studied problem in the field of natural
language processing. Predicting the year of a news article can be useful in
various contexts, such as historical research, sentiment analysis, and media
monitoring. In this work, we investigate the problem of predicting the
publication period of a text document, specifically a news article, based on
its textual content. In order to do so, we created our own extensive labeled
dataset of over 350,000 news articles published by The New York Times over six
decades. In our approach, we use a pretrained BERT model fine-tuned for the
task of text classification, specifically for time period prediction.This model
exceeds our expectations and provides some very impressive results in terms of
accurately classifying news articles into their respective publication decades.
The results beat the performance of the baseline model for this relatively
unexplored task of time prediction from text.
| [
{
"version": "v1",
"created": "Fri, 21 Apr 2023 10:05:03 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 03:56:03 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Gunasekaran",
"Karthick Prasad",
""
],
[
"Babrich",
"B Chase",
""
],
[
"Shirodkar",
"Saurabh",
""
],
[
"Hwang",
"Hee",
""
]
] | TITLE: Text2Time: Transformer-based Article Time Period Prediction
ABSTRACT: The task of predicting the publication period of text documents, such as news
articles, is an important but less studied problem in the field of natural
language processing. Predicting the year of a news article can be useful in
various contexts, such as historical research, sentiment analysis, and media
monitoring. In this work, we investigate the problem of predicting the
publication period of a text document, specifically a news article, based on
its textual content. In order to do so, we created our own extensive labeled
dataset of over 350,000 news articles published by The New York Times over six
decades. In our approach, we use a pretrained BERT model fine-tuned for the
task of text classification, specifically for time period prediction.This model
exceeds our expectations and provides some very impressive results in terms of
accurately classifying news articles into their respective publication decades.
The results beat the performance of the baseline model for this relatively
unexplored task of time prediction from text.
|
2304.11779 | Karthick Prasad Gunasekaran | Karthick Prasad Gunasekaran, Nikita Jaiman | Now You See Me: Robust approach to Partial Occlusions | 6 pages | null | 10.13140/RG.2.2.25839.92323 | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Occlusions of objects is one of the indispensable problems in Computer
vision. While Convolutional Neural Net-works (CNNs) provide various state of
the art approaches for regular image classification, they however, prove to be
not as effective for the classification of images with partial occlusions.
Partial occlusion is scenario where an object is occluded partially by some
other object/space. This problem when solved,holds tremendous potential to
facilitate various scenarios. We in particular are interested in autonomous
driving scenario and its implications in the same. Autonomous vehicle research
is one of the hot topics of this decade, there are ample situations of partial
occlusions of a driving sign or a person or other objects at different angles.
Considering its prime importance in situations which can be further extended to
video analytics of traffic data to handle crimes, anticipate income levels of
various groups etc.,this holds the potential to be exploited in many ways. In
this paper, we introduce our own synthetically created dataset by utilising
Stanford Car Dataset and adding occlusions of various sizes and nature to it.
On this created dataset, we conducted a comprehensive analysis using various
state of the art CNN models such as VGG-19, ResNet 50/101, GoogleNet, DenseNet
121. We further in depth study the effect of varying occlusion proportions and
nature on the performance of these models by fine tuning and training these
from scratch on dataset and how is it likely to perform when trained in
different scenarios, i.e., performance when training with occluded images and
unoccluded images, which model is more robust to partial occlusions and soon.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 00:31:49 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 11:45:50 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Gunasekaran",
"Karthick Prasad",
""
],
[
"Jaiman",
"Nikita",
""
]
] | TITLE: Now You See Me: Robust approach to Partial Occlusions
ABSTRACT: Occlusions of objects is one of the indispensable problems in Computer
vision. While Convolutional Neural Net-works (CNNs) provide various state of
the art approaches for regular image classification, they however, prove to be
not as effective for the classification of images with partial occlusions.
Partial occlusion is scenario where an object is occluded partially by some
other object/space. This problem when solved,holds tremendous potential to
facilitate various scenarios. We in particular are interested in autonomous
driving scenario and its implications in the same. Autonomous vehicle research
is one of the hot topics of this decade, there are ample situations of partial
occlusions of a driving sign or a person or other objects at different angles.
Considering its prime importance in situations which can be further extended to
video analytics of traffic data to handle crimes, anticipate income levels of
various groups etc.,this holds the potential to be exploited in many ways. In
this paper, we introduce our own synthetically created dataset by utilising
Stanford Car Dataset and adding occlusions of various sizes and nature to it.
On this created dataset, we conducted a comprehensive analysis using various
state of the art CNN models such as VGG-19, ResNet 50/101, GoogleNet, DenseNet
121. We further in depth study the effect of varying occlusion proportions and
nature on the performance of these models by fine tuning and training these
from scratch on dataset and how is it likely to perform when trained in
different scenarios, i.e., performance when training with occluded images and
unoccluded images, which model is more robust to partial occlusions and soon.
|
2304.12329 | George Papadakis | Alexandros Zeakis, George Papadakis, Dimitrios Skoutas, Manolis
Koubarakis | Pre-trained Embeddings for Entity Resolution: An Experimental Analysis
[Experiment, Analysis & Benchmark] | null | null | null | null | cs.DB cs.CL | http://creativecommons.org/licenses/by/4.0/ | Many recent works on Entity Resolution (ER) leverage Deep Learning techniques
involving language models to improve effectiveness. This is applied to both
main steps of ER, i.e., blocking and matching. Several pre-trained embeddings
have been tested, with the most popular ones being fastText and variants of the
BERT model. However, there is no detailed analysis of their pros and cons. To
cover this gap, we perform a thorough experimental analysis of 12 popular
language models over 17 established benchmark datasets. First, we assess their
vectorization overhead for converting all input entities into dense embeddings
vectors. Second, we investigate their blocking performance, performing a
detailed scalability analysis, and comparing them with the state-of-the-art
deep learning-based blocking method. Third, we conclude with their relative
performance for both supervised and unsupervised matching. Our experimental
results provide novel insights into the strengths and weaknesses of the main
language models, facilitating researchers and practitioners to select the most
suitable ones in practice.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 08:53:54 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Zeakis",
"Alexandros",
""
],
[
"Papadakis",
"George",
""
],
[
"Skoutas",
"Dimitrios",
""
],
[
"Koubarakis",
"Manolis",
""
]
] | TITLE: Pre-trained Embeddings for Entity Resolution: An Experimental Analysis
[Experiment, Analysis & Benchmark]
ABSTRACT: Many recent works on Entity Resolution (ER) leverage Deep Learning techniques
involving language models to improve effectiveness. This is applied to both
main steps of ER, i.e., blocking and matching. Several pre-trained embeddings
have been tested, with the most popular ones being fastText and variants of the
BERT model. However, there is no detailed analysis of their pros and cons. To
cover this gap, we perform a thorough experimental analysis of 12 popular
language models over 17 established benchmark datasets. First, we assess their
vectorization overhead for converting all input entities into dense embeddings
vectors. Second, we investigate their blocking performance, performing a
detailed scalability analysis, and comparing them with the state-of-the-art
deep learning-based blocking method. Third, we conclude with their relative
performance for both supervised and unsupervised matching. Our experimental
results provide novel insights into the strengths and weaknesses of the main
language models, facilitating researchers and practitioners to select the most
suitable ones in practice.
|
2304.12331 | Arthur Vervaet | Arthur Vervaet (ISEP), Raja Chiky (ISEP), Mar Callau-Zori | USTEP: Structuration des logs en flux gr{\^a}ce {\`a} un arbre de
recherche {\'e}volutif | in French language, Extraction et Gestion des Connaissances:
EGC'2022, 2022 | null | null | null | cs.CL cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Logs record valuable system information at runtime. They are widely used by
data-driven approaches for development and monitoring purposes. Parsing log
messages to structure their format is a classic preliminary step for log-mining
tasks. As they appear upstream, parsing operations can become a processing time
bottleneck for downstream applications. The quality of parsing also has a
direct influence on their efficiency. Here, we propose USTEP, an online log
parsing method based on an evolving tree structure. Evaluation results on a
wide panel of datasets coming from different real-world systems demonstrate
USTEP superiority in terms of both effectiveness and robustness when compared
to other online methods.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 09:12:00 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Vervaet",
"Arthur",
"",
"ISEP"
],
[
"Chiky",
"Raja",
"",
"ISEP"
],
[
"Callau-Zori",
"Mar",
""
]
] | TITLE: USTEP: Structuration des logs en flux gr{\^a}ce {\`a} un arbre de
recherche {\'e}volutif
ABSTRACT: Logs record valuable system information at runtime. They are widely used by
data-driven approaches for development and monitoring purposes. Parsing log
messages to structure their format is a classic preliminary step for log-mining
tasks. As they appear upstream, parsing operations can become a processing time
bottleneck for downstream applications. The quality of parsing also has a
direct influence on their efficiency. Here, we propose USTEP, an online log
parsing method based on an evolving tree structure. Evaluation results on a
wide panel of datasets coming from different real-world systems demonstrate
USTEP superiority in terms of both effectiveness and robustness when compared
to other online methods.
|
2304.12332 | \'Angel L\'opez-Oriona | \'Angel L\'opez Oriona and Jos\'e Antonio Vilar Fern\'andez | Analyzing categorical time series with the R package ctsfeatures | arXiv admin note: text overlap with arXiv:2304.12251 | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | Time series data are ubiquitous nowadays. Whereas most of the literature on
the topic deals with real-valued time series, categorical time series have
received much less attention. However, the development of data mining
techniques for this kind of data has substantially increased in recent years.
The R package ctsfeatures offers users a set of useful tools for analyzing
categorical time series. In particular, several functions allowing the
extraction of well-known statistical features and the construction of
illustrative graphs describing underlying temporal patterns are provided in the
package. The output of some functions can be employed to perform traditional
machine learning tasks including clustering, classification and outlier
detection. The package also includes two datasets of biological sequences
introduced in the literature for clustering purposes, as well as three
interesting synthetic databases. In this work, the main characteristics of the
package are described and its use is illustrated through various examples.
Practitioners from a wide variety of fields could benefit from the valuable
tools provided by ctsfeatures.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 16:16:56 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Oriona",
"Ángel López",
""
],
[
"Fernández",
"José Antonio Vilar",
""
]
] | TITLE: Analyzing categorical time series with the R package ctsfeatures
ABSTRACT: Time series data are ubiquitous nowadays. Whereas most of the literature on
the topic deals with real-valued time series, categorical time series have
received much less attention. However, the development of data mining
techniques for this kind of data has substantially increased in recent years.
The R package ctsfeatures offers users a set of useful tools for analyzing
categorical time series. In particular, several functions allowing the
extraction of well-known statistical features and the construction of
illustrative graphs describing underlying temporal patterns are provided in the
package. The output of some functions can be employed to perform traditional
machine learning tasks including clustering, classification and outlier
detection. The package also includes two datasets of biological sequences
introduced in the literature for clustering purposes, as well as three
interesting synthetic databases. In this work, the main characteristics of the
package are described and its use is illustrated through various examples.
Practitioners from a wide variety of fields could benefit from the valuable
tools provided by ctsfeatures.
|
2304.12370 | Yudhanjaya Wijeratne | Yudhanjaya Wijeratne, Ishan Marikar | Better Question-Answering Models on a Budget | 5 pages, code and datasets available | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Low-rank adaptation (LoRA) and question-answer datasets from large language
models have made it much easier for much smaller models to be finetuned to the
point where they display sophisticated conversational abilities. In this paper,
we present Eluwa, a family of LoRA models that use the Stanford Alpaca dataset
and massively improve the capabilities of Facebook's OPT 1.3B, 2.7B and 6.7B
models. We benchmark these models in multiple ways, including letting GPT-4
judge their answers to prompts that span general knowledge, writing,
programming and other tasks. We show that smaller models here can be fine-tuned
to be as performant as models 3x larger - all for as little as 40 USD in
compute.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 18:06:27 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Wijeratne",
"Yudhanjaya",
""
],
[
"Marikar",
"Ishan",
""
]
] | TITLE: Better Question-Answering Models on a Budget
ABSTRACT: Low-rank adaptation (LoRA) and question-answer datasets from large language
models have made it much easier for much smaller models to be finetuned to the
point where they display sophisticated conversational abilities. In this paper,
we present Eluwa, a family of LoRA models that use the Stanford Alpaca dataset
and massively improve the capabilities of Facebook's OPT 1.3B, 2.7B and 6.7B
models. We benchmark these models in multiple ways, including letting GPT-4
judge their answers to prompts that span general knowledge, writing,
programming and other tasks. We show that smaller models here can be fine-tuned
to be as performant as models 3x larger - all for as little as 40 USD in
compute.
|
2304.12376 | Yaroslav Golubev | Anastasiia Birillo, Elizaveta Artser, Yaroslav Golubev, Maria Tigina,
Hieke Keuning, Nikolay Vyahhi, Timofey Bryksin | Detecting Code Quality Issues in Pre-written Templates of Programming
Tasks in Online Courses | Accepted to ITiCSE'23, 7 pages, 3 figures | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we developed an algorithm for detecting code quality issues in
the templates of online programming tasks, validated it, and conducted an
empirical study on the dataset of student solutions. The algorithm consists of
analyzing recurring unfixed issues in solutions of different students, matching
them with the code of the template, and then filtering the results. Our manual
validation on a subset of tasks demonstrated a precision of 80.8% and a recall
of 73.3%. We used the algorithm on 415 Java tasks from the JetBrains Academy
platform and discovered that as much as 14.7% of tasks have at least one issue
in their template, thus making it harder for students to learn good code
quality practices. We describe our results in detail, provide several
motivating examples and specific cases, and share the feedback of the
developers of the platform, who fixed 51 issues based on the output of our
approach.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 18:13:27 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Birillo",
"Anastasiia",
""
],
[
"Artser",
"Elizaveta",
""
],
[
"Golubev",
"Yaroslav",
""
],
[
"Tigina",
"Maria",
""
],
[
"Keuning",
"Hieke",
""
],
[
"Vyahhi",
"Nikolay",
""
],
[
"Bryksin",
"Timofey",
""
]
] | TITLE: Detecting Code Quality Issues in Pre-written Templates of Programming
Tasks in Online Courses
ABSTRACT: In this work, we developed an algorithm for detecting code quality issues in
the templates of online programming tasks, validated it, and conducted an
empirical study on the dataset of student solutions. The algorithm consists of
analyzing recurring unfixed issues in solutions of different students, matching
them with the code of the template, and then filtering the results. Our manual
validation on a subset of tasks demonstrated a precision of 80.8% and a recall
of 73.3%. We used the algorithm on 415 Java tasks from the JetBrains Academy
platform and discovered that as much as 14.7% of tasks have at least one issue
in their template, thus making it harder for students to learn good code
quality practices. We describe our results in detail, provide several
motivating examples and specific cases, and share the feedback of the
developers of the platform, who fixed 51 issues based on the output of our
approach.
|
2304.12389 | Maria Ercsey-Ravasz | Botond Moln\'ar, Ildik\'o-Be\'ata M\'arton, Szabolcs Horv\'at, M\'aria
Ercsey-Ravasz | Community Detection in Directed Weighted Networks using Voronoi
Partitioning | 22 pages, 8 figures | null | null | null | physics.soc-ph cs.SI physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Community detection is a ubiquitous problem in applied network analysis, yet
efficient techniques do not yet exist for all types of network data. Most
techniques have been developed for undirected graphs, and very few exist that
handle directed and weighted networks effectively. Here we present such an
algorithm based on Voronoi partitionings. As an added benefit, this method can
directly employ edge weights that represent lengths, in contrast to algorithms
that operate with connection strengths, requiring ad-hoc transformations of
length data. We demonstrate the method on inter-areal brain connectivity, air
transportation networks, as well as on randomly generated benchmark networks.
The algorithm can handle dense graphs where weights are the main factor
determining communities. The hierarchical structure of networks can also be
detected, as shown for the brain. Its time efficiency is comparable with other
state-of-the-art algorithms, the most costly part being Dijkstra's shortest
paths algorithm.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 18:34:10 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Molnár",
"Botond",
""
],
[
"Márton",
"Ildikó-Beáta",
""
],
[
"Horvát",
"Szabolcs",
""
],
[
"Ercsey-Ravasz",
"Mária",
""
]
] | TITLE: Community Detection in Directed Weighted Networks using Voronoi
Partitioning
ABSTRACT: Community detection is a ubiquitous problem in applied network analysis, yet
efficient techniques do not yet exist for all types of network data. Most
techniques have been developed for undirected graphs, and very few exist that
handle directed and weighted networks effectively. Here we present such an
algorithm based on Voronoi partitionings. As an added benefit, this method can
directly employ edge weights that represent lengths, in contrast to algorithms
that operate with connection strengths, requiring ad-hoc transformations of
length data. We demonstrate the method on inter-areal brain connectivity, air
transportation networks, as well as on randomly generated benchmark networks.
The algorithm can handle dense graphs where weights are the main factor
determining communities. The hierarchical structure of networks can also be
detected, as shown for the brain. Its time efficiency is comparable with other
state-of-the-art algorithms, the most costly part being Dijkstra's shortest
paths algorithm.
|
2304.12397 | Luiza Pozzobon | Luiza Pozzobon, Beyza Ermis, Patrick Lewis, Sara Hooker | On the Challenges of Using Black-Box APIs for Toxicity Evaluation in
Research | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Perception of toxicity evolves over time and often differs between
geographies and cultural backgrounds. Similarly, black-box commercially
available APIs for detecting toxicity, such as the Perspective API, are not
static, but frequently retrained to address any unattended weaknesses and
biases. We evaluate the implications of these changes on the reproducibility of
findings that compare the relative merits of models and methods that aim to
curb toxicity. Our findings suggest that research that relied on inherited
automatic toxicity scores to compare models and techniques may have resulted in
inaccurate findings. Rescoring all models from HELM, a widely respected living
benchmark, for toxicity with the recent version of the API led to a different
ranking of widely used foundation models. We suggest caution in applying
apples-to-apples comparisons between studies and lay recommendations for a more
structured approach to evaluating toxicity over time. Code and data are
available at https://github.com/for-ai/black-box-api-challenges.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 19:11:51 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Pozzobon",
"Luiza",
""
],
[
"Ermis",
"Beyza",
""
],
[
"Lewis",
"Patrick",
""
],
[
"Hooker",
"Sara",
""
]
] | TITLE: On the Challenges of Using Black-Box APIs for Toxicity Evaluation in
Research
ABSTRACT: Perception of toxicity evolves over time and often differs between
geographies and cultural backgrounds. Similarly, black-box commercially
available APIs for detecting toxicity, such as the Perspective API, are not
static, but frequently retrained to address any unattended weaknesses and
biases. We evaluate the implications of these changes on the reproducibility of
findings that compare the relative merits of models and methods that aim to
curb toxicity. Our findings suggest that research that relied on inherited
automatic toxicity scores to compare models and techniques may have resulted in
inaccurate findings. Rescoring all models from HELM, a widely respected living
benchmark, for toxicity with the recent version of the API led to a different
ranking of widely used foundation models. We suggest caution in applying
apples-to-apples comparisons between studies and lay recommendations for a more
structured approach to evaluating toxicity over time. Code and data are
available at https://github.com/for-ai/black-box-api-challenges.
|
2304.12398 | Pere Verges Boncompte | Pere Verg\'es, Mike Heddes, Igor Nunes, Tony Givargis, Alexandru
Nicolau | HDCC: A Hyperdimensional Computing compiler for classification on
embedded systems and high-performance computing | 8 pages, 3 figures | null | null | null | cs.LG cs.DC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Hyperdimensional Computing (HDC) is a bio-inspired computing framework that
has gained increasing attention, especially as a more efficient approach to
machine learning (ML). This work introduces the \name{} compiler, the first
open-source compiler that translates high-level descriptions of HDC
classification methods into optimized C code. The code generated by the
proposed compiler has three main features for embedded systems and
High-Performance Computing: (1) it is self-contained and has no library or
platform dependencies; (2) it supports multithreading and single instruction
multiple data (SIMD) instructions using C intrinsics; (3) it is optimized for
maximum performance and minimal memory usage. \name{} is designed like a modern
compiler, featuring an intuitive and descriptive input language, an
intermediate representation (IR), and a retargetable backend. This makes
\name{} a valuable tool for research and applications exploring HDC for
classification tasks on embedded systems and High-Performance Computing. To
substantiate these claims, we conducted experiments with HDCC on several of the
most popular datasets in the HDC literature. The experiments were run on four
different machines, including different hyperparameter configurations, and the
results were compared to a popular prototyping library built on PyTorch. The
results show a training and inference speedup of up to 132x, averaging 25x
across all datasets and machines. Regarding memory usage, using
10240-dimensional hypervectors, the average reduction was 5x, reaching up to
14x. When considering vectors of 64 dimensions, the average reduction was 85x,
with a maximum of 158x less memory utilization.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 19:16:03 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Vergés",
"Pere",
""
],
[
"Heddes",
"Mike",
""
],
[
"Nunes",
"Igor",
""
],
[
"Givargis",
"Tony",
""
],
[
"Nicolau",
"Alexandru",
""
]
] | TITLE: HDCC: A Hyperdimensional Computing compiler for classification on
embedded systems and high-performance computing
ABSTRACT: Hyperdimensional Computing (HDC) is a bio-inspired computing framework that
has gained increasing attention, especially as a more efficient approach to
machine learning (ML). This work introduces the \name{} compiler, the first
open-source compiler that translates high-level descriptions of HDC
classification methods into optimized C code. The code generated by the
proposed compiler has three main features for embedded systems and
High-Performance Computing: (1) it is self-contained and has no library or
platform dependencies; (2) it supports multithreading and single instruction
multiple data (SIMD) instructions using C intrinsics; (3) it is optimized for
maximum performance and minimal memory usage. \name{} is designed like a modern
compiler, featuring an intuitive and descriptive input language, an
intermediate representation (IR), and a retargetable backend. This makes
\name{} a valuable tool for research and applications exploring HDC for
classification tasks on embedded systems and High-Performance Computing. To
substantiate these claims, we conducted experiments with HDCC on several of the
most popular datasets in the HDC literature. The experiments were run on four
different machines, including different hyperparameter configurations, and the
results were compared to a popular prototyping library built on PyTorch. The
results show a training and inference speedup of up to 132x, averaging 25x
across all datasets and machines. Regarding memory usage, using
10240-dimensional hypervectors, the average reduction was 5x, reaching up to
14x. When considering vectors of 64 dimensions, the average reduction was 85x,
with a maximum of 158x less memory utilization.
|
2304.12431 | Maximilien Le Clei | Maximilien Le Clei, Pierre Bellec | Neuroevolution of Recurrent Architectures on Control Tasks | null | null | null | null | cs.NE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern artificial intelligence works typically train the parameters of
fixed-sized deep neural networks using gradient-based optimization techniques.
Simple evolutionary algorithms have recently been shown to also be capable of
optimizing deep neural network parameters, at times matching the performance of
gradient-based techniques, e.g. in reinforcement learning settings. In addition
to optimizing network parameters, many evolutionary computation techniques are
also capable of progressively constructing network architectures. However,
constructing network architectures from elementary evolution rules has not yet
been shown to scale to modern reinforcement learning benchmarks. In this paper
we therefore propose a new approach in which the architectures of recurrent
neural networks dynamically evolve according to a small set of mutation rules.
We implement a massively parallel evolutionary algorithm and run experiments on
all 19 OpenAI Gym state-based reinforcement learning control tasks. We find
that in most cases, dynamic agents match or exceed the performance of
gradient-based agents while utilizing orders of magnitude fewer parameters. We
believe our work to open avenues for real-life applications where network
compactness and autonomous design are of critical importance. We provide our
source code, final model checkpoints and full results at
github.com/MaximilienLC/nra.
| [
{
"version": "v1",
"created": "Mon, 3 Apr 2023 16:29:18 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Clei",
"Maximilien Le",
""
],
[
"Bellec",
"Pierre",
""
]
] | TITLE: Neuroevolution of Recurrent Architectures on Control Tasks
ABSTRACT: Modern artificial intelligence works typically train the parameters of
fixed-sized deep neural networks using gradient-based optimization techniques.
Simple evolutionary algorithms have recently been shown to also be capable of
optimizing deep neural network parameters, at times matching the performance of
gradient-based techniques, e.g. in reinforcement learning settings. In addition
to optimizing network parameters, many evolutionary computation techniques are
also capable of progressively constructing network architectures. However,
constructing network architectures from elementary evolution rules has not yet
been shown to scale to modern reinforcement learning benchmarks. In this paper
we therefore propose a new approach in which the architectures of recurrent
neural networks dynamically evolve according to a small set of mutation rules.
We implement a massively parallel evolutionary algorithm and run experiments on
all 19 OpenAI Gym state-based reinforcement learning control tasks. We find
that in most cases, dynamic agents match or exceed the performance of
gradient-based agents while utilizing orders of magnitude fewer parameters. We
believe our work to open avenues for real-life applications where network
compactness and autonomous design are of critical importance. We provide our
source code, final model checkpoints and full results at
github.com/MaximilienLC/nra.
|
2304.12432 | Maximilien Le Clei | Maximilien Le Clei, Pierre Bellec | Generative Adversarial Neuroevolution for Control Behaviour Imitation | null | null | null | null | cs.NE cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There is a recent surge in interest for imitation learning, with large human
video-game and robotic manipulation datasets being used to train agents on very
complex tasks. While deep neuroevolution has recently been shown to match the
performance of gradient-based techniques on various reinforcement learning
problems, the application of deep neuroevolution techniques to imitation
learning remains relatively unexplored. In this work, we propose to explore
whether deep neuroevolution can be used for behaviour imitation on popular
simulation environments. We introduce a simple co-evolutionary adversarial
generation framework, and evaluate its capabilities by evolving standard deep
recurrent networks to imitate state-of-the-art pre-trained agents on 8 OpenAI
Gym state-based control tasks. Across all tasks, we find the final elite actor
agents capable of achieving scores as high as those obtained by the pre-trained
agents, all the while closely following their score trajectories. Our results
suggest that neuroevolution could be a valuable addition to deep learning
techniques to produce accurate emulation of behavioural agents. We believe that
the generality and simplicity of our approach opens avenues for imitating
increasingly complex behaviours in increasingly complex settings, e.g. human
behaviour in real-world settings. We provide our source code, model checkpoints
and results at github.com/MaximilienLC/gane.
| [
{
"version": "v1",
"created": "Mon, 3 Apr 2023 16:33:22 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Clei",
"Maximilien Le",
""
],
[
"Bellec",
"Pierre",
""
]
] | TITLE: Generative Adversarial Neuroevolution for Control Behaviour Imitation
ABSTRACT: There is a recent surge in interest for imitation learning, with large human
video-game and robotic manipulation datasets being used to train agents on very
complex tasks. While deep neuroevolution has recently been shown to match the
performance of gradient-based techniques on various reinforcement learning
problems, the application of deep neuroevolution techniques to imitation
learning remains relatively unexplored. In this work, we propose to explore
whether deep neuroevolution can be used for behaviour imitation on popular
simulation environments. We introduce a simple co-evolutionary adversarial
generation framework, and evaluate its capabilities by evolving standard deep
recurrent networks to imitate state-of-the-art pre-trained agents on 8 OpenAI
Gym state-based control tasks. Across all tasks, we find the final elite actor
agents capable of achieving scores as high as those obtained by the pre-trained
agents, all the while closely following their score trajectories. Our results
suggest that neuroevolution could be a valuable addition to deep learning
techniques to produce accurate emulation of behavioural agents. We believe that
the generality and simplicity of our approach opens avenues for imitating
increasingly complex behaviours in increasingly complex settings, e.g. human
behaviour in real-world settings. We provide our source code, model checkpoints
and results at github.com/MaximilienLC/gane.
|
2304.12437 | Thomas Simpson | Thomas Simpson, Konstantinos Vlachas, Anthony Garland, Nikolaos
Dervilis, Eleni Chatzi | VpROM: A novel Variational AutoEncoder-boosted Reduced Order Model for
the treatment of parametric dependencies in nonlinear systems | null | null | null | null | math.NA cs.CE cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Reduced Order Models (ROMs) are of considerable importance in many areas of
engineering in which computational time presents difficulties. Established
approaches employ projection-based reduction such as Proper Orthogonal
Decomposition, however, such methods can become inefficient or fail in the case
of parameteric or strongly nonlinear models. Such limitations are usually
tackled via a library of local reduction bases each of which being valid for a
given parameter vector. The success of such methods, however, is strongly
reliant upon the method used to relate the parameter vectors to the local
bases, this is typically achieved using clustering or interpolation methods. We
propose the replacement of these methods with a Variational Autoencoder (VAE)
to be used as a generative model which can infer the local basis corresponding
to a given parameter vector in a probabilistic manner. The resulting
VAE-boosted parametric ROM \emph{VpROM} still retains the physical insights of
a projection-based method but also allows for better treatment of problems
where model dependencies or excitation traits cause the dynamic behavior to
span multiple response regimes. Moreover, the probabilistic treatment of the
VAE representation allows for uncertainty quantification on the reduction bases
which may then be propagated to the ROM response. The performance of the
proposed approach is validated on an open-source simulation benchmark featuring
hysteresis and multi-parametric dependencies, and on a large-scale wind turbine
tower characterised by nonlinear material behavior and model uncertainty.
| [
{
"version": "v1",
"created": "Tue, 11 Apr 2023 19:24:35 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Simpson",
"Thomas",
""
],
[
"Vlachas",
"Konstantinos",
""
],
[
"Garland",
"Anthony",
""
],
[
"Dervilis",
"Nikolaos",
""
],
[
"Chatzi",
"Eleni",
""
]
] | TITLE: VpROM: A novel Variational AutoEncoder-boosted Reduced Order Model for
the treatment of parametric dependencies in nonlinear systems
ABSTRACT: Reduced Order Models (ROMs) are of considerable importance in many areas of
engineering in which computational time presents difficulties. Established
approaches employ projection-based reduction such as Proper Orthogonal
Decomposition, however, such methods can become inefficient or fail in the case
of parameteric or strongly nonlinear models. Such limitations are usually
tackled via a library of local reduction bases each of which being valid for a
given parameter vector. The success of such methods, however, is strongly
reliant upon the method used to relate the parameter vectors to the local
bases, this is typically achieved using clustering or interpolation methods. We
propose the replacement of these methods with a Variational Autoencoder (VAE)
to be used as a generative model which can infer the local basis corresponding
to a given parameter vector in a probabilistic manner. The resulting
VAE-boosted parametric ROM \emph{VpROM} still retains the physical insights of
a projection-based method but also allows for better treatment of problems
where model dependencies or excitation traits cause the dynamic behavior to
span multiple response regimes. Moreover, the probabilistic treatment of the
VAE representation allows for uncertainty quantification on the reduction bases
which may then be propagated to the ROM response. The performance of the
proposed approach is validated on an open-source simulation benchmark featuring
hysteresis and multi-parametric dependencies, and on a large-scale wind turbine
tower characterised by nonlinear material behavior and model uncertainty.
|
2304.12443 | Nan-Jiang Jiang | Nan-Jiang Jiang, Chenhao Tan, Marie-Catherine de Marneffe | Understanding and Predicting Human Label Variation in Natural Language
Inference through Explanation | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | Human label variation (Plank 2022), or annotation disagreement, exists in
many natural language processing (NLP) tasks. To be robust and trusted, NLP
models need to identify such variation and be able to explain it. To this end,
we created the first ecologically valid explanation dataset with diverse
reasoning, LiveNLI. LiveNLI contains annotators' highlights and free-text
explanations for the label(s) of their choice for 122 English Natural Language
Inference items, each with at least 10 annotations. We used its explanations
for chain-of-thought prompting, and found there is still room for improvement
in GPT-3's ability to predict label distribution with in-context learning.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 20:45:09 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Jiang",
"Nan-Jiang",
""
],
[
"Tan",
"Chenhao",
""
],
[
"de Marneffe",
"Marie-Catherine",
""
]
] | TITLE: Understanding and Predicting Human Label Variation in Natural Language
Inference through Explanation
ABSTRACT: Human label variation (Plank 2022), or annotation disagreement, exists in
many natural language processing (NLP) tasks. To be robust and trusted, NLP
models need to identify such variation and be able to explain it. To this end,
we created the first ecologically valid explanation dataset with diverse
reasoning, LiveNLI. LiveNLI contains annotators' highlights and free-text
explanations for the label(s) of their choice for 122 English Natural Language
Inference items, each with at least 10 annotations. We used its explanations
for chain-of-thought prompting, and found there is still room for improvement
in GPT-3's ability to predict label distribution with in-context learning.
|
2304.12447 | Praveen Kumar Pandian Shanmuganathan | Eashan Kosaraju, Praveen Kumar Pandian Shanmuganathan | Predicting Pulmonary Hypertension by Electrocardiograms Using Machine
Learning | 19 pages, 5 figures | null | null | null | eess.SP cs.LG | http://creativecommons.org/licenses/by/4.0/ | Pulmonary hypertension (PH) is a condition of high blood pressure that
affects the arteries in the lungs and the right side of the heart (Mayo Clinic,
2017). A mean pulmonary artery pressure greater than 25 mmHg is defined as
Pulmonary hypertension. The estimated 5-year survival rate from the time of
diagnosis of pulmonary hypertension is only 57% without therapy and patients
with right heart failure only survive for approximately 1 year without
treatment (Benza et al., 2012). Given the indolent nature of the disease, early
detection of PH remains a challenge leading to delays in therapy.
Echocardiography is currently used as a screening tool for diagnosing PH.
However, electrocardiography (ECG), a more accessible, simple to use, and
cost-effective tool compared to echocardiography, is less studied and explored
for screening at-risk patients for PH. The goal of this project is to create a
neural network model which can process an ECG signal and detect the presence of
PH with a confidence probability. I created a dense neural network (DNN) model
that has an accuracy of 98% over the available training sample. For future
steps, the current model will be updated with a model suited for time-series
data. To balance the dataset with proper training samples, I will generate
additional data using data augmentation techniques. Through early and accurate
detection of conditions such as PH, we widen the spectrum of innovation in
detecting chronic life-threatening health conditions and reduce associated
mortality and morbidity.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 21:00:16 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Kosaraju",
"Eashan",
""
],
[
"Shanmuganathan",
"Praveen Kumar Pandian",
""
]
] | TITLE: Predicting Pulmonary Hypertension by Electrocardiograms Using Machine
Learning
ABSTRACT: Pulmonary hypertension (PH) is a condition of high blood pressure that
affects the arteries in the lungs and the right side of the heart (Mayo Clinic,
2017). A mean pulmonary artery pressure greater than 25 mmHg is defined as
Pulmonary hypertension. The estimated 5-year survival rate from the time of
diagnosis of pulmonary hypertension is only 57% without therapy and patients
with right heart failure only survive for approximately 1 year without
treatment (Benza et al., 2012). Given the indolent nature of the disease, early
detection of PH remains a challenge leading to delays in therapy.
Echocardiography is currently used as a screening tool for diagnosing PH.
However, electrocardiography (ECG), a more accessible, simple to use, and
cost-effective tool compared to echocardiography, is less studied and explored
for screening at-risk patients for PH. The goal of this project is to create a
neural network model which can process an ECG signal and detect the presence of
PH with a confidence probability. I created a dense neural network (DNN) model
that has an accuracy of 98% over the available training sample. For future
steps, the current model will be updated with a model suited for time-series
data. To balance the dataset with proper training samples, I will generate
additional data using data augmentation techniques. Through early and accurate
detection of conditions such as PH, we widen the spectrum of innovation in
detecting chronic life-threatening health conditions and reduce associated
mortality and morbidity.
|
2304.12455 | Heng Yu | Heng Yu, Zoltan A. Milacski, Laszlo A. Jeni | Unsupervised Style-based Explicit 3D Face Reconstruction from Single
Image | CVPR workshop | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Inferring 3D object structures from a single image is an ill-posed task due
to depth ambiguity and occlusion. Typical resolutions in the literature include
leveraging 2D or 3D ground truth for supervised learning, as well as imposing
hand-crafted symmetry priors or using an implicit representation to hallucinate
novel viewpoints for unsupervised methods. In this work, we propose a general
adversarial learning framework for solving Unsupervised 2D to Explicit 3D Style
Transfer (UE3DST). Specifically, we merge two architectures: the unsupervised
explicit 3D reconstruction network of Wu et al.\ and the Generative Adversarial
Network (GAN) named StarGAN-v2. We experiment across three facial datasets
(Basel Face Model, 3DFAW and CelebA-HQ) and show that our solution is able to
outperform well established solutions such as DepthNet in 3D reconstruction and
Pix2NeRF in conditional style transfer, while we also justify the individual
contributions of our model components via ablation. In contrast to the
aforementioned baselines, our scheme produces features for explicit 3D
rendering, which can be manipulated and utilized in downstream tasks.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 21:25:06 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Yu",
"Heng",
""
],
[
"Milacski",
"Zoltan A.",
""
],
[
"Jeni",
"Laszlo A.",
""
]
] | TITLE: Unsupervised Style-based Explicit 3D Face Reconstruction from Single
Image
ABSTRACT: Inferring 3D object structures from a single image is an ill-posed task due
to depth ambiguity and occlusion. Typical resolutions in the literature include
leveraging 2D or 3D ground truth for supervised learning, as well as imposing
hand-crafted symmetry priors or using an implicit representation to hallucinate
novel viewpoints for unsupervised methods. In this work, we propose a general
adversarial learning framework for solving Unsupervised 2D to Explicit 3D Style
Transfer (UE3DST). Specifically, we merge two architectures: the unsupervised
explicit 3D reconstruction network of Wu et al.\ and the Generative Adversarial
Network (GAN) named StarGAN-v2. We experiment across three facial datasets
(Basel Face Model, 3DFAW and CelebA-HQ) and show that our solution is able to
outperform well established solutions such as DepthNet in 3D reconstruction and
Pix2NeRF in conditional style transfer, while we also justify the individual
contributions of our model components via ablation. In contrast to the
aforementioned baselines, our scheme produces features for explicit 3D
rendering, which can be manipulated and utilized in downstream tasks.
|
2304.12492 | Lucas Pascotti Valem | Lucas Pascotti Valem, Daniel Carlos Guimar\~aes Pedronette, Longin Jan
Latecki | Graph Convolutional Networks based on Manifold Learning for
Semi-Supervised Image Classification | null | null | 10.1016/j.cviu.2022.103618 | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Due to a huge volume of information in many domains, the need for
classification methods is imperious. In spite of many advances, most of the
approaches require a large amount of labeled data, which is often not
available, due to costs and difficulties of manual labeling processes. In this
scenario, unsupervised and semi-supervised approaches have been gaining
increasing attention. The GCNs (Graph Convolutional Neural Networks) represent
a promising solution since they encode the neighborhood information and have
achieved state-of-the-art results on scenarios with limited labeled data.
However, since GCNs require graph-structured data, their use for
semi-supervised image classification is still scarce in the literature. In this
work, we propose a novel approach, the Manifold-GCN, based on GCNs for
semi-supervised image classification. The main hypothesis of this paper is that
the use of manifold learning to model the graph structure can further improve
the GCN classification. To the best of our knowledge, this is the first
framework that allows the combination of GCNs with different types of manifold
learning approaches for image classification. All manifold learning algorithms
employed are completely unsupervised, which is especially useful for scenarios
where the availability of labeled data is a concern. A broad experimental
evaluation was conducted considering 5 GCN models, 3 manifold learning
approaches, 3 image datasets, and 5 deep features. The results reveal that our
approach presents better accuracy than traditional and recent state-of-the-art
methods with very efficient run times for both training and testing.
| [
{
"version": "v1",
"created": "Mon, 24 Apr 2023 23:24:46 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Valem",
"Lucas Pascotti",
""
],
[
"Pedronette",
"Daniel Carlos Guimarães",
""
],
[
"Latecki",
"Longin Jan",
""
]
] | TITLE: Graph Convolutional Networks based on Manifold Learning for
Semi-Supervised Image Classification
ABSTRACT: Due to a huge volume of information in many domains, the need for
classification methods is imperious. In spite of many advances, most of the
approaches require a large amount of labeled data, which is often not
available, due to costs and difficulties of manual labeling processes. In this
scenario, unsupervised and semi-supervised approaches have been gaining
increasing attention. The GCNs (Graph Convolutional Neural Networks) represent
a promising solution since they encode the neighborhood information and have
achieved state-of-the-art results on scenarios with limited labeled data.
However, since GCNs require graph-structured data, their use for
semi-supervised image classification is still scarce in the literature. In this
work, we propose a novel approach, the Manifold-GCN, based on GCNs for
semi-supervised image classification. The main hypothesis of this paper is that
the use of manifold learning to model the graph structure can further improve
the GCN classification. To the best of our knowledge, this is the first
framework that allows the combination of GCNs with different types of manifold
learning approaches for image classification. All manifold learning algorithms
employed are completely unsupervised, which is especially useful for scenarios
where the availability of labeled data is a concern. A broad experimental
evaluation was conducted considering 5 GCN models, 3 manifold learning
approaches, 3 image datasets, and 5 deep features. The results reveal that our
approach presents better accuracy than traditional and recent state-of-the-art
methods with very efficient run times for both training and testing.
|
2304.12506 | Haoran Xie | Jiahao Weng and Xusheng Du and Haoran Xie | DualSlide: Global-to-Local Sketching Interface for Slide Content and
Layout Design | 8 pages, 16 figures. accepted in NICOGRAPH International 2023 | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | Online learning and academic conferences have become pervasive and essential
for education and professional development, especially since the onset of
pandemics. Academic presentations usually require well-designed slides that are
easily understood. Sketches that visually represent design intentions and are
readily accessible to the average users. To assist non-expert users in creating
visually appealing academic slides, we propose DualSlide, a global and local
two-stage sketching interface system that provides image retrieval and user
guidance. At the global stage, DualSlide provides a heat map canvas to display
the distribution of all slide layouts in a dataset, allowing users to explore
the reference slides efficiently. At the local stage of the system, detailed
references and guidance for designing slide content, such as diagrams and
fonts, can be provided. We further propose a sketch-matching algorithm to
compare the user's input sketch and similar diagrams. All user guidance can be
adapted in real-time editing, and users can design slides with a high degree of
freedom. We conducted a user study to verify the effectiveness and usability of
the proposed DualSlide system confirming that DualSlide provides high retrieval
accuracy and satisfactory design results with a good user experience. Video:
https://youtu.be/lUI1zjxCdM0
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 01:03:43 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Weng",
"Jiahao",
""
],
[
"Du",
"Xusheng",
""
],
[
"Xie",
"Haoran",
""
]
] | TITLE: DualSlide: Global-to-Local Sketching Interface for Slide Content and
Layout Design
ABSTRACT: Online learning and academic conferences have become pervasive and essential
for education and professional development, especially since the onset of
pandemics. Academic presentations usually require well-designed slides that are
easily understood. Sketches that visually represent design intentions and are
readily accessible to the average users. To assist non-expert users in creating
visually appealing academic slides, we propose DualSlide, a global and local
two-stage sketching interface system that provides image retrieval and user
guidance. At the global stage, DualSlide provides a heat map canvas to display
the distribution of all slide layouts in a dataset, allowing users to explore
the reference slides efficiently. At the local stage of the system, detailed
references and guidance for designing slide content, such as diagrams and
fonts, can be provided. We further propose a sketch-matching algorithm to
compare the user's input sketch and similar diagrams. All user guidance can be
adapted in real-time editing, and users can design slides with a high degree of
freedom. We conducted a user study to verify the effectiveness and usability of
the proposed DualSlide system confirming that DualSlide provides high retrieval
accuracy and satisfactory design results with a good user experience. Video:
https://youtu.be/lUI1zjxCdM0
|
2304.12508 | Pengyuan Lu | Mengyu Liu and Pengyuan Lu and Xin Chen and Fanxin Kong and Oleg
Sokolsky and Insup Lee | Fulfilling Formal Specifications ASAP by Model-free Reinforcement
Learning | null | null | null | null | cs.LG cs.AI cs.FL cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | We propose a model-free reinforcement learning solution, namely the ASAP-Phi
framework, to encourage an agent to fulfill a formal specification ASAP. The
framework leverages a piece-wise reward function that assigns quantitative
semantic reward to traces not satisfying the specification, and a high constant
reward to the remaining. Then, it trains an agent with an actor-critic-based
algorithm, such as soft actor-critic (SAC), or deep deterministic policy
gradient (DDPG). Moreover, we prove that ASAP-Phi produces policies that
prioritize fulfilling a specification ASAP. Extensive experiments are run,
including ablation studies, on state-of-the-art benchmarks. Results show that
our framework succeeds in finding sufficiently fast trajectories for up to 97\%
test cases and defeats baselines.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 01:16:42 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Liu",
"Mengyu",
""
],
[
"Lu",
"Pengyuan",
""
],
[
"Chen",
"Xin",
""
],
[
"Kong",
"Fanxin",
""
],
[
"Sokolsky",
"Oleg",
""
],
[
"Lee",
"Insup",
""
]
] | TITLE: Fulfilling Formal Specifications ASAP by Model-free Reinforcement
Learning
ABSTRACT: We propose a model-free reinforcement learning solution, namely the ASAP-Phi
framework, to encourage an agent to fulfill a formal specification ASAP. The
framework leverages a piece-wise reward function that assigns quantitative
semantic reward to traces not satisfying the specification, and a high constant
reward to the remaining. Then, it trains an agent with an actor-critic-based
algorithm, such as soft actor-critic (SAC), or deep deterministic policy
gradient (DDPG). Moreover, we prove that ASAP-Phi produces policies that
prioritize fulfilling a specification ASAP. Extensive experiments are run,
including ablation studies, on state-of-the-art benchmarks. Results show that
our framework succeeds in finding sufficiently fast trajectories for up to 97\%
test cases and defeats baselines.
|
2304.12518 | Vimal Mollyn | Vimal Mollyn, Riku Arakawa, Mayank Goel, Chris Harrison, Karan Ahuja | IMUPoser: Full-Body Pose Estimation using IMUs in Phones, Watches, and
Earbuds | null | null | 10.1145/3544548.3581392 | null | cs.HC cs.CV | http://creativecommons.org/licenses/by/4.0/ | Tracking body pose on-the-go could have powerful uses in fitness, mobile
gaming, context-aware virtual assistants, and rehabilitation. However, users
are unlikely to buy and wear special suits or sensor arrays to achieve this
end. Instead, in this work, we explore the feasibility of estimating body pose
using IMUs already in devices that many users own -- namely smartphones,
smartwatches, and earbuds. This approach has several challenges, including
noisy data from low-cost commodity IMUs, and the fact that the number of
instrumentation points on a users body is both sparse and in flux. Our pipeline
receives whatever subset of IMU data is available, potentially from just a
single device, and produces a best-guess pose. To evaluate our model, we
created the IMUPoser Dataset, collected from 10 participants wearing or holding
off-the-shelf consumer devices and across a variety of activity contexts. We
provide a comprehensive evaluation of our system, benchmarking it on both our
own and existing IMU datasets.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 02:13:24 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Mollyn",
"Vimal",
""
],
[
"Arakawa",
"Riku",
""
],
[
"Goel",
"Mayank",
""
],
[
"Harrison",
"Chris",
""
],
[
"Ahuja",
"Karan",
""
]
] | TITLE: IMUPoser: Full-Body Pose Estimation using IMUs in Phones, Watches, and
Earbuds
ABSTRACT: Tracking body pose on-the-go could have powerful uses in fitness, mobile
gaming, context-aware virtual assistants, and rehabilitation. However, users
are unlikely to buy and wear special suits or sensor arrays to achieve this
end. Instead, in this work, we explore the feasibility of estimating body pose
using IMUs already in devices that many users own -- namely smartphones,
smartwatches, and earbuds. This approach has several challenges, including
noisy data from low-cost commodity IMUs, and the fact that the number of
instrumentation points on a users body is both sparse and in flux. Our pipeline
receives whatever subset of IMU data is available, potentially from just a
single device, and produces a best-guess pose. To evaluate our model, we
created the IMUPoser Dataset, collected from 10 participants wearing or holding
off-the-shelf consumer devices and across a variety of activity contexts. We
provide a comprehensive evaluation of our system, benchmarking it on both our
own and existing IMU datasets.
|
2304.12523 | Payam Dibaeinia | Payam Dibaeinia, Saurabh Sinha | CIMLA: Interpretable AI for inference of differential causal networks | null | null | null | null | cs.LG q-bio.QM stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The discovery of causal relationships from high-dimensional data is a major
open problem in bioinformatics. Machine learning and feature attribution models
have shown great promise in this context but lack causal interpretation. Here,
we show that a popular feature attribution model estimates a causal quantity
reflecting the influence of one variable on another, under certain assumptions.
We leverage this insight to implement a new tool, CIMLA, for discovering
condition-dependent changes in causal relationships. We then use CIMLA to
identify differences in gene regulatory networks between biological conditions,
a problem that has received great attention in recent years. Using extensive
benchmarking on simulated data sets, we show that CIMLA is more robust to
confounding variables and is more accurate than leading methods. Finally, we
employ CIMLA to analyze a previously published single-cell RNA-seq data set
collected from subjects with and without Alzheimer's disease (AD), discovering
several potential regulators of AD.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 02:31:12 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Dibaeinia",
"Payam",
""
],
[
"Sinha",
"Saurabh",
""
]
] | TITLE: CIMLA: Interpretable AI for inference of differential causal networks
ABSTRACT: The discovery of causal relationships from high-dimensional data is a major
open problem in bioinformatics. Machine learning and feature attribution models
have shown great promise in this context but lack causal interpretation. Here,
we show that a popular feature attribution model estimates a causal quantity
reflecting the influence of one variable on another, under certain assumptions.
We leverage this insight to implement a new tool, CIMLA, for discovering
condition-dependent changes in causal relationships. We then use CIMLA to
identify differences in gene regulatory networks between biological conditions,
a problem that has received great attention in recent years. Using extensive
benchmarking on simulated data sets, we show that CIMLA is more robust to
confounding variables and is more accurate than leading methods. Finally, we
employ CIMLA to analyze a previously published single-cell RNA-seq data set
collected from subjects with and without Alzheimer's disease (AD), discovering
several potential regulators of AD.
|
2304.12561 | Yakun Yu | Yakun Yu, Jiuding Yang, Weidong Guo, Hui Liu, Yu Xu, and Di Niu | TCR: Short Video Title Generation and Cover Selection with Attention
Refinement | Accepted by PAKDD23 | null | null | null | cs.CV cs.MM | http://creativecommons.org/licenses/by/4.0/ | With the widespread popularity of user-generated short videos, it becomes
increasingly challenging for content creators to promote their content to
potential viewers. Automatically generating appealing titles and covers for
short videos can help grab viewers' attention. Existing studies on video
captioning mostly focus on generating factual descriptions of actions, which do
not conform to video titles intended for catching viewer attention.
Furthermore, research for cover selection based on multimodal information is
sparse. These problems motivate the need for tailored methods to specifically
support the joint task of short video title generation and cover selection
(TG-CS) as well as the demand for creating corresponding datasets to support
the studies. In this paper, we first collect and present a real-world dataset
named Short Video Title Generation (SVTG) that contains videos with appealing
titles and covers. We then propose a Title generation and Cover selection with
attention Refinement (TCR) method for TG-CS. The refinement procedure
progressively selects high-quality samples and highly relevant frames and text
tokens within each sample to refine model training. Extensive experiments show
that our TCR method is superior to various existing video captioning methods in
generating titles and is able to select better covers for noisy real-world
short videos.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 04:08:19 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Yu",
"Yakun",
""
],
[
"Yang",
"Jiuding",
""
],
[
"Guo",
"Weidong",
""
],
[
"Liu",
"Hui",
""
],
[
"Xu",
"Yu",
""
],
[
"Niu",
"Di",
""
]
] | TITLE: TCR: Short Video Title Generation and Cover Selection with Attention
Refinement
ABSTRACT: With the widespread popularity of user-generated short videos, it becomes
increasingly challenging for content creators to promote their content to
potential viewers. Automatically generating appealing titles and covers for
short videos can help grab viewers' attention. Existing studies on video
captioning mostly focus on generating factual descriptions of actions, which do
not conform to video titles intended for catching viewer attention.
Furthermore, research for cover selection based on multimodal information is
sparse. These problems motivate the need for tailored methods to specifically
support the joint task of short video title generation and cover selection
(TG-CS) as well as the demand for creating corresponding datasets to support
the studies. In this paper, we first collect and present a real-world dataset
named Short Video Title Generation (SVTG) that contains videos with appealing
titles and covers. We then propose a Title generation and Cover selection with
attention Refinement (TCR) method for TG-CS. The refinement procedure
progressively selects high-quality samples and highly relevant frames and text
tokens within each sample to refine model training. Extensive experiments show
that our TCR method is superior to various existing video captioning methods in
generating titles and is able to select better covers for noisy real-world
short videos.
|
2304.12570 | Leigang Qu | Leigang Qu, Meng Liu, Wenjie Wang, Zhedong Zheng, Liqiang Nie,
Tat-Seng Chua | Learnable Pillar-based Re-ranking for Image-Text Retrieval | Accepted by SIGIR'2023 | Proceedings of the 46th International ACM SIGIR Conference on
Research and Development in Information Retrieval (SIGIR 2023) | 10.1145/3539618.3591712 | null | cs.CV cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image-text retrieval aims to bridge the modality gap and retrieve cross-modal
content based on semantic similarities. Prior work usually focuses on the
pairwise relations (i.e., whether a data sample matches another) but ignores
the higher-order neighbor relations (i.e., a matching structure among multiple
data samples). Re-ranking, a popular post-processing practice, has revealed the
superiority of capturing neighbor relations in single-modality retrieval tasks.
However, it is ineffective to directly extend existing re-ranking algorithms to
image-text retrieval. In this paper, we analyze the reason from four
perspectives, i.e., generalization, flexibility, sparsity, and asymmetry, and
propose a novel learnable pillar-based re-ranking paradigm. Concretely, we
first select top-ranked intra- and inter-modal neighbors as pillars, and then
reconstruct data samples with the neighbor relations between them and the
pillars. In this way, each sample can be mapped into a multimodal pillar space
only using similarities, ensuring generalization. After that, we design a
neighbor-aware graph reasoning module to flexibly exploit the relations and
excavate the sparse positive items within a neighborhood. We also present a
structure alignment constraint to promote cross-modal collaboration and align
the asymmetric modalities. On top of various base backbones, we carry out
extensive experiments on two benchmark datasets, i.e., Flickr30K and MS-COCO,
demonstrating the effectiveness, superiority, generalization, and
transferability of our proposed re-ranking paradigm.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 04:33:27 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Qu",
"Leigang",
""
],
[
"Liu",
"Meng",
""
],
[
"Wang",
"Wenjie",
""
],
[
"Zheng",
"Zhedong",
""
],
[
"Nie",
"Liqiang",
""
],
[
"Chua",
"Tat-Seng",
""
]
] | TITLE: Learnable Pillar-based Re-ranking for Image-Text Retrieval
ABSTRACT: Image-text retrieval aims to bridge the modality gap and retrieve cross-modal
content based on semantic similarities. Prior work usually focuses on the
pairwise relations (i.e., whether a data sample matches another) but ignores
the higher-order neighbor relations (i.e., a matching structure among multiple
data samples). Re-ranking, a popular post-processing practice, has revealed the
superiority of capturing neighbor relations in single-modality retrieval tasks.
However, it is ineffective to directly extend existing re-ranking algorithms to
image-text retrieval. In this paper, we analyze the reason from four
perspectives, i.e., generalization, flexibility, sparsity, and asymmetry, and
propose a novel learnable pillar-based re-ranking paradigm. Concretely, we
first select top-ranked intra- and inter-modal neighbors as pillars, and then
reconstruct data samples with the neighbor relations between them and the
pillars. In this way, each sample can be mapped into a multimodal pillar space
only using similarities, ensuring generalization. After that, we design a
neighbor-aware graph reasoning module to flexibly exploit the relations and
excavate the sparse positive items within a neighborhood. We also present a
structure alignment constraint to promote cross-modal collaboration and align
the asymmetric modalities. On top of various base backbones, we carry out
extensive experiments on two benchmark datasets, i.e., Flickr30K and MS-COCO,
demonstrating the effectiveness, superiority, generalization, and
transferability of our proposed re-ranking paradigm.
|
2304.12573 | Saravanan Thirumuruganathan | Simone Lazier, Saravanan Thirumuruganathan, Hadis Anahideh | Fairness and Bias in Truth Discovery Algorithms: An Experimental
Analysis | Accepted in Algorithmic Fairness in Artificial intelligence, Machine
learning and Decision Making workshop at SDM 2023 | null | null | null | cs.LG cs.CY cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning (ML) based approaches are increasingly being used in a
number of applications with societal impact. Training ML models often require
vast amounts of labeled data, and crowdsourcing is a dominant paradigm for
obtaining labels from multiple workers. Crowd workers may sometimes provide
unreliable labels, and to address this, truth discovery (TD) algorithms such as
majority voting are applied to determine the consensus labels from conflicting
worker responses. However, it is important to note that these consensus labels
may still be biased based on sensitive attributes such as gender, race, or
political affiliation. Even when sensitive attributes are not involved, the
labels can be biased due to different perspectives of subjective aspects such
as toxicity. In this paper, we conduct a systematic study of the bias and
fairness of TD algorithms. Our findings using two existing crowd-labeled
datasets, reveal that a non-trivial proportion of workers provide biased
results, and using simple approaches for TD is sub-optimal. Our study also
demonstrates that popular TD algorithms are not a panacea. Additionally, we
quantify the impact of these unfair workers on downstream ML tasks and show
that conventional methods for achieving fairness and correcting label biases
are ineffective in this setting. We end the paper with a plea for the design of
novel bias-aware truth discovery algorithms that can ameliorate these issues.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 04:56:35 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Lazier",
"Simone",
""
],
[
"Thirumuruganathan",
"Saravanan",
""
],
[
"Anahideh",
"Hadis",
""
]
] | TITLE: Fairness and Bias in Truth Discovery Algorithms: An Experimental
Analysis
ABSTRACT: Machine learning (ML) based approaches are increasingly being used in a
number of applications with societal impact. Training ML models often require
vast amounts of labeled data, and crowdsourcing is a dominant paradigm for
obtaining labels from multiple workers. Crowd workers may sometimes provide
unreliable labels, and to address this, truth discovery (TD) algorithms such as
majority voting are applied to determine the consensus labels from conflicting
worker responses. However, it is important to note that these consensus labels
may still be biased based on sensitive attributes such as gender, race, or
political affiliation. Even when sensitive attributes are not involved, the
labels can be biased due to different perspectives of subjective aspects such
as toxicity. In this paper, we conduct a systematic study of the bias and
fairness of TD algorithms. Our findings using two existing crowd-labeled
datasets, reveal that a non-trivial proportion of workers provide biased
results, and using simple approaches for TD is sub-optimal. Our study also
demonstrates that popular TD algorithms are not a panacea. Additionally, we
quantify the impact of these unfair workers on downstream ML tasks and show
that conventional methods for achieving fairness and correcting label biases
are ineffective in this setting. We end the paper with a plea for the design of
novel bias-aware truth discovery algorithms that can ameliorate these issues.
|
2304.12577 | Hyungtae Lim | Hyungtae Lim and Daebeom Kim and Beomsoo Kim and Hyun Myung | AdaLIO: Robust Adaptive LiDAR-Inertial Odometry in Degenerate Indoor
Environments | 6 pages, 5 figures | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In recent years, the demand for mapping construction sites or buildings using
light detection and ranging~(LiDAR) sensors has been increased to model
environments for efficient site management. However, it is observed that
sometimes LiDAR-based approaches diverge in narrow and confined environments,
such as spiral stairs and corridors, caused by fixed parameters regardless of
the changes in the environments. That is, the parameters of LiDAR (-inertial)
odometry are mostly set for open space; thus, if the same parameters suitable
for the open space are applied in a corridor-like scene, it results in
divergence of odometry methods, which is referred to as \textit{degeneracy}. To
tackle this degeneracy problem, we propose a robust LiDAR inertial odometry
called \textit{AdaLIO}, which employs an adaptive parameter setting strategy.
To this end, we first check the degeneracy by checking whether the surroundings
are corridor-like environments. If so, the parameters relevant to voxelization
and normal vector estimation are adaptively changed to increase the number of
correspondences. As verified in a public dataset, our proposed method showed
promising performance in narrow and cramped environments, avoiding the
degeneracy problem.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 05:05:47 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Lim",
"Hyungtae",
""
],
[
"Kim",
"Daebeom",
""
],
[
"Kim",
"Beomsoo",
""
],
[
"Myung",
"Hyun",
""
]
] | TITLE: AdaLIO: Robust Adaptive LiDAR-Inertial Odometry in Degenerate Indoor
Environments
ABSTRACT: In recent years, the demand for mapping construction sites or buildings using
light detection and ranging~(LiDAR) sensors has been increased to model
environments for efficient site management. However, it is observed that
sometimes LiDAR-based approaches diverge in narrow and confined environments,
such as spiral stairs and corridors, caused by fixed parameters regardless of
the changes in the environments. That is, the parameters of LiDAR (-inertial)
odometry are mostly set for open space; thus, if the same parameters suitable
for the open space are applied in a corridor-like scene, it results in
divergence of odometry methods, which is referred to as \textit{degeneracy}. To
tackle this degeneracy problem, we propose a robust LiDAR inertial odometry
called \textit{AdaLIO}, which employs an adaptive parameter setting strategy.
To this end, we first check the degeneracy by checking whether the surroundings
are corridor-like environments. If so, the parameters relevant to voxelization
and normal vector estimation are adaptively changed to increase the number of
correspondences. As verified in a public dataset, our proposed method showed
promising performance in narrow and cramped environments, avoiding the
degeneracy problem.
|
2304.12592 | Han Wang | Han Wang, Jiayuan Zhang, Lipeng Wan, Xingyu Chen, Xuguang Lan, Nanning
Zheng | MMRDN: Consistent Representation for Multi-View Manipulation
Relationship Detection in Object-Stacked Scenes | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Manipulation relationship detection (MRD) aims to guide the robot to grasp
objects in the right order, which is important to ensure the safety and
reliability of grasping in object stacked scenes. Previous works infer
manipulation relationship by deep neural network trained with data collected
from a predefined view, which has limitation in visual dislocation in
unstructured environments. Multi-view data provide more comprehensive
information in space, while a challenge of multi-view MRD is domain shift. In
this paper, we propose a novel multi-view fusion framework, namely multi-view
MRD network (MMRDN), which is trained by 2D and 3D multi-view data. We project
the 2D data from different views into a common hidden space and fit the
embeddings with a set of Von-Mises-Fisher distributions to learn the consistent
representations. Besides, taking advantage of position information within the
3D data, we select a set of $K$ Maximum Vertical Neighbors (KMVN) points from
the point cloud of each object pair, which encodes the relative position of
these two objects. Finally, the features of multi-view 2D and 3D data are
concatenated to predict the pairwise relationship of objects. Experimental
results on the challenging REGRAD dataset show that MMRDN outperforms the
state-of-the-art methods in multi-view MRD tasks. The results also demonstrate
that our model trained by synthetic data is capable to transfer to real-world
scenarios.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 05:55:29 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Wang",
"Han",
""
],
[
"Zhang",
"Jiayuan",
""
],
[
"Wan",
"Lipeng",
""
],
[
"Chen",
"Xingyu",
""
],
[
"Lan",
"Xuguang",
""
],
[
"Zheng",
"Nanning",
""
]
] | TITLE: MMRDN: Consistent Representation for Multi-View Manipulation
Relationship Detection in Object-Stacked Scenes
ABSTRACT: Manipulation relationship detection (MRD) aims to guide the robot to grasp
objects in the right order, which is important to ensure the safety and
reliability of grasping in object stacked scenes. Previous works infer
manipulation relationship by deep neural network trained with data collected
from a predefined view, which has limitation in visual dislocation in
unstructured environments. Multi-view data provide more comprehensive
information in space, while a challenge of multi-view MRD is domain shift. In
this paper, we propose a novel multi-view fusion framework, namely multi-view
MRD network (MMRDN), which is trained by 2D and 3D multi-view data. We project
the 2D data from different views into a common hidden space and fit the
embeddings with a set of Von-Mises-Fisher distributions to learn the consistent
representations. Besides, taking advantage of position information within the
3D data, we select a set of $K$ Maximum Vertical Neighbors (KMVN) points from
the point cloud of each object pair, which encodes the relative position of
these two objects. Finally, the features of multi-view 2D and 3D data are
concatenated to predict the pairwise relationship of objects. Experimental
results on the challenging REGRAD dataset show that MMRDN outperforms the
state-of-the-art methods in multi-view MRD tasks. The results also demonstrate
that our model trained by synthetic data is capable to transfer to real-world
scenarios.
|
2304.12596 | Lin Zhang | Yu Zhang and Lin Zhang | Detection of Pavement Cracks by Deep Learning Models of Transformer and
UNet | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fracture is one of the main failure modes of engineering structures such as
buildings and roads. Effective detection of surface cracks is significant for
damage evaluation and structure maintenance. In recent years, the emergence and
development of deep learning techniques have shown great potential to
facilitate surface crack detection. Currently, most reported tasks were
performed by a convolutional neural network (CNN), while the limitation of CNN
may be improved by the transformer architecture introduced recently. In this
study, we investigated nine promising models to evaluate their performance in
pavement surface crack detection by model accuracy, computational complexity,
and model stability. We created 711 images of 224 by 224 pixels with crack
labels, selected an optimal loss function, compared the evaluation metrics of
the validation dataset and test dataset, analyzed the data details, and checked
the segmentation outcomes of each model. We find that transformer-based models
generally are easier to converge during the training process and have higher
accuracy, but usually exhibit more memory consumption and low processing
efficiency. Among nine models, SwinUNet outperforms the other two transformers
and shows the highest accuracy among nine models. The results should shed light
on surface crack detection by various deep-learning models and provide a
guideline for future applications in this field.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 06:07:49 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Zhang",
"Yu",
""
],
[
"Zhang",
"Lin",
""
]
] | TITLE: Detection of Pavement Cracks by Deep Learning Models of Transformer and
UNet
ABSTRACT: Fracture is one of the main failure modes of engineering structures such as
buildings and roads. Effective detection of surface cracks is significant for
damage evaluation and structure maintenance. In recent years, the emergence and
development of deep learning techniques have shown great potential to
facilitate surface crack detection. Currently, most reported tasks were
performed by a convolutional neural network (CNN), while the limitation of CNN
may be improved by the transformer architecture introduced recently. In this
study, we investigated nine promising models to evaluate their performance in
pavement surface crack detection by model accuracy, computational complexity,
and model stability. We created 711 images of 224 by 224 pixels with crack
labels, selected an optimal loss function, compared the evaluation metrics of
the validation dataset and test dataset, analyzed the data details, and checked
the segmentation outcomes of each model. We find that transformer-based models
generally are easier to converge during the training process and have higher
accuracy, but usually exhibit more memory consumption and low processing
efficiency. Among nine models, SwinUNet outperforms the other two transformers
and shows the highest accuracy among nine models. The results should shed light
on surface crack detection by various deep-learning models and provide a
guideline for future applications in this field.
|
2304.12604 | Hao Dong | Hao Dong, Zhiyuan Ning, Pengyang Wang, Ziyue Qiao, Pengfei Wang,
Yuanchun Zhou, Yanjie Fu | Adaptive Path-Memory Network for Temporal Knowledge Graph Reasoning | Accepted to IJCAI 2023 | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Temporal knowledge graph (TKG) reasoning aims to predict the future missing
facts based on historical information and has gained increasing research
interest recently. Lots of works have been made to model the historical
structural and temporal characteristics for the reasoning task. Most existing
works model the graph structure mainly depending on entity representation.
However, the magnitude of TKG entities in real-world scenarios is considerable,
and an increasing number of new entities will arise as time goes on. Therefore,
we propose a novel architecture modeling with relation feature of TKG, namely
aDAptivE path-MemOry Network (DaeMon), which adaptively models the temporal
path information between query subject and each object candidate across history
time. It models the historical information without depending on entity
representation. Specifically, DaeMon uses path memory to record the temporal
path information derived from path aggregation unit across timeline considering
the memory passing strategy between adjacent timestamps. Extensive experiments
conducted on four real-world TKG datasets demonstrate that our proposed model
obtains substantial performance improvement and outperforms the
state-of-the-art up to 4.8% absolute in MRR.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 06:33:08 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Dong",
"Hao",
""
],
[
"Ning",
"Zhiyuan",
""
],
[
"Wang",
"Pengyang",
""
],
[
"Qiao",
"Ziyue",
""
],
[
"Wang",
"Pengfei",
""
],
[
"Zhou",
"Yuanchun",
""
],
[
"Fu",
"Yanjie",
""
]
] | TITLE: Adaptive Path-Memory Network for Temporal Knowledge Graph Reasoning
ABSTRACT: Temporal knowledge graph (TKG) reasoning aims to predict the future missing
facts based on historical information and has gained increasing research
interest recently. Lots of works have been made to model the historical
structural and temporal characteristics for the reasoning task. Most existing
works model the graph structure mainly depending on entity representation.
However, the magnitude of TKG entities in real-world scenarios is considerable,
and an increasing number of new entities will arise as time goes on. Therefore,
we propose a novel architecture modeling with relation feature of TKG, namely
aDAptivE path-MemOry Network (DaeMon), which adaptively models the temporal
path information between query subject and each object candidate across history
time. It models the historical information without depending on entity
representation. Specifically, DaeMon uses path memory to record the temporal
path information derived from path aggregation unit across timeline considering
the memory passing strategy between adjacent timestamps. Extensive experiments
conducted on four real-world TKG datasets demonstrate that our proposed model
obtains substantial performance improvement and outperforms the
state-of-the-art up to 4.8% absolute in MRR.
|
2304.12609 | Subhayan De | Subhayan De and Patrick T. Brewick | A Bi-fidelity DeepONet Approach for Modeling Uncertain and Degrading
Hysteretic Systems | 23 pages, 15 figures | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Nonlinear systems, such as with degrading hysteretic behavior, are often
encountered in engineering applications. In addition, due to the ubiquitous
presence of uncertainty and the modeling of such systems becomes increasingly
difficult. On the other hand, datasets from pristine models developed without
knowing the nature of the degrading effects can be easily obtained. In this
paper, we use datasets from pristine models without considering the degrading
effects of hysteretic systems as low-fidelity representations that capture many
of the important characteristics of the true system's behavior to train a deep
operator network (DeepONet). Three numerical examples are used to show that the
proposed use of the DeepONets to model the discrepancies between the
low-fidelity model and the true system's response leads to significant
improvements in the prediction error in the presence of uncertainty in the
model parameters for degrading hysteretic systems.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 06:49:56 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"De",
"Subhayan",
""
],
[
"Brewick",
"Patrick T.",
""
]
] | TITLE: A Bi-fidelity DeepONet Approach for Modeling Uncertain and Degrading
Hysteretic Systems
ABSTRACT: Nonlinear systems, such as with degrading hysteretic behavior, are often
encountered in engineering applications. In addition, due to the ubiquitous
presence of uncertainty and the modeling of such systems becomes increasingly
difficult. On the other hand, datasets from pristine models developed without
knowing the nature of the degrading effects can be easily obtained. In this
paper, we use datasets from pristine models without considering the degrading
effects of hysteretic systems as low-fidelity representations that capture many
of the important characteristics of the true system's behavior to train a deep
operator network (DeepONet). Three numerical examples are used to show that the
proposed use of the DeepONets to model the discrepancies between the
low-fidelity model and the true system's response leads to significant
improvements in the prediction error in the presence of uncertainty in the
model parameters for degrading hysteretic systems.
|
2304.12616 | Guozhang Li | Guozhang Li, De Cheng, Xinpeng Ding, Nannan Wang, Jie Li, Xinbo Gao | Weakly-Supervised Temporal Action Localization with Bidirectional
Semantic Consistency Constraint | accepted by TNNLS | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Weakly Supervised Temporal Action Localization (WTAL) aims to classify and
localize temporal boundaries of actions for the video, given only video-level
category labels in the training datasets. Due to the lack of boundary
information during training, existing approaches formulate WTAL as a
classificationproblem, i.e., generating the temporal class activation map
(T-CAM) for localization. However, with only classification loss, the model
would be sub-optimized, i.e., the action-related scenes are enough to
distinguish different class labels. Regarding other actions in the
action-related scene ( i.e., the scene same as positive actions) as co-scene
actions, this sub-optimized model would misclassify the co-scene actions as
positive actions. To address this misclassification, we propose a simple yet
efficient method, named bidirectional semantic consistency constraint (Bi-SCC),
to discriminate the positive actions from co-scene actions. The proposed Bi-SCC
firstly adopts a temporal context augmentation to generate an augmented video
that breaks the correlation between positive actions and their co-scene actions
in the inter-video; Then, a semantic consistency constraint (SCC) is used to
enforce the predictions of the original video and augmented video to be
consistent, hence suppressing the co-scene actions. However, we find that this
augmented video would destroy the original temporal context. Simply applying
the consistency constraint would affect the completeness of localized positive
actions. Hence, we boost the SCC in a bidirectional way to suppress co-scene
actions while ensuring the integrity of positive actions, by cross-supervising
the original and augmented videos. Finally, our proposed Bi-SCC can be applied
to current WTAL approaches, and improve their performance. Experimental results
show that our approach outperforms the state-of-the-art methods on THUMOS14 and
ActivityNet.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 07:20:33 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Li",
"Guozhang",
""
],
[
"Cheng",
"De",
""
],
[
"Ding",
"Xinpeng",
""
],
[
"Wang",
"Nannan",
""
],
[
"Li",
"Jie",
""
],
[
"Gao",
"Xinbo",
""
]
] | TITLE: Weakly-Supervised Temporal Action Localization with Bidirectional
Semantic Consistency Constraint
ABSTRACT: Weakly Supervised Temporal Action Localization (WTAL) aims to classify and
localize temporal boundaries of actions for the video, given only video-level
category labels in the training datasets. Due to the lack of boundary
information during training, existing approaches formulate WTAL as a
classificationproblem, i.e., generating the temporal class activation map
(T-CAM) for localization. However, with only classification loss, the model
would be sub-optimized, i.e., the action-related scenes are enough to
distinguish different class labels. Regarding other actions in the
action-related scene ( i.e., the scene same as positive actions) as co-scene
actions, this sub-optimized model would misclassify the co-scene actions as
positive actions. To address this misclassification, we propose a simple yet
efficient method, named bidirectional semantic consistency constraint (Bi-SCC),
to discriminate the positive actions from co-scene actions. The proposed Bi-SCC
firstly adopts a temporal context augmentation to generate an augmented video
that breaks the correlation between positive actions and their co-scene actions
in the inter-video; Then, a semantic consistency constraint (SCC) is used to
enforce the predictions of the original video and augmented video to be
consistent, hence suppressing the co-scene actions. However, we find that this
augmented video would destroy the original temporal context. Simply applying
the consistency constraint would affect the completeness of localized positive
actions. Hence, we boost the SCC in a bidirectional way to suppress co-scene
actions while ensuring the integrity of positive actions, by cross-supervising
the original and augmented videos. Finally, our proposed Bi-SCC can be applied
to current WTAL approaches, and improve their performance. Experimental results
show that our approach outperforms the state-of-the-art methods on THUMOS14 and
ActivityNet.
|
2304.12624 | Mizuki Tabata | Mizuki Tabata, Kana Kurata, Junichiro Tamamatsu | Shape-Net: Room Layout Estimation from Panoramic Images Robust to
Occlusion using Knowledge Distillation with 3D Shapes as Additional Inputs | Accepted by CVPR2023 workshop (CIVILS) | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Estimating the layout of a room from a single-shot panoramic image is
important in virtual/augmented reality and furniture layout simulation. This
involves identifying three-dimensional (3D) geometry, such as the location of
corners and boundaries, and performing 3D reconstruction. However, occlusion is
a common issue that can negatively impact room layout estimation, and this has
not been thoroughly studied to date. It is possible to obtain 3D shape
information of rooms as drawings of buildings and coordinates of corners from
image datasets, thus we propose providing both 2D panoramic and 3D information
to a model to effectively deal with occlusion. However, simply feeding 3D
information to a model is not sufficient to utilize the shape information for
an occluded area. Therefore, we improve the model by introducing 3D
Intersection over Union (IoU) loss to effectively use 3D information. In some
cases, drawings are not available or the construction deviates from a drawing.
Considering such practical cases, we propose a method for distilling knowledge
from a model trained with both images and 3D information to a model that takes
only images as input. The proposed model, which is called Shape-Net, achieves
state-of-the-art (SOTA) performance on benchmark datasets. We also confirmed
its effectiveness in dealing with occlusion through significantly improved
accuracy on images with occlusion compared with existing models.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 07:45:43 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Tabata",
"Mizuki",
""
],
[
"Kurata",
"Kana",
""
],
[
"Tamamatsu",
"Junichiro",
""
]
] | TITLE: Shape-Net: Room Layout Estimation from Panoramic Images Robust to
Occlusion using Knowledge Distillation with 3D Shapes as Additional Inputs
ABSTRACT: Estimating the layout of a room from a single-shot panoramic image is
important in virtual/augmented reality and furniture layout simulation. This
involves identifying three-dimensional (3D) geometry, such as the location of
corners and boundaries, and performing 3D reconstruction. However, occlusion is
a common issue that can negatively impact room layout estimation, and this has
not been thoroughly studied to date. It is possible to obtain 3D shape
information of rooms as drawings of buildings and coordinates of corners from
image datasets, thus we propose providing both 2D panoramic and 3D information
to a model to effectively deal with occlusion. However, simply feeding 3D
information to a model is not sufficient to utilize the shape information for
an occluded area. Therefore, we improve the model by introducing 3D
Intersection over Union (IoU) loss to effectively use 3D information. In some
cases, drawings are not available or the construction deviates from a drawing.
Considering such practical cases, we propose a method for distilling knowledge
from a model trained with both images and 3D information to a model that takes
only images as input. The proposed model, which is called Shape-Net, achieves
state-of-the-art (SOTA) performance on benchmark datasets. We also confirmed
its effectiveness in dealing with occlusion through significantly improved
accuracy on images with occlusion compared with existing models.
|
2304.12630 | Van-Duc Le | Van-Duc Le | Spatiotemporal Graph Convolutional Recurrent Neural Network Model for
Citywide Air Pollution Forecasting | null | null | null | null | cs.CV cs.LG eess.SP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Citywide Air Pollution Forecasting tries to precisely predict the air quality
multiple hours ahead for the entire city. This topic is challenged since air
pollution varies in a spatiotemporal manner and depends on many complicated
factors. Our previous research has solved the problem by considering the whole
city as an image and leveraged a Convolutional Long Short-Term Memory
(ConvLSTM) model to learn the spatiotemporal features. However, an image-based
representation may not be ideal as air pollution and other impact factors have
natural graph structures. In this research, we argue that a Graph Convolutional
Network (GCN) can efficiently represent the spatial features of air quality
readings in the whole city. Specially, we extend the ConvLSTM model to a
Spatiotemporal Graph Convolutional Recurrent Neural Network (Spatiotemporal
GCRNN) model by tightly integrating a GCN architecture into an RNN structure
for efficient learning spatiotemporal characteristics of air quality values and
their influential factors. Our extensive experiments prove the proposed model
has a better performance compare to the state-of-the-art ConvLSTM model for air
pollution predicting while the number of parameters is much smaller. Moreover,
our approach is also superior to a hybrid GCN-based method in a real-world air
pollution dataset.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 07:57:07 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Le",
"Van-Duc",
""
]
] | TITLE: Spatiotemporal Graph Convolutional Recurrent Neural Network Model for
Citywide Air Pollution Forecasting
ABSTRACT: Citywide Air Pollution Forecasting tries to precisely predict the air quality
multiple hours ahead for the entire city. This topic is challenged since air
pollution varies in a spatiotemporal manner and depends on many complicated
factors. Our previous research has solved the problem by considering the whole
city as an image and leveraged a Convolutional Long Short-Term Memory
(ConvLSTM) model to learn the spatiotemporal features. However, an image-based
representation may not be ideal as air pollution and other impact factors have
natural graph structures. In this research, we argue that a Graph Convolutional
Network (GCN) can efficiently represent the spatial features of air quality
readings in the whole city. Specially, we extend the ConvLSTM model to a
Spatiotemporal Graph Convolutional Recurrent Neural Network (Spatiotemporal
GCRNN) model by tightly integrating a GCN architecture into an RNN structure
for efficient learning spatiotemporal characteristics of air quality values and
their influential factors. Our extensive experiments prove the proposed model
has a better performance compare to the state-of-the-art ConvLSTM model for air
pollution predicting while the number of parameters is much smaller. Moreover,
our approach is also superior to a hybrid GCN-based method in a real-world air
pollution dataset.
|
2304.12636 | Nolwenn Bernard | Nolwenn Bernard and Krisztian Balog | MG-ShopDial: A Multi-Goal Conversational Dataset for e-Commerce | Proceedings of the 46th International ACM SIGIR Conference on
Research and Development in Information Retrieval (SIGIR '23), July 23--27,
2023, Taipei, Taiwan | null | 10.1145/3539618.3591883 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conversational systems can be particularly effective in supporting complex
information seeking scenarios with evolving information needs. Finding the
right products on an e-commerce platform is one such scenario, where a
conversational agent would need to be able to provide search capabilities over
the item catalog, understand and make recommendations based on the user's
preferences, and answer a range of questions related to items and their usage.
Yet, existing conversational datasets do not fully support the idea of mixing
different conversational goals (i.e., search, recommendation, and question
answering) and instead focus on a single goal. To address this, we introduce
MG-ShopDial: a dataset of conversations mixing different goals in the domain of
e-commerce. Specifically, we make the following contributions. First, we
develop a coached human-human data collection protocol where each dialogue
participant is given a set of instructions, instead of a specific script or
answers to choose from. Second, we implement a data collection tool to
facilitate the collection of multi-goal conversations via a web chat interface,
using the above protocol. Third, we create the MG-ShopDial collection, which
contains 64 high-quality dialogues with a total of 2,196 utterances for
e-commerce scenarios of varying complexity. The dataset is additionally
annotated with both intents and goals on the utterance level. Finally, we
present an analysis of this dataset and identify multi-goal conversational
patterns.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 08:07:21 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Bernard",
"Nolwenn",
""
],
[
"Balog",
"Krisztian",
""
]
] | TITLE: MG-ShopDial: A Multi-Goal Conversational Dataset for e-Commerce
ABSTRACT: Conversational systems can be particularly effective in supporting complex
information seeking scenarios with evolving information needs. Finding the
right products on an e-commerce platform is one such scenario, where a
conversational agent would need to be able to provide search capabilities over
the item catalog, understand and make recommendations based on the user's
preferences, and answer a range of questions related to items and their usage.
Yet, existing conversational datasets do not fully support the idea of mixing
different conversational goals (i.e., search, recommendation, and question
answering) and instead focus on a single goal. To address this, we introduce
MG-ShopDial: a dataset of conversations mixing different goals in the domain of
e-commerce. Specifically, we make the following contributions. First, we
develop a coached human-human data collection protocol where each dialogue
participant is given a set of instructions, instead of a specific script or
answers to choose from. Second, we implement a data collection tool to
facilitate the collection of multi-goal conversations via a web chat interface,
using the above protocol. Third, we create the MG-ShopDial collection, which
contains 64 high-quality dialogues with a total of 2,196 utterances for
e-commerce scenarios of varying complexity. The dataset is additionally
annotated with both intents and goals on the utterance level. Finally, we
present an analysis of this dataset and identify multi-goal conversational
patterns.
|
2304.12645 | Peng Qian | Peng Qian, Jianting He, Lingling Lu, Siwei Wu, Zhipeng Lu, Lei Wu,
Yajin Zhou, Qinming He | Demystifying Random Number in Ethereum Smart Contract: Taxonomy,
Vulnerability Identification, and Attack Detection | This is the preprint of the paper that has been accepted by IEEE
Transactions on Software Engineering (TSE) | null | null | null | cs.SE cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent years have witnessed explosive growth in blockchain smart contract
applications. As smart contracts become increasingly popular and carry trillion
dollars worth of digital assets, they become more of an appealing target for
attackers, who have exploited vulnerabilities in smart contracts to cause
catastrophic economic losses. Notwithstanding a proliferation of work that has
been developed to detect an impressive list of vulnerabilities, the bad
randomness vulnerability is overlooked by many existing tools. In this paper,
we make the first attempt to provide a systematic analysis of random numbers in
Ethereum smart contracts, by investigating the principles behind pseudo-random
number generation and organizing them into a taxonomy. We also lucubrate
various attacks against bad random numbers and group them into four categories.
Furthermore, we present RNVulDet - a tool that incorporates taint analysis
techniques to automatically identify bad randomness vulnerabilities and detect
corresponding attack transactions. To extensively verify the effectiveness of
RNVulDet, we construct three new datasets: i) 34 well-known contracts that are
reported to possess bad randomness vulnerabilities, ii) 214 popular contracts
that have been rigorously audited before launch and are regarded as free of bad
randomness vulnerabilities, and iii) a dataset consisting of 47,668 smart
contracts and 49,951 suspicious transactions. We compare RNVulDet with three
state-of-the-art smart contract vulnerability detectors, and our tool
significantly outperforms them. Meanwhile, RNVulDet spends 2.98s per contract
on average, in most cases orders-of-magnitude faster than other tools. RNVulDet
successfully reveals 44,264 attack transactions. Our implementation and
datasets are released, hoping to inspire others.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 08:23:11 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Qian",
"Peng",
""
],
[
"He",
"Jianting",
""
],
[
"Lu",
"Lingling",
""
],
[
"Wu",
"Siwei",
""
],
[
"Lu",
"Zhipeng",
""
],
[
"Wu",
"Lei",
""
],
[
"Zhou",
"Yajin",
""
],
[
"He",
"Qinming",
""
]
] | TITLE: Demystifying Random Number in Ethereum Smart Contract: Taxonomy,
Vulnerability Identification, and Attack Detection
ABSTRACT: Recent years have witnessed explosive growth in blockchain smart contract
applications. As smart contracts become increasingly popular and carry trillion
dollars worth of digital assets, they become more of an appealing target for
attackers, who have exploited vulnerabilities in smart contracts to cause
catastrophic economic losses. Notwithstanding a proliferation of work that has
been developed to detect an impressive list of vulnerabilities, the bad
randomness vulnerability is overlooked by many existing tools. In this paper,
we make the first attempt to provide a systematic analysis of random numbers in
Ethereum smart contracts, by investigating the principles behind pseudo-random
number generation and organizing them into a taxonomy. We also lucubrate
various attacks against bad random numbers and group them into four categories.
Furthermore, we present RNVulDet - a tool that incorporates taint analysis
techniques to automatically identify bad randomness vulnerabilities and detect
corresponding attack transactions. To extensively verify the effectiveness of
RNVulDet, we construct three new datasets: i) 34 well-known contracts that are
reported to possess bad randomness vulnerabilities, ii) 214 popular contracts
that have been rigorously audited before launch and are regarded as free of bad
randomness vulnerabilities, and iii) a dataset consisting of 47,668 smart
contracts and 49,951 suspicious transactions. We compare RNVulDet with three
state-of-the-art smart contract vulnerability detectors, and our tool
significantly outperforms them. Meanwhile, RNVulDet spends 2.98s per contract
on average, in most cases orders-of-magnitude faster than other tools. RNVulDet
successfully reveals 44,264 attack transactions. Our implementation and
datasets are released, hoping to inspire others.
|
2304.12664 | Ban Chen | Ban Chen, Xin Jin, Youxin Chen, Longhai Wu, Jie Chen, Jayoon Koo,
Cheul-hee Hahm | Dynamic Video Frame Interpolation with integrated Difficulty
Pre-Assessment | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video frame interpolation(VFI) has witnessed great progress in recent years.
While existing VFI models still struggle to achieve a good trade-off between
accuracy and efficiency: fast models often have inferior accuracy; accurate
models typically run slowly. However, easy samples with small motion or clear
texture can achieve competitive results with simple models and do not require
heavy computation. In this paper, we present an integrated pipeline which
combines difficulty assessment with video frame interpolation. Specifically, it
firstly leverages a pre-assessment model to measure the interpolation
difficulty level of input frames, and then dynamically selects an appropriate
VFI model to generate interpolation results. Furthermore, a large-scale VFI
difficulty assessment dataset is collected and annotated to train our
pre-assessment model. Extensive experiments show that easy samples pass through
fast models while difficult samples inference with heavy models, and our
proposed pipeline can improve the accuracy-efficiency trade-off for VFI.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 09:11:20 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Chen",
"Ban",
""
],
[
"Jin",
"Xin",
""
],
[
"Chen",
"Youxin",
""
],
[
"Wu",
"Longhai",
""
],
[
"Chen",
"Jie",
""
],
[
"Koo",
"Jayoon",
""
],
[
"Hahm",
"Cheul-hee",
""
]
] | TITLE: Dynamic Video Frame Interpolation with integrated Difficulty
Pre-Assessment
ABSTRACT: Video frame interpolation(VFI) has witnessed great progress in recent years.
While existing VFI models still struggle to achieve a good trade-off between
accuracy and efficiency: fast models often have inferior accuracy; accurate
models typically run slowly. However, easy samples with small motion or clear
texture can achieve competitive results with simple models and do not require
heavy computation. In this paper, we present an integrated pipeline which
combines difficulty assessment with video frame interpolation. Specifically, it
firstly leverages a pre-assessment model to measure the interpolation
difficulty level of input frames, and then dynamically selects an appropriate
VFI model to generate interpolation results. Furthermore, a large-scale VFI
difficulty assessment dataset is collected and annotated to train our
pre-assessment model. Extensive experiments show that easy samples pass through
fast models while difficult samples inference with heavy models, and our
proposed pipeline can improve the accuracy-efficiency trade-off for VFI.
|
2304.12667 | Dieter Brughmans | Dieter Brughmans, Lissa Melis, David Martens | Disagreement amongst counterfactual explanations: How transparency can
be deceptive | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Counterfactual explanations are increasingly used as an Explainable
Artificial Intelligence (XAI) technique to provide stakeholders of complex
machine learning algorithms with explanations for data-driven decisions. The
popularity of counterfactual explanations resulted in a boom in the algorithms
generating them. However, not every algorithm creates uniform explanations for
the same instance. Even though in some contexts multiple possible explanations
are beneficial, there are circumstances where diversity amongst counterfactual
explanations results in a potential disagreement problem among stakeholders.
Ethical issues arise when for example, malicious agents use this diversity to
fairwash an unfair machine learning model by hiding sensitive features. As
legislators worldwide tend to start including the right to explanations for
data-driven, high-stakes decisions in their policies, these ethical issues
should be understood and addressed. Our literature review on the disagreement
problem in XAI reveals that this problem has never been empirically assessed
for counterfactual explanations. Therefore, in this work, we conduct a
large-scale empirical analysis, on 40 datasets, using 12 explanation-generating
methods, for two black-box models, yielding over 192.0000 explanations. Our
study finds alarmingly high disagreement levels between the methods tested. A
malicious user is able to both exclude and include desired features when
multiple counterfactual explanations are available. This disagreement seems to
be driven mainly by the dataset characteristics and the type of counterfactual
algorithm. XAI centers on the transparency of algorithmic decision-making, but
our analysis advocates for transparency about this self-proclaimed transparency
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 09:15:37 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Brughmans",
"Dieter",
""
],
[
"Melis",
"Lissa",
""
],
[
"Martens",
"David",
""
]
] | TITLE: Disagreement amongst counterfactual explanations: How transparency can
be deceptive
ABSTRACT: Counterfactual explanations are increasingly used as an Explainable
Artificial Intelligence (XAI) technique to provide stakeholders of complex
machine learning algorithms with explanations for data-driven decisions. The
popularity of counterfactual explanations resulted in a boom in the algorithms
generating them. However, not every algorithm creates uniform explanations for
the same instance. Even though in some contexts multiple possible explanations
are beneficial, there are circumstances where diversity amongst counterfactual
explanations results in a potential disagreement problem among stakeholders.
Ethical issues arise when for example, malicious agents use this diversity to
fairwash an unfair machine learning model by hiding sensitive features. As
legislators worldwide tend to start including the right to explanations for
data-driven, high-stakes decisions in their policies, these ethical issues
should be understood and addressed. Our literature review on the disagreement
problem in XAI reveals that this problem has never been empirically assessed
for counterfactual explanations. Therefore, in this work, we conduct a
large-scale empirical analysis, on 40 datasets, using 12 explanation-generating
methods, for two black-box models, yielding over 192.0000 explanations. Our
study finds alarmingly high disagreement levels between the methods tested. A
malicious user is able to both exclude and include desired features when
multiple counterfactual explanations are available. This disagreement seems to
be driven mainly by the dataset characteristics and the type of counterfactual
algorithm. XAI centers on the transparency of algorithmic decision-making, but
our analysis advocates for transparency about this self-proclaimed transparency
|
2304.12671 | Raquel Blanco | Raquel Blanco, Javier Tuya, Ruben V. Seco | Test adequacy evaluation for the user-database interaction: a
specification-based approach | IEEE Fifth International Conference on Software Testing, Verification
and Validation (ICST 2012), Montreal, QC, Canada, 2012 | null | 10.1109/ICST.2012.87 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Testing a database application is a challenging process where both the
database and the user interaction have to be considered in the design of test
cases. This paper describes a specification-based approach to guide the design
of test inputs (both the test database and the user inputs) for a database
application and to automatically evaluate the test adequacy. First, the system
specification of the application is modelled: (1) the structure of the database
and the user interface are represented in a single model, called Integrated
Data Model (IDM), (2) the functional requirements are expressed as a set of
business rules, written in terms of the IDM. Then, a MCDC-based criterion is
applied over the business rules to automatically derive the situations of
interest to be tested (test requirements), which guide the design of the test
inputs. Finally, the adequacy of these test inputs is automatically evaluated
to determine whether the test requirements are covered. The approach has been
applied to the TPC-C benchmark. The results show that it allows designing test
cases that are able to detect interesting faults which were located in the
procedural code of the implementation.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 09:20:22 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Blanco",
"Raquel",
""
],
[
"Tuya",
"Javier",
""
],
[
"Seco",
"Ruben V.",
""
]
] | TITLE: Test adequacy evaluation for the user-database interaction: a
specification-based approach
ABSTRACT: Testing a database application is a challenging process where both the
database and the user interaction have to be considered in the design of test
cases. This paper describes a specification-based approach to guide the design
of test inputs (both the test database and the user inputs) for a database
application and to automatically evaluate the test adequacy. First, the system
specification of the application is modelled: (1) the structure of the database
and the user interface are represented in a single model, called Integrated
Data Model (IDM), (2) the functional requirements are expressed as a set of
business rules, written in terms of the IDM. Then, a MCDC-based criterion is
applied over the business rules to automatically derive the situations of
interest to be tested (test requirements), which guide the design of the test
inputs. Finally, the adequacy of these test inputs is automatically evaluated
to determine whether the test requirements are covered. The approach has been
applied to the TPC-C benchmark. The results show that it allows designing test
cases that are able to detect interesting faults which were located in the
procedural code of the implementation.
|
2304.12674 | Domagoj Severdija | Domagoj \v{S}everdija, Tomislav Prusina, Antonio Jovanovi\'c, Luka
Borozan, Jurica Maltar, and Domagoj Matijevi\'c | Compressing Sentence Representation with maximum Coding Rate Reduction | 14 pages, 3 figures, accepted on ICT and Electronics Convention
(MIPRO), Croatia | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | In most natural language inference problems, sentence representation is
needed for semantic retrieval tasks. In recent years, pre-trained large
language models have been quite effective for computing such representations.
These models produce high-dimensional sentence embeddings. An evident
performance gap between large and small models exists in practice. Hence, due
to space and time hardware limitations, there is a need to attain comparable
results when using the smaller model, which is usually a distilled version of
the large language model. In this paper, we assess the model distillation of
the sentence representation model Sentence-BERT by augmenting the pre-trained
distilled model with a projection layer additionally learned on the Maximum
Coding Rate Reduction (MCR2)objective, a novel approach developed for
general-purpose manifold clustering. We demonstrate that the new language model
with reduced complexity and sentence embedding size can achieve comparable
results on semantic retrieval benchmarks.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 09:23:43 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Ševerdija",
"Domagoj",
""
],
[
"Prusina",
"Tomislav",
""
],
[
"Jovanović",
"Antonio",
""
],
[
"Borozan",
"Luka",
""
],
[
"Maltar",
"Jurica",
""
],
[
"Matijević",
"Domagoj",
""
]
] | TITLE: Compressing Sentence Representation with maximum Coding Rate Reduction
ABSTRACT: In most natural language inference problems, sentence representation is
needed for semantic retrieval tasks. In recent years, pre-trained large
language models have been quite effective for computing such representations.
These models produce high-dimensional sentence embeddings. An evident
performance gap between large and small models exists in practice. Hence, due
to space and time hardware limitations, there is a need to attain comparable
results when using the smaller model, which is usually a distilled version of
the large language model. In this paper, we assess the model distillation of
the sentence representation model Sentence-BERT by augmenting the pre-trained
distilled model with a projection layer additionally learned on the Maximum
Coding Rate Reduction (MCR2)objective, a novel approach developed for
general-purpose manifold clustering. We demonstrate that the new language model
with reduced complexity and sentence embedding size can achieve comparable
results on semantic retrieval benchmarks.
|
2304.12695 | Mohammadsina Beyraghi Beyraghi | Naser Omrani, Fardin Ghorbani, Sina Beyraghi, Homayoon Oraizi, Hossein
Soleimani | Deep Learning Framework for the Design of Orbital Angular Momentum
Generators Enabled by Leaky-wave Holograms | null | null | null | null | physics.optics cs.LG physics.pop-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present a novel approach for the design of leaky-wave
holographic antennas that generates OAM-carrying electromagnetic waves by
combining Flat Optics (FO) and machine learning (ML) techniques. To improve the
performance of our system, we use a machine learning technique to discover a
mathematical function that can effectively control the entire radiation
pattern, i.e., decrease the side lobe level (SLL) while simultaneously
increasing the central null depth of the radiation pattern. Precise tuning of
the parameters of the impedance equation based on holographic theory is
necessary to achieve optimal results in a variety of scenarios. In this
research, we applied machine learning to determine the approximate values of
the parameters. We can determine the optimal values for each parameter,
resulting in the desired radiation pattern, using a total of 77,000 generated
datasets. Furthermore, the use of ML not only saves time, but also yields more
precise and accurate results than manual parameter tuning and conventional
optimization methods.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 10:01:04 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Omrani",
"Naser",
""
],
[
"Ghorbani",
"Fardin",
""
],
[
"Beyraghi",
"Sina",
""
],
[
"Oraizi",
"Homayoon",
""
],
[
"Soleimani",
"Hossein",
""
]
] | TITLE: Deep Learning Framework for the Design of Orbital Angular Momentum
Generators Enabled by Leaky-wave Holograms
ABSTRACT: In this paper, we present a novel approach for the design of leaky-wave
holographic antennas that generates OAM-carrying electromagnetic waves by
combining Flat Optics (FO) and machine learning (ML) techniques. To improve the
performance of our system, we use a machine learning technique to discover a
mathematical function that can effectively control the entire radiation
pattern, i.e., decrease the side lobe level (SLL) while simultaneously
increasing the central null depth of the radiation pattern. Precise tuning of
the parameters of the impedance equation based on holographic theory is
necessary to achieve optimal results in a variety of scenarios. In this
research, we applied machine learning to determine the approximate values of
the parameters. We can determine the optimal values for each parameter,
resulting in the desired radiation pattern, using a total of 77,000 generated
datasets. Furthermore, the use of ML not only saves time, but also yields more
precise and accurate results than manual parameter tuning and conventional
optimization methods.
|
2304.12702 | Carlos Lassance | Carlos Lassance, Simon Lupart, Herv\'e Dejean, St\'ephane Clinchant,
Nicola Tonellotto | A Static Pruning Study on Sparse Neural Retrievers | Short accepted at SIGIR 2023. Has extra content in the Appendix that
is not present in the ACM version | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sparse neural retrievers, such as DeepImpact, uniCOIL and SPLADE, have been
introduced recently as an efficient and effective way to perform retrieval with
inverted indexes. They aim to learn term importance and, in some cases,
document expansions, to provide a more effective document ranking compared to
traditional bag-of-words retrieval models such as BM25. However, these sparse
neural retrievers have been shown to increase the computational costs and
latency of query processing compared to their classical counterparts. To
mitigate this, we apply a well-known family of techniques for boosting the
efficiency of query processing over inverted indexes: static pruning. We
experiment with three static pruning strategies, namely document-centric,
term-centric and agnostic pruning, and we assess, over diverse datasets, that
these techniques still work with sparse neural retrievers. In particular,
static pruning achieves $2\times$ speedup with negligible effectiveness loss
($\leq 2\%$ drop) and, depending on the use case, even $4\times$ speedup with
minimal impact on the effectiveness ($\leq 8\%$ drop). Moreover, we show that
neural rerankers are robust to candidates from statically pruned indexes.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 10:15:58 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Lassance",
"Carlos",
""
],
[
"Lupart",
"Simon",
""
],
[
"Dejean",
"Hervé",
""
],
[
"Clinchant",
"Stéphane",
""
],
[
"Tonellotto",
"Nicola",
""
]
] | TITLE: A Static Pruning Study on Sparse Neural Retrievers
ABSTRACT: Sparse neural retrievers, such as DeepImpact, uniCOIL and SPLADE, have been
introduced recently as an efficient and effective way to perform retrieval with
inverted indexes. They aim to learn term importance and, in some cases,
document expansions, to provide a more effective document ranking compared to
traditional bag-of-words retrieval models such as BM25. However, these sparse
neural retrievers have been shown to increase the computational costs and
latency of query processing compared to their classical counterparts. To
mitigate this, we apply a well-known family of techniques for boosting the
efficiency of query processing over inverted indexes: static pruning. We
experiment with three static pruning strategies, namely document-centric,
term-centric and agnostic pruning, and we assess, over diverse datasets, that
these techniques still work with sparse neural retrievers. In particular,
static pruning achieves $2\times$ speedup with negligible effectiveness loss
($\leq 2\%$ drop) and, depending on the use case, even $4\times$ speedup with
minimal impact on the effectiveness ($\leq 8\%$ drop). Moreover, we show that
neural rerankers are robust to candidates from statically pruned indexes.
|
2304.12704 | Haolin Zhuang | Haolin Zhuang, Shun Lei, Long Xiao, Weiqin Li, Liyang Chen, Sicheng
Yang, Zhiyong Wu, Shiyin Kang, Helen Meng | GTN-Bailando: Genre Consistent Long-Term 3D Dance Generation based on
Pre-trained Genre Token Network | Accepted by ICASSP2023.Demo page:
https://im1eon.github.io/ICASSP23-GTNB-DG/ | null | null | null | cs.SD cs.MM eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Music-driven 3D dance generation has become an intensive research topic in
recent years with great potential for real-world applications. Most existing
methods lack the consideration of genre, which results in genre inconsistency
in the generated dance movements. In addition, the correlation between the
dance genre and the music has not been investigated. To address these issues,
we propose a genre-consistent dance generation framework, GTN-Bailando. First,
we propose the Genre Token Network (GTN), which infers the genre from music to
enhance the genre consistency of long-term dance generation. Second, to improve
the generalization capability of the model, the strategy of pre-training and
fine-tuning is adopted.Experimental results on the AIST++ dataset show that the
proposed dance generation framework outperforms state-of-the-art methods in
terms of motion quality and genre consistency.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 10:17:29 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Zhuang",
"Haolin",
""
],
[
"Lei",
"Shun",
""
],
[
"Xiao",
"Long",
""
],
[
"Li",
"Weiqin",
""
],
[
"Chen",
"Liyang",
""
],
[
"Yang",
"Sicheng",
""
],
[
"Wu",
"Zhiyong",
""
],
[
"Kang",
"Shiyin",
""
],
[
"Meng",
"Helen",
""
]
] | TITLE: GTN-Bailando: Genre Consistent Long-Term 3D Dance Generation based on
Pre-trained Genre Token Network
ABSTRACT: Music-driven 3D dance generation has become an intensive research topic in
recent years with great potential for real-world applications. Most existing
methods lack the consideration of genre, which results in genre inconsistency
in the generated dance movements. In addition, the correlation between the
dance genre and the music has not been investigated. To address these issues,
we propose a genre-consistent dance generation framework, GTN-Bailando. First,
we propose the Genre Token Network (GTN), which infers the genre from music to
enhance the genre consistency of long-term dance generation. Second, to improve
the generalization capability of the model, the strategy of pre-training and
fine-tuning is adopted.Experimental results on the AIST++ dataset show that the
proposed dance generation framework outperforms state-of-the-art methods in
terms of motion quality and genre consistency.
|
2304.12719 | Mengdi Gao | Hongyang Jiang, Jingqi Huang, Chen Tang, Xiaoqing Zhang, Mengdi Gao,
Jiang Liu | Eye tracking guided deep multiple instance learning with dual
cross-attention for fundus disease detection | 10 pages, 9 figures | null | null | null | eess.IV cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks (DNNs) have promoted the development of computer aided
diagnosis (CAD) systems for fundus diseases, helping ophthalmologists reduce
missed diagnosis and misdiagnosis rate. However, the majority of CAD systems
are data-driven but lack of medical prior knowledge which can be
performance-friendly. In this regard, we innovatively proposed a
human-in-the-loop (HITL) CAD system by leveraging ophthalmologists'
eye-tracking information, which is more efficient and accurate. Concretely, the
HITL CAD system was implemented on the multiple instance learning (MIL), where
eye-tracking gaze maps were beneficial to cherry-pick diagnosis-related
instances. Furthermore, the dual-cross-attention MIL (DCAMIL) network was
utilized to curb the adverse effects of noisy instances. Meanwhile, both
sequence augmentation module and domain adversarial module were introduced to
enrich and standardize instances in the training bag, respectively, thereby
enhancing the robustness of our method. We conduct comparative experiments on
our newly constructed datasets (namely, AMD-Gaze and DR-Gaze), respectively for
the AMD and early DR detection. Rigorous experiments demonstrate the
feasibility of our HITL CAD system and the superiority of the proposed DCAMIL,
fully exploring the ophthalmologists' eye-tracking information. These
investigations indicate that physicians' gaze maps, as medical prior knowledge,
is potential to contribute to the CAD systems of clinical diseases.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 11:06:43 GMT"
}
] | 2023-04-26T00:00:00 | [
[
"Jiang",
"Hongyang",
""
],
[
"Huang",
"Jingqi",
""
],
[
"Tang",
"Chen",
""
],
[
"Zhang",
"Xiaoqing",
""
],
[
"Gao",
"Mengdi",
""
],
[
"Liu",
"Jiang",
""
]
] | TITLE: Eye tracking guided deep multiple instance learning with dual
cross-attention for fundus disease detection
ABSTRACT: Deep neural networks (DNNs) have promoted the development of computer aided
diagnosis (CAD) systems for fundus diseases, helping ophthalmologists reduce
missed diagnosis and misdiagnosis rate. However, the majority of CAD systems
are data-driven but lack of medical prior knowledge which can be
performance-friendly. In this regard, we innovatively proposed a
human-in-the-loop (HITL) CAD system by leveraging ophthalmologists'
eye-tracking information, which is more efficient and accurate. Concretely, the
HITL CAD system was implemented on the multiple instance learning (MIL), where
eye-tracking gaze maps were beneficial to cherry-pick diagnosis-related
instances. Furthermore, the dual-cross-attention MIL (DCAMIL) network was
utilized to curb the adverse effects of noisy instances. Meanwhile, both
sequence augmentation module and domain adversarial module were introduced to
enrich and standardize instances in the training bag, respectively, thereby
enhancing the robustness of our method. We conduct comparative experiments on
our newly constructed datasets (namely, AMD-Gaze and DR-Gaze), respectively for
the AMD and early DR detection. Rigorous experiments demonstrate the
feasibility of our HITL CAD system and the superiority of the proposed DCAMIL,
fully exploring the ophthalmologists' eye-tracking information. These
investigations indicate that physicians' gaze maps, as medical prior knowledge,
is potential to contribute to the CAD systems of clinical diseases.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.