id
large_stringlengths 9
16
| submitter
large_stringlengths 3
64
⌀ | authors
large_stringlengths 5
20.2k
| title
large_stringlengths 5
245
| comments
large_stringlengths 1
482
⌀ | journal-ref
large_stringlengths 4
382
⌀ | doi
large_stringlengths 9
151
⌀ | report-no
large_stringlengths 2
509
⌀ | categories
large_stringlengths 5
78
| license
large_stringclasses 9
values | abstract
large_stringlengths 83
3.55k
| versions
large listlengths 1
21
| update_date
timestamp[ms]date 2007-05-23 00:00:00
2025-05-30 00:00:00
| authors_parsed
large listlengths 1
1.35k
| prompt
large_stringlengths 166
3.64k
| label
stringclasses 2
values | prob
float64 0.5
0.73
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2211.08691 | Neehar Peri | Neehar Peri, Achal Dave, Deva Ramanan, Shu Kong | Towards Long-Tailed 3D Detection | This work has been accepted to the Conference on Robot Learning
(CoRL) 2022 | null | null | null | cs.CV cs.RO | http://creativecommons.org/licenses/by/4.0/ | Contemporary autonomous vehicle (AV) benchmarks have advanced techniques for
training 3D detectors, particularly on large-scale lidar data. Surprisingly,
although semantic class labels naturally follow a long-tailed distribution,
contemporary benchmarks focus on only a few common classes (e.g., pedestrian
and car) and neglect many rare classes in-the-tail (e.g., debris and stroller).
However, AVs must still detect rare classes to ensure safe operation. Moreover,
semantic classes are often organized within a hierarchy, e.g., tail classes
such as child and construction-worker are arguably subclasses of pedestrian.
However, such hierarchical relationships are often ignored, which may lead to
misleading estimates of performance and missed opportunities for algorithmic
innovation. We address these challenges by formally studying the problem of
Long-Tailed 3D Detection (LT3D), which evaluates on all classes, including
those in-the-tail. We evaluate and innovate upon popular 3D detection
codebases, such as CenterPoint and PointPillars, adapting them for LT3D. We
develop hierarchical losses that promote feature sharing across common-vs-rare
classes, as well as improved detection metrics that award partial credit to
"reasonable" mistakes respecting the hierarchy (e.g., mistaking a child for an
adult). Finally, we point out that fine-grained tail class accuracy is
particularly improved via multimodal fusion of RGB images with LiDAR; simply
put, small fine-grained classes are challenging to identify from sparse (lidar)
geometry alone, suggesting that multimodal cues are crucial to long-tailed 3D
detection. Our modifications improve accuracy by 5% AP on average for all
classes, and dramatically improve AP for rare classes (e.g., stroller AP
improves from 3.6 to 31.6)! Our code is available at
https://github.com/neeharperi/LT3D
| [
{
"version": "v1",
"created": "Wed, 16 Nov 2022 06:00:47 GMT"
},
{
"version": "v2",
"created": "Fri, 19 May 2023 20:19:23 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Peri",
"Neehar",
""
],
[
"Dave",
"Achal",
""
],
[
"Ramanan",
"Deva",
""
],
[
"Kong",
"Shu",
""
]
] | TITLE: Towards Long-Tailed 3D Detection
ABSTRACT: Contemporary autonomous vehicle (AV) benchmarks have advanced techniques for
training 3D detectors, particularly on large-scale lidar data. Surprisingly,
although semantic class labels naturally follow a long-tailed distribution,
contemporary benchmarks focus on only a few common classes (e.g., pedestrian
and car) and neglect many rare classes in-the-tail (e.g., debris and stroller).
However, AVs must still detect rare classes to ensure safe operation. Moreover,
semantic classes are often organized within a hierarchy, e.g., tail classes
such as child and construction-worker are arguably subclasses of pedestrian.
However, such hierarchical relationships are often ignored, which may lead to
misleading estimates of performance and missed opportunities for algorithmic
innovation. We address these challenges by formally studying the problem of
Long-Tailed 3D Detection (LT3D), which evaluates on all classes, including
those in-the-tail. We evaluate and innovate upon popular 3D detection
codebases, such as CenterPoint and PointPillars, adapting them for LT3D. We
develop hierarchical losses that promote feature sharing across common-vs-rare
classes, as well as improved detection metrics that award partial credit to
"reasonable" mistakes respecting the hierarchy (e.g., mistaking a child for an
adult). Finally, we point out that fine-grained tail class accuracy is
particularly improved via multimodal fusion of RGB images with LiDAR; simply
put, small fine-grained classes are challenging to identify from sparse (lidar)
geometry alone, suggesting that multimodal cues are crucial to long-tailed 3D
detection. Our modifications improve accuracy by 5% AP on average for all
classes, and dramatically improve AP for rare classes (e.g., stroller AP
improves from 3.6 to 31.6)! Our code is available at
https://github.com/neeharperi/LT3D
| no_new_dataset | 0.710622 |
2211.10665 | Joel Kuepper | Joel Kuepper, Andres Erbsen, Jason Gross, Owen Conoly, Chuyue Sun,
Samuel Tian, David Wu, Adam Chlipala, Chitchanok Chuengsatiansup, Daniel
Genkin, Markus Wagner, Yuval Yarom | CryptOpt: Verified Compilation with Randomized Program Search for
Cryptographic Primitives (full version) | null | null | null | null | cs.CR cs.NE cs.PL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Most software domains rely on compilers to translate high-level code to
multiple different machine languages, with performance not too much worse than
what developers would have the patience to write directly in assembly language.
However, cryptography has been an exception, where many performance-critical
routines have been written directly in assembly (sometimes through
metaprogramming layers). Some past work has shown how to do formal verification
of that assembly, and other work has shown how to generate C code automatically
along with formal proof, but with consequent performance penalties vs. the
best-known assembly. We present CryptOpt, the first compilation pipeline that
specializes high-level cryptographic functional programs into assembly code
significantly faster than what GCC or Clang produce, with mechanized proof (in
Coq) whose final theorem statement mentions little beyond the input functional
program and the operational semantics of x86-64 assembly. On the optimization
side, we apply randomized search through the space of assembly programs, with
repeated automatic benchmarking on target CPUs. On the formal-verification
side, we connect to the Fiat Cryptography framework (which translates
functional programs into C-like IR code) and extend it with a new formally
verified program-equivalence checker, incorporating a modest subset of known
features of SMT solvers and symbolic-execution engines. The overall prototype
is quite practical, e.g. producing new fastest-known implementations of
finite-field arithmetic for both Curve25519 (part of the TLS standard) and the
Bitcoin elliptic curve secp256k1 for the Intel $12^{th}$ and $13^{th}$
generations.
| [
{
"version": "v1",
"created": "Sat, 19 Nov 2022 11:07:39 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Apr 2023 05:57:42 GMT"
},
{
"version": "v3",
"created": "Mon, 22 May 2023 00:53:32 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Kuepper",
"Joel",
""
],
[
"Erbsen",
"Andres",
""
],
[
"Gross",
"Jason",
""
],
[
"Conoly",
"Owen",
""
],
[
"Sun",
"Chuyue",
""
],
[
"Tian",
"Samuel",
""
],
[
"Wu",
"David",
""
],
[
"Chlipala",
"Adam",
""
],
[
"Chuengsatiansup",
"Chitchanok",
""
],
[
"Genkin",
"Daniel",
""
],
[
"Wagner",
"Markus",
""
],
[
"Yarom",
"Yuval",
""
]
] | TITLE: CryptOpt: Verified Compilation with Randomized Program Search for
Cryptographic Primitives (full version)
ABSTRACT: Most software domains rely on compilers to translate high-level code to
multiple different machine languages, with performance not too much worse than
what developers would have the patience to write directly in assembly language.
However, cryptography has been an exception, where many performance-critical
routines have been written directly in assembly (sometimes through
metaprogramming layers). Some past work has shown how to do formal verification
of that assembly, and other work has shown how to generate C code automatically
along with formal proof, but with consequent performance penalties vs. the
best-known assembly. We present CryptOpt, the first compilation pipeline that
specializes high-level cryptographic functional programs into assembly code
significantly faster than what GCC or Clang produce, with mechanized proof (in
Coq) whose final theorem statement mentions little beyond the input functional
program and the operational semantics of x86-64 assembly. On the optimization
side, we apply randomized search through the space of assembly programs, with
repeated automatic benchmarking on target CPUs. On the formal-verification
side, we connect to the Fiat Cryptography framework (which translates
functional programs into C-like IR code) and extend it with a new formally
verified program-equivalence checker, incorporating a modest subset of known
features of SMT solvers and symbolic-execution engines. The overall prototype
is quite practical, e.g. producing new fastest-known implementations of
finite-field arithmetic for both Curve25519 (part of the TLS standard) and the
Bitcoin elliptic curve secp256k1 for the Intel $12^{th}$ and $13^{th}$
generations.
| no_new_dataset | 0.706577 |
2211.11534 | Yuni Lai | Yuni Lai, Yulin Zhu, Wenqi Fan, Xiaoge Zhang, Kai Zhou | Towards Adversarially Robust Recommendation from Adaptive Fraudster
Detection | null | null | null | null | cs.IR cs.AI cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The robustness of recommender systems under node injection attacks has
garnered significant attention. Recently, GraphRfi, a GNN-based recommender
system, was proposed and shown to effectively mitigate the impact of injected
fake users. However, we demonstrate that GraphRfi remains vulnerable to attacks
due to the supervised nature of its fraudster detection component, where
obtaining clean labels is challenging in practice. In particular, we propose a
powerful poisoning attack, MetaC, against both GNN-based and MF-based
recommender systems. Furthermore, we analyze why GraphRfi fails under such an
attack. Then, based on our insights obtained from vulnerability analysis, we
design an adaptive fraudster detection module that explicitly considers label
uncertainty. This module can serve as a plug-in for different recommender
systems, resulting in a robust framework named PDR. Comprehensive experiments
show that our defense approach outperforms other benchmark methods under
attacks. Overall, our research presents an effective framework for integrating
fraudster detection into recommendation systems to achieve adversarial
robustness.
| [
{
"version": "v1",
"created": "Tue, 8 Nov 2022 15:38:43 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Nov 2022 06:58:14 GMT"
},
{
"version": "v3",
"created": "Sat, 20 May 2023 08:30:21 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Lai",
"Yuni",
""
],
[
"Zhu",
"Yulin",
""
],
[
"Fan",
"Wenqi",
""
],
[
"Zhang",
"Xiaoge",
""
],
[
"Zhou",
"Kai",
""
]
] | TITLE: Towards Adversarially Robust Recommendation from Adaptive Fraudster
Detection
ABSTRACT: The robustness of recommender systems under node injection attacks has
garnered significant attention. Recently, GraphRfi, a GNN-based recommender
system, was proposed and shown to effectively mitigate the impact of injected
fake users. However, we demonstrate that GraphRfi remains vulnerable to attacks
due to the supervised nature of its fraudster detection component, where
obtaining clean labels is challenging in practice. In particular, we propose a
powerful poisoning attack, MetaC, against both GNN-based and MF-based
recommender systems. Furthermore, we analyze why GraphRfi fails under such an
attack. Then, based on our insights obtained from vulnerability analysis, we
design an adaptive fraudster detection module that explicitly considers label
uncertainty. This module can serve as a plug-in for different recommender
systems, resulting in a robust framework named PDR. Comprehensive experiments
show that our defense approach outperforms other benchmark methods under
attacks. Overall, our research presents an effective framework for integrating
fraudster detection into recommendation systems to achieve adversarial
robustness.
| no_new_dataset | 0.705151 |
2211.16675 | Xuhang Chen | Xuhang Chen, Xiaodong Cun, Chi-Man Pun, Shuqiang Wang | ShaDocNet: Learning Spatial-Aware Tokens in Transformer for Document
Shadow Removal | null | null | 10.1109/ICASSP49357.2023.10095403 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Shadow removal improves the visual quality and legibility of digital copies
of documents. However, document shadow removal remains an unresolved subject.
Traditional techniques rely on heuristics that vary from situation to
situation. Given the quality and quantity of current public datasets, the
majority of neural network models are ill-equipped for this task. In this
paper, we propose a Transformer-based model for document shadow removal that
utilizes shadow context encoding and decoding in both shadow and shadow-free
regions. Additionally, shadow detection and pixel-level enhancement are
included in the whole coarse-to-fine process. On the basis of comprehensive
benchmark evaluations, it is competitive with state-of-the-art methods.
| [
{
"version": "v1",
"created": "Wed, 30 Nov 2022 01:46:29 GMT"
},
{
"version": "v2",
"created": "Wed, 22 Feb 2023 01:29:11 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chen",
"Xuhang",
""
],
[
"Cun",
"Xiaodong",
""
],
[
"Pun",
"Chi-Man",
""
],
[
"Wang",
"Shuqiang",
""
]
] | TITLE: ShaDocNet: Learning Spatial-Aware Tokens in Transformer for Document
Shadow Removal
ABSTRACT: Shadow removal improves the visual quality and legibility of digital copies
of documents. However, document shadow removal remains an unresolved subject.
Traditional techniques rely on heuristics that vary from situation to
situation. Given the quality and quantity of current public datasets, the
majority of neural network models are ill-equipped for this task. In this
paper, we propose a Transformer-based model for document shadow removal that
utilizes shadow context encoding and decoding in both shadow and shadow-free
regions. Additionally, shadow detection and pixel-level enhancement are
included in the whole coarse-to-fine process. On the basis of comprehensive
benchmark evaluations, it is competitive with state-of-the-art methods.
| no_new_dataset | 0.710879 |
2211.17196 | Dongji Gao | Dongji Gao and Jiatong Shi and Shun-Po Chuang and Leibny Paola Garcia
and Hung-yi Lee and Shinji Watanabe and Sanjeev Khudanpur | EURO: ESPnet Unsupervised ASR Open-source Toolkit | null | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper describes the ESPnet Unsupervised ASR Open-source Toolkit (EURO),
an end-to-end open-source toolkit for unsupervised automatic speech recognition
(UASR). EURO adopts the state-of-the-art UASR learning method introduced by the
Wav2vec-U, originally implemented at FAIRSEQ, which leverages self-supervised
speech representations and adversarial training. In addition to wav2vec2, EURO
extends the functionality and promotes reproducibility for UASR tasks by
integrating S3PRL and k2, resulting in flexible frontends from 27
self-supervised models and various graph-based decoding strategies. EURO is
implemented in ESPnet and follows its unified pipeline to provide UASR recipes
with a complete setup. This improves the pipeline's efficiency and allows EURO
to be easily applied to existing datasets in ESPnet. Extensive experiments on
three mainstream self-supervised models demonstrate the toolkit's effectiveness
and achieve state-of-the-art UASR performance on TIMIT and LibriSpeech
datasets. EURO will be publicly available at https://github.com/espnet/espnet,
aiming to promote this exciting and emerging research area based on UASR
through open-source activity.
| [
{
"version": "v1",
"created": "Wed, 30 Nov 2022 17:40:54 GMT"
},
{
"version": "v2",
"created": "Thu, 1 Dec 2022 14:56:38 GMT"
},
{
"version": "v3",
"created": "Sun, 21 May 2023 00:56:05 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Gao",
"Dongji",
""
],
[
"Shi",
"Jiatong",
""
],
[
"Chuang",
"Shun-Po",
""
],
[
"Garcia",
"Leibny Paola",
""
],
[
"Lee",
"Hung-yi",
""
],
[
"Watanabe",
"Shinji",
""
],
[
"Khudanpur",
"Sanjeev",
""
]
] | TITLE: EURO: ESPnet Unsupervised ASR Open-source Toolkit
ABSTRACT: This paper describes the ESPnet Unsupervised ASR Open-source Toolkit (EURO),
an end-to-end open-source toolkit for unsupervised automatic speech recognition
(UASR). EURO adopts the state-of-the-art UASR learning method introduced by the
Wav2vec-U, originally implemented at FAIRSEQ, which leverages self-supervised
speech representations and adversarial training. In addition to wav2vec2, EURO
extends the functionality and promotes reproducibility for UASR tasks by
integrating S3PRL and k2, resulting in flexible frontends from 27
self-supervised models and various graph-based decoding strategies. EURO is
implemented in ESPnet and follows its unified pipeline to provide UASR recipes
with a complete setup. This improves the pipeline's efficiency and allows EURO
to be easily applied to existing datasets in ESPnet. Extensive experiments on
three mainstream self-supervised models demonstrate the toolkit's effectiveness
and achieve state-of-the-art UASR performance on TIMIT and LibriSpeech
datasets. EURO will be publicly available at https://github.com/espnet/espnet,
aiming to promote this exciting and emerging research area based on UASR
through open-source activity.
| no_new_dataset | 0.710069 |
2212.00506 | Alberto Pozanco | Alberto Pozanco, Daniel Borrajo | Fairness in Multi-Agent Planning | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In cooperative Multi-Agent Planning (MAP), a set of goals has to be achieved
by a set of agents. Independently of whether they perform a pre-assignment of
goals to agents or they directly search for a solution without any goal
assignment, most previous works did not focus on a fair
distribution/achievement of goals by agents. This paper adapts well-known
fairness schemes to MAP, and introduces two novel approaches to generate
cost-aware fair plans. The first one solves an optimization problem to
pre-assign goals to agents, and then solves a centralized MAP task using that
assignment. The second one consists of a planning-based compilation that allows
solving the joint problem of goal assignment and planning while taking into
account the given fairness scheme. Empirical results in several standard MAP
benchmarks show that these approaches outperform different baselines. They also
show that there is no need to sacrifice much plan cost to generate fair plans.
| [
{
"version": "v1",
"created": "Thu, 1 Dec 2022 13:58:46 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 10:55:25 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Pozanco",
"Alberto",
""
],
[
"Borrajo",
"Daniel",
""
]
] | TITLE: Fairness in Multi-Agent Planning
ABSTRACT: In cooperative Multi-Agent Planning (MAP), a set of goals has to be achieved
by a set of agents. Independently of whether they perform a pre-assignment of
goals to agents or they directly search for a solution without any goal
assignment, most previous works did not focus on a fair
distribution/achievement of goals by agents. This paper adapts well-known
fairness schemes to MAP, and introduces two novel approaches to generate
cost-aware fair plans. The first one solves an optimization problem to
pre-assign goals to agents, and then solves a centralized MAP task using that
assignment. The second one consists of a planning-based compilation that allows
solving the joint problem of goal assignment and planning while taking into
account the given fairness scheme. Empirical results in several standard MAP
benchmarks show that these approaches outperform different baselines. They also
show that there is no need to sacrifice much plan cost to generate fair plans.
| no_new_dataset | 0.710226 |
2212.02762 | Zonghai Yao | Zonghai Yao and Jack Tsai and Weisong Liu and David A. Levy and Emily
Druhl and Joel I Reisman and Hong Yu | Automated Identification of Eviction Status from Electronic Health
Record Notes | This article has been accepted for publication in Journal of the
American Medical Informatics Association Published by Oxford University
Press. https://doi.org/10.1093/jamia/ocad081 | Journal of the American Medical Informatics Association, ocad081,
2023 | 10.1093/jamia/ocad081 | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Objective: Evictions are important social and behavioral determinants of
health. Evictions are associated with a cascade of negative events that can
lead to unemployment, housing insecurity/homelessness, long-term poverty, and
mental health problems. In this study, we developed a natural language
processing system to automatically detect eviction status from electronic
health record (EHR) notes.
Materials and Methods: We first defined eviction status (eviction presence
and eviction period) and then annotated eviction status in 5000 EHR notes from
the Veterans Health Administration (VHA). We developed a novel model, KIRESH,
that has shown to substantially outperform other state-of-the-art models such
as fine-tuning pre-trained language models like BioBERT and BioClinicalBERT.
Moreover, we designed a novel prompt to further improve the model performance
by using the intrinsic connection between the two sub-tasks of eviction
presence and period prediction. Finally, we used the Temperature Scaling-based
Calibration on our KIRESH-Prompt method to avoid over-confidence issues arising
from the imbalance dataset.
Results: KIRESH-Prompt substantially outperformed strong baseline models
including fine-tuning the BioClinicalBERT model to achieve 0.74672 MCC, 0.71153
Macro-F1, and 0.83396 Micro-F1 in predicting eviction period and 0.66827 MCC,
0.62734 Macro-F1, and 0.7863 Micro-F1 in predicting eviction presence. We also
conducted additional experiments on a benchmark social determinants of health
(SBDH) dataset to demonstrate the generalizability of our methods.
Conclusion and Future Work: KIRESH-Prompt has substantially improved eviction
status classification. We plan to deploy KIRESH-Prompt to the VHA EHRs as an
eviction surveillance system to help address the US Veterans' housing
insecurity.
| [
{
"version": "v1",
"created": "Tue, 6 Dec 2022 05:25:32 GMT"
},
{
"version": "v2",
"created": "Fri, 7 Apr 2023 15:17:30 GMT"
},
{
"version": "v3",
"created": "Sat, 20 May 2023 05:03:57 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Yao",
"Zonghai",
""
],
[
"Tsai",
"Jack",
""
],
[
"Liu",
"Weisong",
""
],
[
"Levy",
"David A.",
""
],
[
"Druhl",
"Emily",
""
],
[
"Reisman",
"Joel I",
""
],
[
"Yu",
"Hong",
""
]
] | TITLE: Automated Identification of Eviction Status from Electronic Health
Record Notes
ABSTRACT: Objective: Evictions are important social and behavioral determinants of
health. Evictions are associated with a cascade of negative events that can
lead to unemployment, housing insecurity/homelessness, long-term poverty, and
mental health problems. In this study, we developed a natural language
processing system to automatically detect eviction status from electronic
health record (EHR) notes.
Materials and Methods: We first defined eviction status (eviction presence
and eviction period) and then annotated eviction status in 5000 EHR notes from
the Veterans Health Administration (VHA). We developed a novel model, KIRESH,
that has shown to substantially outperform other state-of-the-art models such
as fine-tuning pre-trained language models like BioBERT and BioClinicalBERT.
Moreover, we designed a novel prompt to further improve the model performance
by using the intrinsic connection between the two sub-tasks of eviction
presence and period prediction. Finally, we used the Temperature Scaling-based
Calibration on our KIRESH-Prompt method to avoid over-confidence issues arising
from the imbalance dataset.
Results: KIRESH-Prompt substantially outperformed strong baseline models
including fine-tuning the BioClinicalBERT model to achieve 0.74672 MCC, 0.71153
Macro-F1, and 0.83396 Micro-F1 in predicting eviction period and 0.66827 MCC,
0.62734 Macro-F1, and 0.7863 Micro-F1 in predicting eviction presence. We also
conducted additional experiments on a benchmark social determinants of health
(SBDH) dataset to demonstrate the generalizability of our methods.
Conclusion and Future Work: KIRESH-Prompt has substantially improved eviction
status classification. We plan to deploy KIRESH-Prompt to the VHA EHRs as an
eviction surveillance system to help address the US Veterans' housing
insecurity.
| no_new_dataset | 0.672385 |
2212.05422 | Shitong Shao | Shitong Shao and Huanran Chen and Zhen Huang and Linrui Gong and Shuai
Wang and Xinxiao Wu | Teaching What You Should Teach: A Data-Based Distillation Method | 13 pages, 4 figures Accepted by IJCAI2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In real teaching scenarios, an excellent teacher always teaches what he (or
she) is good at but the student is not. This gives the student the best
assistance in making up for his (or her) weaknesses and becoming a good one
overall. Enlightened by this, we introduce the "Teaching what you Should Teach"
strategy into a knowledge distillation framework, and propose a data-based
distillation method named "TST" that searches for desirable augmented samples
to assist in distilling more efficiently and rationally. To be specific, we
design a neural network-based data augmentation module with priori bias, which
assists in finding what meets the teacher's strengths but the student's
weaknesses, by learning magnitudes and probabilities to generate suitable data
samples. By training the data augmentation module and the generalized
distillation paradigm in turn, a student model is learned with excellent
generalization ability. To verify the effectiveness of our method, we conducted
extensive comparative experiments on object recognition, detection, and
segmentation tasks. The results on the CIFAR-10, ImageNet-1k, MS-COCO, and
Cityscapes datasets demonstrate that our method achieves state-of-the-art
performance on almost all teacher-student pairs. Furthermore, we conduct
visualization studies to explore what magnitudes and probabilities are needed
for the distillation process.
| [
{
"version": "v1",
"created": "Sun, 11 Dec 2022 06:22:14 GMT"
},
{
"version": "v2",
"created": "Tue, 27 Dec 2022 12:18:07 GMT"
},
{
"version": "v3",
"created": "Sat, 28 Jan 2023 08:17:02 GMT"
},
{
"version": "v4",
"created": "Wed, 5 Apr 2023 02:32:48 GMT"
},
{
"version": "v5",
"created": "Fri, 21 Apr 2023 12:40:51 GMT"
},
{
"version": "v6",
"created": "Sat, 20 May 2023 11:55:10 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Shao",
"Shitong",
""
],
[
"Chen",
"Huanran",
""
],
[
"Huang",
"Zhen",
""
],
[
"Gong",
"Linrui",
""
],
[
"Wang",
"Shuai",
""
],
[
"Wu",
"Xinxiao",
""
]
] | TITLE: Teaching What You Should Teach: A Data-Based Distillation Method
ABSTRACT: In real teaching scenarios, an excellent teacher always teaches what he (or
she) is good at but the student is not. This gives the student the best
assistance in making up for his (or her) weaknesses and becoming a good one
overall. Enlightened by this, we introduce the "Teaching what you Should Teach"
strategy into a knowledge distillation framework, and propose a data-based
distillation method named "TST" that searches for desirable augmented samples
to assist in distilling more efficiently and rationally. To be specific, we
design a neural network-based data augmentation module with priori bias, which
assists in finding what meets the teacher's strengths but the student's
weaknesses, by learning magnitudes and probabilities to generate suitable data
samples. By training the data augmentation module and the generalized
distillation paradigm in turn, a student model is learned with excellent
generalization ability. To verify the effectiveness of our method, we conducted
extensive comparative experiments on object recognition, detection, and
segmentation tasks. The results on the CIFAR-10, ImageNet-1k, MS-COCO, and
Cityscapes datasets demonstrate that our method achieves state-of-the-art
performance on almost all teacher-student pairs. Furthermore, we conduct
visualization studies to explore what magnitudes and probabilities are needed
for the distillation process.
| no_new_dataset | 0.711249 |
2212.06711 | Qixiang Fang | Qixiang Fang, Anastasia Giachanou, Ayoub Bagheri, Laura Boeschoten,
Erik-Jan van Kesteren, Mahdi Shafiee Kamalabad, Daniel L Oberski | On Text-based Personality Computing: Challenges and Future Directions | Findings of ACL 2023. Long paper | null | null | null | cs.CL cs.CY | http://creativecommons.org/licenses/by/4.0/ | Text-based personality computing (TPC) has gained many research interests in
NLP. In this paper, we describe 15 challenges that we consider deserving the
attention of the research community. These challenges are organized by the
following topics: personality taxonomies, measurement quality, datasets,
performance evaluation, modelling choices, as well as ethics and fairness. When
addressing each challenge, not only do we combine perspectives from both NLP
and social sciences, but also offer concrete suggestions. We hope to inspire
more valid and reliable TPC research.
| [
{
"version": "v1",
"created": "Tue, 13 Dec 2022 16:29:51 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Dec 2022 13:54:19 GMT"
},
{
"version": "v3",
"created": "Thu, 18 May 2023 20:52:50 GMT"
},
{
"version": "v4",
"created": "Mon, 22 May 2023 17:40:14 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Fang",
"Qixiang",
""
],
[
"Giachanou",
"Anastasia",
""
],
[
"Bagheri",
"Ayoub",
""
],
[
"Boeschoten",
"Laura",
""
],
[
"van Kesteren",
"Erik-Jan",
""
],
[
"Kamalabad",
"Mahdi Shafiee",
""
],
[
"Oberski",
"Daniel L",
""
]
] | TITLE: On Text-based Personality Computing: Challenges and Future Directions
ABSTRACT: Text-based personality computing (TPC) has gained many research interests in
NLP. In this paper, we describe 15 challenges that we consider deserving the
attention of the research community. These challenges are organized by the
following topics: personality taxonomies, measurement quality, datasets,
performance evaluation, modelling choices, as well as ethics and fairness. When
addressing each challenge, not only do we combine perspectives from both NLP
and social sciences, but also offer concrete suggestions. We hope to inspire
more valid and reliable TPC research.
| no_new_dataset | 0.711638 |
2212.08775 | David Uthus | David Uthus and Jianmo Ni | RISE: Leveraging Retrieval Techniques for Summarization Evaluation | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Evaluating automatically-generated text summaries is a challenging task.
While there have been many interesting approaches, they still fall short of
human evaluations. We present RISE, a new approach for evaluating summaries by
leveraging techniques from information retrieval. RISE is first trained as a
retrieval task using a dual-encoder retrieval setup, and can then be
subsequently utilized for evaluating a generated summary given an input
document, without gold reference summaries. RISE is especially well suited when
working on new datasets where one may not have reference summaries available
for evaluation. We conduct comprehensive experiments on the SummEval benchmark
(Fabbri et al., 2021) and the results show that RISE has higher correlation
with human evaluations compared to many past approaches to summarization
evaluation. Furthermore, RISE also demonstrates data-efficiency and
generalizability across languages.
| [
{
"version": "v1",
"created": "Sat, 17 Dec 2022 01:09:22 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 16:53:58 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Uthus",
"David",
""
],
[
"Ni",
"Jianmo",
""
]
] | TITLE: RISE: Leveraging Retrieval Techniques for Summarization Evaluation
ABSTRACT: Evaluating automatically-generated text summaries is a challenging task.
While there have been many interesting approaches, they still fall short of
human evaluations. We present RISE, a new approach for evaluating summaries by
leveraging techniques from information retrieval. RISE is first trained as a
retrieval task using a dual-encoder retrieval setup, and can then be
subsequently utilized for evaluating a generated summary given an input
document, without gold reference summaries. RISE is especially well suited when
working on new datasets where one may not have reference summaries available
for evaluation. We conduct comprehensive experiments on the SummEval benchmark
(Fabbri et al., 2021) and the results show that RISE has higher correlation
with human evaluations compared to many past approaches to summarization
evaluation. Furthermore, RISE also demonstrates data-efficiency and
generalizability across languages.
| no_new_dataset | 0.70844 |
2212.09721 | Aaron Chan | Aaron Chan, Zhiyuan Zeng, Wyatt Lake, Brihi Joshi, Hanjie Chen, Xiang
Ren | KNIFE: Distilling Reasoning Knowledge From Free-Text Rationales | TrustML-(un)Limited Workshop at ICLR 2023 | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Language models (LMs) have yielded impressive results on many language
reasoning tasks, but their unexpected errors raise doubts about their reasoning
abilities. In light of this, there is growing interest in finetuning/prompting
LMs with both task instances and their associated free-text rationales (FTRs),
which explain the correct reasoning process for predicting the correct task
output (i.e., how to be "right for the right reasons"). However, existing
finetuning methods fail to improve LM performance, while prompting needs
prohibitively large (i.e., >50B) LMs to work well. We propose KNIFE, which
shows that reasoning knowledge can be effectively distilled from FTRs into a
small (i.e., <1B) LM and improve the LM's performance. First, KNIFE finetunes a
teacher LM (given task input and FTR) to predict the task output, transferring
reasoning knowledge from the FTRs to the teacher's hidden states. Second, KNIFE
finetunes a student LM (given task input only) such that its hidden states are
aligned with the teacher's. Thus, the student is endowed with reasoning
knowledge but can be used for inference without direct FTR input. On two
question-answering datasets, KNIFE outperforms various finetuning and prompting
baselines in fully-supervised and low-resource settings. Also, we observe that
FTR quality is crucial to KNIFE's performance.
| [
{
"version": "v1",
"created": "Mon, 19 Dec 2022 18:49:09 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 00:19:21 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chan",
"Aaron",
""
],
[
"Zeng",
"Zhiyuan",
""
],
[
"Lake",
"Wyatt",
""
],
[
"Joshi",
"Brihi",
""
],
[
"Chen",
"Hanjie",
""
],
[
"Ren",
"Xiang",
""
]
] | TITLE: KNIFE: Distilling Reasoning Knowledge From Free-Text Rationales
ABSTRACT: Language models (LMs) have yielded impressive results on many language
reasoning tasks, but their unexpected errors raise doubts about their reasoning
abilities. In light of this, there is growing interest in finetuning/prompting
LMs with both task instances and their associated free-text rationales (FTRs),
which explain the correct reasoning process for predicting the correct task
output (i.e., how to be "right for the right reasons"). However, existing
finetuning methods fail to improve LM performance, while prompting needs
prohibitively large (i.e., >50B) LMs to work well. We propose KNIFE, which
shows that reasoning knowledge can be effectively distilled from FTRs into a
small (i.e., <1B) LM and improve the LM's performance. First, KNIFE finetunes a
teacher LM (given task input and FTR) to predict the task output, transferring
reasoning knowledge from the FTRs to the teacher's hidden states. Second, KNIFE
finetunes a student LM (given task input only) such that its hidden states are
aligned with the teacher's. Thus, the student is endowed with reasoning
knowledge but can be used for inference without direct FTR input. On two
question-answering datasets, KNIFE outperforms various finetuning and prompting
baselines in fully-supervised and low-resource settings. Also, we observe that
FTR quality is crucial to KNIFE's performance.
| no_new_dataset | 0.70878 |
2212.10557 | Prakhar Gupta | Prakhar Gupta, Yang Liu, Di Jin, Behnam Hedayatnia, Spandana Gella,
Sijia Liu, Patrick Lange, Julia Hirschberg, Dilek Hakkani-Tur | DialGuide: Aligning Dialogue Model Behavior with Developer Guidelines | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Dialogue models are able to generate coherent and fluent responses, but they
can still be challenging to control and may produce non-engaging, unsafe
results. This unpredictability diminishes user trust and can hinder the use of
the models in the real world. To address this, we introduce DialGuide, a novel
framework for controlling dialogue model behavior using natural language rules,
or guidelines. These guidelines provide information about the context they are
applicable to and what should be included in the response, allowing the models
to generate responses that are more closely aligned with the developer's
expectations and intent. We evaluate DialGuide on three tasks in open-domain
dialogue response generation: guideline selection, response generation, and
response entailment verification. Our dataset contains 10,737 positive and
15,467 negative dialogue context-response-guideline triplets across two domains
- chit-chat and safety. We provide baseline models for the tasks and benchmark
their performance. We also demonstrate that DialGuide is effective in the
dialogue safety domain, producing safe and engaging responses that follow
developer guidelines.
| [
{
"version": "v1",
"created": "Tue, 20 Dec 2022 18:57:18 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 19:38:01 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Gupta",
"Prakhar",
""
],
[
"Liu",
"Yang",
""
],
[
"Jin",
"Di",
""
],
[
"Hedayatnia",
"Behnam",
""
],
[
"Gella",
"Spandana",
""
],
[
"Liu",
"Sijia",
""
],
[
"Lange",
"Patrick",
""
],
[
"Hirschberg",
"Julia",
""
],
[
"Hakkani-Tur",
"Dilek",
""
]
] | TITLE: DialGuide: Aligning Dialogue Model Behavior with Developer Guidelines
ABSTRACT: Dialogue models are able to generate coherent and fluent responses, but they
can still be challenging to control and may produce non-engaging, unsafe
results. This unpredictability diminishes user trust and can hinder the use of
the models in the real world. To address this, we introduce DialGuide, a novel
framework for controlling dialogue model behavior using natural language rules,
or guidelines. These guidelines provide information about the context they are
applicable to and what should be included in the response, allowing the models
to generate responses that are more closely aligned with the developer's
expectations and intent. We evaluate DialGuide on three tasks in open-domain
dialogue response generation: guideline selection, response generation, and
response entailment verification. Our dataset contains 10,737 positive and
15,467 negative dialogue context-response-guideline triplets across two domains
- chit-chat and safety. We provide baseline models for the tasks and benchmark
their performance. We also demonstrate that DialGuide is effective in the
dialogue safety domain, producing safe and engaging responses that follow
developer guidelines.
| new_dataset | 0.713276 |
2301.00979 | Fangyu Li | Fangyu Li and Shenbao Yu and Feng Zeng and Fang Yang | Improving Sequential Recommendation Models with an Enhanced Loss
Function | null | null | null | null | cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There has been a growing interest in benchmarking sequential recommendation
models and reproducing/improving existing models. For example, Rendle et al.
improved matrix factorization models by tuning their parameters and
hyperparameters. Petrov and Macdonald developed a more efficient and effective
implementation of BERT4Rec, which resolved inconsistencies in performance
comparison between BERT4Rec and SASRec in previous works. In particular,
BERT4Rec and SASRec share a similar network structure, with the main difference
lying in their training objective/loss function. Therefore, we analyzed the
advantages and disadvantages of commonly used loss functions in sequential
recommendation and proposed an improved loss function that leverages their
strengths. We conduct extensive experiments on two influential open-source
libraries, and the results demonstrate that our improved loss function
significantly enhances the performance of GRU4Rec, SASRec, SR-GNN, and S3Rec
models, improving their benchmarks significantly. Furthermore, the improved
SASRec benchmark outperforms BERT4Rec on the ML-1M and Beauty datasets and
achieves similar results to BERT4Rec on the ML-20M and Steam datasets. We also
reproduce the results of the BERT4Rec model on the Beauty dataset. Finally, we
provide a comprehensive explanation of the effectiveness of our improved loss
function through experiments. Our code is publicly available at
https://github.com/Li-fAngyU/sequential_rec.
| [
{
"version": "v1",
"created": "Tue, 3 Jan 2023 07:18:54 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 02:48:16 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Li",
"Fangyu",
""
],
[
"Yu",
"Shenbao",
""
],
[
"Zeng",
"Feng",
""
],
[
"Yang",
"Fang",
""
]
] | TITLE: Improving Sequential Recommendation Models with an Enhanced Loss
Function
ABSTRACT: There has been a growing interest in benchmarking sequential recommendation
models and reproducing/improving existing models. For example, Rendle et al.
improved matrix factorization models by tuning their parameters and
hyperparameters. Petrov and Macdonald developed a more efficient and effective
implementation of BERT4Rec, which resolved inconsistencies in performance
comparison between BERT4Rec and SASRec in previous works. In particular,
BERT4Rec and SASRec share a similar network structure, with the main difference
lying in their training objective/loss function. Therefore, we analyzed the
advantages and disadvantages of commonly used loss functions in sequential
recommendation and proposed an improved loss function that leverages their
strengths. We conduct extensive experiments on two influential open-source
libraries, and the results demonstrate that our improved loss function
significantly enhances the performance of GRU4Rec, SASRec, SR-GNN, and S3Rec
models, improving their benchmarks significantly. Furthermore, the improved
SASRec benchmark outperforms BERT4Rec on the ML-1M and Beauty datasets and
achieves similar results to BERT4Rec on the ML-20M and Steam datasets. We also
reproduce the results of the BERT4Rec model on the Beauty dataset. Finally, we
provide a comprehensive explanation of the effectiveness of our improved loss
function through experiments. Our code is publicly available at
https://github.com/Li-fAngyU/sequential_rec.
| no_new_dataset | 0.709856 |
2301.01380 | Huiyu Wang | Huiyu Wang, Mitesh Kumar Singh, Lorenzo Torresani | Ego-Only: Egocentric Action Detection without Exocentric Transferring | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present Ego-Only, the first approach that enables state-of-the-art action
detection on egocentric (first-person) videos without any form of exocentric
(third-person) transferring. Despite the content and appearance gap separating
the two domains, large-scale exocentric transferring has been the default
choice for egocentric action detection. This is because prior works found that
egocentric models are difficult to train from scratch and that transferring
from exocentric representations leads to improved accuracy. However, in this
paper, we revisit this common belief. Motivated by the large gap separating the
two domains, we propose a strategy that enables effective training of
egocentric models without exocentric transferring. Our Ego-Only approach is
simple. It trains the video representation with a masked autoencoder finetuned
for temporal segmentation. The learned features are then fed to an
off-the-shelf temporal action localization method to detect actions. We find
that this renders exocentric transferring unnecessary by showing remarkably
strong results achieved by this simple Ego-Only approach on three established
egocentric video datasets: Ego4D, EPIC-Kitchens-100, and Charades-Ego. On both
action detection and action recognition, Ego-Only outperforms previous best
exocentric transferring methods that use orders of magnitude more labels.
Ego-Only sets new state-of-the-art results on these datasets and benchmarks
without exocentric data.
| [
{
"version": "v1",
"created": "Tue, 3 Jan 2023 22:22:34 GMT"
},
{
"version": "v2",
"created": "Fri, 19 May 2023 22:23:48 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Wang",
"Huiyu",
""
],
[
"Singh",
"Mitesh Kumar",
""
],
[
"Torresani",
"Lorenzo",
""
]
] | TITLE: Ego-Only: Egocentric Action Detection without Exocentric Transferring
ABSTRACT: We present Ego-Only, the first approach that enables state-of-the-art action
detection on egocentric (first-person) videos without any form of exocentric
(third-person) transferring. Despite the content and appearance gap separating
the two domains, large-scale exocentric transferring has been the default
choice for egocentric action detection. This is because prior works found that
egocentric models are difficult to train from scratch and that transferring
from exocentric representations leads to improved accuracy. However, in this
paper, we revisit this common belief. Motivated by the large gap separating the
two domains, we propose a strategy that enables effective training of
egocentric models without exocentric transferring. Our Ego-Only approach is
simple. It trains the video representation with a masked autoencoder finetuned
for temporal segmentation. The learned features are then fed to an
off-the-shelf temporal action localization method to detect actions. We find
that this renders exocentric transferring unnecessary by showing remarkably
strong results achieved by this simple Ego-Only approach on three established
egocentric video datasets: Ego4D, EPIC-Kitchens-100, and Charades-Ego. On both
action detection and action recognition, Ego-Only outperforms previous best
exocentric transferring methods that use orders of magnitude more labels.
Ego-Only sets new state-of-the-art results on these datasets and benchmarks
without exocentric data.
| no_new_dataset | 0.71081 |
2301.04829 | Liangqi Yuan | Liangqi Yuan, Lu Su, Ziran Wang | Federated Transfer-Ordered-Personalized Learning for Driver Monitoring
Application | IEEE Internet of Things Journal | null | 10.1109/JIOT.2023.3279273 | null | cs.LG cs.CR | http://creativecommons.org/licenses/by/4.0/ | Federated learning (FL) shines through in the internet of things (IoT) with
its ability to realize collaborative learning and improve learning efficiency
by sharing client model parameters trained on local data. Although FL has been
successfully applied to various domains, including driver monitoring
applications (DMAs) on the internet of vehicles (IoV), its usages still face
some open issues, such as data and system heterogeneity, large-scale
parallelism communication resources, malicious attacks, and data poisoning.
This paper proposes a federated transfer-ordered-personalized learning (FedTOP)
framework to address the above problems and test on two real-world datasets
with and without system heterogeneity. The performance of the three extensions,
transfer, ordered, and personalized, is compared by an ablation study and
achieves 92.32% and 95.96% accuracy on the test clients of two datasets,
respectively. Compared to the baseline, there is a 462% improvement in accuracy
and a 37.46% reduction in communication resource consumption. The results
demonstrate that the proposed FedTOP can be used as a highly accurate,
streamlined, privacy-preserving, cybersecurity-oriented, and personalized
framework for DMA.
| [
{
"version": "v1",
"created": "Thu, 12 Jan 2023 06:12:04 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 05:52:56 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Yuan",
"Liangqi",
""
],
[
"Su",
"Lu",
""
],
[
"Wang",
"Ziran",
""
]
] | TITLE: Federated Transfer-Ordered-Personalized Learning for Driver Monitoring
Application
ABSTRACT: Federated learning (FL) shines through in the internet of things (IoT) with
its ability to realize collaborative learning and improve learning efficiency
by sharing client model parameters trained on local data. Although FL has been
successfully applied to various domains, including driver monitoring
applications (DMAs) on the internet of vehicles (IoV), its usages still face
some open issues, such as data and system heterogeneity, large-scale
parallelism communication resources, malicious attacks, and data poisoning.
This paper proposes a federated transfer-ordered-personalized learning (FedTOP)
framework to address the above problems and test on two real-world datasets
with and without system heterogeneity. The performance of the three extensions,
transfer, ordered, and personalized, is compared by an ablation study and
achieves 92.32% and 95.96% accuracy on the test clients of two datasets,
respectively. Compared to the baseline, there is a 462% improvement in accuracy
and a 37.46% reduction in communication resource consumption. The results
demonstrate that the proposed FedTOP can be used as a highly accurate,
streamlined, privacy-preserving, cybersecurity-oriented, and personalized
framework for DMA.
| no_new_dataset | 0.712445 |
2301.11233 | Haotong Qin | Haotong Qin, Mingyuan Zhang, Yifu Ding, Aoyu Li, Zhongang Cai, Ziwei
Liu, Fisher Yu, Xianglong Liu | BiBench: Benchmarking and Analyzing Network Binarization | null | 2023 International Conference on Machine Learning | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Network binarization emerges as one of the most promising compression
approaches offering extraordinary computation and memory savings by minimizing
the bit-width. However, recent research has shown that applying existing
binarization algorithms to diverse tasks, architectures, and hardware in
realistic scenarios is still not straightforward. Common challenges of
binarization, such as accuracy degradation and efficiency limitation, suggest
that its attributes are not fully understood. To close this gap, we present
BiBench, a rigorously designed benchmark with in-depth analysis for network
binarization. We first carefully scrutinize the requirements of binarization in
the actual production and define evaluation tracks and metrics for a
comprehensive and fair investigation. Then, we evaluate and analyze a series of
milestone binarization algorithms that function at the operator level and with
extensive influence. Our benchmark reveals that 1) the binarized operator has a
crucial impact on the performance and deployability of binarized networks; 2)
the accuracy of binarization varies significantly across different learning
tasks and neural architectures; 3) binarization has demonstrated promising
efficiency potential on edge devices despite the limited hardware support. The
results and analysis also lead to a promising paradigm for accurate and
efficient binarization. We believe that BiBench will contribute to the broader
adoption of binarization and serve as a foundation for future research. The
code for our BiBench is released https://github.com/htqin/BiBench .
| [
{
"version": "v1",
"created": "Thu, 26 Jan 2023 17:17:16 GMT"
},
{
"version": "v2",
"created": "Sat, 20 May 2023 11:04:19 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Qin",
"Haotong",
""
],
[
"Zhang",
"Mingyuan",
""
],
[
"Ding",
"Yifu",
""
],
[
"Li",
"Aoyu",
""
],
[
"Cai",
"Zhongang",
""
],
[
"Liu",
"Ziwei",
""
],
[
"Yu",
"Fisher",
""
],
[
"Liu",
"Xianglong",
""
]
] | TITLE: BiBench: Benchmarking and Analyzing Network Binarization
ABSTRACT: Network binarization emerges as one of the most promising compression
approaches offering extraordinary computation and memory savings by minimizing
the bit-width. However, recent research has shown that applying existing
binarization algorithms to diverse tasks, architectures, and hardware in
realistic scenarios is still not straightforward. Common challenges of
binarization, such as accuracy degradation and efficiency limitation, suggest
that its attributes are not fully understood. To close this gap, we present
BiBench, a rigorously designed benchmark with in-depth analysis for network
binarization. We first carefully scrutinize the requirements of binarization in
the actual production and define evaluation tracks and metrics for a
comprehensive and fair investigation. Then, we evaluate and analyze a series of
milestone binarization algorithms that function at the operator level and with
extensive influence. Our benchmark reveals that 1) the binarized operator has a
crucial impact on the performance and deployability of binarized networks; 2)
the accuracy of binarization varies significantly across different learning
tasks and neural architectures; 3) binarization has demonstrated promising
efficiency potential on edge devices despite the limited hardware support. The
results and analysis also lead to a promising paradigm for accurate and
efficient binarization. We believe that BiBench will contribute to the broader
adoption of binarization and serve as a foundation for future research. The
code for our BiBench is released https://github.com/htqin/BiBench .
| new_dataset | 0.692265 |
2301.12941 | Antonia Tsili | Antonia Tsili, Matthew Dyer, Vladimir Gusev, Piotr Krysta, Rahul
Savani | First Order Methods for Geometric Optimization of Crystal Structures | null | null | null | null | math.OC cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The geometric optimization of crystal structures is a procedure widely used
in Chemistry that changes the geometrical placement of the particles inside a
structure. It is called structural relaxation and constitutes a local
minimization problem with a non-convex objective function whose domain
complexity increases according to the number of particles involved. In this
work we study the performance of the two most popular first order optimization
methods in structural relaxation. Although frequently employed, there is a lack
of their study in this context from an algorithmic point of view. We run each
algorithm in combination with a constant step size, which provides a benchmark
for the methods' analysis and direct comparison. We also design dynamic step
size rules and study how these improve the two algorithms' performance. Our
results show that there is a trade-off between convergence rate and the
possibility of an experiment to succeed, hence we construct a function to
assign utility to each method based on our respective preference. The function
is built according to a recently introduced model of preference indication
concerning algorithms with deadline and their run time. Finally, building on
all our insights from the experimental results, we provide algorithmic recipes
that best correspond to each of the presented preferences and select one recipe
as the optimal for equally weighted preferences.
Alongside our results we present our open source Python software veltiCRYS,
which was used to perform the geometric optimization experiments. Our
implementation, can be easily edited to accommodate other energy functions and
is especially targeted for testing different methods in structural relaxation.
| [
{
"version": "v1",
"created": "Thu, 26 Jan 2023 16:27:56 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 17:33:58 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Tsili",
"Antonia",
""
],
[
"Dyer",
"Matthew",
""
],
[
"Gusev",
"Vladimir",
""
],
[
"Krysta",
"Piotr",
""
],
[
"Savani",
"Rahul",
""
]
] | TITLE: First Order Methods for Geometric Optimization of Crystal Structures
ABSTRACT: The geometric optimization of crystal structures is a procedure widely used
in Chemistry that changes the geometrical placement of the particles inside a
structure. It is called structural relaxation and constitutes a local
minimization problem with a non-convex objective function whose domain
complexity increases according to the number of particles involved. In this
work we study the performance of the two most popular first order optimization
methods in structural relaxation. Although frequently employed, there is a lack
of their study in this context from an algorithmic point of view. We run each
algorithm in combination with a constant step size, which provides a benchmark
for the methods' analysis and direct comparison. We also design dynamic step
size rules and study how these improve the two algorithms' performance. Our
results show that there is a trade-off between convergence rate and the
possibility of an experiment to succeed, hence we construct a function to
assign utility to each method based on our respective preference. The function
is built according to a recently introduced model of preference indication
concerning algorithms with deadline and their run time. Finally, building on
all our insights from the experimental results, we provide algorithmic recipes
that best correspond to each of the presented preferences and select one recipe
as the optimal for equally weighted preferences.
Alongside our results we present our open source Python software veltiCRYS,
which was used to perform the geometric optimization experiments. Our
implementation, can be easily edited to accommodate other energy functions and
is especially targeted for testing different methods in structural relaxation.
| no_new_dataset | 0.708427 |
2302.01825 | Wangmeng Xiang | Hanyuan Chen, Jun-Yan He, Wangmeng Xiang, Zhi-Qi Cheng, Wei Liu,
Hanbing Liu, Bin Luo, Yifeng Geng, Xuansong Xie | HDFormer: High-order Directed Transformer for 3D Human Pose Estimation | Accepted to IJCAI 2023; 9 pages, 5 figures, 7 tables; the code is at
https://github.com/hyer/HDFormer | In the 32nd international Joint Conference on Artificial
Intelligence (IJCAI 2023) | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human pose estimation is a challenging task due to its structured data
sequence nature. Existing methods primarily focus on pair-wise interaction of
body joints, which is insufficient for scenarios involving overlapping joints
and rapidly changing poses. To overcome these issues, we introduce a novel
approach, the High-order Directed Transformer (HDFormer), which leverages
high-order bone and joint relationships for improved pose estimation.
Specifically, HDFormer incorporates both self-attention and high-order
attention to formulate a multi-order attention module. This module facilitates
first-order "joint$\leftrightarrow$joint", second-order
"bone$\leftrightarrow$joint", and high-order "hyperbone$\leftrightarrow$joint"
interactions, effectively addressing issues in complex and occlusion-heavy
situations. In addition, modern CNN techniques are integrated into the
transformer-based architecture, balancing the trade-off between performance and
efficiency. HDFormer significantly outperforms state-of-the-art (SOTA) models
on Human3.6M and MPI-INF-3DHP datasets, requiring only 1/10 of the parameters
and significantly lower computational costs. Moreover, HDFormer demonstrates
broad real-world applicability, enabling real-time, accurate 3D pose
estimation. The source code is in https://github.com/hyer/HDFormer
| [
{
"version": "v1",
"created": "Fri, 3 Feb 2023 16:00:48 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 06:32:17 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chen",
"Hanyuan",
""
],
[
"He",
"Jun-Yan",
""
],
[
"Xiang",
"Wangmeng",
""
],
[
"Cheng",
"Zhi-Qi",
""
],
[
"Liu",
"Wei",
""
],
[
"Liu",
"Hanbing",
""
],
[
"Luo",
"Bin",
""
],
[
"Geng",
"Yifeng",
""
],
[
"Xie",
"Xuansong",
""
]
] | TITLE: HDFormer: High-order Directed Transformer for 3D Human Pose Estimation
ABSTRACT: Human pose estimation is a challenging task due to its structured data
sequence nature. Existing methods primarily focus on pair-wise interaction of
body joints, which is insufficient for scenarios involving overlapping joints
and rapidly changing poses. To overcome these issues, we introduce a novel
approach, the High-order Directed Transformer (HDFormer), which leverages
high-order bone and joint relationships for improved pose estimation.
Specifically, HDFormer incorporates both self-attention and high-order
attention to formulate a multi-order attention module. This module facilitates
first-order "joint$\leftrightarrow$joint", second-order
"bone$\leftrightarrow$joint", and high-order "hyperbone$\leftrightarrow$joint"
interactions, effectively addressing issues in complex and occlusion-heavy
situations. In addition, modern CNN techniques are integrated into the
transformer-based architecture, balancing the trade-off between performance and
efficiency. HDFormer significantly outperforms state-of-the-art (SOTA) models
on Human3.6M and MPI-INF-3DHP datasets, requiring only 1/10 of the parameters
and significantly lower computational costs. Moreover, HDFormer demonstrates
broad real-world applicability, enabling real-time, accurate 3D pose
estimation. The source code is in https://github.com/hyer/HDFormer
| no_new_dataset | 0.709233 |
2302.06037 | Arman Asgharpoor Golroudbari | Arman Asgharpoor Golroudbari, Mohammad Hossein Sabour | Generalizable End-to-End Deep Learning Frameworks for Real-Time Attitude
Estimation Using 6DoF Inertial Measurement Units | null | null | null | null | cs.RO cs.AI eess.SP | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel end-to-end deep learning framework for real-time
inertial attitude estimation using 6DoF IMU measurements. Inertial Measurement
Units are widely used in various applications, including engineering and
medical sciences. However, traditional filters used for attitude estimation
suffer from poor generalization over different motion patterns and
environmental disturbances. To address this problem, we propose two deep
learning models that incorporate accelerometer and gyroscope readings as
inputs. These models are designed to be generalized to different motion
patterns, sampling rates, and environmental disturbances. Our models consist of
convolutional neural network layers combined with Bi-Directional Long-Short
Term Memory followed by a Fully Forward Neural Network to estimate the
quaternion. We evaluate the proposed method on seven publicly available
datasets, totaling more than 120 hours and 200 kilometers of IMU measurements.
Our results show that the proposed method outperforms state-of-the-art methods
in terms of accuracy and robustness. Additionally, our framework demonstrates
superior generalization over various motion characteristics and sensor sampling
rates. Overall, this paper provides a comprehensive and reliable solution for
real-time inertial attitude estimation using 6DoF IMUs, which has significant
implications for a wide range of applications.
| [
{
"version": "v1",
"created": "Mon, 13 Feb 2023 00:41:49 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 13:00:14 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Golroudbari",
"Arman Asgharpoor",
""
],
[
"Sabour",
"Mohammad Hossein",
""
]
] | TITLE: Generalizable End-to-End Deep Learning Frameworks for Real-Time Attitude
Estimation Using 6DoF Inertial Measurement Units
ABSTRACT: This paper presents a novel end-to-end deep learning framework for real-time
inertial attitude estimation using 6DoF IMU measurements. Inertial Measurement
Units are widely used in various applications, including engineering and
medical sciences. However, traditional filters used for attitude estimation
suffer from poor generalization over different motion patterns and
environmental disturbances. To address this problem, we propose two deep
learning models that incorporate accelerometer and gyroscope readings as
inputs. These models are designed to be generalized to different motion
patterns, sampling rates, and environmental disturbances. Our models consist of
convolutional neural network layers combined with Bi-Directional Long-Short
Term Memory followed by a Fully Forward Neural Network to estimate the
quaternion. We evaluate the proposed method on seven publicly available
datasets, totaling more than 120 hours and 200 kilometers of IMU measurements.
Our results show that the proposed method outperforms state-of-the-art methods
in terms of accuracy and robustness. Additionally, our framework demonstrates
superior generalization over various motion characteristics and sensor sampling
rates. Overall, this paper provides a comprehensive and reliable solution for
real-time inertial attitude estimation using 6DoF IMUs, which has significant
implications for a wide range of applications.
| no_new_dataset | 0.710616 |
2302.06605 | Mingyu Ding | Haoyu Lu, Yuqi Huo, Guoxing Yang, Zhiwu Lu, Wei Zhan, Masayoshi
Tomizuka, Mingyu Ding | UniAdapter: Unified Parameter-Efficient Transfer Learning for
Cross-modal Modeling | null | null | null | null | cs.CV cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large-scale vision-language pre-trained models have shown promising
transferability to various downstream tasks. As the size of these foundation
models and the number of downstream tasks grow, the standard full fine-tuning
paradigm becomes unsustainable due to heavy computational and storage costs.
This paper proposes UniAdapter, which unifies unimodal and multimodal adapters
for parameter-efficient cross-modal adaptation on pre-trained vision-language
models. Specifically, adapters are distributed to different modalities and
their interactions, with the total number of tunable parameters reduced by
partial weight sharing. The unified and knowledge-sharing design enables
powerful cross-modal representations that can benefit various downstream tasks,
requiring only 1.0%-2.0% tunable parameters of the pre-trained model. Extensive
experiments on 6 cross-modal downstream benchmarks (including video-text
retrieval, image-text retrieval, VideoQA, and VQA) show that in most cases,
UniAdapter not only outperforms the state-of-the-arts, but even beats the full
fine-tuning strategy. Particularly, on the MSRVTT retrieval task, UniAdapter
achieves 49.7% recall@1 with 2.2% model parameters, outperforming the latest
competitors by 2.0%. The code and models are available at
https://github.com/RERV/UniAdapter.
| [
{
"version": "v1",
"created": "Mon, 13 Feb 2023 18:59:10 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 17:50:30 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Lu",
"Haoyu",
""
],
[
"Huo",
"Yuqi",
""
],
[
"Yang",
"Guoxing",
""
],
[
"Lu",
"Zhiwu",
""
],
[
"Zhan",
"Wei",
""
],
[
"Tomizuka",
"Masayoshi",
""
],
[
"Ding",
"Mingyu",
""
]
] | TITLE: UniAdapter: Unified Parameter-Efficient Transfer Learning for
Cross-modal Modeling
ABSTRACT: Large-scale vision-language pre-trained models have shown promising
transferability to various downstream tasks. As the size of these foundation
models and the number of downstream tasks grow, the standard full fine-tuning
paradigm becomes unsustainable due to heavy computational and storage costs.
This paper proposes UniAdapter, which unifies unimodal and multimodal adapters
for parameter-efficient cross-modal adaptation on pre-trained vision-language
models. Specifically, adapters are distributed to different modalities and
their interactions, with the total number of tunable parameters reduced by
partial weight sharing. The unified and knowledge-sharing design enables
powerful cross-modal representations that can benefit various downstream tasks,
requiring only 1.0%-2.0% tunable parameters of the pre-trained model. Extensive
experiments on 6 cross-modal downstream benchmarks (including video-text
retrieval, image-text retrieval, VideoQA, and VQA) show that in most cases,
UniAdapter not only outperforms the state-of-the-arts, but even beats the full
fine-tuning strategy. Particularly, on the MSRVTT retrieval task, UniAdapter
achieves 49.7% recall@1 with 2.2% model parameters, outperforming the latest
competitors by 2.0%. The code and models are available at
https://github.com/RERV/UniAdapter.
| no_new_dataset | 0.708244 |
2303.02475 | Edmond Adib | Edmond Adib, Amanda Fernandez, Fatemeh Afghah and John Jeff Prevost | Synthetic ECG Signal Generation using Probabilistic Diffusion Models | null | null | null | null | eess.SP cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning image processing models have had remarkable success in recent
years in generating high quality images. Particularly, the Improved Denoising
Diffusion Probabilistic Models (DDPM) have shown superiority in image quality
to the state-of-the-art generative models, which motivated us to investigate
their capability in the generation of the synthetic electrocardiogram (ECG)
signals. In this work, synthetic ECG signals are generated by the Improved DDPM
and by the Wasserstein GAN with Gradient Penalty (WGAN-GP) models and then
compared. To this end, we devise a pipeline to utilize DDPM in its original
$2D$ form. First, the $1D$ ECG time series data are embedded into the $2D$
space, for which we employed the Gramian Angular Summation/Difference Fields
(GASF/GADF) as well as Markov Transition Fields (MTF) to generate three $2D$
matrices from each ECG time series, which when put together, form a $3$-channel
$2D$ datum. Then $2D$ DDPM is used to generate $2D$ $3$-channel synthetic ECG
images. The $1D$ ECG signals are created by de-embedding the $2D$ generated
image files back into the $1D$ space. This work focuses on unconditional models
and the generation of \emph{Normal Sinus Beat} ECG signals exclusively, where
the Normal Sinus Beat class from the MIT-BIH Arrhythmia dataset is used in the
training phase. The \emph{quality}, \emph{distribution}, and the
\emph{authenticity} of the generated ECG signals by each model are
quantitatively evaluated and compared. Our results show that in the proposed
pipeline and in the particular setting of this paper, the WGAN-GP model is
consistently superior to DDPM in all the considered metrics.
| [
{
"version": "v1",
"created": "Sat, 4 Mar 2023 18:24:45 GMT"
},
{
"version": "v2",
"created": "Thu, 16 Mar 2023 15:47:32 GMT"
},
{
"version": "v3",
"created": "Wed, 17 May 2023 23:17:52 GMT"
},
{
"version": "v4",
"created": "Mon, 22 May 2023 16:15:54 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Adib",
"Edmond",
""
],
[
"Fernandez",
"Amanda",
""
],
[
"Afghah",
"Fatemeh",
""
],
[
"Prevost",
"John Jeff",
""
]
] | TITLE: Synthetic ECG Signal Generation using Probabilistic Diffusion Models
ABSTRACT: Deep learning image processing models have had remarkable success in recent
years in generating high quality images. Particularly, the Improved Denoising
Diffusion Probabilistic Models (DDPM) have shown superiority in image quality
to the state-of-the-art generative models, which motivated us to investigate
their capability in the generation of the synthetic electrocardiogram (ECG)
signals. In this work, synthetic ECG signals are generated by the Improved DDPM
and by the Wasserstein GAN with Gradient Penalty (WGAN-GP) models and then
compared. To this end, we devise a pipeline to utilize DDPM in its original
$2D$ form. First, the $1D$ ECG time series data are embedded into the $2D$
space, for which we employed the Gramian Angular Summation/Difference Fields
(GASF/GADF) as well as Markov Transition Fields (MTF) to generate three $2D$
matrices from each ECG time series, which when put together, form a $3$-channel
$2D$ datum. Then $2D$ DDPM is used to generate $2D$ $3$-channel synthetic ECG
images. The $1D$ ECG signals are created by de-embedding the $2D$ generated
image files back into the $1D$ space. This work focuses on unconditional models
and the generation of \emph{Normal Sinus Beat} ECG signals exclusively, where
the Normal Sinus Beat class from the MIT-BIH Arrhythmia dataset is used in the
training phase. The \emph{quality}, \emph{distribution}, and the
\emph{authenticity} of the generated ECG signals by each model are
quantitatively evaluated and compared. Our results show that in the proposed
pipeline and in the particular setting of this paper, the WGAN-GP model is
consistently superior to DDPM in all the considered metrics.
| no_new_dataset | 0.711055 |
2303.06060 | Liwei Huang | Liwei Huang, Zhengyu Ma, Liutao Yu, Huihui Zhou, Yonghong Tian | Deep Spiking Neural Networks with High Representation Similarity Model
Visual Pathways of Macaque and Mouse | Accepted by Proceedings of the 37th AAAI Conference on Artificial
Intelligence (AAAI-23) | null | null | null | q-bio.NC cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep artificial neural networks (ANNs) play a major role in modeling the
visual pathways of primate and rodent. However, they highly simplify the
computational properties of neurons compared to their biological counterparts.
Instead, Spiking Neural Networks (SNNs) are more biologically plausible models
since spiking neurons encode information with time sequences of spikes, just
like biological neurons do. However, there is a lack of studies on visual
pathways with deep SNNs models. In this study, we model the visual cortex with
deep SNNs for the first time, and also with a wide range of state-of-the-art
deep CNNs and ViTs for comparison. Using three similarity metrics, we conduct
neural representation similarity experiments on three neural datasets collected
from two species under three types of stimuli. Based on extensive similarity
analyses, we further investigate the functional hierarchy and mechanisms across
species. Almost all similarity scores of SNNs are higher than their
counterparts of CNNs with an average of 6.6%. Depths of the layers with the
highest similarity scores exhibit little differences across mouse cortical
regions, but vary significantly across macaque regions, suggesting that the
visual processing structure of mice is more regionally homogeneous than that of
macaques. Besides, the multi-branch structures observed in some top mouse
brain-like neural networks provide computational evidence of parallel
processing streams in mice, and the different performance in fitting macaque
neural representations under different stimuli exhibits the functional
specialization of information processing in macaques. Taken together, our study
demonstrates that SNNs could serve as promising candidates to better model and
explain the functional hierarchy and mechanisms of the visual system.
| [
{
"version": "v1",
"created": "Thu, 9 Mar 2023 13:07:30 GMT"
},
{
"version": "v2",
"created": "Mon, 27 Mar 2023 10:12:45 GMT"
},
{
"version": "v3",
"created": "Wed, 5 Apr 2023 12:29:38 GMT"
},
{
"version": "v4",
"created": "Fri, 12 May 2023 09:16:39 GMT"
},
{
"version": "v5",
"created": "Mon, 22 May 2023 04:03:46 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Huang",
"Liwei",
""
],
[
"Ma",
"Zhengyu",
""
],
[
"Yu",
"Liutao",
""
],
[
"Zhou",
"Huihui",
""
],
[
"Tian",
"Yonghong",
""
]
] | TITLE: Deep Spiking Neural Networks with High Representation Similarity Model
Visual Pathways of Macaque and Mouse
ABSTRACT: Deep artificial neural networks (ANNs) play a major role in modeling the
visual pathways of primate and rodent. However, they highly simplify the
computational properties of neurons compared to their biological counterparts.
Instead, Spiking Neural Networks (SNNs) are more biologically plausible models
since spiking neurons encode information with time sequences of spikes, just
like biological neurons do. However, there is a lack of studies on visual
pathways with deep SNNs models. In this study, we model the visual cortex with
deep SNNs for the first time, and also with a wide range of state-of-the-art
deep CNNs and ViTs for comparison. Using three similarity metrics, we conduct
neural representation similarity experiments on three neural datasets collected
from two species under three types of stimuli. Based on extensive similarity
analyses, we further investigate the functional hierarchy and mechanisms across
species. Almost all similarity scores of SNNs are higher than their
counterparts of CNNs with an average of 6.6%. Depths of the layers with the
highest similarity scores exhibit little differences across mouse cortical
regions, but vary significantly across macaque regions, suggesting that the
visual processing structure of mice is more regionally homogeneous than that of
macaques. Besides, the multi-branch structures observed in some top mouse
brain-like neural networks provide computational evidence of parallel
processing streams in mice, and the different performance in fitting macaque
neural representations under different stimuli exhibits the functional
specialization of information processing in macaques. Taken together, our study
demonstrates that SNNs could serve as promising candidates to better model and
explain the functional hierarchy and mechanisms of the visual system.
| no_new_dataset | 0.709435 |
2303.09124 | Wan Liu Wan | Wan Liu, Yuqian Chen, Chuyang Ye, Nikos Makris, Yogesh Rathi, Weidong
Cai, Fan Zhang, Lauren J. O'Donnell | Fiber Tract Shape Measures Inform Prediction of Non-Imaging Phenotypes | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neuroimaging measures of the brain's white matter connections can enable the
prediction of non-imaging phenotypes, such as demographic and cognitive
measures. Existing works have investigated traditional microstructure and
connectivity measures from diffusion MRI tractography, without considering the
shape of the connections reconstructed by tractography. In this paper, we
investigate the potential of fiber tract shape features for predicting
non-imaging phenotypes, both individually and in combination with traditional
features. We focus on three basic shape features: length, diameter, and
elongation. Two different prediction methods are used, including a traditional
regression method and a deep-learning-based prediction method. Experiments use
an efficient two-stage fusion strategy for prediction using microstructure,
connectivity, and shape measures. To reduce predictive bias due to brain size,
normalized shape features are also investigated. Experimental results on the
Human Connectome Project (HCP) young adult dataset (n=1065) demonstrate that
individual shape features are predictive of non-imaging phenotypes. When
combined with microstructure and connectivity features, shape features
significantly improve performance for predicting the cognitive score TPVT (NIH
Toolbox picture vocabulary test). Overall, this study demonstrates that the
shape of fiber tracts contains useful information for the description and study
of the living human brain using machine learning.
| [
{
"version": "v1",
"created": "Thu, 16 Mar 2023 07:39:07 GMT"
},
{
"version": "v2",
"created": "Sat, 20 May 2023 15:22:38 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Liu",
"Wan",
""
],
[
"Chen",
"Yuqian",
""
],
[
"Ye",
"Chuyang",
""
],
[
"Makris",
"Nikos",
""
],
[
"Rathi",
"Yogesh",
""
],
[
"Cai",
"Weidong",
""
],
[
"Zhang",
"Fan",
""
],
[
"O'Donnell",
"Lauren J.",
""
]
] | TITLE: Fiber Tract Shape Measures Inform Prediction of Non-Imaging Phenotypes
ABSTRACT: Neuroimaging measures of the brain's white matter connections can enable the
prediction of non-imaging phenotypes, such as demographic and cognitive
measures. Existing works have investigated traditional microstructure and
connectivity measures from diffusion MRI tractography, without considering the
shape of the connections reconstructed by tractography. In this paper, we
investigate the potential of fiber tract shape features for predicting
non-imaging phenotypes, both individually and in combination with traditional
features. We focus on three basic shape features: length, diameter, and
elongation. Two different prediction methods are used, including a traditional
regression method and a deep-learning-based prediction method. Experiments use
an efficient two-stage fusion strategy for prediction using microstructure,
connectivity, and shape measures. To reduce predictive bias due to brain size,
normalized shape features are also investigated. Experimental results on the
Human Connectome Project (HCP) young adult dataset (n=1065) demonstrate that
individual shape features are predictive of non-imaging phenotypes. When
combined with microstructure and connectivity features, shape features
significantly improve performance for predicting the cognitive score TPVT (NIH
Toolbox picture vocabulary test). Overall, this study demonstrates that the
shape of fiber tracts contains useful information for the description and study
of the living human brain using machine learning.
| no_new_dataset | 0.710998 |
2303.10902 | Shuai Wang | Shuai Wang, Daoan Zhang, Zipei Yan, Jianguo Zhang, Rui Li | Feature Alignment and Uniformity for Test Time Adaptation | Accepted by CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Test time adaptation (TTA) aims to adapt deep neural networks when receiving
out of distribution test domain samples. In this setting, the model can only
access online unlabeled test samples and pre-trained models on the training
domains. We first address TTA as a feature revision problem due to the domain
gap between source domains and target domains. After that, we follow the two
measurements alignment and uniformity to discuss the test time feature
revision. For test time feature uniformity, we propose a test time
self-distillation strategy to guarantee the consistency of uniformity between
representations of the current batch and all the previous batches. For test
time feature alignment, we propose a memorized spatial local clustering
strategy to align the representations among the neighborhood samples for the
upcoming batch. To deal with the common noisy label problem, we propound the
entropy and consistency filters to select and drop the possible noisy labels.
To prove the scalability and efficacy of our method, we conduct experiments on
four domain generalization benchmarks and four medical image segmentation tasks
with various backbones. Experiment results show that our method not only
improves baseline stably but also outperforms existing state-of-the-art test
time adaptation methods. Code is available at
\href{https://github.com/SakurajimaMaiii/TSD}{https://github.com/SakurajimaMaiii/TSD}.
| [
{
"version": "v1",
"created": "Mon, 20 Mar 2023 06:44:49 GMT"
},
{
"version": "v2",
"created": "Sat, 29 Apr 2023 10:05:59 GMT"
},
{
"version": "v3",
"created": "Sun, 21 May 2023 04:42:08 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Wang",
"Shuai",
""
],
[
"Zhang",
"Daoan",
""
],
[
"Yan",
"Zipei",
""
],
[
"Zhang",
"Jianguo",
""
],
[
"Li",
"Rui",
""
]
] | TITLE: Feature Alignment and Uniformity for Test Time Adaptation
ABSTRACT: Test time adaptation (TTA) aims to adapt deep neural networks when receiving
out of distribution test domain samples. In this setting, the model can only
access online unlabeled test samples and pre-trained models on the training
domains. We first address TTA as a feature revision problem due to the domain
gap between source domains and target domains. After that, we follow the two
measurements alignment and uniformity to discuss the test time feature
revision. For test time feature uniformity, we propose a test time
self-distillation strategy to guarantee the consistency of uniformity between
representations of the current batch and all the previous batches. For test
time feature alignment, we propose a memorized spatial local clustering
strategy to align the representations among the neighborhood samples for the
upcoming batch. To deal with the common noisy label problem, we propound the
entropy and consistency filters to select and drop the possible noisy labels.
To prove the scalability and efficacy of our method, we conduct experiments on
four domain generalization benchmarks and four medical image segmentation tasks
with various backbones. Experiment results show that our method not only
improves baseline stably but also outperforms existing state-of-the-art test
time adaptation methods. Code is available at
\href{https://github.com/SakurajimaMaiii/TSD}{https://github.com/SakurajimaMaiii/TSD}.
| no_new_dataset | 0.709265 |
2303.15212 | Sebastian Pineda Arango | Abdus Salam Khazi, Sebastian Pineda Arango, Josif Grabocka | Deep Ranking Ensembles for Hyperparameter Optimization | Published in ICLR 2023 | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Automatically optimizing the hyperparameters of Machine Learning algorithms
is one of the primary open questions in AI. Existing work in Hyperparameter
Optimization (HPO) trains surrogate models for approximating the response
surface of hyperparameters as a regression task. In contrast, we hypothesize
that the optimal strategy for training surrogates is to preserve the ranks of
the performances of hyperparameter configurations as a Learning to Rank
problem. As a result, we present a novel method that meta-learns neural network
surrogates optimized for ranking the configurations' performances while
modeling their uncertainty via ensembling. In a large-scale experimental
protocol comprising 12 baselines, 16 HPO search spaces and 86 datasets/tasks,
we demonstrate that our method achieves new state-of-the-art results in HPO.
| [
{
"version": "v1",
"created": "Mon, 27 Mar 2023 13:52:40 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 13:31:47 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Khazi",
"Abdus Salam",
""
],
[
"Arango",
"Sebastian Pineda",
""
],
[
"Grabocka",
"Josif",
""
]
] | TITLE: Deep Ranking Ensembles for Hyperparameter Optimization
ABSTRACT: Automatically optimizing the hyperparameters of Machine Learning algorithms
is one of the primary open questions in AI. Existing work in Hyperparameter
Optimization (HPO) trains surrogate models for approximating the response
surface of hyperparameters as a regression task. In contrast, we hypothesize
that the optimal strategy for training surrogates is to preserve the ranks of
the performances of hyperparameter configurations as a Learning to Rank
problem. As a result, we present a novel method that meta-learns neural network
surrogates optimized for ranking the configurations' performances while
modeling their uncertainty via ensembling. In a large-scale experimental
protocol comprising 12 baselines, 16 HPO search spaces and 86 datasets/tasks,
we demonstrate that our method achieves new state-of-the-art results in HPO.
| no_new_dataset | 0.708168 |
2303.16509 | Animesh Karnewar | Animesh Karnewar, Andrea Vedaldi, David Novotny, Niloy Mitra | HoloDiffusion: Training a 3D Diffusion Model using 2D Images | CVPR 2023 conference; project page at:
https://holodiffusion.github.io/ | null | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion models have emerged as the best approach for generative modeling of
2D images. Part of their success is due to the possibility of training them on
millions if not billions of images with a stable learning objective. However,
extending these models to 3D remains difficult for two reasons. First, finding
a large quantity of 3D training data is much more complex than for 2D images.
Second, while it is conceptually trivial to extend the models to operate on 3D
rather than 2D grids, the associated cubic growth in memory and compute
complexity makes this infeasible. We address the first challenge by introducing
a new diffusion setup that can be trained, end-to-end, with only posed 2D
images for supervision; and the second challenge by proposing an image
formation model that decouples model memory from spatial memory. We evaluate
our method on real-world data, using the CO3D dataset which has not been used
to train 3D generative models before. We show that our diffusion models are
scalable, train robustly, and are competitive in terms of sample quality and
fidelity to existing approaches for 3D generative modeling.
| [
{
"version": "v1",
"created": "Wed, 29 Mar 2023 07:35:56 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 22:38:07 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Karnewar",
"Animesh",
""
],
[
"Vedaldi",
"Andrea",
""
],
[
"Novotny",
"David",
""
],
[
"Mitra",
"Niloy",
""
]
] | TITLE: HoloDiffusion: Training a 3D Diffusion Model using 2D Images
ABSTRACT: Diffusion models have emerged as the best approach for generative modeling of
2D images. Part of their success is due to the possibility of training them on
millions if not billions of images with a stable learning objective. However,
extending these models to 3D remains difficult for two reasons. First, finding
a large quantity of 3D training data is much more complex than for 2D images.
Second, while it is conceptually trivial to extend the models to operate on 3D
rather than 2D grids, the associated cubic growth in memory and compute
complexity makes this infeasible. We address the first challenge by introducing
a new diffusion setup that can be trained, end-to-end, with only posed 2D
images for supervision; and the second challenge by proposing an image
formation model that decouples model memory from spatial memory. We evaluate
our method on real-world data, using the CO3D dataset which has not been used
to train 3D generative models before. We show that our diffusion models are
scalable, train robustly, and are competitive in terms of sample quality and
fidelity to existing approaches for 3D generative modeling.
| no_new_dataset | 0.710666 |
2303.17144 | Zhi-Qi Cheng | Jun-Yan He, Zhi-Qi Cheng, Chenyang Li, Wangmeng Xiang, Binghui Chen,
Bin Luo, Yifeng Geng, Xuansong Xie | DAMO-StreamNet: Optimizing Streaming Perception in Autonomous Driving | Accepted to IJCAI 2023; 9 pages, 4 figures, 6 tables; the code is at
https://github.com/zhiqic/DAMO-StreamNet | In the 32nd International Joint Conference on Artificial
Intelligence (IJCAI 2023) | null | null | cs.CV cs.AI cs.MM cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Real-time perception, or streaming perception, is a crucial aspect of
autonomous driving that has yet to be thoroughly explored in existing research.
To address this gap, we present DAMO-StreamNet, an optimized framework that
combines recent advances from the YOLO series with a comprehensive analysis of
spatial and temporal perception mechanisms, delivering a cutting-edge solution.
The key innovations of DAMO-StreamNet are (1) A robust neck structure
incorporating deformable convolution, enhancing the receptive field and feature
alignment capabilities (2) A dual-branch structure that integrates short-path
semantic features and long-path temporal features, improving motion state
prediction accuracy. (3) Logits-level distillation for efficient optimization,
aligning the logits of teacher and student networks in semantic space. (4) A
real-time forecasting mechanism that updates support frame features with the
current frame, ensuring seamless streaming perception during inference. Our
experiments demonstrate that DAMO-StreamNet surpasses existing state-of-the-art
methods, achieving 37.8% (normal size (600, 960)) and 43.3% (large size (1200,
1920)) sAP without using extra data. This work not only sets a new benchmark
for real-time perception but also provides valuable insights for future
research. Additionally, DAMO-StreamNet can be applied to various autonomous
systems, such as drones and robots, paving the way for real-time perception.
The code is at https://github.com/zhiqic/DAMO-StreamNet.
| [
{
"version": "v1",
"created": "Thu, 30 Mar 2023 04:34:31 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Apr 2023 00:25:58 GMT"
},
{
"version": "v3",
"created": "Sat, 20 May 2023 21:37:07 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"He",
"Jun-Yan",
""
],
[
"Cheng",
"Zhi-Qi",
""
],
[
"Li",
"Chenyang",
""
],
[
"Xiang",
"Wangmeng",
""
],
[
"Chen",
"Binghui",
""
],
[
"Luo",
"Bin",
""
],
[
"Geng",
"Yifeng",
""
],
[
"Xie",
"Xuansong",
""
]
] | TITLE: DAMO-StreamNet: Optimizing Streaming Perception in Autonomous Driving
ABSTRACT: Real-time perception, or streaming perception, is a crucial aspect of
autonomous driving that has yet to be thoroughly explored in existing research.
To address this gap, we present DAMO-StreamNet, an optimized framework that
combines recent advances from the YOLO series with a comprehensive analysis of
spatial and temporal perception mechanisms, delivering a cutting-edge solution.
The key innovations of DAMO-StreamNet are (1) A robust neck structure
incorporating deformable convolution, enhancing the receptive field and feature
alignment capabilities (2) A dual-branch structure that integrates short-path
semantic features and long-path temporal features, improving motion state
prediction accuracy. (3) Logits-level distillation for efficient optimization,
aligning the logits of teacher and student networks in semantic space. (4) A
real-time forecasting mechanism that updates support frame features with the
current frame, ensuring seamless streaming perception during inference. Our
experiments demonstrate that DAMO-StreamNet surpasses existing state-of-the-art
methods, achieving 37.8% (normal size (600, 960)) and 43.3% (large size (1200,
1920)) sAP without using extra data. This work not only sets a new benchmark
for real-time perception but also provides valuable insights for future
research. Additionally, DAMO-StreamNet can be applied to various autonomous
systems, such as drones and robots, paving the way for real-time perception.
The code is at https://github.com/zhiqic/DAMO-StreamNet.
| no_new_dataset | 0.709617 |
2304.07945 | Yunpu Zhang | Yunpu Zhang, Changsheng You, Weijie Yuan, Fan Liu, Rui Zhang | Joint Beam Scheduling and Power Allocation for SWIPT in Mixed Near- and
Far-Field Channels | We considered the SWIPT in the mixed near- and far-field
communications. This paper has been submitted to IEEE for possible
publication | null | null | null | eess.SP cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extremely large-scale array (XL-array) has emerged as a promising technology
to enhance the spectrum efficiency and spatial resolution in future wireless
networks, leading to a fundamental paradigm shift from conventional far-field
communications towards the new near-field communications. Different from the
existing works that mostly considered simultaneous wireless information and
power transfer (SWIPT) in the far field, we consider in this paper a new and
practical scenario, called mixed near- and far-field SWIPT, in which energy
harvesting (EH) and information decoding (ID) receivers are located in the
near- and far-field regions of the XL-array base station (BS), respectively.
Specifically, we formulate an optimization problem to maximize the weighted
sum-power harvested at all EH receivers by jointly designing the BS beam
scheduling and power allocation, under the constraints on the ID sum-rate and
BS transmit power. To solve this nonconvex optimization problem, an efficient
algorithm is proposed to obtain a suboptimal solution by leveraging the binary
variable elimination and successive convex approximation methods. Numerical
results demonstrate that our proposed joint design achieves substantial
performance gain over other benchmark schemes.
| [
{
"version": "v1",
"created": "Mon, 17 Apr 2023 01:59:24 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 06:18:44 GMT"
},
{
"version": "v3",
"created": "Sun, 21 May 2023 07:42:27 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Zhang",
"Yunpu",
""
],
[
"You",
"Changsheng",
""
],
[
"Yuan",
"Weijie",
""
],
[
"Liu",
"Fan",
""
],
[
"Zhang",
"Rui",
""
]
] | TITLE: Joint Beam Scheduling and Power Allocation for SWIPT in Mixed Near- and
Far-Field Channels
ABSTRACT: Extremely large-scale array (XL-array) has emerged as a promising technology
to enhance the spectrum efficiency and spatial resolution in future wireless
networks, leading to a fundamental paradigm shift from conventional far-field
communications towards the new near-field communications. Different from the
existing works that mostly considered simultaneous wireless information and
power transfer (SWIPT) in the far field, we consider in this paper a new and
practical scenario, called mixed near- and far-field SWIPT, in which energy
harvesting (EH) and information decoding (ID) receivers are located in the
near- and far-field regions of the XL-array base station (BS), respectively.
Specifically, we formulate an optimization problem to maximize the weighted
sum-power harvested at all EH receivers by jointly designing the BS beam
scheduling and power allocation, under the constraints on the ID sum-rate and
BS transmit power. To solve this nonconvex optimization problem, an efficient
algorithm is proposed to obtain a suboptimal solution by leveraging the binary
variable elimination and successive convex approximation methods. Numerical
results demonstrate that our proposed joint design achieves substantial
performance gain over other benchmark schemes.
| no_new_dataset | 0.711262 |
2304.10573 | Philippe Hansen-Estruch | Philippe Hansen-Estruch, Ilya Kostrikov, Michael Janner, Jakub
Grudzien Kuba, Sergey Levine | IDQL: Implicit Q-Learning as an Actor-Critic Method with Diffusion
Policies | 9 Pages, 4 Figures, 3 Tables | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Effective offline RL methods require properly handling out-of-distribution
actions. Implicit Q-learning (IQL) addresses this by training a Q-function
using only dataset actions through a modified Bellman backup. However, it is
unclear which policy actually attains the values represented by this implicitly
trained Q-function. In this paper, we reinterpret IQL as an actor-critic method
by generalizing the critic objective and connecting it to a
behavior-regularized implicit actor. This generalization shows how the induced
actor balances reward maximization and divergence from the behavior policy,
with the specific loss choice determining the nature of this tradeoff. Notably,
this actor can exhibit complex and multimodal characteristics, suggesting
issues with the conditional Gaussian actor fit with advantage weighted
regression (AWR) used in prior methods. Instead, we propose using samples from
a diffusion parameterized behavior policy and weights computed from the critic
to then importance sampled our intended policy. We introduce Implicit Diffusion
Q-learning (IDQL), combining our general IQL critic with the policy extraction
method. IDQL maintains the ease of implementation of IQL while outperforming
prior offline RL methods and demonstrating robustness to hyperparameters. Code
is available at https://github.com/philippe-eecs/IDQL.
| [
{
"version": "v1",
"created": "Thu, 20 Apr 2023 18:04:09 GMT"
},
{
"version": "v2",
"created": "Fri, 19 May 2023 18:31:04 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Hansen-Estruch",
"Philippe",
""
],
[
"Kostrikov",
"Ilya",
""
],
[
"Janner",
"Michael",
""
],
[
"Kuba",
"Jakub Grudzien",
""
],
[
"Levine",
"Sergey",
""
]
] | TITLE: IDQL: Implicit Q-Learning as an Actor-Critic Method with Diffusion
Policies
ABSTRACT: Effective offline RL methods require properly handling out-of-distribution
actions. Implicit Q-learning (IQL) addresses this by training a Q-function
using only dataset actions through a modified Bellman backup. However, it is
unclear which policy actually attains the values represented by this implicitly
trained Q-function. In this paper, we reinterpret IQL as an actor-critic method
by generalizing the critic objective and connecting it to a
behavior-regularized implicit actor. This generalization shows how the induced
actor balances reward maximization and divergence from the behavior policy,
with the specific loss choice determining the nature of this tradeoff. Notably,
this actor can exhibit complex and multimodal characteristics, suggesting
issues with the conditional Gaussian actor fit with advantage weighted
regression (AWR) used in prior methods. Instead, we propose using samples from
a diffusion parameterized behavior policy and weights computed from the critic
to then importance sampled our intended policy. We introduce Implicit Diffusion
Q-learning (IDQL), combining our general IQL critic with the policy extraction
method. IDQL maintains the ease of implementation of IQL while outperforming
prior offline RL methods and demonstrating robustness to hyperparameters. Code
is available at https://github.com/philippe-eecs/IDQL.
| no_new_dataset | 0.708635 |
2304.11719 | Wei Yao | Jie Shao, Wei Yao, Puzuo Wang, Zhiyi He, Lei Luo | Urban GeoBIM construction by integrating semantic LiDAR point clouds
with as-designed BIM models | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Developments in three-dimensional real worlds promote the integration of
geoinformation and building information models (BIM) known as GeoBIM in urban
construction. Light detection and ranging (LiDAR) integrated with global
navigation satellite systems can provide geo-referenced spatial information.
However, constructing detailed urban GeoBIM poses challenges in terms of LiDAR
data quality. BIM models designed from software are rich in geometrical
information but often lack accurate geo-referenced locations. In this paper, we
propose a complementary strategy that integrates LiDAR point clouds with
as-designed BIM models for reconstructing urban scenes. A state-of-the-art deep
learning framework and graph theory are first combined for LiDAR point cloud
segmentation. A coarse-to-fine matching program is then developed to integrate
object point clouds with corresponding BIM models. Results show the overall
segmentation accuracy of LiDAR datasets reaches up to 90%, and average
positioning accuracies of BIM models are 0.023 m for pole-like objects and
0.156 m for buildings, demonstrating the effectiveness of the method in
segmentation and matching processes. This work offers a practical solution for
rapid and accurate urban GeoBIM construction.
| [
{
"version": "v1",
"created": "Sun, 23 Apr 2023 18:16:14 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 05:21:06 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Shao",
"Jie",
""
],
[
"Yao",
"Wei",
""
],
[
"Wang",
"Puzuo",
""
],
[
"He",
"Zhiyi",
""
],
[
"Luo",
"Lei",
""
]
] | TITLE: Urban GeoBIM construction by integrating semantic LiDAR point clouds
with as-designed BIM models
ABSTRACT: Developments in three-dimensional real worlds promote the integration of
geoinformation and building information models (BIM) known as GeoBIM in urban
construction. Light detection and ranging (LiDAR) integrated with global
navigation satellite systems can provide geo-referenced spatial information.
However, constructing detailed urban GeoBIM poses challenges in terms of LiDAR
data quality. BIM models designed from software are rich in geometrical
information but often lack accurate geo-referenced locations. In this paper, we
propose a complementary strategy that integrates LiDAR point clouds with
as-designed BIM models for reconstructing urban scenes. A state-of-the-art deep
learning framework and graph theory are first combined for LiDAR point cloud
segmentation. A coarse-to-fine matching program is then developed to integrate
object point clouds with corresponding BIM models. Results show the overall
segmentation accuracy of LiDAR datasets reaches up to 90%, and average
positioning accuracies of BIM models are 0.023 m for pole-like objects and
0.156 m for buildings, demonstrating the effectiveness of the method in
segmentation and matching processes. This work offers a practical solution for
rapid and accurate urban GeoBIM construction.
| no_new_dataset | 0.709629 |
2304.14378 | Mar\'ia Barroso | Mar\'ia Barroso, Carlos Mar\'ia Ala\'iz, \'Angela Fern\'andez and Jose
Luis Torrecilla | Functional Diffusion Maps | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Nowadays many real-world datasets can be considered as functional, in the
sense that the processes which generate them are continuous. A fundamental
property of this type of data is that in theory they belong to an
infinite-dimensional space. Although in practice we usually receive finite
observations, they are still high-dimensional and hence dimensionality
reduction methods are crucial. In this vein, the main state-of-the-art method
for functional data analysis is Functional PCA. Nevertheless, this classic
technique assumes that the data lie in a linear manifold, and hence it could
have problems when this hypothesis is not fulfilled. In this research,
attention has been placed on a non-linear manifold learning method: Diffusion
Maps. The article explains how to extend this multivariate method to functional
data and compares its behavior against Functional PCA over different simulated
and real examples.
| [
{
"version": "v1",
"created": "Thu, 27 Apr 2023 17:48:15 GMT"
},
{
"version": "v2",
"created": "Sat, 20 May 2023 15:59:31 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Barroso",
"María",
""
],
[
"Alaíz",
"Carlos María",
""
],
[
"Fernández",
"Ángela",
""
],
[
"Torrecilla",
"Jose Luis",
""
]
] | TITLE: Functional Diffusion Maps
ABSTRACT: Nowadays many real-world datasets can be considered as functional, in the
sense that the processes which generate them are continuous. A fundamental
property of this type of data is that in theory they belong to an
infinite-dimensional space. Although in practice we usually receive finite
observations, they are still high-dimensional and hence dimensionality
reduction methods are crucial. In this vein, the main state-of-the-art method
for functional data analysis is Functional PCA. Nevertheless, this classic
technique assumes that the data lie in a linear manifold, and hence it could
have problems when this hypothesis is not fulfilled. In this research,
attention has been placed on a non-linear manifold learning method: Diffusion
Maps. The article explains how to extend this multivariate method to functional
data and compares its behavior against Functional PCA over different simulated
and real examples.
| no_new_dataset | 0.709629 |
2305.02795 | Ming-Kun Xie | Ming-Kun Xie, Jia-Hao Xiao, Hao-Zhe Liu, Gang Niu, Masashi Sugiyama,
Sheng-Jun Huang | Class-Distribution-Aware Pseudo Labeling for Semi-Supervised Multi-Label
Learning | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pseudo-labeling has emerged as a popular and effective approach for utilizing
unlabeled data. However, in the context of semi-supervised multi-label learning
(SSMLL), conventional pseudo-labeling methods encounter difficulties when
dealing with instances associated with multiple labels and an unknown label
count. These limitations often result in the introduction of false positive
labels or the neglect of true positive ones. To overcome these challenges, this
paper proposes a novel solution called Class-Aware Pseudo-Labeling (CAP) that
performs pseudo-labeling in a class-aware manner. The proposed approach
introduces a regularized learning framework incorporating class-aware
thresholds, which effectively control the assignment of positive and negative
pseudo-labels for each class. Notably, even with a small proportion of labeled
examples, our observations demonstrate that the estimated class distribution
serves as a reliable approximation. Motivated by this finding, we develop a
class-distribution-aware thresholding strategy to ensure the alignment of
pseudo-label distribution with the true distribution. The correctness of the
estimated class distribution is theoretically verified, and a generalization
error bound is provided for our proposed method. Extensive experiments on
multiple benchmark datasets confirm the efficacy of CAP in addressing the
challenges of SSMLL problems.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 12:52:18 GMT"
},
{
"version": "v2",
"created": "Sat, 20 May 2023 06:18:28 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Xie",
"Ming-Kun",
""
],
[
"Xiao",
"Jia-Hao",
""
],
[
"Liu",
"Hao-Zhe",
""
],
[
"Niu",
"Gang",
""
],
[
"Sugiyama",
"Masashi",
""
],
[
"Huang",
"Sheng-Jun",
""
]
] | TITLE: Class-Distribution-Aware Pseudo Labeling for Semi-Supervised Multi-Label
Learning
ABSTRACT: Pseudo-labeling has emerged as a popular and effective approach for utilizing
unlabeled data. However, in the context of semi-supervised multi-label learning
(SSMLL), conventional pseudo-labeling methods encounter difficulties when
dealing with instances associated with multiple labels and an unknown label
count. These limitations often result in the introduction of false positive
labels or the neglect of true positive ones. To overcome these challenges, this
paper proposes a novel solution called Class-Aware Pseudo-Labeling (CAP) that
performs pseudo-labeling in a class-aware manner. The proposed approach
introduces a regularized learning framework incorporating class-aware
thresholds, which effectively control the assignment of positive and negative
pseudo-labels for each class. Notably, even with a small proportion of labeled
examples, our observations demonstrate that the estimated class distribution
serves as a reliable approximation. Motivated by this finding, we develop a
class-distribution-aware thresholding strategy to ensure the alignment of
pseudo-label distribution with the true distribution. The correctness of the
estimated class distribution is theoretically verified, and a generalization
error bound is provided for our proposed method. Extensive experiments on
multiple benchmark datasets confirm the efficacy of CAP in addressing the
challenges of SSMLL problems.
| no_new_dataset | 0.709831 |
2305.03117 | Bingsheng Yao | Bingsheng Yao, Prithviraj Sen, Lucian Popa, James Hendler and Dakuo
Wang | Are Human Explanations Always Helpful? Towards Objective Evaluation of
Human Natural Language Explanations | Accepted to ACL2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Human-annotated labels and explanations are critical for training explainable
NLP models. However, unlike human-annotated labels whose quality is easier to
calibrate (e.g., with a majority vote), human-crafted free-form explanations
can be quite subjective. Before blindly using them as ground truth to train ML
models, a vital question needs to be asked: How do we evaluate a
human-annotated explanation's quality? In this paper, we build on the view that
the quality of a human-annotated explanation can be measured based on its
helpfulness (or impairment) to the ML models' performance for the desired NLP
tasks for which the annotations were collected. In comparison to the commonly
used Simulatability score, we define a new metric that can take into
consideration the helpfulness of an explanation for model performance at both
fine-tuning and inference. With the help of a unified dataset format, we
evaluated the proposed metric on five datasets (e.g., e-SNLI) against two model
architectures (T5 and BART), and the results show that our proposed metric can
objectively evaluate the quality of human-annotated explanations, while
Simulatability falls short.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 19:31:50 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 05:20:04 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Yao",
"Bingsheng",
""
],
[
"Sen",
"Prithviraj",
""
],
[
"Popa",
"Lucian",
""
],
[
"Hendler",
"James",
""
],
[
"Wang",
"Dakuo",
""
]
] | TITLE: Are Human Explanations Always Helpful? Towards Objective Evaluation of
Human Natural Language Explanations
ABSTRACT: Human-annotated labels and explanations are critical for training explainable
NLP models. However, unlike human-annotated labels whose quality is easier to
calibrate (e.g., with a majority vote), human-crafted free-form explanations
can be quite subjective. Before blindly using them as ground truth to train ML
models, a vital question needs to be asked: How do we evaluate a
human-annotated explanation's quality? In this paper, we build on the view that
the quality of a human-annotated explanation can be measured based on its
helpfulness (or impairment) to the ML models' performance for the desired NLP
tasks for which the annotations were collected. In comparison to the commonly
used Simulatability score, we define a new metric that can take into
consideration the helpfulness of an explanation for model performance at both
fine-tuning and inference. With the help of a unified dataset format, we
evaluated the proposed metric on five datasets (e.g., e-SNLI) against two model
architectures (T5 and BART), and the results show that our proposed metric can
objectively evaluate the quality of human-annotated explanations, while
Simulatability falls short.
| no_new_dataset | 0.686751 |
2305.03169 | Kai Zhang | Kai Zhang and Xiaoqian Jiang | Sensitive Data Detection with High-Throughput Machine Learning Models in
Electrical Health Records | Add fugire axis label | null | null | null | cs.CR cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the era of big data, there is an increasing need for healthcare providers,
communities, and researchers to share data and collaborate to improve health
outcomes, generate valuable insights, and advance research. The Health
Insurance Portability and Accountability Act of 1996 (HIPAA) is a federal law
designed to protect sensitive health information by defining regulations for
protected health information (PHI). However, it does not provide efficient
tools for detecting or removing PHI before data sharing. One of the challenges
in this area of research is the heterogeneous nature of PHI fields in data
across different parties. This variability makes rule-based sensitive variable
identification systems that work on one database fail on another. To address
this issue, our paper explores the use of machine learning algorithms to
identify sensitive variables in structured data, thus facilitating the
de-identification process. We made a key observation that the distributions of
metadata of PHI fields and non-PHI fields are very different. Based on this
novel finding, we engineered over 30 features from the metadata of the original
features and used machine learning to build classification models to
automatically identify PHI fields in structured Electronic Health Record (EHR)
data. We trained the model on a variety of large EHR databases from different
data sources and found that our algorithm achieves 99% accuracy when detecting
PHI-related fields for unseen datasets. The implications of our study are
significant and can benefit industries that handle sensitive data.
| [
{
"version": "v1",
"created": "Sun, 30 Apr 2023 16:14:23 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 00:07:33 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Zhang",
"Kai",
""
],
[
"Jiang",
"Xiaoqian",
""
]
] | TITLE: Sensitive Data Detection with High-Throughput Machine Learning Models in
Electrical Health Records
ABSTRACT: In the era of big data, there is an increasing need for healthcare providers,
communities, and researchers to share data and collaborate to improve health
outcomes, generate valuable insights, and advance research. The Health
Insurance Portability and Accountability Act of 1996 (HIPAA) is a federal law
designed to protect sensitive health information by defining regulations for
protected health information (PHI). However, it does not provide efficient
tools for detecting or removing PHI before data sharing. One of the challenges
in this area of research is the heterogeneous nature of PHI fields in data
across different parties. This variability makes rule-based sensitive variable
identification systems that work on one database fail on another. To address
this issue, our paper explores the use of machine learning algorithms to
identify sensitive variables in structured data, thus facilitating the
de-identification process. We made a key observation that the distributions of
metadata of PHI fields and non-PHI fields are very different. Based on this
novel finding, we engineered over 30 features from the metadata of the original
features and used machine learning to build classification models to
automatically identify PHI fields in structured Electronic Health Record (EHR)
data. We trained the model on a variety of large EHR databases from different
data sources and found that our algorithm achieves 99% accuracy when detecting
PHI-related fields for unseen datasets. The implications of our study are
significant and can benefit industries that handle sensitive data.
| no_new_dataset | 0.708648 |
2305.04160 | Feilong Chen | Feilong Chen, Minglun Han, Haozhi Zhao, Qingyang Zhang, Jing Shi,
Shuang Xu, Bo Xu | X-LLM: Bootstrapping Advanced Large Language Models by Treating
Multi-Modalities as Foreign Languages | null | null | null | null | cs.CL cs.AI cs.CV eess.AS | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Large language models (LLMs) have demonstrated remarkable language abilities.
GPT-4, based on advanced LLMs, exhibits extraordinary multimodal capabilities
beyond previous visual language models. We attribute this to the use of more
advanced LLMs compared with previous multimodal models. Unfortunately, the
model architecture and training strategies of GPT-4 are unknown. To endow LLMs
with multimodal capabilities, we propose X-LLM, which converts Multi-modalities
(images, speech, videos) into foreign languages using X2L interfaces and inputs
them into a large Language model (ChatGLM). Specifically, X-LLM aligns multiple
frozen single-modal encoders and a frozen LLM using X2L interfaces, where ``X''
denotes multi-modalities such as image, speech, and videos, and ``L'' denotes
languages. X-LLM's training consists of three stages: (1) Converting Multimodal
Information: The first stage trains each X2L interface to align with its
respective single-modal encoder separately to convert multimodal information
into languages. (2) Aligning X2L representations with the LLM: single-modal
encoders are aligned with the LLM through X2L interfaces independently. (3)
Integrating multiple modalities: all single-modal encoders are aligned with the
LLM through X2L interfaces to integrate multimodal capabilities into the LLM.
Our experiments show that X-LLM demonstrates impressive multimodel chat
abilities, sometimes exhibiting the behaviors of multimodal GPT-4 on unseen
images/instructions, and yields a 84.5\% relative score compared with GPT-4 on
a synthetic multimodal instruction-following dataset. And we also conduct
quantitative tests on using LLM for ASR and multimodal ASR, hoping to promote
the era of LLM-based speech recognition.
| [
{
"version": "v1",
"created": "Sun, 7 May 2023 02:25:42 GMT"
},
{
"version": "v2",
"created": "Wed, 10 May 2023 02:11:40 GMT"
},
{
"version": "v3",
"created": "Mon, 22 May 2023 02:37:02 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chen",
"Feilong",
""
],
[
"Han",
"Minglun",
""
],
[
"Zhao",
"Haozhi",
""
],
[
"Zhang",
"Qingyang",
""
],
[
"Shi",
"Jing",
""
],
[
"Xu",
"Shuang",
""
],
[
"Xu",
"Bo",
""
]
] | TITLE: X-LLM: Bootstrapping Advanced Large Language Models by Treating
Multi-Modalities as Foreign Languages
ABSTRACT: Large language models (LLMs) have demonstrated remarkable language abilities.
GPT-4, based on advanced LLMs, exhibits extraordinary multimodal capabilities
beyond previous visual language models. We attribute this to the use of more
advanced LLMs compared with previous multimodal models. Unfortunately, the
model architecture and training strategies of GPT-4 are unknown. To endow LLMs
with multimodal capabilities, we propose X-LLM, which converts Multi-modalities
(images, speech, videos) into foreign languages using X2L interfaces and inputs
them into a large Language model (ChatGLM). Specifically, X-LLM aligns multiple
frozen single-modal encoders and a frozen LLM using X2L interfaces, where ``X''
denotes multi-modalities such as image, speech, and videos, and ``L'' denotes
languages. X-LLM's training consists of three stages: (1) Converting Multimodal
Information: The first stage trains each X2L interface to align with its
respective single-modal encoder separately to convert multimodal information
into languages. (2) Aligning X2L representations with the LLM: single-modal
encoders are aligned with the LLM through X2L interfaces independently. (3)
Integrating multiple modalities: all single-modal encoders are aligned with the
LLM through X2L interfaces to integrate multimodal capabilities into the LLM.
Our experiments show that X-LLM demonstrates impressive multimodel chat
abilities, sometimes exhibiting the behaviors of multimodal GPT-4 on unseen
images/instructions, and yields a 84.5\% relative score compared with GPT-4 on
a synthetic multimodal instruction-following dataset. And we also conduct
quantitative tests on using LLM for ASR and multimodal ASR, hoping to promote
the era of LLM-based speech recognition.
| no_new_dataset | 0.708988 |
2305.04540 | Miroslav Mitev | Miroslav Mitev, Arsenia Chorti, Gerhard Fettweis | Physical Layer Secret Key Generation with Kalman Filter Detrending | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | The massive deployment of low-end wireless Internet of things (IoT) devices
opens the challenge of finding de-centralized and lightweight alternatives for
secret key distribution. A possible solution, coming from the physical layer,
is the secret key generation (SKG) from channel state information (CSI) during
the channel's coherence time. This work acknowledges the fact that the CSI
consists of deterministic (predictable) and stochastic (unpredictable)
components, loosely captured through the terms large-scale and small-scale
fading, respectively. Hence, keys must be generated using only the random and
unpredictable part. To detrend CSI measurements from deterministic components,
a simple and lightweight approach based on Kalman filters is proposed and is
evaluated using an implementation of the complete SKG protocol (including
privacy amplification that is typically missing in many published works). In
our study we use a massive multiple input multiple output (mMIMO) orthogonal
frequency division multiplexing outdoor measured CSI dataset. The threat model
assumes a passive eavesdropper in the vicinity (at 1 meter distance or less)
from one of the legitimate nodes and the Kalman filter is parameterized to
maximize the achievable key rate.
| [
{
"version": "v1",
"created": "Mon, 8 May 2023 08:28:46 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 14:11:28 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Mitev",
"Miroslav",
""
],
[
"Chorti",
"Arsenia",
""
],
[
"Fettweis",
"Gerhard",
""
]
] | TITLE: Physical Layer Secret Key Generation with Kalman Filter Detrending
ABSTRACT: The massive deployment of low-end wireless Internet of things (IoT) devices
opens the challenge of finding de-centralized and lightweight alternatives for
secret key distribution. A possible solution, coming from the physical layer,
is the secret key generation (SKG) from channel state information (CSI) during
the channel's coherence time. This work acknowledges the fact that the CSI
consists of deterministic (predictable) and stochastic (unpredictable)
components, loosely captured through the terms large-scale and small-scale
fading, respectively. Hence, keys must be generated using only the random and
unpredictable part. To detrend CSI measurements from deterministic components,
a simple and lightweight approach based on Kalman filters is proposed and is
evaluated using an implementation of the complete SKG protocol (including
privacy amplification that is typically missing in many published works). In
our study we use a massive multiple input multiple output (mMIMO) orthogonal
frequency division multiplexing outdoor measured CSI dataset. The threat model
assumes a passive eavesdropper in the vicinity (at 1 meter distance or less)
from one of the legitimate nodes and the Kalman filter is parameterized to
maximize the achievable key rate.
| no_new_dataset | 0.708603 |
2305.07922 | Yue Wang | Yue Wang, Hung Le, Akhilesh Deepak Gotmare, Nghi D.Q. Bui, Junnan Li,
Steven C.H. Hoi | CodeT5+: Open Code Large Language Models for Code Understanding and
Generation | 26 pages, preprint | null | null | null | cs.CL cs.LG cs.PL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) pretrained on vast source code have achieved
prominent progress in code intelligence. However, existing code LLMs have two
main limitations in terms of architecture and pretraining tasks. First, they
often adopt a specific architecture (encoder-only or decoder-only) or rely on a
unified encoder-decoder network for different downstream tasks. The former
paradigm is limited by inflexibility in applications while in the latter, the
model is treated as a single system for all tasks, leading to suboptimal
performance on a subset of tasks. Secondly, they often employ a limited set of
pretraining objectives which might not be relevant to some downstream tasks and
hence result in substantial performance degrade. To address these limitations,
we propose ``CodeT5+'', a family of encoder-decoder LLMs for code in which
component modules can be flexibly combined to suit a wide range of downstream
code tasks. Such flexibility is enabled by our proposed mixture of pretraining
objectives to mitigate the pretrain-finetune discrepancy. These objectives
cover span denoising, contrastive learning, text-code matching, and causal LM
pretraining tasks, on both unimodal and bimodal multilingual code corpora.
Furthermore, we propose to initialize CodeT5+ with frozen off-the-shelf LLMs
without training from scratch to efficiently scale up our models, and explore
instruction-tuning to align with natural language instructions. We extensively
evaluate CodeT5+ on over 20 code-related benchmarks in different settings,
including zero-shot, finetuning, and instruction-tuning. We observe
state-of-the-art (SoTA) model performance on various code-related tasks, such
as code generation and completion, math programming, and text-to-code retrieval
tasks. Particularly, our instruction-tuned CodeT5+ 16B achieves new SoTA
results on HumanEval code generation task against other open code LLMs.
| [
{
"version": "v1",
"created": "Sat, 13 May 2023 14:23:07 GMT"
},
{
"version": "v2",
"created": "Sat, 20 May 2023 07:27:15 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Wang",
"Yue",
""
],
[
"Le",
"Hung",
""
],
[
"Gotmare",
"Akhilesh Deepak",
""
],
[
"Bui",
"Nghi D. Q.",
""
],
[
"Li",
"Junnan",
""
],
[
"Hoi",
"Steven C. H.",
""
]
] | TITLE: CodeT5+: Open Code Large Language Models for Code Understanding and
Generation
ABSTRACT: Large language models (LLMs) pretrained on vast source code have achieved
prominent progress in code intelligence. However, existing code LLMs have two
main limitations in terms of architecture and pretraining tasks. First, they
often adopt a specific architecture (encoder-only or decoder-only) or rely on a
unified encoder-decoder network for different downstream tasks. The former
paradigm is limited by inflexibility in applications while in the latter, the
model is treated as a single system for all tasks, leading to suboptimal
performance on a subset of tasks. Secondly, they often employ a limited set of
pretraining objectives which might not be relevant to some downstream tasks and
hence result in substantial performance degrade. To address these limitations,
we propose ``CodeT5+'', a family of encoder-decoder LLMs for code in which
component modules can be flexibly combined to suit a wide range of downstream
code tasks. Such flexibility is enabled by our proposed mixture of pretraining
objectives to mitigate the pretrain-finetune discrepancy. These objectives
cover span denoising, contrastive learning, text-code matching, and causal LM
pretraining tasks, on both unimodal and bimodal multilingual code corpora.
Furthermore, we propose to initialize CodeT5+ with frozen off-the-shelf LLMs
without training from scratch to efficiently scale up our models, and explore
instruction-tuning to align with natural language instructions. We extensively
evaluate CodeT5+ on over 20 code-related benchmarks in different settings,
including zero-shot, finetuning, and instruction-tuning. We observe
state-of-the-art (SoTA) model performance on various code-related tasks, such
as code generation and completion, math programming, and text-to-code retrieval
tasks. Particularly, our instruction-tuned CodeT5+ 16B achieves new SoTA
results on HumanEval code generation task against other open code LLMs.
| no_new_dataset | 0.710478 |
2305.08135 | Qianglong Chen | Qianglong Chen, Guohai Xu, Ming Yan, Ji Zhang, Fei Huang, Luo Si and
Yin Zhang | Distinguish Before Answer: Generating Contrastive Explanation as
Knowledge for Commonsense Question Answering | Accepted to ACL2023(Findings). The Camera-ready Version | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Existing knowledge-enhanced methods have achieved remarkable results in
certain QA tasks via obtaining diverse knowledge from different knowledge
bases. However, limited by the properties of retrieved knowledge, they still
have trouble benefiting from both the knowledge relevance and distinguishment
simultaneously. To address the challenge, we propose CPACE, a Concept-centric
Prompt-bAsed Contrastive Explanation Generation model, which aims to convert
obtained symbolic knowledge into a contrastive explanation for better
distinguishing the differences among given candidates. Firstly, following
previous works, we retrieve different types of symbolic knowledge with a
concept-centric knowledge extraction module. After that, we generate
corresponding contrastive explanations using acquired symbolic knowledge and
explanation prompts as guidance for better modeling the knowledge
distinguishment and interpretability. Finally, we regard the generated
contrastive explanation as external knowledge for downstream task enhancement.
We conduct a series of experiments on three widely-used question-answering
datasets: CSQA, QASC, and OBQA. Experimental results demonstrate that with the
help of generated contrastive explanation, our CPACE model achieves new SOTA on
CSQA (89.8% on the testing set, 0.9% higher than human performance), and gains
impressive improvement on QASC and OBQA (4.2% and 3.5%, respectively).
| [
{
"version": "v1",
"created": "Sun, 14 May 2023 12:12:24 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 15:07:23 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chen",
"Qianglong",
""
],
[
"Xu",
"Guohai",
""
],
[
"Yan",
"Ming",
""
],
[
"Zhang",
"Ji",
""
],
[
"Huang",
"Fei",
""
],
[
"Si",
"Luo",
""
],
[
"Zhang",
"Yin",
""
]
] | TITLE: Distinguish Before Answer: Generating Contrastive Explanation as
Knowledge for Commonsense Question Answering
ABSTRACT: Existing knowledge-enhanced methods have achieved remarkable results in
certain QA tasks via obtaining diverse knowledge from different knowledge
bases. However, limited by the properties of retrieved knowledge, they still
have trouble benefiting from both the knowledge relevance and distinguishment
simultaneously. To address the challenge, we propose CPACE, a Concept-centric
Prompt-bAsed Contrastive Explanation Generation model, which aims to convert
obtained symbolic knowledge into a contrastive explanation for better
distinguishing the differences among given candidates. Firstly, following
previous works, we retrieve different types of symbolic knowledge with a
concept-centric knowledge extraction module. After that, we generate
corresponding contrastive explanations using acquired symbolic knowledge and
explanation prompts as guidance for better modeling the knowledge
distinguishment and interpretability. Finally, we regard the generated
contrastive explanation as external knowledge for downstream task enhancement.
We conduct a series of experiments on three widely-used question-answering
datasets: CSQA, QASC, and OBQA. Experimental results demonstrate that with the
help of generated contrastive explanation, our CPACE model achieves new SOTA on
CSQA (89.8% on the testing set, 0.9% higher than human performance), and gains
impressive improvement on QASC and OBQA (4.2% and 3.5%, respectively).
| no_new_dataset | 0.70844 |
2305.08348 | YanLing Li | Yanling Li, Bowei Zou, Yifan Fan, Mengxing Dong, Yu Hong | Coreference-aware Double-channel Attention Network for Multi-party
Dialogue Reading Comprehension | IJCNN2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We tackle Multi-party Dialogue Reading Comprehension (abbr., MDRC). MDRC
stands for an extractive reading comprehension task grounded on a batch of
dialogues among multiple interlocutors. It is challenging due to the
requirement of understanding cross-utterance contexts and relationships in a
multi-turn multi-party conversation. Previous studies have made great efforts
on the utterance profiling of a single interlocutor and graph-based interaction
modeling. The corresponding solutions contribute to the answer-oriented
reasoning on a series of well-organized and thread-aware conversational
contexts. However, the current MDRC models still suffer from two bottlenecks.
On the one hand, a pronoun like "it" most probably produces multi-skip
reasoning throughout the utterances of different interlocutors. On the other
hand, an MDRC encoder is potentially puzzled by fuzzy features, i.e., the
mixture of inner linguistic features in utterances and external interactive
features among utterances. To overcome the bottlenecks, we propose a
coreference-aware attention modeling method to strengthen the reasoning
ability. In addition, we construct a two-channel encoding network. It
separately encodes utterance profiles and interactive relationships, so as to
relieve the confusion among heterogeneous features. We experiment on the
benchmark corpora Molweni and FriendsQA. Experimental results demonstrate that
our approach yields substantial improvements on both corpora, compared to the
fine-tuned BERT and ELECTRA baselines. The maximum performance gain is about
2.5\% F1-score. Besides, our MDRC models outperform the state-of-the-art in
most cases.
| [
{
"version": "v1",
"created": "Mon, 15 May 2023 05:01:29 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 04:42:40 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Li",
"Yanling",
""
],
[
"Zou",
"Bowei",
""
],
[
"Fan",
"Yifan",
""
],
[
"Dong",
"Mengxing",
""
],
[
"Hong",
"Yu",
""
]
] | TITLE: Coreference-aware Double-channel Attention Network for Multi-party
Dialogue Reading Comprehension
ABSTRACT: We tackle Multi-party Dialogue Reading Comprehension (abbr., MDRC). MDRC
stands for an extractive reading comprehension task grounded on a batch of
dialogues among multiple interlocutors. It is challenging due to the
requirement of understanding cross-utterance contexts and relationships in a
multi-turn multi-party conversation. Previous studies have made great efforts
on the utterance profiling of a single interlocutor and graph-based interaction
modeling. The corresponding solutions contribute to the answer-oriented
reasoning on a series of well-organized and thread-aware conversational
contexts. However, the current MDRC models still suffer from two bottlenecks.
On the one hand, a pronoun like "it" most probably produces multi-skip
reasoning throughout the utterances of different interlocutors. On the other
hand, an MDRC encoder is potentially puzzled by fuzzy features, i.e., the
mixture of inner linguistic features in utterances and external interactive
features among utterances. To overcome the bottlenecks, we propose a
coreference-aware attention modeling method to strengthen the reasoning
ability. In addition, we construct a two-channel encoding network. It
separately encodes utterance profiles and interactive relationships, so as to
relieve the confusion among heterogeneous features. We experiment on the
benchmark corpora Molweni and FriendsQA. Experimental results demonstrate that
our approach yields substantial improvements on both corpora, compared to the
fine-tuned BERT and ELECTRA baselines. The maximum performance gain is about
2.5\% F1-score. Besides, our MDRC models outperform the state-of-the-art in
most cases.
| no_new_dataset | 0.709202 |
2305.08719 | Hiuyi Cheng | Hiuyi Cheng, Peirong Zhang, Sihang Wu, Jiaxin Zhang, Qiyuan Zhu,
Zecheng Xie, Jing Li, Kai Ding, and Lianwen Jin | M$^{6}$Doc: A Large-Scale Multi-Format, Multi-Type, Multi-Layout,
Multi-Language, Multi-Annotation Category Dataset for Modern Document Layout
Analysis | Accepted by CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Document layout analysis is a crucial prerequisite for document
understanding, including document retrieval and conversion. Most public
datasets currently contain only PDF documents and lack realistic documents.
Models trained on these datasets may not generalize well to real-world
scenarios. Therefore, this paper introduces a large and diverse document layout
analysis dataset called $M^{6}Doc$. The $M^6$ designation represents six
properties: (1) Multi-Format (including scanned, photographed, and PDF
documents); (2) Multi-Type (such as scientific articles, textbooks, books, test
papers, magazines, newspapers, and notes); (3) Multi-Layout (rectangular,
Manhattan, non-Manhattan, and multi-column Manhattan); (4) Multi-Language
(Chinese and English); (5) Multi-Annotation Category (74 types of annotation
labels with 237,116 annotation instances in 9,080 manually annotated pages);
and (6) Modern documents. Additionally, we propose a transformer-based document
layout analysis method called TransDLANet, which leverages an adaptive element
matching mechanism that enables query embedding to better match ground truth to
improve recall, and constructs a segmentation branch for more precise document
image instance segmentation. We conduct a comprehensive evaluation of
$M^{6}Doc$ with various layout analysis methods and demonstrate its
effectiveness. TransDLANet achieves state-of-the-art performance on $M^{6}Doc$
with 64.5% mAP. The $M^{6}Doc$ dataset will be available at
https://github.com/HCIILAB/M6Doc.
| [
{
"version": "v1",
"created": "Mon, 15 May 2023 15:29:06 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 14:22:39 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Cheng",
"Hiuyi",
""
],
[
"Zhang",
"Peirong",
""
],
[
"Wu",
"Sihang",
""
],
[
"Zhang",
"Jiaxin",
""
],
[
"Zhu",
"Qiyuan",
""
],
[
"Xie",
"Zecheng",
""
],
[
"Li",
"Jing",
""
],
[
"Ding",
"Kai",
""
],
[
"Jin",
"Lianwen",
""
]
] | TITLE: M$^{6}$Doc: A Large-Scale Multi-Format, Multi-Type, Multi-Layout,
Multi-Language, Multi-Annotation Category Dataset for Modern Document Layout
Analysis
ABSTRACT: Document layout analysis is a crucial prerequisite for document
understanding, including document retrieval and conversion. Most public
datasets currently contain only PDF documents and lack realistic documents.
Models trained on these datasets may not generalize well to real-world
scenarios. Therefore, this paper introduces a large and diverse document layout
analysis dataset called $M^{6}Doc$. The $M^6$ designation represents six
properties: (1) Multi-Format (including scanned, photographed, and PDF
documents); (2) Multi-Type (such as scientific articles, textbooks, books, test
papers, magazines, newspapers, and notes); (3) Multi-Layout (rectangular,
Manhattan, non-Manhattan, and multi-column Manhattan); (4) Multi-Language
(Chinese and English); (5) Multi-Annotation Category (74 types of annotation
labels with 237,116 annotation instances in 9,080 manually annotated pages);
and (6) Modern documents. Additionally, we propose a transformer-based document
layout analysis method called TransDLANet, which leverages an adaptive element
matching mechanism that enables query embedding to better match ground truth to
improve recall, and constructs a segmentation branch for more precise document
image instance segmentation. We conduct a comprehensive evaluation of
$M^{6}Doc$ with various layout analysis methods and demonstrate its
effectiveness. TransDLANet achieves state-of-the-art performance on $M^{6}Doc$
with 64.5% mAP. The $M^{6}Doc$ dataset will be available at
https://github.com/HCIILAB/M6Doc.
| new_dataset | 0.71823 |
2305.09210 | Shuichiro Shimizu | Shuichiro Shimizu, Chenhui Chu, Sheng Li and Sadao Kurohashi | Towards Speech Dialogue Translation Mediating Speakers of Different
Languages | 11 pages, 4 figures. Accepted to ACL 2023 Findings. Dataset:
https://github.com/ku-nlp/speechBSD | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | We present a new task, speech dialogue translation mediating speakers of
different languages. We construct the SpeechBSD dataset for the task and
conduct baseline experiments. Furthermore, we consider context to be an
important aspect that needs to be addressed in this task and propose two ways
of utilizing context, namely monolingual context and bilingual context. We
conduct cascaded speech translation experiments using Whisper and mBART, and
show that bilingual context performs better in our settings.
| [
{
"version": "v1",
"created": "Tue, 16 May 2023 06:37:55 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 04:45:29 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Shimizu",
"Shuichiro",
""
],
[
"Chu",
"Chenhui",
""
],
[
"Li",
"Sheng",
""
],
[
"Kurohashi",
"Sadao",
""
]
] | TITLE: Towards Speech Dialogue Translation Mediating Speakers of Different
Languages
ABSTRACT: We present a new task, speech dialogue translation mediating speakers of
different languages. We construct the SpeechBSD dataset for the task and
conduct baseline experiments. Furthermore, we consider context to be an
important aspect that needs to be addressed in this task and propose two ways
of utilizing context, namely monolingual context and bilingual context. We
conduct cascaded speech translation experiments using Whisper and mBART, and
show that bilingual context performs better in our settings.
| new_dataset | 0.71365 |
2305.09418 | Dominic Williams | Dominic Williams, Fraser Macfarlane, Avril Britten | Leaf Only SAM: A Segment Anything Pipeline for Zero-Shot Automated Leaf
Segmentation | 9 pages, 4 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Segment Anything Model (SAM) is a new foundation model that can be used as a
zero-shot object segmentation method with the use of either guide prompts such
as bounding boxes, polygons, or points. Alternatively, additional post
processing steps can be used to identify objects of interest after segmenting
everything in an image. Here we present a method using segment anything
together with a series of post processing steps to segment potato leaves,
called Leaf Only SAM. The advantage of this proposed method is that it does not
require any training data to produce its results so has many applications
across the field of plant phenotyping where there is limited high quality
annotated data available. We compare the performance of Leaf Only SAM to a Mask
R-CNN model which has been fine-tuned on our small novel potato leaf dataset.
On the evaluation dataset, Leaf Only SAM finds an average recall of 63.2 and an
average precision of 60.3, compared to recall of 78.7 and precision of 74.7 for
Mask R-CNN. Leaf Only SAM does not perform better than the fine-tuned Mask
R-CNN model on our data, but the SAM based model does not require any extra
training or annotation of our new dataset. This shows there is potential to use
SAM as a zero-shot classifier with the addition of post processing steps.
| [
{
"version": "v1",
"created": "Tue, 16 May 2023 13:16:33 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 09:53:21 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Williams",
"Dominic",
""
],
[
"Macfarlane",
"Fraser",
""
],
[
"Britten",
"Avril",
""
]
] | TITLE: Leaf Only SAM: A Segment Anything Pipeline for Zero-Shot Automated Leaf
Segmentation
ABSTRACT: Segment Anything Model (SAM) is a new foundation model that can be used as a
zero-shot object segmentation method with the use of either guide prompts such
as bounding boxes, polygons, or points. Alternatively, additional post
processing steps can be used to identify objects of interest after segmenting
everything in an image. Here we present a method using segment anything
together with a series of post processing steps to segment potato leaves,
called Leaf Only SAM. The advantage of this proposed method is that it does not
require any training data to produce its results so has many applications
across the field of plant phenotyping where there is limited high quality
annotated data available. We compare the performance of Leaf Only SAM to a Mask
R-CNN model which has been fine-tuned on our small novel potato leaf dataset.
On the evaluation dataset, Leaf Only SAM finds an average recall of 63.2 and an
average precision of 60.3, compared to recall of 78.7 and precision of 74.7 for
Mask R-CNN. Leaf Only SAM does not perform better than the fine-tuned Mask
R-CNN model on our data, but the SAM based model does not require any extra
training or annotation of our new dataset. This shows there is potential to use
SAM as a zero-shot classifier with the addition of post processing steps.
| new_dataset | 0.713606 |
2305.10263 | Yuqi Ren | Chuang Liu, Renren Jin, Yuqi Ren, Linhao Yu, Tianyu Dong, Xiaohan
Peng, Shuting Zhang, Jianxiang Peng, Peiyi Zhang, Qingqing Lyu, Xiaowen Su,
Qun Liu, Deyi Xiong | M3KE: A Massive Multi-Level Multi-Subject Knowledge Evaluation Benchmark
for Chinese Large Language Models | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models have recently made tremendous progress in a variety of
aspects, e.g., cross-task generalization, instruction following.
Comprehensively evaluating the capability of large language models in multiple
tasks is of great importance. In this paper, we propose M3KE, a Massive
Multi-Level Multi-Subject Knowledge Evaluation benchmark, which is developed to
measure knowledge acquired by Chinese large language models by testing their
multitask accuracy in zero- and few-shot settings. We have collected 20,477
questions from 71 tasks. Our selection covers all major levels of Chinese
education system, ranging from the primary school to college, as well as a wide
variety of subjects, including humanities, history, politics, law, education,
psychology, science, technology, art and religion. All questions are
multiple-choice questions with four options, hence guaranteeing a standardized
and unified assessment process. We've assessed a number of state-of-the-art
open-source Chinese large language models on the proposed benchmark. The size
of these models varies from 335M to 130B parameters. Experiment results
demonstrate that they perform significantly worse than GPT-3.5 that reaches an
accuracy of ~ 48% on M3KE. The dataset is available at
https://github.com/tjunlp-lab/M3KE.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 14:56:31 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 03:57:11 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Liu",
"Chuang",
""
],
[
"Jin",
"Renren",
""
],
[
"Ren",
"Yuqi",
""
],
[
"Yu",
"Linhao",
""
],
[
"Dong",
"Tianyu",
""
],
[
"Peng",
"Xiaohan",
""
],
[
"Zhang",
"Shuting",
""
],
[
"Peng",
"Jianxiang",
""
],
[
"Zhang",
"Peiyi",
""
],
[
"Lyu",
"Qingqing",
""
],
[
"Su",
"Xiaowen",
""
],
[
"Liu",
"Qun",
""
],
[
"Xiong",
"Deyi",
""
]
] | TITLE: M3KE: A Massive Multi-Level Multi-Subject Knowledge Evaluation Benchmark
for Chinese Large Language Models
ABSTRACT: Large language models have recently made tremendous progress in a variety of
aspects, e.g., cross-task generalization, instruction following.
Comprehensively evaluating the capability of large language models in multiple
tasks is of great importance. In this paper, we propose M3KE, a Massive
Multi-Level Multi-Subject Knowledge Evaluation benchmark, which is developed to
measure knowledge acquired by Chinese large language models by testing their
multitask accuracy in zero- and few-shot settings. We have collected 20,477
questions from 71 tasks. Our selection covers all major levels of Chinese
education system, ranging from the primary school to college, as well as a wide
variety of subjects, including humanities, history, politics, law, education,
psychology, science, technology, art and religion. All questions are
multiple-choice questions with four options, hence guaranteeing a standardized
and unified assessment process. We've assessed a number of state-of-the-art
open-source Chinese large language models on the proposed benchmark. The size
of these models varies from 335M to 130B parameters. Experiment results
demonstrate that they perform significantly worse than GPT-3.5 that reaches an
accuracy of ~ 48% on M3KE. The dataset is available at
https://github.com/tjunlp-lab/M3KE.
| new_dataset | 0.708767 |
2305.10306 | JunYu Lu | Ping Yang, Junyu Lu, Ruyi Gan, Junjie Wang, Yuxiang Zhang, Jiaxing
Zhang, Pingjian Zhang | UniEX: An Effective and Efficient Framework for Unified Information
Extraction via a Span-extractive Perspective | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We propose a new paradigm for universal information extraction (IE) that is
compatible with any schema format and applicable to a list of IE tasks, such as
named entity recognition, relation extraction, event extraction and sentiment
analysis. Our approach converts the text-based IE tasks as the token-pair
problem, which uniformly disassembles all extraction targets into joint span
detection, classification and association problems with a unified extractive
framework, namely UniEX. UniEX can synchronously encode schema-based prompt and
textual information, and collaboratively learn the generalized knowledge from
pre-defined information using the auto-encoder language models. We develop a
traffine attention mechanism to integrate heterogeneous factors including
tasks, labels and inside tokens, and obtain the extraction target via a scoring
matrix. Experiment results show that UniEX can outperform generative universal
IE models in terms of performance and inference-speed on $14$ benchmarks IE
datasets with the supervised setting. The state-of-the-art performance in
low-resource scenarios also verifies the transferability and effectiveness of
UniEX.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 15:44:12 GMT"
},
{
"version": "v2",
"created": "Fri, 19 May 2023 04:08:16 GMT"
},
{
"version": "v3",
"created": "Mon, 22 May 2023 05:25:33 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Yang",
"Ping",
""
],
[
"Lu",
"Junyu",
""
],
[
"Gan",
"Ruyi",
""
],
[
"Wang",
"Junjie",
""
],
[
"Zhang",
"Yuxiang",
""
],
[
"Zhang",
"Jiaxing",
""
],
[
"Zhang",
"Pingjian",
""
]
] | TITLE: UniEX: An Effective and Efficient Framework for Unified Information
Extraction via a Span-extractive Perspective
ABSTRACT: We propose a new paradigm for universal information extraction (IE) that is
compatible with any schema format and applicable to a list of IE tasks, such as
named entity recognition, relation extraction, event extraction and sentiment
analysis. Our approach converts the text-based IE tasks as the token-pair
problem, which uniformly disassembles all extraction targets into joint span
detection, classification and association problems with a unified extractive
framework, namely UniEX. UniEX can synchronously encode schema-based prompt and
textual information, and collaboratively learn the generalized knowledge from
pre-defined information using the auto-encoder language models. We develop a
traffine attention mechanism to integrate heterogeneous factors including
tasks, labels and inside tokens, and obtain the extraction target via a scoring
matrix. Experiment results show that UniEX can outperform generative universal
IE models in terms of performance and inference-speed on $14$ benchmarks IE
datasets with the supervised setting. The state-of-the-art performance in
low-resource scenarios also verifies the transferability and effectiveness of
UniEX.
| no_new_dataset | 0.707443 |
2305.10308 | Soma Onishi | Soma Onishi and Shoya Meguro | Rethinking Data Augmentation for Tabular Data in Deep Learning | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tabular data is the most widely used data format in machine learning (ML).
While tree-based methods outperform DL-based methods in supervised learning,
recent literature reports that self-supervised learning with Transformer-based
models outperforms tree-based methods. In the existing literature on
self-supervised learning for tabular data, contrastive learning is the
predominant method. In contrastive learning, data augmentation is important to
generate different views. However, data augmentation for tabular data has been
difficult due to the unique structure and high complexity of tabular data. In
addition, three main components are proposed together in existing methods:
model structure, self-supervised learning methods, and data augmentation.
Therefore, previous works have compared the performance without comprehensively
considering these components, and it is not clear how each component affects
the actual performance.
In this study, we focus on data augmentation to address these issues. We
propose a novel data augmentation method, $\textbf{M}$ask $\textbf{T}$oken
$\textbf{R}$eplacement ($\texttt{MTR}$), which replaces the mask token with a
portion of each tokenized column; $\texttt{MTR}$ takes advantage of the
properties of Transformer, which is becoming the predominant DL-based
architecture for tabular data, to perform data augmentation for each column
embedding. Through experiments with 13 diverse public datasets in both
supervised and self-supervised learning scenarios, we show that $\texttt{MTR}$
achieves competitive performance against existing data augmentation methods and
improves model performance. In addition, we discuss specific scenarios in which
$\texttt{MTR}$ is most effective and identify the scope of its application. The
code is available at https://github.com/somaonishi/MTR/.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 15:46:03 GMT"
},
{
"version": "v2",
"created": "Mon, 22 May 2023 13:02:40 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Onishi",
"Soma",
""
],
[
"Meguro",
"Shoya",
""
]
] | TITLE: Rethinking Data Augmentation for Tabular Data in Deep Learning
ABSTRACT: Tabular data is the most widely used data format in machine learning (ML).
While tree-based methods outperform DL-based methods in supervised learning,
recent literature reports that self-supervised learning with Transformer-based
models outperforms tree-based methods. In the existing literature on
self-supervised learning for tabular data, contrastive learning is the
predominant method. In contrastive learning, data augmentation is important to
generate different views. However, data augmentation for tabular data has been
difficult due to the unique structure and high complexity of tabular data. In
addition, three main components are proposed together in existing methods:
model structure, self-supervised learning methods, and data augmentation.
Therefore, previous works have compared the performance without comprehensively
considering these components, and it is not clear how each component affects
the actual performance.
In this study, we focus on data augmentation to address these issues. We
propose a novel data augmentation method, $\textbf{M}$ask $\textbf{T}$oken
$\textbf{R}$eplacement ($\texttt{MTR}$), which replaces the mask token with a
portion of each tokenized column; $\texttt{MTR}$ takes advantage of the
properties of Transformer, which is becoming the predominant DL-based
architecture for tabular data, to perform data augmentation for each column
embedding. Through experiments with 13 diverse public datasets in both
supervised and self-supervised learning scenarios, we show that $\texttt{MTR}$
achieves competitive performance against existing data augmentation methods and
improves model performance. In addition, we discuss specific scenarios in which
$\texttt{MTR}$ is most effective and identify the scope of its application. The
code is available at https://github.com/somaonishi/MTR/.
| no_new_dataset | 0.709806 |
2305.10352 | Adrien Petralia | Adrien Petralia and Philippe Charpentier and Paul Boniol and Themis
Palpanas | Appliance Detection Using Very Low-Frequency Smart Meter Time Series | 11 pages, 7 figures. This paper appeared in ACM e-Energy 2023 | null | 10.1145/3575813.3595198 | null | eess.SP cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, smart meters have been widely adopted by electricity
suppliers to improve the management of the smart grid system. These meters
usually collect energy consumption data at a very low frequency (every 30min),
enabling utilities to bill customers more accurately. To provide more
personalized recommendations, the next step is to detect the appliances owned
by customers, which is a challenging problem, due to the very-low meter reading
frequency. Even though the appliance detection problem can be cast as a time
series classification problem, with many such classifiers having been proposed
in the literature, no study has applied and compared them on this specific
problem. This paper presents an in-depth evaluation and comparison of
state-of-the-art time series classifiers applied to detecting the
presence/absence of diverse appliances in very low-frequency smart meter data.
We report results with five real datasets. We first study the impact of the
detection quality of 13 different appliances using 30min sampled data, and we
subsequently propose an analysis of the possible detection performance gain by
using a higher meter reading frequency. The results indicate that the
performance of current time series classifiers varies significantly. Some of
them, namely deep learning-based classifiers, provide promising results in
terms of accuracy (especially for certain appliances), even using 30min sampled
data, and are scalable to the large smart meter time series collections of
energy consumption data currently available to electricity suppliers.
Nevertheless, our study shows that more work is needed in this area to further
improve the accuracy of the proposed solutions. This paper appeared in ACM
e-Energy 2023.
| [
{
"version": "v1",
"created": "Wed, 10 May 2023 10:26:43 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 20:23:55 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Petralia",
"Adrien",
""
],
[
"Charpentier",
"Philippe",
""
],
[
"Boniol",
"Paul",
""
],
[
"Palpanas",
"Themis",
""
]
] | TITLE: Appliance Detection Using Very Low-Frequency Smart Meter Time Series
ABSTRACT: In recent years, smart meters have been widely adopted by electricity
suppliers to improve the management of the smart grid system. These meters
usually collect energy consumption data at a very low frequency (every 30min),
enabling utilities to bill customers more accurately. To provide more
personalized recommendations, the next step is to detect the appliances owned
by customers, which is a challenging problem, due to the very-low meter reading
frequency. Even though the appliance detection problem can be cast as a time
series classification problem, with many such classifiers having been proposed
in the literature, no study has applied and compared them on this specific
problem. This paper presents an in-depth evaluation and comparison of
state-of-the-art time series classifiers applied to detecting the
presence/absence of diverse appliances in very low-frequency smart meter data.
We report results with five real datasets. We first study the impact of the
detection quality of 13 different appliances using 30min sampled data, and we
subsequently propose an analysis of the possible detection performance gain by
using a higher meter reading frequency. The results indicate that the
performance of current time series classifiers varies significantly. Some of
them, namely deep learning-based classifiers, provide promising results in
terms of accuracy (especially for certain appliances), even using 30min sampled
data, and are scalable to the large smart meter time series collections of
energy consumption data currently available to electricity suppliers.
Nevertheless, our study shows that more work is needed in this area to further
improve the accuracy of the proposed solutions. This paper appeared in ACM
e-Energy 2023.
| no_new_dataset | 0.71081 |
2305.10431 | Guangxuan Xiao | Guangxuan Xiao, Tianwei Yin, William T. Freeman, Fr\'edo Durand, Song
Han | FastComposer: Tuning-Free Multi-Subject Image Generation with Localized
Attention | The first two authors contributed equally to this work | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Diffusion models excel at text-to-image generation, especially in
subject-driven generation for personalized images. However, existing methods
are inefficient due to the subject-specific fine-tuning, which is
computationally intensive and hampers efficient deployment. Moreover, existing
methods struggle with multi-subject generation as they often blend features
among subjects. We present FastComposer which enables efficient, personalized,
multi-subject text-to-image generation without fine-tuning. FastComposer uses
subject embeddings extracted by an image encoder to augment the generic text
conditioning in diffusion models, enabling personalized image generation based
on subject images and textual instructions with only forward passes. To address
the identity blending problem in the multi-subject generation, FastComposer
proposes cross-attention localization supervision during training, enforcing
the attention of reference subjects localized to the correct regions in the
target images. Naively conditioning on subject embeddings results in subject
overfitting. FastComposer proposes delayed subject conditioning in the
denoising step to maintain both identity and editability in subject-driven
image generation. FastComposer generates images of multiple unseen individuals
with different styles, actions, and contexts. It achieves
300$\times$-2500$\times$ speedup compared to fine-tuning-based methods and
requires zero extra storage for new subjects. FastComposer paves the way for
efficient, personalized, and high-quality multi-subject image creation. Code,
model, and dataset are available at
https://github.com/mit-han-lab/fastcomposer.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 17:59:55 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 17:26:40 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Xiao",
"Guangxuan",
""
],
[
"Yin",
"Tianwei",
""
],
[
"Freeman",
"William T.",
""
],
[
"Durand",
"Frédo",
""
],
[
"Han",
"Song",
""
]
] | TITLE: FastComposer: Tuning-Free Multi-Subject Image Generation with Localized
Attention
ABSTRACT: Diffusion models excel at text-to-image generation, especially in
subject-driven generation for personalized images. However, existing methods
are inefficient due to the subject-specific fine-tuning, which is
computationally intensive and hampers efficient deployment. Moreover, existing
methods struggle with multi-subject generation as they often blend features
among subjects. We present FastComposer which enables efficient, personalized,
multi-subject text-to-image generation without fine-tuning. FastComposer uses
subject embeddings extracted by an image encoder to augment the generic text
conditioning in diffusion models, enabling personalized image generation based
on subject images and textual instructions with only forward passes. To address
the identity blending problem in the multi-subject generation, FastComposer
proposes cross-attention localization supervision during training, enforcing
the attention of reference subjects localized to the correct regions in the
target images. Naively conditioning on subject embeddings results in subject
overfitting. FastComposer proposes delayed subject conditioning in the
denoising step to maintain both identity and editability in subject-driven
image generation. FastComposer generates images of multiple unseen individuals
with different styles, actions, and contexts. It achieves
300$\times$-2500$\times$ speedup compared to fine-tuning-based methods and
requires zero extra storage for new subjects. FastComposer paves the way for
efficient, personalized, and high-quality multi-subject image creation. Code,
model, and dataset are available at
https://github.com/mit-han-lab/fastcomposer.
| no_new_dataset | 0.709648 |
2305.10552 | Georg W\"olflein | Georg W\"olflein and Lucie Charlotte Magister and Pietro Li\`o and
David J. Harrison and Ognjen Arandjelovi\'c | Deep Multiple Instance Learning with Distance-Aware Self-Attention | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Traditional supervised learning tasks require a label for every instance in
the training set, but in many real-world applications, labels are only
available for collections (bags) of instances. This problem setting, known as
multiple instance learning (MIL), is particularly relevant in the medical
domain, where high-resolution images are split into smaller patches, but labels
apply to the image as a whole. Recent MIL models are able to capture
correspondences between patches by employing self-attention, allowing them to
weigh each patch differently based on all other patches in the bag. However,
these approaches still do not consider the relative spatial relationships
between patches within the larger image, which is especially important in
computational pathology. To this end, we introduce a novel MIL model with
distance-aware self-attention (DAS-MIL), which explicitly takes into account
relative spatial information when modelling the interactions between patches.
Unlike existing relative position representations for self-attention which are
discrete, our approach introduces continuous distance-dependent terms into the
computation of the attention weights, and is the first to apply relative
position representations in the context of MIL. We evaluate our model on a
custom MNIST-based MIL dataset that requires the consideration of relative
spatial information, as well as on CAMELYON16, a publicly available cancer
metastasis detection dataset, where we achieve a test AUROC score of 0.91. On
both datasets, our model outperforms existing MIL approaches that employ
absolute positional encodings, as well as existing relative position
representation schemes applied to MIL. Our code is available at
https://anonymous.4open.science/r/das-mil.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 20:11:43 GMT"
},
{
"version": "v2",
"created": "Sat, 20 May 2023 12:45:27 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Wölflein",
"Georg",
""
],
[
"Magister",
"Lucie Charlotte",
""
],
[
"Liò",
"Pietro",
""
],
[
"Harrison",
"David J.",
""
],
[
"Arandjelović",
"Ognjen",
""
]
] | TITLE: Deep Multiple Instance Learning with Distance-Aware Self-Attention
ABSTRACT: Traditional supervised learning tasks require a label for every instance in
the training set, but in many real-world applications, labels are only
available for collections (bags) of instances. This problem setting, known as
multiple instance learning (MIL), is particularly relevant in the medical
domain, where high-resolution images are split into smaller patches, but labels
apply to the image as a whole. Recent MIL models are able to capture
correspondences between patches by employing self-attention, allowing them to
weigh each patch differently based on all other patches in the bag. However,
these approaches still do not consider the relative spatial relationships
between patches within the larger image, which is especially important in
computational pathology. To this end, we introduce a novel MIL model with
distance-aware self-attention (DAS-MIL), which explicitly takes into account
relative spatial information when modelling the interactions between patches.
Unlike existing relative position representations for self-attention which are
discrete, our approach introduces continuous distance-dependent terms into the
computation of the attention weights, and is the first to apply relative
position representations in the context of MIL. We evaluate our model on a
custom MNIST-based MIL dataset that requires the consideration of relative
spatial information, as well as on CAMELYON16, a publicly available cancer
metastasis detection dataset, where we achieve a test AUROC score of 0.91. On
both datasets, our model outperforms existing MIL approaches that employ
absolute positional encodings, as well as existing relative position
representation schemes applied to MIL. Our code is available at
https://anonymous.4open.science/r/das-mil.
| no_new_dataset | 0.712057 |
2305.10853 | Gabriela Ben Melech | Gabriela Ben Melech Stan, Diana Wofk, Scottie Fox, Alex Redden, Will
Saxton, Jean Yu, Estelle Aflalo, Shao-Yen Tseng, Fabio Nonato, Matthias
Muller, Vasudev Lal | LDM3D: Latent Diffusion Model for 3D | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This research paper proposes a Latent Diffusion Model for 3D (LDM3D) that
generates both image and depth map data from a given text prompt, allowing
users to generate RGBD images from text prompts. The LDM3D model is fine-tuned
on a dataset of tuples containing an RGB image, depth map and caption, and
validated through extensive experiments. We also develop an application called
DepthFusion, which uses the generated RGB images and depth maps to create
immersive and interactive 360-degree-view experiences using TouchDesigner. This
technology has the potential to transform a wide range of industries, from
entertainment and gaming to architecture and design. Overall, this paper
presents a significant contribution to the field of generative AI and computer
vision, and showcases the potential of LDM3D and DepthFusion to revolutionize
content creation and digital experiences. A short video summarizing the
approach can be found at https://t.ly/tdi2.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 10:15:06 GMT"
},
{
"version": "v2",
"created": "Sun, 21 May 2023 20:26:30 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Stan",
"Gabriela Ben Melech",
""
],
[
"Wofk",
"Diana",
""
],
[
"Fox",
"Scottie",
""
],
[
"Redden",
"Alex",
""
],
[
"Saxton",
"Will",
""
],
[
"Yu",
"Jean",
""
],
[
"Aflalo",
"Estelle",
""
],
[
"Tseng",
"Shao-Yen",
""
],
[
"Nonato",
"Fabio",
""
],
[
"Muller",
"Matthias",
""
],
[
"Lal",
"Vasudev",
""
]
] | TITLE: LDM3D: Latent Diffusion Model for 3D
ABSTRACT: This research paper proposes a Latent Diffusion Model for 3D (LDM3D) that
generates both image and depth map data from a given text prompt, allowing
users to generate RGBD images from text prompts. The LDM3D model is fine-tuned
on a dataset of tuples containing an RGB image, depth map and caption, and
validated through extensive experiments. We also develop an application called
DepthFusion, which uses the generated RGB images and depth maps to create
immersive and interactive 360-degree-view experiences using TouchDesigner. This
technology has the potential to transform a wide range of industries, from
entertainment and gaming to architecture and design. Overall, this paper
presents a significant contribution to the field of generative AI and computer
vision, and showcases the potential of LDM3D and DepthFusion to revolutionize
content creation and digital experiences. A short video summarizing the
approach can be found at https://t.ly/tdi2.
| no_new_dataset | 0.668204 |
2305.11204 | Mithil Sai Jakka | Mithil Sai Jakka | Assessing Exoplanet Habitability through Data-driven Approaches: A
Comprehensive Literature Review | null | null | null | null | cs.OH | http://creativecommons.org/licenses/by/4.0/ | The exploration and study of exoplanets remain at the frontier of
astronomical research, challenging scientists to continuously innovate and
refine methodologies to navigate the vast, complex data these celestial bodies
produce. This literature the review aims to illuminate the emerging trends and
advancements within this sphere, specifically focusing on the interplay between
exoplanet detection, classification, and visualization, and the the
increasingly pivotal role of machine learning and computational models. Our
journey through this realm of exploration commences with a comprehensive
analysis of fifteen meticulously selected, seminal papers in the field. These
papers, each representing a distinct facet of exoplanet research, collectively
offer a multi-dimensional perspective on the current state of the field. They
provide valuable insights into the innovative application of machine learning
techniques to overcome the challenges posed by the analysis and interpretation
of astronomical data. From the application of Support Vector Machines (SVM) to
Deep Learning models, the review encapsulates the broad spectrum of machine
learning approaches employed in exoplanet research. The review also seeks to
unravel the story woven by the data within these papers, detailing the triumphs
and tribulations of the field. It highlights the increasing reliance on diverse
datasets, such as Kepler and TESS, and the push for improved accuracy in
exoplanet detection and classification models. The narrative concludes with key
takeaways and insights, drawing together the threads of research to present a
cohesive picture of the direction in which the field is moving. This literature
review, therefore, serves not just as an academic exploration, but also as a
narrative of scientific discovery and innovation in the quest to understand our
cosmic neighborhood.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 17:18:15 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Jakka",
"Mithil Sai",
""
]
] | TITLE: Assessing Exoplanet Habitability through Data-driven Approaches: A
Comprehensive Literature Review
ABSTRACT: The exploration and study of exoplanets remain at the frontier of
astronomical research, challenging scientists to continuously innovate and
refine methodologies to navigate the vast, complex data these celestial bodies
produce. This literature the review aims to illuminate the emerging trends and
advancements within this sphere, specifically focusing on the interplay between
exoplanet detection, classification, and visualization, and the the
increasingly pivotal role of machine learning and computational models. Our
journey through this realm of exploration commences with a comprehensive
analysis of fifteen meticulously selected, seminal papers in the field. These
papers, each representing a distinct facet of exoplanet research, collectively
offer a multi-dimensional perspective on the current state of the field. They
provide valuable insights into the innovative application of machine learning
techniques to overcome the challenges posed by the analysis and interpretation
of astronomical data. From the application of Support Vector Machines (SVM) to
Deep Learning models, the review encapsulates the broad spectrum of machine
learning approaches employed in exoplanet research. The review also seeks to
unravel the story woven by the data within these papers, detailing the triumphs
and tribulations of the field. It highlights the increasing reliance on diverse
datasets, such as Kepler and TESS, and the push for improved accuracy in
exoplanet detection and classification models. The narrative concludes with key
takeaways and insights, drawing together the threads of research to present a
cohesive picture of the direction in which the field is moving. This literature
review, therefore, serves not just as an academic exploration, but also as a
narrative of scientific discovery and innovation in the quest to understand our
cosmic neighborhood.
| no_new_dataset | 0.702925 |
2305.11700 | Ruyu Li | Ruyu Li, Wenhao Deng, Yu Cheng, Zheng Yuan, Jiaqi Zhang, Fajie Yuan | Exploring the Upper Limits of Text-Based Collaborative Filtering Using
Large Language Models: Discoveries and Insights | null | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-based collaborative filtering (TCF) has become the mainstream approach
for text and news recommendation, utilizing text encoders, also known as
language models (LMs), to represent items. However, existing TCF models
primarily focus on using small or medium-sized LMs. It remains uncertain what
impact replacing the item encoder with one of the largest and most powerful
LMs, such as the 175-billion parameter GPT-3 model, would have on
recommendation performance. Can we expect unprecedented results? To this end,
we conduct an extensive series of experiments aimed at exploring the
performance limits of the TCF paradigm. Specifically, we increase the size of
item encoders from one hundred million to one hundred billion to reveal the
scaling limits of the TCF paradigm. We then examine whether these extremely
large LMs could enable a universal item representation for the recommendation
task. Furthermore, we compare the performance of the TCF paradigm utilizing the
most powerful LMs to the currently dominant ID embedding-based paradigm and
investigate the transferability of this TCF paradigm. Finally, we compare TCF
with the recently popularized prompt-based recommendation using ChatGPT. Our
research findings have not only yielded positive results but also uncovered
some surprising and previously unknown negative outcomes, which can inspire
deeper reflection and innovative thinking regarding text-based recommender
systems. Codes and datasets will be released for further research.
| [
{
"version": "v1",
"created": "Fri, 19 May 2023 14:24:00 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Li",
"Ruyu",
""
],
[
"Deng",
"Wenhao",
""
],
[
"Cheng",
"Yu",
""
],
[
"Yuan",
"Zheng",
""
],
[
"Zhang",
"Jiaqi",
""
],
[
"Yuan",
"Fajie",
""
]
] | TITLE: Exploring the Upper Limits of Text-Based Collaborative Filtering Using
Large Language Models: Discoveries and Insights
ABSTRACT: Text-based collaborative filtering (TCF) has become the mainstream approach
for text and news recommendation, utilizing text encoders, also known as
language models (LMs), to represent items. However, existing TCF models
primarily focus on using small or medium-sized LMs. It remains uncertain what
impact replacing the item encoder with one of the largest and most powerful
LMs, such as the 175-billion parameter GPT-3 model, would have on
recommendation performance. Can we expect unprecedented results? To this end,
we conduct an extensive series of experiments aimed at exploring the
performance limits of the TCF paradigm. Specifically, we increase the size of
item encoders from one hundred million to one hundred billion to reveal the
scaling limits of the TCF paradigm. We then examine whether these extremely
large LMs could enable a universal item representation for the recommendation
task. Furthermore, we compare the performance of the TCF paradigm utilizing the
most powerful LMs to the currently dominant ID embedding-based paradigm and
investigate the transferability of this TCF paradigm. Finally, we compare TCF
with the recently popularized prompt-based recommendation using ChatGPT. Our
research findings have not only yielded positive results but also uncovered
some surprising and previously unknown negative outcomes, which can inspire
deeper reflection and innovative thinking regarding text-based recommender
systems. Codes and datasets will be released for further research.
| no_new_dataset | 0.709868 |
2305.11871 | Bala Murugan MS | Srija Santhanam, Kavipriya P, Balamurugan MS, Manoj Kumar Rajagopal | Amity -- A Hybrid Mental Health Application | eighteen pages and seven figure | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Wellness in trivial terms combines physical, social, and mental wellbeing.
While mental health is neglected, long-term success in a person life is mostly
determined by his psychological health and contentment. For a person in
distress, professional mental health services are quite expensive, unpopular,
and invite a lot of hesitation. Hence, it would be effective to use an Android
application that can offer day to day therapeutic assistance, meditation
sessions, and guidance since it can cater to a massive community instantly. In
this paper, we propose a mobile and web application AMITY with a chat group and
chatbot created using a machine learning approach. We have also built a dataset
to train the chatbot model that we propose in this paper. We briefly introduce
the dataset and the machine learning model in section 3. In section 4, we
include the architecture and the development details of the Hybrid application.
Next, we present our results on usability and the efficiency of the idea we
propose.
| [
{
"version": "v1",
"created": "Tue, 18 Apr 2023 06:26:53 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Santhanam",
"Srija",
""
],
[
"P",
"Kavipriya",
""
],
[
"MS",
"Balamurugan",
""
],
[
"Rajagopal",
"Manoj Kumar",
""
]
] | TITLE: Amity -- A Hybrid Mental Health Application
ABSTRACT: Wellness in trivial terms combines physical, social, and mental wellbeing.
While mental health is neglected, long-term success in a person life is mostly
determined by his psychological health and contentment. For a person in
distress, professional mental health services are quite expensive, unpopular,
and invite a lot of hesitation. Hence, it would be effective to use an Android
application that can offer day to day therapeutic assistance, meditation
sessions, and guidance since it can cater to a massive community instantly. In
this paper, we propose a mobile and web application AMITY with a chat group and
chatbot created using a machine learning approach. We have also built a dataset
to train the chatbot model that we propose in this paper. We briefly introduce
the dataset and the machine learning model in section 3. In section 4, we
include the architecture and the development details of the Hybrid application.
Next, we present our results on usability and the efficiency of the idea we
propose.
| new_dataset | 0.715206 |
2305.11905 | Pascal Pernot | Pascal Pernot | Properties of the ENCE and other MAD-based calibration metrics | null | null | null | null | cs.LG physics.chem-ph physics.data-an stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Expected Normalized Calibration Error (ENCE) is a popular calibration
statistic used in Machine Learning to assess the quality of prediction
uncertainties for regression problems. Estimation of the ENCE is based on the
binning of calibration data. In this short note, I illustrate an annoying
property of the ENCE, i.e. its proportionality to the square root of the number
of bins for well calibrated or nearly calibrated datasets. A similar behavior
affects the calibration error based on the variance of z-scores (ZVE), and in
both cases this property is a consequence of the use of a Mean Absolute
Deviation (MAD) statistic to estimate calibration errors. Hence, the question
arises of which number of bins to choose for a reliable estimation of
calibration error statistics. A solution is proposed to infer ENCE and ZVE
values that do not depend on the number of bins for datasets assumed to be
calibrated, providing simultaneously a statistical calibration test. It is also
shown that the ZVE is less sensitive than the ENCE to outstanding errors or
uncertainties.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 08:51:42 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Pernot",
"Pascal",
""
]
] | TITLE: Properties of the ENCE and other MAD-based calibration metrics
ABSTRACT: The Expected Normalized Calibration Error (ENCE) is a popular calibration
statistic used in Machine Learning to assess the quality of prediction
uncertainties for regression problems. Estimation of the ENCE is based on the
binning of calibration data. In this short note, I illustrate an annoying
property of the ENCE, i.e. its proportionality to the square root of the number
of bins for well calibrated or nearly calibrated datasets. A similar behavior
affects the calibration error based on the variance of z-scores (ZVE), and in
both cases this property is a consequence of the use of a Mean Absolute
Deviation (MAD) statistic to estimate calibration errors. Hence, the question
arises of which number of bins to choose for a reliable estimation of
calibration error statistics. A solution is proposed to infer ENCE and ZVE
values that do not depend on the number of bins for datasets assumed to be
calibrated, providing simultaneously a statistical calibration test. It is also
shown that the ZVE is less sensitive than the ENCE to outstanding errors or
uncertainties.
| no_new_dataset | 0.709466 |
2305.11921 | Ali Ismail-Fawaz | Ali Ismail-Fawaz, Angus Dempster, Chang Wei Tan, Matthieu Herrmann,
Lynn Miller, Daniel F. Schmidt, Stefano Berretti, Jonathan Weber, Maxime
Devanne, Germain Forestier, Geoffrey I. Webb | An Approach to Multiple Comparison Benchmark Evaluations that is Stable
Under Manipulation of the Comparate Set | null | null | null | null | stat.ME cs.AI cs.LG cs.PF | http://creativecommons.org/licenses/by/4.0/ | The measurement of progress using benchmarks evaluations is ubiquitous in
computer science and machine learning. However, common approaches to analyzing
and presenting the results of benchmark comparisons of multiple algorithms over
multiple datasets, such as the critical difference diagram introduced by
Dem\v{s}ar (2006), have important shortcomings and, we show, are open to both
inadvertent and intentional manipulation. To address these issues, we propose a
new approach to presenting the results of benchmark comparisons, the Multiple
Comparison Matrix (MCM), that prioritizes pairwise comparisons and precludes
the means of manipulating experimental results in existing approaches. MCM can
be used to show the results of an all-pairs comparison, or to show the results
of a comparison between one or more selected algorithms and the state of the
art. MCM is implemented in Python and is publicly available.
| [
{
"version": "v1",
"created": "Fri, 19 May 2023 08:58:55 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Ismail-Fawaz",
"Ali",
""
],
[
"Dempster",
"Angus",
""
],
[
"Tan",
"Chang Wei",
""
],
[
"Herrmann",
"Matthieu",
""
],
[
"Miller",
"Lynn",
""
],
[
"Schmidt",
"Daniel F.",
""
],
[
"Berretti",
"Stefano",
""
],
[
"Weber",
"Jonathan",
""
],
[
"Devanne",
"Maxime",
""
],
[
"Forestier",
"Germain",
""
],
[
"Webb",
"Geoffrey I.",
""
]
] | TITLE: An Approach to Multiple Comparison Benchmark Evaluations that is Stable
Under Manipulation of the Comparate Set
ABSTRACT: The measurement of progress using benchmarks evaluations is ubiquitous in
computer science and machine learning. However, common approaches to analyzing
and presenting the results of benchmark comparisons of multiple algorithms over
multiple datasets, such as the critical difference diagram introduced by
Dem\v{s}ar (2006), have important shortcomings and, we show, are open to both
inadvertent and intentional manipulation. To address these issues, we propose a
new approach to presenting the results of benchmark comparisons, the Multiple
Comparison Matrix (MCM), that prioritizes pairwise comparisons and precludes
the means of manipulating experimental results in existing approaches. MCM can
be used to show the results of an all-pairs comparison, or to show the results
of a comparison between one or more selected algorithms and the state of the
art. MCM is implemented in Python and is publicly available.
| no_new_dataset | 0.709038 |
2305.11965 | Zi-Hao Qiu | Zi-Hao Qiu, Quanqi Hu, Zhuoning Yuan, Denny Zhou, Lijun Zhang, Tianbao
Yang | Not All Semantics are Created Equal: Contrastive Self-supervised
Learning with Automatic Temperature Individualization | 33 pages, 11 figures, accepted by ICML2023 | null | null | null | cs.LG cs.AI math.OC stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we aim to optimize a contrastive loss with individualized
temperatures in a principled and systematic manner for self-supervised
learning. The common practice of using a global temperature parameter $\tau$
ignores the fact that ``not all semantics are created equal", meaning that
different anchor data may have different numbers of samples with similar
semantics, especially when data exhibits long-tails. First, we propose a new
robust contrastive loss inspired by distributionally robust optimization (DRO),
providing us an intuition about the effect of $\tau$ and a mechanism for
automatic temperature individualization. Then, we propose an efficient
stochastic algorithm for optimizing the robust contrastive loss with a provable
convergence guarantee without using large mini-batch sizes. Theoretical and
experimental results show that our algorithm automatically learns a suitable
$\tau$ for each sample. Specifically, samples with frequent semantics use large
temperatures to keep local semantic structures, while samples with rare
semantics use small temperatures to induce more separable features. Our method
not only outperforms prior strong baselines (e.g., SimCLR, CLIP) on unimodal
and bimodal datasets with larger improvements on imbalanced data but also is
less sensitive to hyper-parameters. To our best knowledge, this is the first
methodical approach to optimizing a contrastive loss with individualized
temperatures.
| [
{
"version": "v1",
"created": "Fri, 19 May 2023 19:25:56 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Qiu",
"Zi-Hao",
""
],
[
"Hu",
"Quanqi",
""
],
[
"Yuan",
"Zhuoning",
""
],
[
"Zhou",
"Denny",
""
],
[
"Zhang",
"Lijun",
""
],
[
"Yang",
"Tianbao",
""
]
] | TITLE: Not All Semantics are Created Equal: Contrastive Self-supervised
Learning with Automatic Temperature Individualization
ABSTRACT: In this paper, we aim to optimize a contrastive loss with individualized
temperatures in a principled and systematic manner for self-supervised
learning. The common practice of using a global temperature parameter $\tau$
ignores the fact that ``not all semantics are created equal", meaning that
different anchor data may have different numbers of samples with similar
semantics, especially when data exhibits long-tails. First, we propose a new
robust contrastive loss inspired by distributionally robust optimization (DRO),
providing us an intuition about the effect of $\tau$ and a mechanism for
automatic temperature individualization. Then, we propose an efficient
stochastic algorithm for optimizing the robust contrastive loss with a provable
convergence guarantee without using large mini-batch sizes. Theoretical and
experimental results show that our algorithm automatically learns a suitable
$\tau$ for each sample. Specifically, samples with frequent semantics use large
temperatures to keep local semantic structures, while samples with rare
semantics use small temperatures to induce more separable features. Our method
not only outperforms prior strong baselines (e.g., SimCLR, CLIP) on unimodal
and bimodal datasets with larger improvements on imbalanced data but also is
less sensitive to hyper-parameters. To our best knowledge, this is the first
methodical approach to optimizing a contrastive loss with individualized
temperatures.
| no_new_dataset | 0.710239 |
2305.11979 | Siddharth Varia | Robert Vacareanu, Siddharth Varia, Kishaloy Halder, Shuai Wang,
Giovanni Paolini, Neha Anna John, Miguel Ballesteros, Smaranda Muresan | A Weak Supervision Approach for Few-Shot Aspect Based Sentiment | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We explore how weak supervision on abundant unlabeled data can be leveraged
to improve few-shot performance in aspect-based sentiment analysis (ABSA)
tasks. We propose a pipeline approach to construct a noisy ABSA dataset, and we
use it to adapt a pre-trained sequence-to-sequence model to the ABSA tasks. We
test the resulting model on three widely used ABSA datasets, before and after
fine-tuning. Our proposed method preserves the full fine-tuning performance
while showing significant improvements (15.84% absolute F1) in the few-shot
learning scenario for the harder tasks. In zero-shot (i.e., without
fine-tuning), our method outperforms the previous state of the art on the
aspect extraction sentiment classification (AESC) task and is, additionally,
capable of performing the harder aspect sentiment triplet extraction (ASTE)
task.
| [
{
"version": "v1",
"created": "Fri, 19 May 2023 19:53:54 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Vacareanu",
"Robert",
""
],
[
"Varia",
"Siddharth",
""
],
[
"Halder",
"Kishaloy",
""
],
[
"Wang",
"Shuai",
""
],
[
"Paolini",
"Giovanni",
""
],
[
"John",
"Neha Anna",
""
],
[
"Ballesteros",
"Miguel",
""
],
[
"Muresan",
"Smaranda",
""
]
] | TITLE: A Weak Supervision Approach for Few-Shot Aspect Based Sentiment
ABSTRACT: We explore how weak supervision on abundant unlabeled data can be leveraged
to improve few-shot performance in aspect-based sentiment analysis (ABSA)
tasks. We propose a pipeline approach to construct a noisy ABSA dataset, and we
use it to adapt a pre-trained sequence-to-sequence model to the ABSA tasks. We
test the resulting model on three widely used ABSA datasets, before and after
fine-tuning. Our proposed method preserves the full fine-tuning performance
while showing significant improvements (15.84% absolute F1) in the few-shot
learning scenario for the harder tasks. In zero-shot (i.e., without
fine-tuning), our method outperforms the previous state of the art on the
aspect extraction sentiment classification (AESC) task and is, additionally,
capable of performing the harder aspect sentiment triplet extraction (ASTE)
task.
| no_new_dataset | 0.703295 |
2305.11999 | Tal Kadosh | Tal Kadosh, Nadav Schneider, Niranjan Hasabnis, Timothy Mattson, Yuval
Pinter, and Gal Oren | Advising OpenMP Parallelization via a Graph-Based Approach with
Transformers | null | null | null | null | cs.DC cs.AI cs.LG cs.PF | http://creativecommons.org/licenses/by/4.0/ | There is an ever-present need for shared memory parallelization schemes to
exploit the full potential of multi-core architectures. The most common
parallelization API addressing this need today is OpenMP. Nevertheless, writing
parallel code manually is complex and effort-intensive. Thus, many
deterministic source-to-source (S2S) compilers have emerged, intending to
automate the process of translating serial to parallel code. However, recent
studies have shown that these compilers are impractical in many scenarios. In
this work, we combine the latest advancements in the field of AI and natural
language processing (NLP) with the vast amount of open-source code to address
the problem of automatic parallelization. Specifically, we propose a novel
approach, called OMPify, to detect and predict the OpenMP pragmas and
shared-memory attributes in parallel code, given its serial version. OMPify is
based on a Transformer-based model that leverages a graph-based representation
of source code that exploits the inherent structure of code. We evaluated our
tool by predicting the parallelization pragmas and attributes of a large corpus
of (over 54,000) snippets of serial code written in C and C++ languages
(Open-OMP-Plus). Our results demonstrate that OMPify outperforms existing
approaches, the general-purposed and popular ChatGPT and targeted PragFormer
models, in terms of F1 score and accuracy. Specifically, OMPify achieves up to
90% accuracy on commonly-used OpenMP benchmark tests such as NAS, SPEC, and
PolyBench. Additionally, we performed an ablation study to assess the impact of
different model components and present interesting insights derived from the
study. Lastly, we also explored the potential of using data augmentation and
curriculum learning techniques to improve the model's robustness and
generalization capabilities.
| [
{
"version": "v1",
"created": "Tue, 16 May 2023 16:56:10 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Kadosh",
"Tal",
""
],
[
"Schneider",
"Nadav",
""
],
[
"Hasabnis",
"Niranjan",
""
],
[
"Mattson",
"Timothy",
""
],
[
"Pinter",
"Yuval",
""
],
[
"Oren",
"Gal",
""
]
] | TITLE: Advising OpenMP Parallelization via a Graph-Based Approach with
Transformers
ABSTRACT: There is an ever-present need for shared memory parallelization schemes to
exploit the full potential of multi-core architectures. The most common
parallelization API addressing this need today is OpenMP. Nevertheless, writing
parallel code manually is complex and effort-intensive. Thus, many
deterministic source-to-source (S2S) compilers have emerged, intending to
automate the process of translating serial to parallel code. However, recent
studies have shown that these compilers are impractical in many scenarios. In
this work, we combine the latest advancements in the field of AI and natural
language processing (NLP) with the vast amount of open-source code to address
the problem of automatic parallelization. Specifically, we propose a novel
approach, called OMPify, to detect and predict the OpenMP pragmas and
shared-memory attributes in parallel code, given its serial version. OMPify is
based on a Transformer-based model that leverages a graph-based representation
of source code that exploits the inherent structure of code. We evaluated our
tool by predicting the parallelization pragmas and attributes of a large corpus
of (over 54,000) snippets of serial code written in C and C++ languages
(Open-OMP-Plus). Our results demonstrate that OMPify outperforms existing
approaches, the general-purposed and popular ChatGPT and targeted PragFormer
models, in terms of F1 score and accuracy. Specifically, OMPify achieves up to
90% accuracy on commonly-used OpenMP benchmark tests such as NAS, SPEC, and
PolyBench. Additionally, we performed an ablation study to assess the impact of
different model components and present interesting insights derived from the
study. Lastly, we also explored the potential of using data augmentation and
curriculum learning techniques to improve the model's robustness and
generalization capabilities.
| no_new_dataset | 0.707777 |
2305.12000 | Kai North | Kai North, Tharindu Ranasinghe, Matthew Shardlow, Marcos Zampieri | Deep Learning Approaches to Lexical Simplification: A Survey | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Lexical Simplification (LS) is the task of replacing complex for simpler
words in a sentence whilst preserving the sentence's original meaning. LS is
the lexical component of Text Simplification (TS) with the aim of making texts
more accessible to various target populations. A past survey (Paetzold and
Specia, 2017) has provided a detailed overview of LS. Since this survey,
however, the AI/NLP community has been taken by storm by recent advances in
deep learning, particularly with the introduction of large language models
(LLM) and prompt learning. The high performance of these models sparked renewed
interest in LS. To reflect these recent advances, we present a comprehensive
survey of papers published between 2017 and 2023 on LS and its sub-tasks with a
special focus on deep learning. We also present benchmark datasets for the
future development of LS systems.
| [
{
"version": "v1",
"created": "Fri, 19 May 2023 20:56:22 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"North",
"Kai",
""
],
[
"Ranasinghe",
"Tharindu",
""
],
[
"Shardlow",
"Matthew",
""
],
[
"Zampieri",
"Marcos",
""
]
] | TITLE: Deep Learning Approaches to Lexical Simplification: A Survey
ABSTRACT: Lexical Simplification (LS) is the task of replacing complex for simpler
words in a sentence whilst preserving the sentence's original meaning. LS is
the lexical component of Text Simplification (TS) with the aim of making texts
more accessible to various target populations. A past survey (Paetzold and
Specia, 2017) has provided a detailed overview of LS. Since this survey,
however, the AI/NLP community has been taken by storm by recent advances in
deep learning, particularly with the introduction of large language models
(LLM) and prompt learning. The high performance of these models sparked renewed
interest in LS. To reflect these recent advances, we present a comprehensive
survey of papers published between 2017 and 2023 on LS and its sub-tasks with a
special focus on deep learning. We also present benchmark datasets for the
future development of LS systems.
| new_dataset | 0.717396 |
2305.12019 | Srivathsan Amruth | Srivathsan Amruth, Xin Yee Lam | A Foray into Parallel Optimisation Algorithms for High Dimension Low
Sample Space Generalized Distance Weighted Discrimination problems | 34 pages. arXiv admin note: substantial text overlap with
arXiv:1604.05473 | null | null | null | math.OC cs.DC | http://creativecommons.org/licenses/by/4.0/ | In many modern data sets, High dimension low sample size (HDLSS) data is
prevalent in many fields of studies. There has been an increased focus recently
on using machine learning and statistical methods to mine valuable information
out of these data sets. Thus, there has been an increased interest in efficient
learning in high dimensions. Naturally, as the dimension of the input data
increases, the learning task will become more difficult, due to increasing
computational and statistical complexities. This makes it crucial to overcome
the curse of dimensionality in a given dataset, within a reasonable time frame,
in a bid to obtain the insights required to keep a competitive edge. To solve
HDLSS problems, classical methods such as support vector machines can be
utilised to alleviate data piling at the margin. However, when we question
geometric domains and their assumptions on input data, we are naturally lead to
convex optimisation problems and this gives rise to the development of
solutions like distance weighted discrimination (DWD), which can be modelled as
a second-order cone programming problem and solved by interior-point methods
when sample size and feature dimensions of the data is moderate. In this paper,
our focus is on designing an even more scalable and robust algorithm for
solving large-scale generalized DWD problems.
| [
{
"version": "v1",
"created": "Fri, 19 May 2023 22:06:22 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Amruth",
"Srivathsan",
""
],
[
"Lam",
"Xin Yee",
""
]
] | TITLE: A Foray into Parallel Optimisation Algorithms for High Dimension Low
Sample Space Generalized Distance Weighted Discrimination problems
ABSTRACT: In many modern data sets, High dimension low sample size (HDLSS) data is
prevalent in many fields of studies. There has been an increased focus recently
on using machine learning and statistical methods to mine valuable information
out of these data sets. Thus, there has been an increased interest in efficient
learning in high dimensions. Naturally, as the dimension of the input data
increases, the learning task will become more difficult, due to increasing
computational and statistical complexities. This makes it crucial to overcome
the curse of dimensionality in a given dataset, within a reasonable time frame,
in a bid to obtain the insights required to keep a competitive edge. To solve
HDLSS problems, classical methods such as support vector machines can be
utilised to alleviate data piling at the margin. However, when we question
geometric domains and their assumptions on input data, we are naturally lead to
convex optimisation problems and this gives rise to the development of
solutions like distance weighted discrimination (DWD), which can be modelled as
a second-order cone programming problem and solved by interior-point methods
when sample size and feature dimensions of the data is moderate. In this paper,
our focus is on designing an even more scalable and robust algorithm for
solving large-scale generalized DWD problems.
| no_new_dataset | 0.707992 |
2305.12058 | Menglin Kong | Menglin Kong, Muzhou Hou, Shaojie Zhao, Feng Liu, Ri Su and Yinghao
Chen | DADIN: Domain Adversarial Deep Interest Network for Cross Domain
Recommender Systems | null | null | null | null | cs.IR cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Click-Through Rate (CTR) prediction is one of the main tasks of the
recommendation system, which is conducted by a user for different items to give
the recommendation results. Cross-domain CTR prediction models have been
proposed to overcome problems of data sparsity, long tail distribution of
user-item interactions, and cold start of items or users. In order to make
knowledge transfer from source domain to target domain more smoothly, an
innovative deep learning cross-domain CTR prediction model, Domain Adversarial
Deep Interest Network (DADIN) is proposed to convert the cross-domain
recommendation task into a domain adaptation problem. The joint distribution
alignment of two domains is innovatively realized by introducing domain
agnostic layers and specially designed loss, and optimized together with CTR
prediction loss in a way of adversarial training. It is found that the Area
Under Curve (AUC) of DADIN is 0.08% higher than the most competitive baseline
on Huawei dataset and is 0.71% higher than its competitors on Amazon dataset,
achieving the state-of-the-art results on the basis of the evaluation of this
model performance on two real datasets. The ablation study shows that by
introducing adversarial method, this model has respectively led to the AUC
improvements of 2.34% on Huawei dataset and 16.67% on Amazon dataset.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 01:56:29 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Kong",
"Menglin",
""
],
[
"Hou",
"Muzhou",
""
],
[
"Zhao",
"Shaojie",
""
],
[
"Liu",
"Feng",
""
],
[
"Su",
"Ri",
""
],
[
"Chen",
"Yinghao",
""
]
] | TITLE: DADIN: Domain Adversarial Deep Interest Network for Cross Domain
Recommender Systems
ABSTRACT: Click-Through Rate (CTR) prediction is one of the main tasks of the
recommendation system, which is conducted by a user for different items to give
the recommendation results. Cross-domain CTR prediction models have been
proposed to overcome problems of data sparsity, long tail distribution of
user-item interactions, and cold start of items or users. In order to make
knowledge transfer from source domain to target domain more smoothly, an
innovative deep learning cross-domain CTR prediction model, Domain Adversarial
Deep Interest Network (DADIN) is proposed to convert the cross-domain
recommendation task into a domain adaptation problem. The joint distribution
alignment of two domains is innovatively realized by introducing domain
agnostic layers and specially designed loss, and optimized together with CTR
prediction loss in a way of adversarial training. It is found that the Area
Under Curve (AUC) of DADIN is 0.08% higher than the most competitive baseline
on Huawei dataset and is 0.71% higher than its competitors on Amazon dataset,
achieving the state-of-the-art results on the basis of the evaluation of this
model performance on two real datasets. The ablation study shows that by
introducing adversarial method, this model has respectively led to the AUC
improvements of 2.34% on Huawei dataset and 16.67% on Amazon dataset.
| no_new_dataset | 0.711481 |
2305.12068 | Hui Li | Hui Li, Carlos A. Pena Solorzano, Susan Wei, Davis J. McCarthy | Technical outlier detection via convolutional variational autoencoder
for the ADMANI breast mammogram dataset | null | null | null | null | eess.IV cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | The ADMANI datasets (annotated digital mammograms and associated non-image
datasets) from the Transforming Breast Cancer Screening with AI programme
(BRAIx) run by BreastScreen Victoria in Australia are multi-centre, large
scale, clinically curated, real-world databases. The datasets are expected to
aid in the development of clinically relevant Artificial Intelligence (AI)
algorithms for breast cancer detection, early diagnosis, and other
applications. To ensure high data quality, technical outliers must be removed
before any downstream algorithm development. As a first step, we randomly
select 30,000 individual mammograms and use Convolutional Variational
Autoencoder (CVAE), a deep generative neural network, to detect outliers. CVAE
is expected to detect all sorts of outliers, although its detection performance
differs among different types of outliers. Traditional image processing
techniques such as erosion and pectoral muscle analysis can compensate for the
poor performance of CVAE in certain outlier types. We identify seven types of
technical outliers: implant, pacemaker, cardiac loop recorder, improper
radiography, atypical lesion/calcification, incorrect exposure parameter and
improper placement. The outlier recall rate for the test set is 61% if CVAE,
erosion and pectoral muscle analysis each select the top 1% images ranked in
ascending or descending order according to image outlier score under each
detection method, and 83% if each selects the top 5% images. This study offers
an overview of technical outliers in the ADMANI dataset and suggests future
directions to improve outlier detection effectiveness.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 03:08:42 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Li",
"Hui",
""
],
[
"Solorzano",
"Carlos A. Pena",
""
],
[
"Wei",
"Susan",
""
],
[
"McCarthy",
"Davis J.",
""
]
] | TITLE: Technical outlier detection via convolutional variational autoencoder
for the ADMANI breast mammogram dataset
ABSTRACT: The ADMANI datasets (annotated digital mammograms and associated non-image
datasets) from the Transforming Breast Cancer Screening with AI programme
(BRAIx) run by BreastScreen Victoria in Australia are multi-centre, large
scale, clinically curated, real-world databases. The datasets are expected to
aid in the development of clinically relevant Artificial Intelligence (AI)
algorithms for breast cancer detection, early diagnosis, and other
applications. To ensure high data quality, technical outliers must be removed
before any downstream algorithm development. As a first step, we randomly
select 30,000 individual mammograms and use Convolutional Variational
Autoencoder (CVAE), a deep generative neural network, to detect outliers. CVAE
is expected to detect all sorts of outliers, although its detection performance
differs among different types of outliers. Traditional image processing
techniques such as erosion and pectoral muscle analysis can compensate for the
poor performance of CVAE in certain outlier types. We identify seven types of
technical outliers: implant, pacemaker, cardiac loop recorder, improper
radiography, atypical lesion/calcification, incorrect exposure parameter and
improper placement. The outlier recall rate for the test set is 61% if CVAE,
erosion and pectoral muscle analysis each select the top 1% images ranked in
ascending or descending order according to image outlier score under each
detection method, and 83% if each selects the top 5% images. This study offers
an overview of technical outliers in the ADMANI dataset and suggests future
directions to improve outlier detection effectiveness.
| no_new_dataset | 0.710019 |
2305.12070 | Weizhi Nie | Weizhi Nie, Chen Zhang, Dan song, Yunpeng Bai, Keliang Xie, Anan Liu | Instrumental Variable Learning for Chest X-ray Classification | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The chest X-ray (CXR) is commonly employed to diagnose thoracic illnesses,
but the challenge of achieving accurate automatic diagnosis through this method
persists due to the complex relationship between pathology. In recent years,
various deep learning-based approaches have been suggested to tackle this
problem but confounding factors such as image resolution or noise problems
often damage model performance. In this paper, we focus on the chest X-ray
classification task and proposed an interpretable instrumental variable (IV)
learning framework, to eliminate the spurious association and obtain accurate
causal representation. Specifically, we first construct a structural causal
model (SCM) for our task and learn the confounders and the preliminary
representations of IV, we then leverage electronic health record (EHR) as
auxiliary information and we fuse the above feature with our transformer-based
semantic fusion module, so the IV has the medical semantic. Meanwhile, the
reliability of IV is further guaranteed via the constraints of mutual
information between related causal variables. Finally, our approach's
performance is demonstrated using the MIMIC-CXR, NIH ChestX-ray 14, and
CheXpert datasets, and we achieve competitive results.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 03:12:23 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Nie",
"Weizhi",
""
],
[
"Zhang",
"Chen",
""
],
[
"song",
"Dan",
""
],
[
"Bai",
"Yunpeng",
""
],
[
"Xie",
"Keliang",
""
],
[
"Liu",
"Anan",
""
]
] | TITLE: Instrumental Variable Learning for Chest X-ray Classification
ABSTRACT: The chest X-ray (CXR) is commonly employed to diagnose thoracic illnesses,
but the challenge of achieving accurate automatic diagnosis through this method
persists due to the complex relationship between pathology. In recent years,
various deep learning-based approaches have been suggested to tackle this
problem but confounding factors such as image resolution or noise problems
often damage model performance. In this paper, we focus on the chest X-ray
classification task and proposed an interpretable instrumental variable (IV)
learning framework, to eliminate the spurious association and obtain accurate
causal representation. Specifically, we first construct a structural causal
model (SCM) for our task and learn the confounders and the preliminary
representations of IV, we then leverage electronic health record (EHR) as
auxiliary information and we fuse the above feature with our transformer-based
semantic fusion module, so the IV has the medical semantic. Meanwhile, the
reliability of IV is further guaranteed via the constraints of mutual
information between related causal variables. Finally, our approach's
performance is demonstrated using the MIMIC-CXR, NIH ChestX-ray 14, and
CheXpert datasets, and we achieve competitive results.
| no_new_dataset | 0.708584 |
2305.12087 | Gang Liu | Gang Liu, Tong Zhao, Eric Inae, Tengfei Luo, Meng Jiang | Semi-Supervised Graph Imbalanced Regression | Accepted by KDD 2023. 17 pages, 5 figures, 10 tables | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Data imbalance is easily found in annotated data when the observations of
certain continuous label values are difficult to collect for regression tasks.
When they come to molecule and polymer property predictions, the annotated
graph datasets are often small because labeling them requires expensive
equipment and effort. To address the lack of examples of rare label values in
graph regression tasks, we propose a semi-supervised framework to progressively
balance training data and reduce model bias via self-training. The training
data balance is achieved by (1) pseudo-labeling more graphs for
under-represented labels with a novel regression confidence measurement and (2)
augmenting graph examples in latent space for remaining rare labels after data
balancing with pseudo-labels. The former is to identify quality examples from
unlabeled data whose labels are confidently predicted and sample a subset of
them with a reverse distribution from the imbalanced annotated data. The latter
collaborates with the former to target a perfect balance using a novel
label-anchored mixup algorithm. We perform experiments in seven regression
tasks on graph datasets. Results demonstrate that the proposed framework
significantly reduces the error of predicted graph properties, especially in
under-represented label areas.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 04:11:00 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Liu",
"Gang",
""
],
[
"Zhao",
"Tong",
""
],
[
"Inae",
"Eric",
""
],
[
"Luo",
"Tengfei",
""
],
[
"Jiang",
"Meng",
""
]
] | TITLE: Semi-Supervised Graph Imbalanced Regression
ABSTRACT: Data imbalance is easily found in annotated data when the observations of
certain continuous label values are difficult to collect for regression tasks.
When they come to molecule and polymer property predictions, the annotated
graph datasets are often small because labeling them requires expensive
equipment and effort. To address the lack of examples of rare label values in
graph regression tasks, we propose a semi-supervised framework to progressively
balance training data and reduce model bias via self-training. The training
data balance is achieved by (1) pseudo-labeling more graphs for
under-represented labels with a novel regression confidence measurement and (2)
augmenting graph examples in latent space for remaining rare labels after data
balancing with pseudo-labels. The former is to identify quality examples from
unlabeled data whose labels are confidently predicted and sample a subset of
them with a reverse distribution from the imbalanced annotated data. The latter
collaborates with the former to target a perfect balance using a novel
label-anchored mixup algorithm. We perform experiments in seven regression
tasks on graph datasets. Results demonstrate that the proposed framework
significantly reduces the error of predicted graph properties, especially in
under-represented label areas.
| no_new_dataset | 0.713213 |
2305.12092 | Mike Zhang | Mike Zhang and Rob van der Goot and Barbara Plank | ESCOXLM-R: Multilingual Taxonomy-driven Pre-training for the Job Market
Domain | Accepted at ACL2023 (Main) | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The increasing number of benchmarks for Natural Language Processing (NLP)
tasks in the computational job market domain highlights the demand for methods
that can handle job-related tasks such as skill extraction, skill
classification, job title classification, and de-identification. While some
approaches have been developed that are specific to the job market domain,
there is a lack of generalized, multilingual models and benchmarks for these
tasks. In this study, we introduce a language model called ESCOXLM-R, based on
XLM-R, which uses domain-adaptive pre-training on the European Skills,
Competences, Qualifications and Occupations (ESCO) taxonomy, covering 27
languages. The pre-training objectives for ESCOXLM-R include dynamic masked
language modeling and a novel additional objective for inducing multilingual
taxonomical ESCO relations. We comprehensively evaluate the performance of
ESCOXLM-R on 6 sequence labeling and 3 classification tasks in 4 languages and
find that it achieves state-of-the-art results on 6 out of 9 datasets. Our
analysis reveals that ESCOXLM-R performs better on short spans and outperforms
XLM-R on entity-level and surface-level span-F1, likely due to ESCO containing
short skill and occupation titles, and encoding information on the
entity-level.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 04:50:20 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Zhang",
"Mike",
""
],
[
"van der Goot",
"Rob",
""
],
[
"Plank",
"Barbara",
""
]
] | TITLE: ESCOXLM-R: Multilingual Taxonomy-driven Pre-training for the Job Market
Domain
ABSTRACT: The increasing number of benchmarks for Natural Language Processing (NLP)
tasks in the computational job market domain highlights the demand for methods
that can handle job-related tasks such as skill extraction, skill
classification, job title classification, and de-identification. While some
approaches have been developed that are specific to the job market domain,
there is a lack of generalized, multilingual models and benchmarks for these
tasks. In this study, we introduce a language model called ESCOXLM-R, based on
XLM-R, which uses domain-adaptive pre-training on the European Skills,
Competences, Qualifications and Occupations (ESCO) taxonomy, covering 27
languages. The pre-training objectives for ESCOXLM-R include dynamic masked
language modeling and a novel additional objective for inducing multilingual
taxonomical ESCO relations. We comprehensively evaluate the performance of
ESCOXLM-R on 6 sequence labeling and 3 classification tasks in 4 languages and
find that it achieves state-of-the-art results on 6 out of 9 datasets. Our
analysis reveals that ESCOXLM-R performs better on short spans and outperforms
XLM-R on entity-level and surface-level span-F1, likely due to ESCO containing
short skill and occupation titles, and encoding information on the
entity-level.
| no_new_dataset | 0.708421 |
2305.12111 | Xiaomin Zeng | Xiao-Min Zeng, Yan Song, Zhu Zhuo, Yu Zhou, Yu-Hong Li, Hui Xue,
Li-Rong Dai, Ian McLoughlin | Joint Generative-Contrastive Representation Learning for Anomalous Sound
Detection | Accepted by ICASSP2023 | null | null | null | eess.AS cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we propose a joint generative and contrastive representation
learning method (GeCo) for anomalous sound detection (ASD). GeCo exploits a
Predictive AutoEncoder (PAE) equipped with self-attention as a generative model
to perform frame-level prediction. The output of the PAE together with original
normal samples, are used for supervised contrastive representative learning in
a multi-task framework. Besides cross-entropy loss between classes, contrastive
loss is used to separate PAE output and original samples within each class.
GeCo aims to better capture context information among frames, thanks to the
self-attention mechanism for PAE model. Furthermore, GeCo combines generative
and contrastive learning from which we aim to yield more effective and
informative representations, compared to existing methods. Extensive
experiments have been conducted on the DCASE2020 Task2 development dataset,
showing that GeCo outperforms state-of-the-art generative and discriminative
methods.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 06:10:53 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Zeng",
"Xiao-Min",
""
],
[
"Song",
"Yan",
""
],
[
"Zhuo",
"Zhu",
""
],
[
"Zhou",
"Yu",
""
],
[
"Li",
"Yu-Hong",
""
],
[
"Xue",
"Hui",
""
],
[
"Dai",
"Li-Rong",
""
],
[
"McLoughlin",
"Ian",
""
]
] | TITLE: Joint Generative-Contrastive Representation Learning for Anomalous Sound
Detection
ABSTRACT: In this paper, we propose a joint generative and contrastive representation
learning method (GeCo) for anomalous sound detection (ASD). GeCo exploits a
Predictive AutoEncoder (PAE) equipped with self-attention as a generative model
to perform frame-level prediction. The output of the PAE together with original
normal samples, are used for supervised contrastive representative learning in
a multi-task framework. Besides cross-entropy loss between classes, contrastive
loss is used to separate PAE output and original samples within each class.
GeCo aims to better capture context information among frames, thanks to the
self-attention mechanism for PAE model. Furthermore, GeCo combines generative
and contrastive learning from which we aim to yield more effective and
informative representations, compared to existing methods. Extensive
experiments have been conducted on the DCASE2020 Task2 development dataset,
showing that GeCo outperforms state-of-the-art generative and discriminative
methods.
| no_new_dataset | 0.708994 |
2305.12144 | Zefan Cai | Yufeng He, Zefan Cai, Xu Gan, Baobao Chang | DiffCap: Exploring Continuous Diffusion on Image Captioning | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current image captioning works usually focus on generating descriptions in an
autoregressive manner. However, there are limited works that focus on
generating descriptions non-autoregressively, which brings more decoding
diversity. Inspired by the success of diffusion models on generating
natural-looking images, we propose a novel method DiffCap to apply continuous
diffusions on image captioning. Unlike image generation where the output is
fixed-size and continuous, image description length varies with discrete
tokens. Our method transforms discrete tokens in a natural way and applies
continuous diffusion on them to successfully fuse extracted image features for
diffusion caption generation. Our experiments on COCO dataset demonstrate that
our method uses a much simpler structure to achieve comparable results to the
previous non-autoregressive works. Apart from quality, an intriguing property
of DiffCap is its high diversity during generation, which is missing from many
autoregressive models. We believe our method on fusing multimodal features in
diffusion language generation will inspire more researches on multimodal
language generation tasks for its simplicity and decoding flexibility.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 09:02:10 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"He",
"Yufeng",
""
],
[
"Cai",
"Zefan",
""
],
[
"Gan",
"Xu",
""
],
[
"Chang",
"Baobao",
""
]
] | TITLE: DiffCap: Exploring Continuous Diffusion on Image Captioning
ABSTRACT: Current image captioning works usually focus on generating descriptions in an
autoregressive manner. However, there are limited works that focus on
generating descriptions non-autoregressively, which brings more decoding
diversity. Inspired by the success of diffusion models on generating
natural-looking images, we propose a novel method DiffCap to apply continuous
diffusions on image captioning. Unlike image generation where the output is
fixed-size and continuous, image description length varies with discrete
tokens. Our method transforms discrete tokens in a natural way and applies
continuous diffusion on them to successfully fuse extracted image features for
diffusion caption generation. Our experiments on COCO dataset demonstrate that
our method uses a much simpler structure to achieve comparable results to the
previous non-autoregressive works. Apart from quality, an intriguing property
of DiffCap is its high diversity during generation, which is missing from many
autoregressive models. We believe our method on fusing multimodal features in
diffusion language generation will inspire more researches on multimodal
language generation tasks for its simplicity and decoding flexibility.
| no_new_dataset | 0.713425 |
2305.12158 | Ibrahim Ahmed | Ibrahim Ahmed and Marcos Quinones-Grueiro and Gautam Biswas | Model-based adaptation for sample efficient transfer in reinforcement
learning control of parameter-varying systems | Published to IEEE CoDiT 2023 | null | null | null | eess.SY cs.LG cs.SY | http://creativecommons.org/licenses/by/4.0/ | In this paper, we leverage ideas from model-based control to address the
sample efficiency problem of reinforcement learning (RL) algorithms.
Accelerating learning is an active field of RL highly relevant in the context
of time-varying systems. Traditional transfer learning methods propose to use
prior knowledge of the system behavior to devise a gradual or immediate
data-driven transformation of the control policy obtained through RL. Such
transformation is usually computed by estimating the performance of previous
control policies based on measurements recently collected from the system.
However, such retrospective measures have debatable utility with no guarantees
of positive transfer in most cases. Instead, we propose a model-based
transformation, such that when actions from a control policy are applied to the
target system, a positive transfer is achieved. The transformation can be used
as an initialization for the reinforcement learning process to converge to a
new optimum. We validate the performance of our approach through four benchmark
examples. We demonstrate that our approach is more sample-efficient than
fine-tuning with reinforcement learning alone and achieves comparable
performance to linear-quadratic-regulators and model-predictive control when an
accurate linear model is known in the three cases. If an accurate model is not
known, we empirically show that the proposed approach still guarantees positive
transfer with jump-start improvement.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 10:11:09 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Ahmed",
"Ibrahim",
""
],
[
"Quinones-Grueiro",
"Marcos",
""
],
[
"Biswas",
"Gautam",
""
]
] | TITLE: Model-based adaptation for sample efficient transfer in reinforcement
learning control of parameter-varying systems
ABSTRACT: In this paper, we leverage ideas from model-based control to address the
sample efficiency problem of reinforcement learning (RL) algorithms.
Accelerating learning is an active field of RL highly relevant in the context
of time-varying systems. Traditional transfer learning methods propose to use
prior knowledge of the system behavior to devise a gradual or immediate
data-driven transformation of the control policy obtained through RL. Such
transformation is usually computed by estimating the performance of previous
control policies based on measurements recently collected from the system.
However, such retrospective measures have debatable utility with no guarantees
of positive transfer in most cases. Instead, we propose a model-based
transformation, such that when actions from a control policy are applied to the
target system, a positive transfer is achieved. The transformation can be used
as an initialization for the reinforcement learning process to converge to a
new optimum. We validate the performance of our approach through four benchmark
examples. We demonstrate that our approach is more sample-efficient than
fine-tuning with reinforcement learning alone and achieves comparable
performance to linear-quadratic-regulators and model-predictive control when an
accurate linear model is known in the three cases. If an accurate model is not
known, we empirically show that the proposed approach still guarantees positive
transfer with jump-start improvement.
| no_new_dataset | 0.708975 |
2305.12190 | Zoran Medi\'c | Zoran Medi\'c, Jan \v{S}najder | Paragraph-level Citation Recommendation based on Topic Sentences as
Queries | null | null | null | null | cs.IR cs.CL | http://creativecommons.org/licenses/by/4.0/ | Citation recommendation (CR) models may help authors find relevant articles
at various stages of the paper writing process. Most research has dealt with
either global CR, which produces general recommendations suitable for the
initial writing stage, or local CR, which produces specific recommendations
more fitting for the final writing stages. We propose the task of
paragraph-level CR as a middle ground between the two approaches, where the
paragraph's topic sentence is taken as input and recommendations for citing
within the paragraph are produced at the output. We propose a model for this
task, fine-tune it using the quadruplet loss on the dataset of ACL papers, and
show improvements over the baselines.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 13:28:22 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Medić",
"Zoran",
""
],
[
"Šnajder",
"Jan",
""
]
] | TITLE: Paragraph-level Citation Recommendation based on Topic Sentences as
Queries
ABSTRACT: Citation recommendation (CR) models may help authors find relevant articles
at various stages of the paper writing process. Most research has dealt with
either global CR, which produces general recommendations suitable for the
initial writing stage, or local CR, which produces specific recommendations
more fitting for the final writing stages. We propose the task of
paragraph-level CR as a middle ground between the two approaches, where the
paragraph's topic sentence is taken as input and recommendations for citing
within the paragraph are produced at the output. We propose a model for this
task, fine-tune it using the quadruplet loss on the dataset of ACL papers, and
show improvements over the baselines.
| no_new_dataset | 0.71227 |
2305.12200 | Yuyue Wang | Yuyue Wang, Huan Xiao, Yihan Wu, Ruihua Song | ComedicSpeech: Text To Speech For Stand-up Comedies in Low-Resource
Scenarios | 5 pages, 4 tables, 2 figure | null | null | null | cs.SD cs.AI eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text to Speech (TTS) models can generate natural and high-quality speech, but
it is not expressive enough when synthesizing speech with dramatic
expressiveness, such as stand-up comedies. Considering comedians have diverse
personal speech styles, including personal prosody, rhythm, and fillers, it
requires real-world datasets and strong speech style modeling capabilities,
which brings challenges. In this paper, we construct a new dataset and develop
ComedicSpeech, a TTS system tailored for the stand-up comedy synthesis in
low-resource scenarios. First, we extract prosody representation by the prosody
encoder and condition it to the TTS model in a flexible way. Second, we enhance
the personal rhythm modeling by a conditional duration predictor. Third, we
model the personal fillers by introducing comedian-related special tokens.
Experiments show that ComedicSpeech achieves better expressiveness than
baselines with only ten-minute training data for each comedian. The audio
samples are available at https://xh621.github.io/stand-up-comedy-demo/
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 14:24:45 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Wang",
"Yuyue",
""
],
[
"Xiao",
"Huan",
""
],
[
"Wu",
"Yihan",
""
],
[
"Song",
"Ruihua",
""
]
] | TITLE: ComedicSpeech: Text To Speech For Stand-up Comedies in Low-Resource
Scenarios
ABSTRACT: Text to Speech (TTS) models can generate natural and high-quality speech, but
it is not expressive enough when synthesizing speech with dramatic
expressiveness, such as stand-up comedies. Considering comedians have diverse
personal speech styles, including personal prosody, rhythm, and fillers, it
requires real-world datasets and strong speech style modeling capabilities,
which brings challenges. In this paper, we construct a new dataset and develop
ComedicSpeech, a TTS system tailored for the stand-up comedy synthesis in
low-resource scenarios. First, we extract prosody representation by the prosody
encoder and condition it to the TTS model in a flexible way. Second, we enhance
the personal rhythm modeling by a conditional duration predictor. Third, we
model the personal fillers by introducing comedian-related special tokens.
Experiments show that ComedicSpeech achieves better expressiveness than
baselines with only ten-minute training data for each comedian. The audio
samples are available at https://xh621.github.io/stand-up-comedy-demo/
| new_dataset | 0.714827 |
2305.12217 | Mozhi Zhang | Mozhi Zhang, Hang Yan, Yaqian Zhou, Xipeng Qiu | PromptNER: A Prompting Method for Few-shot Named Entity Recognition via
k Nearest Neighbor Search | work in progress | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Few-shot Named Entity Recognition (NER) is a task aiming to identify named
entities via limited annotated samples. Recently, prototypical networks have
shown promising performance in few-shot NER. Most of prototypical networks will
utilize the entities from the support set to construct label prototypes and use
the query set to compute span-level similarities and optimize these label
prototype representations. However, these methods are usually unsuitable for
fine-tuning in the target domain, where only the support set is available. In
this paper, we propose PromptNER: a novel prompting method for few-shot NER via
k nearest neighbor search. We use prompts that contains entity category
information to construct label prototypes, which enables our model to fine-tune
with only the support set. Our approach achieves excellent transfer learning
ability, and extensive experiments on the Few-NERD and CrossNER datasets
demonstrate that our model achieves superior performance over state-of-the-art
methods.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 15:47:59 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Zhang",
"Mozhi",
""
],
[
"Yan",
"Hang",
""
],
[
"Zhou",
"Yaqian",
""
],
[
"Qiu",
"Xipeng",
""
]
] | TITLE: PromptNER: A Prompting Method for Few-shot Named Entity Recognition via
k Nearest Neighbor Search
ABSTRACT: Few-shot Named Entity Recognition (NER) is a task aiming to identify named
entities via limited annotated samples. Recently, prototypical networks have
shown promising performance in few-shot NER. Most of prototypical networks will
utilize the entities from the support set to construct label prototypes and use
the query set to compute span-level similarities and optimize these label
prototype representations. However, these methods are usually unsuitable for
fine-tuning in the target domain, where only the support set is available. In
this paper, we propose PromptNER: a novel prompting method for few-shot NER via
k nearest neighbor search. We use prompts that contains entity category
information to construct label prototypes, which enables our model to fine-tune
with only the support set. Our approach achieves excellent transfer learning
ability, and extensive experiments on the Few-NERD and CrossNER datasets
demonstrate that our model achieves superior performance over state-of-the-art
methods.
| no_new_dataset | 0.71282 |
2305.12218 | Peng Jin | Peng Jin, Hao Li, Zesen Cheng, Jinfa Huang, Zhennan Wang, Li Yuan,
Chang Liu, Jie Chen | Text-Video Retrieval with Disentangled Conceptualization and Set-to-Set
Alignment | IJCAI 2023 | null | null | null | cs.CV cs.AI cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-video retrieval is a challenging cross-modal task, which aims to align
visual entities with natural language descriptions. Current methods either fail
to leverage the local details or are computationally expensive. What's worse,
they fail to leverage the heterogeneous concepts in data. In this paper, we
propose the Disentangled Conceptualization and Set-to-set Alignment (DiCoSA) to
simulate the conceptualizing and reasoning process of human beings. For
disentangled conceptualization, we divide the coarse feature into multiple
latent factors related to semantic concepts. For set-to-set alignment, where a
set of visual concepts correspond to a set of textual concepts, we propose an
adaptive pooling method to aggregate semantic concepts to address the partial
matching. In particular, since we encode concepts independently in only a few
dimensions, DiCoSA is superior at efficiency and granularity, ensuring
fine-grained interactions using a similar computational complexity as
coarse-grained alignment. Extensive experiments on five datasets, including
MSR-VTT, LSMDC, MSVD, ActivityNet, and DiDeMo, demonstrate that our method
outperforms the existing state-of-the-art methods.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 15:48:47 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Jin",
"Peng",
""
],
[
"Li",
"Hao",
""
],
[
"Cheng",
"Zesen",
""
],
[
"Huang",
"Jinfa",
""
],
[
"Wang",
"Zhennan",
""
],
[
"Yuan",
"Li",
""
],
[
"Liu",
"Chang",
""
],
[
"Chen",
"Jie",
""
]
] | TITLE: Text-Video Retrieval with Disentangled Conceptualization and Set-to-Set
Alignment
ABSTRACT: Text-video retrieval is a challenging cross-modal task, which aims to align
visual entities with natural language descriptions. Current methods either fail
to leverage the local details or are computationally expensive. What's worse,
they fail to leverage the heterogeneous concepts in data. In this paper, we
propose the Disentangled Conceptualization and Set-to-set Alignment (DiCoSA) to
simulate the conceptualizing and reasoning process of human beings. For
disentangled conceptualization, we divide the coarse feature into multiple
latent factors related to semantic concepts. For set-to-set alignment, where a
set of visual concepts correspond to a set of textual concepts, we propose an
adaptive pooling method to aggregate semantic concepts to address the partial
matching. In particular, since we encode concepts independently in only a few
dimensions, DiCoSA is superior at efficiency and granularity, ensuring
fine-grained interactions using a similar computational complexity as
coarse-grained alignment. Extensive experiments on five datasets, including
MSR-VTT, LSMDC, MSVD, ActivityNet, and DiDeMo, demonstrate that our method
outperforms the existing state-of-the-art methods.
| no_new_dataset | 0.709617 |
2305.12221 | Daniela Zaharie | M\u{a}d\u{a}lina-Andreea Mitran, Anna V. Kononova, Fabio Caraffini,
Daniela Zaharie | Patterns of Convergence and Bound Constraint Violation in Differential
Evolution on SBOX-COST Benchmarking Suite | null | null | null | null | cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study investigates the influence of several bound constraint handling
methods (BCHMs) on the search process specific to Differential Evolution (DE),
with a focus on identifying similarities between BCHMs and grouping patterns
with respect to the number of cases when a BCHM is activated. The empirical
analysis is conducted on the SBOX-COST benchmarking test suite, where bound
constraints are enforced on the problem domain. This analysis provides some
insights that might be useful in designing adaptive strategies for handling
such constraints.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 16:04:06 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Mitran",
"Mădălina-Andreea",
""
],
[
"Kononova",
"Anna V.",
""
],
[
"Caraffini",
"Fabio",
""
],
[
"Zaharie",
"Daniela",
""
]
] | TITLE: Patterns of Convergence and Bound Constraint Violation in Differential
Evolution on SBOX-COST Benchmarking Suite
ABSTRACT: This study investigates the influence of several bound constraint handling
methods (BCHMs) on the search process specific to Differential Evolution (DE),
with a focus on identifying similarities between BCHMs and grouping patterns
with respect to the number of cases when a BCHM is activated. The empirical
analysis is conducted on the SBOX-COST benchmarking test suite, where bound
constraints are enforced on the problem domain. This analysis provides some
insights that might be useful in designing adaptive strategies for handling
such constraints.
| no_new_dataset | 0.710019 |
2305.12228 | Yiming Chen | Yiming Chen, Simin Chen, Zexin Li, Wei Yang, Cong Liu, Robby T. Tan,
Haizhou Li | Dynamic Transformers Provide a False Sense of Efficiency | Accepted by ACL2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite much success in natural language processing (NLP), pre-trained
language models typically lead to a high computational cost during inference.
Multi-exit is a mainstream approach to address this issue by making a trade-off
between efficiency and accuracy, where the saving of computation comes from an
early exit. However, whether such saving from early-exiting is robust remains
unknown. Motivated by this, we first show that directly adapting existing
adversarial attack approaches targeting model accuracy cannot significantly
reduce inference efficiency. To this end, we propose a simple yet effective
attacking framework, SAME, a novel slowdown attack framework on multi-exit
models, which is specially tailored to reduce the efficiency of the multi-exit
models. By leveraging the multi-exit models' design characteristics, we utilize
all internal predictions to guide the adversarial sample generation instead of
merely considering the final prediction. Experiments on the GLUE benchmark show
that SAME can effectively diminish the efficiency gain of various multi-exit
models by 80% on average, convincingly validating its effectiveness and
generalization ability.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 16:41:48 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chen",
"Yiming",
""
],
[
"Chen",
"Simin",
""
],
[
"Li",
"Zexin",
""
],
[
"Yang",
"Wei",
""
],
[
"Liu",
"Cong",
""
],
[
"Tan",
"Robby T.",
""
],
[
"Li",
"Haizhou",
""
]
] | TITLE: Dynamic Transformers Provide a False Sense of Efficiency
ABSTRACT: Despite much success in natural language processing (NLP), pre-trained
language models typically lead to a high computational cost during inference.
Multi-exit is a mainstream approach to address this issue by making a trade-off
between efficiency and accuracy, where the saving of computation comes from an
early exit. However, whether such saving from early-exiting is robust remains
unknown. Motivated by this, we first show that directly adapting existing
adversarial attack approaches targeting model accuracy cannot significantly
reduce inference efficiency. To this end, we propose a simple yet effective
attacking framework, SAME, a novel slowdown attack framework on multi-exit
models, which is specially tailored to reduce the efficiency of the multi-exit
models. By leveraging the multi-exit models' design characteristics, we utilize
all internal predictions to guide the adversarial sample generation instead of
merely considering the final prediction. Experiments on the GLUE benchmark show
that SAME can effectively diminish the efficiency gain of various multi-exit
models by 80% on average, convincingly validating its effectiveness and
generalization ability.
| no_new_dataset | 0.706823 |
2305.12238 | Dylan Cope | Nandi Schoots, Dylan Cope | Low-Entropy Latent Variables Hurt Out-of-Distribution Performance | Published as a workshop paper at ICLR 2023 Domain Generalization | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We study the relationship between the entropy of intermediate representations
and a model's robustness to distributional shift. We train models consisting of
two feed-forward networks end-to-end separated by a discrete $n$-bit channel on
an unsupervised contrastive learning task. Different masking strategies are
applied after training that remove a proportion of low-entropy bits,
high-entropy bits, or randomly selected bits, and the effects on performance
are compared to the baseline accuracy with no mask. We hypothesize that the
entropy of a bit serves as a guide to its usefulness out-of-distribution (OOD).
Through experiment on three OOD datasets we demonstrate that the removal of
low-entropy bits can notably benefit OOD performance. Conversely, we find that
top-entropy masking disproportionately harms performance both in-distribution
(InD) and OOD.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 17:09:44 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Schoots",
"Nandi",
""
],
[
"Cope",
"Dylan",
""
]
] | TITLE: Low-Entropy Latent Variables Hurt Out-of-Distribution Performance
ABSTRACT: We study the relationship between the entropy of intermediate representations
and a model's robustness to distributional shift. We train models consisting of
two feed-forward networks end-to-end separated by a discrete $n$-bit channel on
an unsupervised contrastive learning task. Different masking strategies are
applied after training that remove a proportion of low-entropy bits,
high-entropy bits, or randomly selected bits, and the effects on performance
are compared to the baseline accuracy with no mask. We hypothesize that the
entropy of a bit serves as a guide to its usefulness out-of-distribution (OOD).
Through experiment on three OOD datasets we demonstrate that the removal of
low-entropy bits can notably benefit OOD performance. Conversely, we find that
top-entropy masking disproportionately harms performance both in-distribution
(InD) and OOD.
| no_new_dataset | 0.710998 |
2305.12242 | Huili Chen | Qimao Yang, Huili Chen, Qiwei Dong | Comparative Analysis of Deep Learning Models for Brand Logo
Classification in Real-World Scenarios | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This report presents a comprehensive study on deep learning models for brand
logo classification in real-world scenarios. The dataset contains 3,717 labeled
images of logos from ten prominent brands. Two types of models, Convolutional
Neural Networks (CNN) and Vision Transformer (ViT), were evaluated for their
performance. The ViT model, DaViT small, achieved the highest accuracy of
99.60%, while the DenseNet29 achieved the fastest inference speed of 366.62
FPS. The findings suggest that the DaViT model is a suitable choice for offline
applications due to its superior accuracy. This study demonstrates the
practical application of deep learning in brand logo classification tasks.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 17:24:06 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Yang",
"Qimao",
""
],
[
"Chen",
"Huili",
""
],
[
"Dong",
"Qiwei",
""
]
] | TITLE: Comparative Analysis of Deep Learning Models for Brand Logo
Classification in Real-World Scenarios
ABSTRACT: This report presents a comprehensive study on deep learning models for brand
logo classification in real-world scenarios. The dataset contains 3,717 labeled
images of logos from ten prominent brands. Two types of models, Convolutional
Neural Networks (CNN) and Vision Transformer (ViT), were evaluated for their
performance. The ViT model, DaViT small, achieved the highest accuracy of
99.60%, while the DenseNet29 achieved the fastest inference speed of 366.62
FPS. The findings suggest that the DaViT model is a suitable choice for offline
applications due to its superior accuracy. This study demonstrates the
practical application of deep learning in brand logo classification tasks.
| new_dataset | 0.707367 |
2305.12252 | Jie Yang | Jie Yang, Bingliang Li, Fengyu Yang, Ailing Zeng, Lei Zhang, Ruimao
Zhang | Boosting Human-Object Interaction Detection with Text-to-Image Diffusion
Model | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates the problem of the current HOI detection methods and
introduces DiffHOI, a novel HOI detection scheme grounded on a pre-trained
text-image diffusion model, which enhances the detector's performance via
improved data diversity and HOI representation. We demonstrate that the
internal representation space of a frozen text-to-image diffusion model is
highly relevant to verb concepts and their corresponding context. Accordingly,
we propose an adapter-style tuning method to extract the various semantic
associated representation from a frozen diffusion model and CLIP model to
enhance the human and object representations from the pre-trained detector,
further reducing the ambiguity in interaction prediction. Moreover, to fill in
the gaps of HOI datasets, we propose SynHOI, a class-balance, large-scale, and
high-diversity synthetic dataset containing over 140K HOI images with fully
triplet annotations. It is built using an automatic and scalable pipeline
designed to scale up the generation of diverse and high-precision HOI-annotated
data. SynHOI could effectively relieve the long-tail issue in existing datasets
and facilitate learning interaction representations. Extensive experiments
demonstrate that DiffHOI significantly outperforms the state-of-the-art in
regular detection (i.e., 41.50 mAP) and zero-shot detection. Furthermore,
SynHOI can improve the performance of model-agnostic and backbone-agnostic HOI
detection, particularly exhibiting an outstanding 11.55% mAP improvement in
rare classes.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 17:59:23 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Yang",
"Jie",
""
],
[
"Li",
"Bingliang",
""
],
[
"Yang",
"Fengyu",
""
],
[
"Zeng",
"Ailing",
""
],
[
"Zhang",
"Lei",
""
],
[
"Zhang",
"Ruimao",
""
]
] | TITLE: Boosting Human-Object Interaction Detection with Text-to-Image Diffusion
Model
ABSTRACT: This paper investigates the problem of the current HOI detection methods and
introduces DiffHOI, a novel HOI detection scheme grounded on a pre-trained
text-image diffusion model, which enhances the detector's performance via
improved data diversity and HOI representation. We demonstrate that the
internal representation space of a frozen text-to-image diffusion model is
highly relevant to verb concepts and their corresponding context. Accordingly,
we propose an adapter-style tuning method to extract the various semantic
associated representation from a frozen diffusion model and CLIP model to
enhance the human and object representations from the pre-trained detector,
further reducing the ambiguity in interaction prediction. Moreover, to fill in
the gaps of HOI datasets, we propose SynHOI, a class-balance, large-scale, and
high-diversity synthetic dataset containing over 140K HOI images with fully
triplet annotations. It is built using an automatic and scalable pipeline
designed to scale up the generation of diverse and high-precision HOI-annotated
data. SynHOI could effectively relieve the long-tail issue in existing datasets
and facilitate learning interaction representations. Extensive experiments
demonstrate that DiffHOI significantly outperforms the state-of-the-art in
regular detection (i.e., 41.50 mAP) and zero-shot detection. Furthermore,
SynHOI can improve the performance of model-agnostic and backbone-agnostic HOI
detection, particularly exhibiting an outstanding 11.55% mAP improvement in
rare classes.
| no_new_dataset | 0.691979 |
2305.12257 | Ankur Sinha PhD | Ankur Sinha, Satishwar Kedas, Rishu Kumar, Pekka Malo | SEntFiN 1.0: Entity-Aware Sentiment Analysis for Financial News | 32 Pages | null | 10.1002/asi.24634 | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Fine-grained financial sentiment analysis on news headlines is a challenging
task requiring human-annotated datasets to achieve high performance. Limited
studies have tried to address the sentiment extraction task in a setting where
multiple entities are present in a news headline. In an effort to further
research in this area, we make publicly available SEntFiN 1.0, a
human-annotated dataset of 10,753 news headlines with entity-sentiment
annotations, of which 2,847 headlines contain multiple entities, often with
conflicting sentiments. We augment our dataset with a database of over 1,000
financial entities and their various representations in news media amounting to
over 5,000 phrases. We propose a framework that enables the extraction of
entity-relevant sentiments using a feature-based approach rather than an
expression-based approach. For sentiment extraction, we utilize 12 different
learning schemes utilizing lexicon-based and pre-trained sentence
representations and five classification approaches. Our experiments indicate
that lexicon-based n-gram ensembles are above par with pre-trained word
embedding schemes such as GloVe. Overall, RoBERTa and finBERT (domain-specific
BERT) achieve the highest average accuracy of 94.29% and F1-score of 93.27%.
Further, using over 210,000 entity-sentiment predictions, we validate the
economic effect of sentiments on aggregate market movements over a long
duration.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 18:20:39 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Sinha",
"Ankur",
""
],
[
"Kedas",
"Satishwar",
""
],
[
"Kumar",
"Rishu",
""
],
[
"Malo",
"Pekka",
""
]
] | TITLE: SEntFiN 1.0: Entity-Aware Sentiment Analysis for Financial News
ABSTRACT: Fine-grained financial sentiment analysis on news headlines is a challenging
task requiring human-annotated datasets to achieve high performance. Limited
studies have tried to address the sentiment extraction task in a setting where
multiple entities are present in a news headline. In an effort to further
research in this area, we make publicly available SEntFiN 1.0, a
human-annotated dataset of 10,753 news headlines with entity-sentiment
annotations, of which 2,847 headlines contain multiple entities, often with
conflicting sentiments. We augment our dataset with a database of over 1,000
financial entities and their various representations in news media amounting to
over 5,000 phrases. We propose a framework that enables the extraction of
entity-relevant sentiments using a feature-based approach rather than an
expression-based approach. For sentiment extraction, we utilize 12 different
learning schemes utilizing lexicon-based and pre-trained sentence
representations and five classification approaches. Our experiments indicate
that lexicon-based n-gram ensembles are above par with pre-trained word
embedding schemes such as GloVe. Overall, RoBERTa and finBERT (domain-specific
BERT) achieve the highest average accuracy of 94.29% and F1-score of 93.27%.
Further, using over 210,000 entity-sentiment predictions, we validate the
economic effect of sentiments on aggregate market movements over a long
duration.
| new_dataset | 0.718372 |
2305.12266 | Tie Luo | Ronit Das, Tie Luo | LightESD: Fully-Automated and Lightweight Anomaly Detection Framework
for Edge Computing | IEEE EDGE 2023, Chicago, USA, July 2023 | null | null | null | cs.LG cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Anomaly detection is widely used in a broad range of domains from
cybersecurity to manufacturing, finance, and so on. Deep learning based anomaly
detection has recently drawn much attention because of its superior capability
of recognizing complex data patterns and identifying outliers accurately.
However, deep learning models are typically iteratively optimized in a central
server with input data gathered from edge devices, and such data transfer
between edge devices and the central server impose substantial overhead on the
network and incur additional latency and energy consumption. To overcome this
problem, we propose a fully-automated, lightweight, statistical learning based
anomaly detection framework called LightESD. It is an on-device learning method
without the need for data transfer between edge and server, and is extremely
lightweight that most low-end edge devices can easily afford with negligible
delay, CPU/memory utilization, and power consumption. Yet, it achieves highly
competitive detection accuracy. Another salient feature is that it can
auto-adapt to probably any dataset without manually setting or configuring
model parameters or hyperparameters, which is a drawback of most existing
methods. We focus on time series data due to its pervasiveness in edge
applications such as IoT. Our evaluation demonstrates that LightESD outperforms
other SOTA methods on detection accuracy, efficiency, and resource consumption.
Additionally, its fully automated feature gives it another competitive
advantage in terms of practical usability and generalizability.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 18:48:41 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Das",
"Ronit",
""
],
[
"Luo",
"Tie",
""
]
] | TITLE: LightESD: Fully-Automated and Lightweight Anomaly Detection Framework
for Edge Computing
ABSTRACT: Anomaly detection is widely used in a broad range of domains from
cybersecurity to manufacturing, finance, and so on. Deep learning based anomaly
detection has recently drawn much attention because of its superior capability
of recognizing complex data patterns and identifying outliers accurately.
However, deep learning models are typically iteratively optimized in a central
server with input data gathered from edge devices, and such data transfer
between edge devices and the central server impose substantial overhead on the
network and incur additional latency and energy consumption. To overcome this
problem, we propose a fully-automated, lightweight, statistical learning based
anomaly detection framework called LightESD. It is an on-device learning method
without the need for data transfer between edge and server, and is extremely
lightweight that most low-end edge devices can easily afford with negligible
delay, CPU/memory utilization, and power consumption. Yet, it achieves highly
competitive detection accuracy. Another salient feature is that it can
auto-adapt to probably any dataset without manually setting or configuring
model parameters or hyperparameters, which is a drawback of most existing
methods. We focus on time series data due to its pervasiveness in edge
applications such as IoT. Our evaluation demonstrates that LightESD outperforms
other SOTA methods on detection accuracy, efficiency, and resource consumption.
Additionally, its fully automated feature gives it another competitive
advantage in terms of practical usability and generalizability.
| no_new_dataset | 0.708244 |
2305.12268 | Bowen Jin | Bowen Jin, Wentao Zhang, Yu Zhang, Yu Meng, Xinyang Zhang, Qi Zhu,
Jiawei Han | Patton: Language Model Pretraining on Text-Rich Networks | ACL 2023. (Code: https://github.com/PeterGriffinJin/Patton) | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A real-world text corpus sometimes comprises not only text documents but also
semantic links between them (e.g., academic papers in a bibliographic network
are linked by citations and co-authorships). Text documents and semantic
connections form a text-rich network, which empowers a wide range of downstream
tasks such as classification and retrieval. However, pretraining methods for
such structures are still lacking, making it difficult to build one generic
model that can be adapted to various tasks on text-rich networks. Current
pretraining objectives, such as masked language modeling, purely model texts
and do not take inter-document structure information into consideration. To
this end, we propose our PretrAining on TexT-Rich NetwOrk framework Patton.
Patton includes two pretraining strategies: network-contextualized masked
language modeling and masked node prediction, to capture the inherent
dependency between textual attributes and network structure. We conduct
experiments on four downstream tasks in five datasets from both academic and
e-commerce domains, where Patton outperforms baselines significantly and
consistently.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 19:17:10 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Jin",
"Bowen",
""
],
[
"Zhang",
"Wentao",
""
],
[
"Zhang",
"Yu",
""
],
[
"Meng",
"Yu",
""
],
[
"Zhang",
"Xinyang",
""
],
[
"Zhu",
"Qi",
""
],
[
"Han",
"Jiawei",
""
]
] | TITLE: Patton: Language Model Pretraining on Text-Rich Networks
ABSTRACT: A real-world text corpus sometimes comprises not only text documents but also
semantic links between them (e.g., academic papers in a bibliographic network
are linked by citations and co-authorships). Text documents and semantic
connections form a text-rich network, which empowers a wide range of downstream
tasks such as classification and retrieval. However, pretraining methods for
such structures are still lacking, making it difficult to build one generic
model that can be adapted to various tasks on text-rich networks. Current
pretraining objectives, such as masked language modeling, purely model texts
and do not take inter-document structure information into consideration. To
this end, we propose our PretrAining on TexT-Rich NetwOrk framework Patton.
Patton includes two pretraining strategies: network-contextualized masked
language modeling and masked node prediction, to capture the inherent
dependency between textual attributes and network structure. We conduct
experiments on four downstream tasks in five datasets from both academic and
e-commerce domains, where Patton outperforms baselines significantly and
consistently.
| no_new_dataset | 0.706608 |
2305.12287 | Didong Li | Sam Hawke, Hengrui Luo and Didong Li | Contrastive inverse regression for dimension reduction | null | null | null | null | stat.ML cs.LG stat.AP stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Supervised dimension reduction (SDR) has been a topic of growing interest in
data science, as it enables the reduction of high-dimensional covariates while
preserving the functional relation with certain response variables of interest.
However, existing SDR methods are not suitable for analyzing datasets collected
from case-control studies. In this setting, the goal is to learn and exploit
the low-dimensional structure unique to or enriched by the case group, also
known as the foreground group. While some unsupervised techniques such as the
contrastive latent variable model and its variants have been developed for this
purpose, they fail to preserve the functional relationship between the
dimension-reduced covariates and the response variable. In this paper, we
propose a supervised dimension reduction method called contrastive inverse
regression (CIR) specifically designed for the contrastive setting. CIR
introduces an optimization problem defined on the Stiefel manifold with a
non-standard loss function. We prove the convergence of CIR to a local optimum
using a gradient descent-based algorithm, and our numerical study empirically
demonstrates the improved performance over competing methods for
high-dimensional data.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 21:44:11 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Hawke",
"Sam",
""
],
[
"Luo",
"Hengrui",
""
],
[
"Li",
"Didong",
""
]
] | TITLE: Contrastive inverse regression for dimension reduction
ABSTRACT: Supervised dimension reduction (SDR) has been a topic of growing interest in
data science, as it enables the reduction of high-dimensional covariates while
preserving the functional relation with certain response variables of interest.
However, existing SDR methods are not suitable for analyzing datasets collected
from case-control studies. In this setting, the goal is to learn and exploit
the low-dimensional structure unique to or enriched by the case group, also
known as the foreground group. While some unsupervised techniques such as the
contrastive latent variable model and its variants have been developed for this
purpose, they fail to preserve the functional relationship between the
dimension-reduced covariates and the response variable. In this paper, we
propose a supervised dimension reduction method called contrastive inverse
regression (CIR) specifically designed for the contrastive setting. CIR
introduces an optimization problem defined on the Stiefel manifold with a
non-standard loss function. We prove the convergence of CIR to a local optimum
using a gradient descent-based algorithm, and our numerical study empirically
demonstrates the improved performance over competing methods for
high-dimensional data.
| no_new_dataset | 0.709806 |
2305.12289 | Haw-Shiuan Chang | Haw-Shiuan Chang, Zonghai Yao, Alolika Gon, Hong Yu, Andrew McCallum | Revisiting the Architectures like Pointer Networks to Efficiently
Improve the Next Word Distribution, Summarization Factuality, and Beyond | ACL Findings 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Is the output softmax layer, which is adopted by most language models (LMs),
always the best way to compute the next word probability? Given so many
attention layers in a modern transformer-based LM, are the pointer networks
redundant nowadays? In this study, we discover that the answers to both
questions are no. This is because the softmax bottleneck sometimes prevents the
LMs from predicting the desired distribution and the pointer networks can be
used to break the bottleneck efficiently. Based on the finding, we propose
several softmax alternatives by simplifying the pointer networks and
accelerating the word-by-word rerankers. In GPT-2, our proposals are
significantly better and more efficient than mixture of softmax, a
state-of-the-art softmax alternative. In summarization experiments, without
significantly decreasing its training/testing speed, our best method based on
T5-Small improves factCC score by 2 points in CNN/DM and XSUM dataset, and
improves MAUVE scores by 30% in BookSum paragraph-level dataset.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 21:52:24 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chang",
"Haw-Shiuan",
""
],
[
"Yao",
"Zonghai",
""
],
[
"Gon",
"Alolika",
""
],
[
"Yu",
"Hong",
""
],
[
"McCallum",
"Andrew",
""
]
] | TITLE: Revisiting the Architectures like Pointer Networks to Efficiently
Improve the Next Word Distribution, Summarization Factuality, and Beyond
ABSTRACT: Is the output softmax layer, which is adopted by most language models (LMs),
always the best way to compute the next word probability? Given so many
attention layers in a modern transformer-based LM, are the pointer networks
redundant nowadays? In this study, we discover that the answers to both
questions are no. This is because the softmax bottleneck sometimes prevents the
LMs from predicting the desired distribution and the pointer networks can be
used to break the bottleneck efficiently. Based on the finding, we propose
several softmax alternatives by simplifying the pointer networks and
accelerating the word-by-word rerankers. In GPT-2, our proposals are
significantly better and more efficient than mixture of softmax, a
state-of-the-art softmax alternative. In summarization experiments, without
significantly decreasing its training/testing speed, our best method based on
T5-Small improves factCC score by 2 points in CNN/DM and XSUM dataset, and
improves MAUVE scores by 30% in BookSum paragraph-level dataset.
| no_new_dataset | 0.707834 |
2305.12301 | Soujanya Poria | Yi Xuan Tan, Navonil Majumder, Soujanya Poria | Sentence Embedder Guided Utterance Encoder (SEGUE) for Spoken Language
Understanding | Interspeech 2023 | null | null | null | cs.CL cs.AI cs.SD eess.AS | http://creativecommons.org/licenses/by-sa/4.0/ | The pre-trained speech encoder wav2vec 2.0 performs very well on various
spoken language understanding (SLU) tasks. However, on many tasks, it trails
behind text encoders with textual input. To improve the understanding
capability of SLU encoders, various studies have used knowledge distillation to
transfer knowledge from natural language understanding (NLU) encoders. We use a
very simple method of distilling from a textual sentence embedder directly into
wav2vec 2.0 as pre-training, utilizing paired audio-text datasets. We observed
that this method is indeed capable of improving SLU task performance in
fine-tuned settings, as well as full-data and few-shot transfer on a frozen
encoder. However, the model performs worse on certain tasks highlighting the
strengths and weaknesses of our approach.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 23:55:55 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Tan",
"Yi Xuan",
""
],
[
"Majumder",
"Navonil",
""
],
[
"Poria",
"Soujanya",
""
]
] | TITLE: Sentence Embedder Guided Utterance Encoder (SEGUE) for Spoken Language
Understanding
ABSTRACT: The pre-trained speech encoder wav2vec 2.0 performs very well on various
spoken language understanding (SLU) tasks. However, on many tasks, it trails
behind text encoders with textual input. To improve the understanding
capability of SLU encoders, various studies have used knowledge distillation to
transfer knowledge from natural language understanding (NLU) encoders. We use a
very simple method of distilling from a textual sentence embedder directly into
wav2vec 2.0 as pre-training, utilizing paired audio-text datasets. We observed
that this method is indeed capable of improving SLU task performance in
fine-tuned settings, as well as full-data and few-shot transfer on a frozen
encoder. However, the model performs worse on certain tasks highlighting the
strengths and weaknesses of our approach.
| no_new_dataset | 0.708219 |
2305.12311 | Ziyi Yang | Ziyi Yang, Mahmoud Khademi, Yichong Xu, Reid Pryzant, Yuwei Fang,
Chenguang Zhu, Dongdong Chen, Yao Qian, Mei Gao, Yi-Ling Chen, Robert Gmyr,
Naoyuki Kanda, Noel Codella, Bin Xiao, Yu Shi, Lu Yuan, Takuya Yoshioka,
Michael Zeng, Xuedong Huang | i-Code V2: An Autoregressive Generation Framework over Vision, Language,
and Speech Data | null | null | null | null | cs.CL cs.AI cs.CV cs.LG eess.AS | http://creativecommons.org/licenses/by/4.0/ | The convergence of text, visual, and audio data is a key step towards
human-like artificial intelligence, however the current Vision-Language-Speech
landscape is dominated by encoder-only models which lack generative abilities.
We propose closing this gap with i-Code V2, the first model capable of
generating natural language from any combination of Vision, Language, and
Speech data. i-Code V2 is an integrative system that leverages state-of-the-art
single-modality encoders, combining their outputs with a new modality-fusing
encoder in order to flexibly project combinations of modalities into a shared
representational space. Next, language tokens are generated from these
representations via an autoregressive decoder. The whole framework is
pretrained end-to-end on a large collection of dual- and single-modality
datasets using a novel text completion objective that can be generalized across
arbitrary combinations of modalities. i-Code V2 matches or outperforms
state-of-the-art single- and dual-modality baselines on 7 multimodal tasks,
demonstrating the power of generative multimodal pretraining across a diversity
of tasks and signals.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 01:25:44 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Yang",
"Ziyi",
""
],
[
"Khademi",
"Mahmoud",
""
],
[
"Xu",
"Yichong",
""
],
[
"Pryzant",
"Reid",
""
],
[
"Fang",
"Yuwei",
""
],
[
"Zhu",
"Chenguang",
""
],
[
"Chen",
"Dongdong",
""
],
[
"Qian",
"Yao",
""
],
[
"Gao",
"Mei",
""
],
[
"Chen",
"Yi-Ling",
""
],
[
"Gmyr",
"Robert",
""
],
[
"Kanda",
"Naoyuki",
""
],
[
"Codella",
"Noel",
""
],
[
"Xiao",
"Bin",
""
],
[
"Shi",
"Yu",
""
],
[
"Yuan",
"Lu",
""
],
[
"Yoshioka",
"Takuya",
""
],
[
"Zeng",
"Michael",
""
],
[
"Huang",
"Xuedong",
""
]
] | TITLE: i-Code V2: An Autoregressive Generation Framework over Vision, Language,
and Speech Data
ABSTRACT: The convergence of text, visual, and audio data is a key step towards
human-like artificial intelligence, however the current Vision-Language-Speech
landscape is dominated by encoder-only models which lack generative abilities.
We propose closing this gap with i-Code V2, the first model capable of
generating natural language from any combination of Vision, Language, and
Speech data. i-Code V2 is an integrative system that leverages state-of-the-art
single-modality encoders, combining their outputs with a new modality-fusing
encoder in order to flexibly project combinations of modalities into a shared
representational space. Next, language tokens are generated from these
representations via an autoregressive decoder. The whole framework is
pretrained end-to-end on a large collection of dual- and single-modality
datasets using a novel text completion objective that can be generalized across
arbitrary combinations of modalities. i-Code V2 matches or outperforms
state-of-the-art single- and dual-modality baselines on 7 multimodal tasks,
demonstrating the power of generative multimodal pretraining across a diversity
of tasks and signals.
| no_new_dataset | 0.710208 |
2305.12314 | Kazuki Nakajima | Kazuki Nakajima, Kazuyuki Shudo | Random Walk Sampling in Social Networks Involving Private Nodes | 27 pages, 10 figures, 3 tables | ACM Transactions on Knowledge Discovery from Data, Volume 17,
Issue 4, Article No. 51, pp 1-28, 2023 | 10.1145/3561388 | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Analysis of social networks with limited data access is challenging for third
parties. To address this challenge, a number of studies have developed
algorithms that estimate properties of social networks via a simple random
walk. However, most existing algorithms do not assume private nodes that do not
publish their neighbors' data when they are queried in empirical social
networks. Here we propose a practical framework for estimating properties via
random walk-based sampling in social networks involving private nodes. First,
we develop a sampling algorithm by extending a simple random walk to the case
of social networks involving private nodes. Then, we propose estimators with
reduced biases induced by private nodes for the network size, average degree,
and density of the node label. Our results show that the proposed estimators
reduce biases induced by private nodes in the existing estimators by up to
92.6% on social network datasets involving private nodes.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 01:37:50 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Nakajima",
"Kazuki",
""
],
[
"Shudo",
"Kazuyuki",
""
]
] | TITLE: Random Walk Sampling in Social Networks Involving Private Nodes
ABSTRACT: Analysis of social networks with limited data access is challenging for third
parties. To address this challenge, a number of studies have developed
algorithms that estimate properties of social networks via a simple random
walk. However, most existing algorithms do not assume private nodes that do not
publish their neighbors' data when they are queried in empirical social
networks. Here we propose a practical framework for estimating properties via
random walk-based sampling in social networks involving private nodes. First,
we develop a sampling algorithm by extending a simple random walk to the case
of social networks involving private nodes. Then, we propose estimators with
reduced biases induced by private nodes for the network size, average degree,
and density of the node label. Our results show that the proposed estimators
reduce biases induced by private nodes in the existing estimators by up to
92.6% on social network datasets involving private nodes.
| no_new_dataset | 0.712407 |
2305.12316 | Mohamed Elmahallawy | Mohamed Elmahallawy, Tie Luo | One-Shot Federated Learning for LEO Constellations that Reduces
Convergence Time from Days to 90 Minutes | This article belongs to The 24th IEEE International Conference on
Mobile Data Management (MDM 2023) | null | null | null | cs.LG cs.NI | http://creativecommons.org/licenses/by/4.0/ | A Low Earth orbit (LEO) satellite constellation consists of a large number of
small satellites traveling in space with high mobility and collecting vast
amounts of mobility data such as cloud movement for weather forecast, large
herds of animals migrating across geo-regions, spreading of forest fires, and
aircraft tracking. Machine learning can be utilized to analyze these mobility
data to address global challenges, and Federated Learning (FL) is a promising
approach because it eliminates the need for transmitting raw data and hence is
both bandwidth and privacy-friendly. However, FL requires many communication
rounds between clients (satellites) and the parameter server (PS), leading to
substantial delays of up to several days in LEO constellations. In this paper,
we propose a novel one-shot FL approach for LEO satellites, called LEOShot,
that needs only a single communication round to complete the entire learning
process. LEOShot comprises three processes: (i) synthetic data generation, (ii)
knowledge distillation, and (iii) virtual model retraining. We evaluate and
benchmark LEOShot against the state of the art and the results show that it
drastically expedites FL convergence by more than an order of magnitude. Also
surprisingly, despite the one-shot nature, its model accuracy is on par with or
even outperforms regular iterative FL schemes by a large margin
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 01:57:56 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Elmahallawy",
"Mohamed",
""
],
[
"Luo",
"Tie",
""
]
] | TITLE: One-Shot Federated Learning for LEO Constellations that Reduces
Convergence Time from Days to 90 Minutes
ABSTRACT: A Low Earth orbit (LEO) satellite constellation consists of a large number of
small satellites traveling in space with high mobility and collecting vast
amounts of mobility data such as cloud movement for weather forecast, large
herds of animals migrating across geo-regions, spreading of forest fires, and
aircraft tracking. Machine learning can be utilized to analyze these mobility
data to address global challenges, and Federated Learning (FL) is a promising
approach because it eliminates the need for transmitting raw data and hence is
both bandwidth and privacy-friendly. However, FL requires many communication
rounds between clients (satellites) and the parameter server (PS), leading to
substantial delays of up to several days in LEO constellations. In this paper,
we propose a novel one-shot FL approach for LEO satellites, called LEOShot,
that needs only a single communication round to complete the entire learning
process. LEOShot comprises three processes: (i) synthetic data generation, (ii)
knowledge distillation, and (iii) virtual model retraining. We evaluate and
benchmark LEOShot against the state of the art and the results show that it
drastically expedites FL convergence by more than an order of magnitude. Also
surprisingly, despite the one-shot nature, its model accuracy is on par with or
even outperforms regular iterative FL schemes by a large margin
| no_new_dataset | 0.7118 |
2305.12319 | Yue Xu | Yue Xu, Qijie Shen, Jianwen Yin, Zengde Deng, Dimin Wang, Hao Chen,
Lixiang Lai, Tao Zhuang, Junfeng Ge | Multi-channel Integrated Recommendation with Exposure Constraints | null | KDD 2023 | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Integrated recommendation, which aims at jointly recommending heterogeneous
items from different channels in a main feed, has been widely applied to
various online platforms. Though attractive, integrated recommendation requires
the ranking methods to migrate from conventional user-item models to the new
user-channel-item paradigm in order to better capture users' preferences on
both item and channel levels. Moreover, practical feed recommendation systems
usually impose exposure constraints on different channels to ensure user
experience. This leads to greater difficulty in the joint ranking of
heterogeneous items. In this paper, we investigate the integrated
recommendation task with exposure constraints in practical recommender systems.
Our contribution is forth-fold. First, we formulate this task as a binary
online linear programming problem and propose a two-layer framework named
Multi-channel Integrated Recommendation with Exposure Constraints (MIREC) to
obtain the optimal solution. Second, we propose an efficient online allocation
algorithm to determine the optimal exposure assignment of different channels
from a global view of all user requests over the entire time horizon. We prove
that this algorithm reaches the optimal point under a regret bound of $
\mathcal{O}(\sqrt{T}) $ with linear complexity. Third, we propose a series of
collaborative models to determine the optimal layout of heterogeneous items at
each user request. The joint modeling of user interests, cross-channel
correlation, and page context in our models aligns more with the browsing
nature of feed products than existing models. Finally, we conduct extensive
experiments on both offline datasets and online A/B tests to verify the
effectiveness of MIREC. The proposed framework has now been implemented on the
homepage of Taobao to serve the main traffic.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 02:31:29 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Xu",
"Yue",
""
],
[
"Shen",
"Qijie",
""
],
[
"Yin",
"Jianwen",
""
],
[
"Deng",
"Zengde",
""
],
[
"Wang",
"Dimin",
""
],
[
"Chen",
"Hao",
""
],
[
"Lai",
"Lixiang",
""
],
[
"Zhuang",
"Tao",
""
],
[
"Ge",
"Junfeng",
""
]
] | TITLE: Multi-channel Integrated Recommendation with Exposure Constraints
ABSTRACT: Integrated recommendation, which aims at jointly recommending heterogeneous
items from different channels in a main feed, has been widely applied to
various online platforms. Though attractive, integrated recommendation requires
the ranking methods to migrate from conventional user-item models to the new
user-channel-item paradigm in order to better capture users' preferences on
both item and channel levels. Moreover, practical feed recommendation systems
usually impose exposure constraints on different channels to ensure user
experience. This leads to greater difficulty in the joint ranking of
heterogeneous items. In this paper, we investigate the integrated
recommendation task with exposure constraints in practical recommender systems.
Our contribution is forth-fold. First, we formulate this task as a binary
online linear programming problem and propose a two-layer framework named
Multi-channel Integrated Recommendation with Exposure Constraints (MIREC) to
obtain the optimal solution. Second, we propose an efficient online allocation
algorithm to determine the optimal exposure assignment of different channels
from a global view of all user requests over the entire time horizon. We prove
that this algorithm reaches the optimal point under a regret bound of $
\mathcal{O}(\sqrt{T}) $ with linear complexity. Third, we propose a series of
collaborative models to determine the optimal layout of heterogeneous items at
each user request. The joint modeling of user interests, cross-channel
correlation, and page context in our models aligns more with the browsing
nature of feed products than existing models. Finally, we conduct extensive
experiments on both offline datasets and online A/B tests to verify the
effectiveness of MIREC. The proposed framework has now been implemented on the
homepage of Taobao to serve the main traffic.
| no_new_dataset | 0.709453 |
2305.12335 | Tirthankar Roy | Sinan Rasiya Koya and Tirthankar Roy | Temporal Fusion Transformers for Streamflow Prediction: Value of
Combining Attention with Recurrence | null | null | null | null | cs.LG physics.geo-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Over the past few decades, the hydrology community has witnessed notable
advancements in streamflow prediction, particularly with the introduction of
cutting-edge machine-learning algorithms. Recurrent neural networks, especially
Long Short-Term Memory (LSTM) networks, have become popular due to their
capacity to create precise forecasts and realistically mimic the system
dynamics. Attention-based models, such as Transformers, can learn from the
entire data sequence concurrently, a feature that LSTM does not have. This work
tests the hypothesis that combining recurrence with attention can improve
streamflow prediction. We set up the Temporal Fusion Transformer (TFT)
architecture, a model that combines both of these aspects and has never been
applied in hydrology before. We compare the performance of LSTM, Transformers,
and TFT over 2,610 globally distributed catchments from the recently available
Caravan dataset. Our results demonstrate that TFT indeed exceeds the
performance benchmark set by the LSTM and Transformers for streamflow
prediction. Additionally, being an explainable AI method, TFT helps in gaining
insights into the streamflow generation processes.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 03:58:16 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Koya",
"Sinan Rasiya",
""
],
[
"Roy",
"Tirthankar",
""
]
] | TITLE: Temporal Fusion Transformers for Streamflow Prediction: Value of
Combining Attention with Recurrence
ABSTRACT: Over the past few decades, the hydrology community has witnessed notable
advancements in streamflow prediction, particularly with the introduction of
cutting-edge machine-learning algorithms. Recurrent neural networks, especially
Long Short-Term Memory (LSTM) networks, have become popular due to their
capacity to create precise forecasts and realistically mimic the system
dynamics. Attention-based models, such as Transformers, can learn from the
entire data sequence concurrently, a feature that LSTM does not have. This work
tests the hypothesis that combining recurrence with attention can improve
streamflow prediction. We set up the Temporal Fusion Transformer (TFT)
architecture, a model that combines both of these aspects and has never been
applied in hydrology before. We compare the performance of LSTM, Transformers,
and TFT over 2,610 globally distributed catchments from the recently available
Caravan dataset. Our results demonstrate that TFT indeed exceeds the
performance benchmark set by the LSTM and Transformers for streamflow
prediction. Additionally, being an explainable AI method, TFT helps in gaining
insights into the streamflow generation processes.
| no_new_dataset | 0.70939 |
2305.12344 | Wahyu Pebrianto | Wahyu Pebrianto, Panca Mudjirahardjo, Sholeh Hadi Pramono, Rahmadwati,
Raden Arief Setyawan | YOLOv3 with Spatial Pyramid Pooling for Object Detection with Unmanned
Aerial Vehicles | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Object detection with Unmanned Aerial Vehicles (UAVs) has attracted much
attention in the research field of computer vision. However, not easy to
accurately detect objects with data obtained from UAVs, which capture images
from very high altitudes, making the image dominated by small object sizes,
that difficult to detect. Motivated by that challenge, we aim to improve the
performance of the one-stage detector YOLOv3 by adding a Spatial Pyramid
Pooling (SPP) layer on the end of the backbone darknet-53 to obtain more
efficient feature extraction process in object detection tasks with UAVs. We
also conducted an evaluation study on different versions of YOLOv3 methods.
Includes YOLOv3 with SPP, YOLOv3, and YOLOv3-tiny, which we analyzed with the
VisDrone2019-Det dataset. Here we show that YOLOv3 with SPP can get results mAP
0.6% higher than YOLOv3 and 26.6% than YOLOv3-Tiny at 640x640 input scale and
is even able to maintain accuracy at different input image scales than other
versions of the YOLOv3 method. Those results prove that the addition of SPP
layers to YOLOv3 can be an efficient solution for improving the performance of
the object detection method with data obtained from UAVs.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 04:41:52 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Pebrianto",
"Wahyu",
""
],
[
"Mudjirahardjo",
"Panca",
""
],
[
"Pramono",
"Sholeh Hadi",
""
],
[
"Rahmadwati",
"",
""
],
[
"Setyawan",
"Raden Arief",
""
]
] | TITLE: YOLOv3 with Spatial Pyramid Pooling for Object Detection with Unmanned
Aerial Vehicles
ABSTRACT: Object detection with Unmanned Aerial Vehicles (UAVs) has attracted much
attention in the research field of computer vision. However, not easy to
accurately detect objects with data obtained from UAVs, which capture images
from very high altitudes, making the image dominated by small object sizes,
that difficult to detect. Motivated by that challenge, we aim to improve the
performance of the one-stage detector YOLOv3 by adding a Spatial Pyramid
Pooling (SPP) layer on the end of the backbone darknet-53 to obtain more
efficient feature extraction process in object detection tasks with UAVs. We
also conducted an evaluation study on different versions of YOLOv3 methods.
Includes YOLOv3 with SPP, YOLOv3, and YOLOv3-tiny, which we analyzed with the
VisDrone2019-Det dataset. Here we show that YOLOv3 with SPP can get results mAP
0.6% higher than YOLOv3 and 26.6% than YOLOv3-Tiny at 640x640 input scale and
is even able to maintain accuracy at different input image scales than other
versions of the YOLOv3 method. Those results prove that the addition of SPP
layers to YOLOv3 can be an efficient solution for improving the performance of
the object detection method with data obtained from UAVs.
| no_new_dataset | 0.708213 |
2305.12349 | Eli Chien | Eli Chien, Jiong Zhang, Cho-Jui Hsieh, Jyun-Yu Jiang, Wei-Cheng Chang,
Olgica Milenkovic, Hsiang-Fu Yu | PINA: Leveraging Side Information in eXtreme Multi-label Classification
via Predicted Instance Neighborhood Aggregation | ICML 2023 | null | null | null | cs.LG cs.IR | http://creativecommons.org/licenses/by/4.0/ | The eXtreme Multi-label Classification~(XMC) problem seeks to find relevant
labels from an exceptionally large label space. Most of the existing XMC
learners focus on the extraction of semantic features from input query text.
However, conventional XMC studies usually neglect the side information of
instances and labels, which can be of use in many real-world applications such
as recommendation systems and e-commerce product search. We propose Predicted
Instance Neighborhood Aggregation (PINA), a data enhancement method for the
general XMC problem that leverages beneficial side information. Unlike most
existing XMC frameworks that treat labels and input instances as featureless
indicators and independent entries, PINA extracts information from the label
metadata and the correlations among training instances. Extensive experimental
results demonstrate the consistent gain of PINA on various XMC tasks compared
to the state-of-the-art methods: PINA offers a gain in accuracy compared to
standard XR-Transformers on five public benchmark datasets. Moreover, PINA
achieves a $\sim 5\%$ gain in accuracy on the largest dataset
LF-AmazonTitles-1.3M. Our implementation is publicly available.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 05:00:40 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Chien",
"Eli",
""
],
[
"Zhang",
"Jiong",
""
],
[
"Hsieh",
"Cho-Jui",
""
],
[
"Jiang",
"Jyun-Yu",
""
],
[
"Chang",
"Wei-Cheng",
""
],
[
"Milenkovic",
"Olgica",
""
],
[
"Yu",
"Hsiang-Fu",
""
]
] | TITLE: PINA: Leveraging Side Information in eXtreme Multi-label Classification
via Predicted Instance Neighborhood Aggregation
ABSTRACT: The eXtreme Multi-label Classification~(XMC) problem seeks to find relevant
labels from an exceptionally large label space. Most of the existing XMC
learners focus on the extraction of semantic features from input query text.
However, conventional XMC studies usually neglect the side information of
instances and labels, which can be of use in many real-world applications such
as recommendation systems and e-commerce product search. We propose Predicted
Instance Neighborhood Aggregation (PINA), a data enhancement method for the
general XMC problem that leverages beneficial side information. Unlike most
existing XMC frameworks that treat labels and input instances as featureless
indicators and independent entries, PINA extracts information from the label
metadata and the correlations among training instances. Extensive experimental
results demonstrate the consistent gain of PINA on various XMC tasks compared
to the state-of-the-art methods: PINA offers a gain in accuracy compared to
standard XR-Transformers on five public benchmark datasets. Moreover, PINA
achieves a $\sim 5\%$ gain in accuracy on the largest dataset
LF-AmazonTitles-1.3M. Our implementation is publicly available.
| no_new_dataset | 0.709837 |
2305.12358 | Thomas Wendler | Mehdi Astaraki, Francesca De Benetti, Yousef Yeganeh, Iuliana
Toma-Dasu, \"Orjan Smedby, Chunliang Wang, Nassir Navab, Thomas Wendler | AutoPaint: A Self-Inpainting Method for Unsupervised Anomaly Detection | 41 pages, 15 figures, follow-up paper to conference abstract at
yearly meeting of German Nuclear Medicine in 2022 | null | null | null | cs.CV cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Robust and accurate detection and segmentation of heterogenous tumors
appearing in different anatomical organs with supervised methods require
large-scale labeled datasets covering all possible types of diseases. Due to
the unavailability of such rich datasets and the high cost of annotations,
unsupervised anomaly detection (UAD) methods have been developed aiming to
detect the pathologies as deviation from the normality by utilizing the
unlabeled healthy image data. However, developed UAD models are often trained
with an incomplete distribution of healthy anatomies and have difficulties in
preserving anatomical constraints. This work intends to, first, propose a
robust inpainting model to learn the details of healthy anatomies and
reconstruct high-resolution images by preserving anatomical constraints.
Second, we propose an autoinpainting pipeline to automatically detect tumors,
replace their appearance with the learned healthy anatomies, and based on that
segment the tumoral volumes in a purely unsupervised fashion. Three imaging
datasets, including PET, CT, and PET-CT scans of lung tumors and head and neck
tumors, are studied as benchmarks for evaluation. Experimental results
demonstrate the significant superiority of the proposed method over a wide
range of state-of-the-art UAD methods. Moreover, the unsupervised method we
propose produces comparable results to a robust supervised segmentation method
when applied to multimodal images.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 05:45:38 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Astaraki",
"Mehdi",
""
],
[
"De Benetti",
"Francesca",
""
],
[
"Yeganeh",
"Yousef",
""
],
[
"Toma-Dasu",
"Iuliana",
""
],
[
"Smedby",
"Örjan",
""
],
[
"Wang",
"Chunliang",
""
],
[
"Navab",
"Nassir",
""
],
[
"Wendler",
"Thomas",
""
]
] | TITLE: AutoPaint: A Self-Inpainting Method for Unsupervised Anomaly Detection
ABSTRACT: Robust and accurate detection and segmentation of heterogenous tumors
appearing in different anatomical organs with supervised methods require
large-scale labeled datasets covering all possible types of diseases. Due to
the unavailability of such rich datasets and the high cost of annotations,
unsupervised anomaly detection (UAD) methods have been developed aiming to
detect the pathologies as deviation from the normality by utilizing the
unlabeled healthy image data. However, developed UAD models are often trained
with an incomplete distribution of healthy anatomies and have difficulties in
preserving anatomical constraints. This work intends to, first, propose a
robust inpainting model to learn the details of healthy anatomies and
reconstruct high-resolution images by preserving anatomical constraints.
Second, we propose an autoinpainting pipeline to automatically detect tumors,
replace their appearance with the learned healthy anatomies, and based on that
segment the tumoral volumes in a purely unsupervised fashion. Three imaging
datasets, including PET, CT, and PET-CT scans of lung tumors and head and neck
tumors, are studied as benchmarks for evaluation. Experimental results
demonstrate the significant superiority of the proposed method over a wide
range of state-of-the-art UAD methods. Moreover, the unsupervised method we
propose produces comparable results to a robust supervised segmentation method
when applied to multimodal images.
| no_new_dataset | 0.709849 |
2305.12369 | Yubin Kim | Yubin Kim, Dong Won Lee, Paul Pu Liang, Sharifa Algohwinem, Cynthia
Breazeal, Hae Won Park | HIINT: Historical, Intra- and Inter- personal Dynamics Modeling with
Cross-person Memory Transformer | null | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurately modeling affect dynamics, which refers to the changes and
fluctuations in emotions and affective displays during human conversations, is
crucial for understanding human interactions. By analyzing affect dynamics, we
can gain insights into how people communicate, respond to different situations,
and form relationships. However, modeling affect dynamics is challenging due to
contextual factors, such as the complex and nuanced nature of interpersonal
relationships, the situation, and other factors that influence affective
displays. To address this challenge, we propose a Cross-person Memory
Transformer (CPM-T) framework which is able to explicitly model affective
dynamics (intrapersonal and interpersonal influences) by identifying verbal and
non-verbal cues, and with a large language model to utilize the pre-trained
knowledge and perform verbal reasoning. The CPM-T framework maintains memory
modules to store and update the contexts within the conversation window,
enabling the model to capture dependencies between earlier and later parts of a
conversation. Additionally, our framework employs cross-modal attention to
effectively align information from multi-modalities and leverage cross-person
attention to align behaviors in multi-party interactions. We evaluate the
effectiveness and generalizability of our approach on three publicly available
datasets for joint engagement, rapport, and human beliefs prediction tasks.
Remarkably, the CPM-T framework outperforms baseline models in average
F1-scores by up to 7.3%, 9.3%, and 2.0% respectively. Finally, we demonstrate
the importance of each component in the framework via ablation studies with
respect to multimodal temporal behavior.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 06:43:35 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Kim",
"Yubin",
""
],
[
"Lee",
"Dong Won",
""
],
[
"Liang",
"Paul Pu",
""
],
[
"Algohwinem",
"Sharifa",
""
],
[
"Breazeal",
"Cynthia",
""
],
[
"Park",
"Hae Won",
""
]
] | TITLE: HIINT: Historical, Intra- and Inter- personal Dynamics Modeling with
Cross-person Memory Transformer
ABSTRACT: Accurately modeling affect dynamics, which refers to the changes and
fluctuations in emotions and affective displays during human conversations, is
crucial for understanding human interactions. By analyzing affect dynamics, we
can gain insights into how people communicate, respond to different situations,
and form relationships. However, modeling affect dynamics is challenging due to
contextual factors, such as the complex and nuanced nature of interpersonal
relationships, the situation, and other factors that influence affective
displays. To address this challenge, we propose a Cross-person Memory
Transformer (CPM-T) framework which is able to explicitly model affective
dynamics (intrapersonal and interpersonal influences) by identifying verbal and
non-verbal cues, and with a large language model to utilize the pre-trained
knowledge and perform verbal reasoning. The CPM-T framework maintains memory
modules to store and update the contexts within the conversation window,
enabling the model to capture dependencies between earlier and later parts of a
conversation. Additionally, our framework employs cross-modal attention to
effectively align information from multi-modalities and leverage cross-person
attention to align behaviors in multi-party interactions. We evaluate the
effectiveness and generalizability of our approach on three publicly available
datasets for joint engagement, rapport, and human beliefs prediction tasks.
Remarkably, the CPM-T framework outperforms baseline models in average
F1-scores by up to 7.3%, 9.3%, and 2.0% respectively. Finally, we demonstrate
the importance of each component in the framework via ablation studies with
respect to multimodal temporal behavior.
| no_new_dataset | 0.709617 |
2305.12376 | Nadav Borenstein | Nadav Borenstein and Karolina Sta\'nczak and Thea Rolskov and
Nat\'alia da Silva Perez and Natacha Klein K\"afer and Isabelle Augenstein | Measuring Intersectional Biases in Historical Documents | Accepted to Findings of ACL2023 | null | null | null | cs.CL cs.CY cs.LG | http://creativecommons.org/licenses/by/4.0/ | Data-driven analyses of biases in historical texts can help illuminate the
origin and development of biases prevailing in modern society.
However, digitised historical documents pose a challenge for NLP
practitioners as these corpora suffer from errors introduced by optical
character recognition (OCR) and are written in an archaic language. In this
paper, we investigate the continuities and transformations of bias in
historical newspapers published in the Caribbean during the colonial era (18th
to 19th centuries). Our analyses are performed along the axes of gender, race,
and their intersection. We examine these biases by conducting a temporal study
in which we measure the development of lexical associations using
distributional semantics models and word embeddings. Further, we evaluate the
effectiveness of techniques designed to process OCR-generated data and assess
their stability when trained on and applied to the noisy historical newspapers.
We find that there is a trade-off between the stability of the word embeddings
and their compatibility with the historical dataset. We provide evidence that
gender and racial biases are interdependent, and their intersection triggers
distinct effects. These findings align with the theory of intersectionality,
which stresses that biases affecting people with multiple marginalised
identities compound to more than the sum of their constituents.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 07:10:31 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Borenstein",
"Nadav",
""
],
[
"Stańczak",
"Karolina",
""
],
[
"Rolskov",
"Thea",
""
],
[
"Perez",
"Natália da Silva",
""
],
[
"Käfer",
"Natacha Klein",
""
],
[
"Augenstein",
"Isabelle",
""
]
] | TITLE: Measuring Intersectional Biases in Historical Documents
ABSTRACT: Data-driven analyses of biases in historical texts can help illuminate the
origin and development of biases prevailing in modern society.
However, digitised historical documents pose a challenge for NLP
practitioners as these corpora suffer from errors introduced by optical
character recognition (OCR) and are written in an archaic language. In this
paper, we investigate the continuities and transformations of bias in
historical newspapers published in the Caribbean during the colonial era (18th
to 19th centuries). Our analyses are performed along the axes of gender, race,
and their intersection. We examine these biases by conducting a temporal study
in which we measure the development of lexical associations using
distributional semantics models and word embeddings. Further, we evaluate the
effectiveness of techniques designed to process OCR-generated data and assess
their stability when trained on and applied to the noisy historical newspapers.
We find that there is a trade-off between the stability of the word embeddings
and their compatibility with the historical dataset. We provide evidence that
gender and racial biases are interdependent, and their intersection triggers
distinct effects. These findings align with the theory of intersectionality,
which stresses that biases affecting people with multiple marginalised
identities compound to more than the sum of their constituents.
| no_new_dataset | 0.709202 |
2305.12389 | Jun-Yu Ma | Jun-Yu Ma, Jia-Chen Gu, Zhen-Hua Ling, Quan Liu, Cong Liu, Guoping Hu | SHINE: Syntax-augmented Hierarchical Interactive Encoder for Zero-shot
Cross-lingual Information Extraction | 15pages | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Zero-shot cross-lingual information extraction(IE) aims at constructing an IE
model for some low-resource target languages, given annotations exclusively in
some rich-resource languages. Recent studies based on language-universal
features have shown their effectiveness and are attracting increasing
attention. However, prior work has neither explored the potential of
establishing interactions between language-universal features and contextual
representations nor incorporated features that can effectively model
constituent span attributes and relationships between multiple spans. In this
study, a syntax-augmented hierarchical interactive encoder (SHINE) is proposed
to transfer cross-lingual IE knowledge. The proposed encoder is capable of
interactively capturing complementary information between features and
contextual information, to derive language-agnostic representations for various
IE tasks. Concretely, a multi-level interaction network is designed to
hierarchically interact the complementary information to strengthen domain
adaptability. Besides, in addition to the well-studied syntax features of
part-of-speech and dependency relation, a new syntax feature of constituency
structure is introduced to model the constituent span information which is
crucial for IE. Experiments across seven languages on three IE tasks and four
benchmarks verify the effectiveness and generalization ability of the proposed
method.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 08:02:06 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Ma",
"Jun-Yu",
""
],
[
"Gu",
"Jia-Chen",
""
],
[
"Ling",
"Zhen-Hua",
""
],
[
"Liu",
"Quan",
""
],
[
"Liu",
"Cong",
""
],
[
"Hu",
"Guoping",
""
]
] | TITLE: SHINE: Syntax-augmented Hierarchical Interactive Encoder for Zero-shot
Cross-lingual Information Extraction
ABSTRACT: Zero-shot cross-lingual information extraction(IE) aims at constructing an IE
model for some low-resource target languages, given annotations exclusively in
some rich-resource languages. Recent studies based on language-universal
features have shown their effectiveness and are attracting increasing
attention. However, prior work has neither explored the potential of
establishing interactions between language-universal features and contextual
representations nor incorporated features that can effectively model
constituent span attributes and relationships between multiple spans. In this
study, a syntax-augmented hierarchical interactive encoder (SHINE) is proposed
to transfer cross-lingual IE knowledge. The proposed encoder is capable of
interactively capturing complementary information between features and
contextual information, to derive language-agnostic representations for various
IE tasks. Concretely, a multi-level interaction network is designed to
hierarchically interact the complementary information to strengthen domain
adaptability. Besides, in addition to the well-studied syntax features of
part-of-speech and dependency relation, a new syntax feature of constituency
structure is introduced to model the constituent span information which is
crucial for IE. Experiments across seven languages on three IE tasks and four
benchmarks verify the effectiveness and generalization ability of the proposed
method.
| no_new_dataset | 0.708584 |
2305.12398 | Haojun Xu | Haojun Xu, Yan Gao, Zheng Hui, Jie Li, and Xinbo Gao | Language Knowledge-Assisted Representation Learning for Skeleton-Based
Action Recognition | first upload with 13 pages and 8 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | How humans understand and recognize the actions of others is a complex
neuroscientific problem that involves a combination of cognitive mechanisms and
neural networks. Research has shown that humans have brain areas that recognize
actions that process top-down attentional information, such as the
temporoparietal association area. Also, humans have brain regions dedicated to
understanding the minds of others and analyzing their intentions, such as the
medial prefrontal cortex of the temporal lobe. Skeleton-based action
recognition creates mappings for the complex connections between the human
skeleton movement patterns and behaviors. Although existing studies encoded
meaningful node relationships and synthesized action representations for
classification with good results, few of them considered incorporating a priori
knowledge to aid potential representation learning for better performance.
LA-GCN proposes a graph convolution network using large-scale language models
(LLM) knowledge assistance. First, the LLM knowledge is mapped into a priori
global relationship (GPR) topology and a priori category relationship (CPR)
topology between nodes. The GPR guides the generation of new "bone"
representations, aiming to emphasize essential node information from the data
level. The CPR mapping simulates category prior knowledge in human brain
regions, encoded by the PC-AC module and used to add additional
supervision-forcing the model to learn class-distinguishable features. In
addition, to improve information transfer efficiency in topology modeling, we
propose multi-hop attention graph convolution. It aggregates each node's
k-order neighbor simultaneously to speed up model convergence. LA-GCN reaches
state-of-the-art on NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 08:29:16 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Xu",
"Haojun",
""
],
[
"Gao",
"Yan",
""
],
[
"Hui",
"Zheng",
""
],
[
"Li",
"Jie",
""
],
[
"Gao",
"Xinbo",
""
]
] | TITLE: Language Knowledge-Assisted Representation Learning for Skeleton-Based
Action Recognition
ABSTRACT: How humans understand and recognize the actions of others is a complex
neuroscientific problem that involves a combination of cognitive mechanisms and
neural networks. Research has shown that humans have brain areas that recognize
actions that process top-down attentional information, such as the
temporoparietal association area. Also, humans have brain regions dedicated to
understanding the minds of others and analyzing their intentions, such as the
medial prefrontal cortex of the temporal lobe. Skeleton-based action
recognition creates mappings for the complex connections between the human
skeleton movement patterns and behaviors. Although existing studies encoded
meaningful node relationships and synthesized action representations for
classification with good results, few of them considered incorporating a priori
knowledge to aid potential representation learning for better performance.
LA-GCN proposes a graph convolution network using large-scale language models
(LLM) knowledge assistance. First, the LLM knowledge is mapped into a priori
global relationship (GPR) topology and a priori category relationship (CPR)
topology between nodes. The GPR guides the generation of new "bone"
representations, aiming to emphasize essential node information from the data
level. The CPR mapping simulates category prior knowledge in human brain
regions, encoded by the PC-AC module and used to add additional
supervision-forcing the model to learn class-distinguishable features. In
addition, to improve information transfer efficiency in topology modeling, we
propose multi-hop attention graph convolution. It aggregates each node's
k-order neighbor simultaneously to speed up model convergence. LA-GCN reaches
state-of-the-art on NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets.
| no_new_dataset | 0.712038 |
2305.12410 | Shunli Tian | Xiangrong Zhang, Shunli Tian, Guanchun Wang, Huiyu Zhou, and Licheng
Jiao | DiffUCD:Unsupervised Hyperspectral Image Change Detection with Semantic
Correlation Diffusion Model | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hyperspectral image change detection (HSI-CD) has emerged as a crucial
research area in remote sensing due to its ability to detect subtle changes on
the earth's surface. Recently, diffusional denoising probabilistic models
(DDPM) have demonstrated remarkable performance in the generative domain. Apart
from their image generation capability, the denoising process in diffusion
models can comprehensively account for the semantic correlation of
spectral-spatial features in HSI, resulting in the retrieval of semantically
relevant features in the original image. In this work, we extend the diffusion
model's application to the HSI-CD field and propose a novel unsupervised HSI-CD
with semantic correlation diffusion model (DiffUCD). Specifically, the semantic
correlation diffusion model (SCDM) leverages abundant unlabeled samples and
fully accounts for the semantic correlation of spectral-spatial features, which
mitigates pseudo change between multi-temporal images arising from inconsistent
imaging conditions. Besides, objects with the same semantic concept at the same
spatial location may exhibit inconsistent spectral signatures at different
times, resulting in pseudo change. To address this problem, we propose a
cross-temporal contrastive learning (CTCL) mechanism that aligns the spectral
feature representations of unchanged samples. By doing so, the spectral
difference invariant features caused by environmental changes can be obtained.
Experiments conducted on three publicly available datasets demonstrate that the
proposed method outperforms the other state-of-the-art unsupervised methods in
terms of Overall Accuracy (OA), Kappa Coefficient (KC), and F1 scores,
achieving improvements of approximately 3.95%, 8.13%, and 4.45%, respectively.
Notably, our method can achieve comparable results to those fully supervised
methods requiring numerous annotated samples.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 09:21:41 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Zhang",
"Xiangrong",
""
],
[
"Tian",
"Shunli",
""
],
[
"Wang",
"Guanchun",
""
],
[
"Zhou",
"Huiyu",
""
],
[
"Jiao",
"Licheng",
""
]
] | TITLE: DiffUCD:Unsupervised Hyperspectral Image Change Detection with Semantic
Correlation Diffusion Model
ABSTRACT: Hyperspectral image change detection (HSI-CD) has emerged as a crucial
research area in remote sensing due to its ability to detect subtle changes on
the earth's surface. Recently, diffusional denoising probabilistic models
(DDPM) have demonstrated remarkable performance in the generative domain. Apart
from their image generation capability, the denoising process in diffusion
models can comprehensively account for the semantic correlation of
spectral-spatial features in HSI, resulting in the retrieval of semantically
relevant features in the original image. In this work, we extend the diffusion
model's application to the HSI-CD field and propose a novel unsupervised HSI-CD
with semantic correlation diffusion model (DiffUCD). Specifically, the semantic
correlation diffusion model (SCDM) leverages abundant unlabeled samples and
fully accounts for the semantic correlation of spectral-spatial features, which
mitigates pseudo change between multi-temporal images arising from inconsistent
imaging conditions. Besides, objects with the same semantic concept at the same
spatial location may exhibit inconsistent spectral signatures at different
times, resulting in pseudo change. To address this problem, we propose a
cross-temporal contrastive learning (CTCL) mechanism that aligns the spectral
feature representations of unchanged samples. By doing so, the spectral
difference invariant features caused by environmental changes can be obtained.
Experiments conducted on three publicly available datasets demonstrate that the
proposed method outperforms the other state-of-the-art unsupervised methods in
terms of Overall Accuracy (OA), Kappa Coefficient (KC), and F1 scores,
achieving improvements of approximately 3.95%, 8.13%, and 4.45%, respectively.
Notably, our method can achieve comparable results to those fully supervised
methods requiring numerous annotated samples.
| no_new_dataset | 0.710396 |
2305.12412 | Yiyang Li | Yiyang Li, Hai Zhao | EM Pre-training for Multi-party Dialogue Response Generation | Accepted by ACL 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Dialogue response generation requires an agent to generate a response
according to the current dialogue history, in terms of which two-party
dialogues have been well studied, but leaving a great gap for multi-party
dialogues at the same time. Different from two-party dialogues where each
response is a direct reply to its previous utterance, the addressee of a
response utterance should be specified before it is generated in the
multi-party scenario. Thanks to the huge amount of two-party conversational
data, various pre-trained language models for two-party dialogue response
generation have been proposed. However, due to the lack of annotated addressee
labels in multi-party dialogue datasets, it is hard to use them to pre-train a
response generation model for multi-party dialogues. To tackle this obstacle,
we propose an Expectation-Maximization (EM) approach that iteratively performs
the expectation steps to generate addressee labels, and the maximization steps
to optimize a response generation model. Theoretical analyses and extensive
experiments have justified the feasibility and effectiveness of our proposed
method.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 09:22:41 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Li",
"Yiyang",
""
],
[
"Zhao",
"Hai",
""
]
] | TITLE: EM Pre-training for Multi-party Dialogue Response Generation
ABSTRACT: Dialogue response generation requires an agent to generate a response
according to the current dialogue history, in terms of which two-party
dialogues have been well studied, but leaving a great gap for multi-party
dialogues at the same time. Different from two-party dialogues where each
response is a direct reply to its previous utterance, the addressee of a
response utterance should be specified before it is generated in the
multi-party scenario. Thanks to the huge amount of two-party conversational
data, various pre-trained language models for two-party dialogue response
generation have been proposed. However, due to the lack of annotated addressee
labels in multi-party dialogue datasets, it is hard to use them to pre-train a
response generation model for multi-party dialogues. To tackle this obstacle,
we propose an Expectation-Maximization (EM) approach that iteratively performs
the expectation steps to generate addressee labels, and the maximization steps
to optimize a response generation model. Theoretical analyses and extensive
experiments have justified the feasibility and effectiveness of our proposed
method.
| no_new_dataset | 0.709856 |
2305.12417 | Juan M. Gandarias | Juan M. Gandarias (1), Alfonso J. Garc\'ia-Cerezo (1), Jes\'us M.
G\'omez-de-Gabriel (1) ((1) Robotics and Mechatronics, Systems Engineering
and Automation Department, University of Malaga) | CNN-based Methods for Object Recognition with High-Resolution Tactile
Sensors | null | null | 10.1109/JSEN.2019.2912968 | null | cs.CV cs.AI cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Novel high-resolution pressure-sensor arrays allow treating pressure readings
as standard images. Computer vision algorithms and methods such as
Convolutional Neural Networks (CNN) can be used to identify contact objects. In
this paper, a high-resolution tactile sensor has been attached to a robotic
end-effector to identify contacted objects. Two CNN-based approaches have been
employed to classify pressure images. These methods include a transfer learning
approach using a pre-trained CNN on an RGB-images dataset and a custom-made CNN
(TactNet) trained from scratch with tactile information. The transfer learning
approach can be carried out by retraining the classification layers of the
network or replacing these layers with an SVM. Overall, 11 configurations based
on these methods have been tested: 8 transfer learning-based, and 3
TactNet-based. Moreover, a study of the performance of the methods and a
comparative discussion with the current state-of-the-art on tactile object
recognition is presented.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 09:54:12 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Gandarias",
"Juan M.",
""
],
[
"García-Cerezo",
"Alfonso J.",
""
],
[
"Gómez-de-Gabriel",
"Jesús M.",
""
]
] | TITLE: CNN-based Methods for Object Recognition with High-Resolution Tactile
Sensors
ABSTRACT: Novel high-resolution pressure-sensor arrays allow treating pressure readings
as standard images. Computer vision algorithms and methods such as
Convolutional Neural Networks (CNN) can be used to identify contact objects. In
this paper, a high-resolution tactile sensor has been attached to a robotic
end-effector to identify contacted objects. Two CNN-based approaches have been
employed to classify pressure images. These methods include a transfer learning
approach using a pre-trained CNN on an RGB-images dataset and a custom-made CNN
(TactNet) trained from scratch with tactile information. The transfer learning
approach can be carried out by retraining the classification layers of the
network or replacing these layers with an SVM. Overall, 11 configurations based
on these methods have been tested: 8 transfer learning-based, and 3
TactNet-based. Moreover, a study of the performance of the methods and a
comparative discussion with the current state-of-the-art on tactile object
recognition is presented.
| no_new_dataset | 0.71022 |
2305.12424 | Tetsuya Kobayashi J | Mengji Zhang, Yusuke Hiki, Akira Funahashi, Tetsuya J. Kobayashi | Mol-PECO: a deep learning model to predict human olfactory perception
from molecular structures | 17 pages, 8 figures | null | null | null | cs.LG cs.AI q-bio.BM q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | While visual and auditory information conveyed by wavelength of light and
frequency of sound have been decoded, predicting olfactory information encoded
by the combination of odorants remains challenging due to the unknown and
potentially discontinuous perceptual space of smells and odorants. Herein, we
develop a deep learning model called Mol-PECO (Molecular Representation by
Positional Encoding of Coulomb Matrix) to predict olfactory perception from
molecular structures. Mol-PECO updates the learned atom embedding by
directional graph convolutional networks (GCN), which model the Laplacian
eigenfunctions as positional encoding, and Coulomb matrix, which encodes atomic
coordinates and charges. With a comprehensive dataset of 8,503 molecules,
Mol-PECO directly achieves an area-under-the-receiver-operating-characteristic
(AUROC) of 0.813 in 118 odor descriptors, superior to the machine learning of
molecular fingerprints (AUROC of 0.761) and GCN of adjacency matrix (AUROC of
0.678). The learned embeddings by Mol-PECO also capture a meaningful odor space
with global clustering of descriptors and local retrieval of similar odorants.
Our work may promote the understanding and decoding of the olfactory sense and
mechanisms.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 10:44:02 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Zhang",
"Mengji",
""
],
[
"Hiki",
"Yusuke",
""
],
[
"Funahashi",
"Akira",
""
],
[
"Kobayashi",
"Tetsuya J.",
""
]
] | TITLE: Mol-PECO: a deep learning model to predict human olfactory perception
from molecular structures
ABSTRACT: While visual and auditory information conveyed by wavelength of light and
frequency of sound have been decoded, predicting olfactory information encoded
by the combination of odorants remains challenging due to the unknown and
potentially discontinuous perceptual space of smells and odorants. Herein, we
develop a deep learning model called Mol-PECO (Molecular Representation by
Positional Encoding of Coulomb Matrix) to predict olfactory perception from
molecular structures. Mol-PECO updates the learned atom embedding by
directional graph convolutional networks (GCN), which model the Laplacian
eigenfunctions as positional encoding, and Coulomb matrix, which encodes atomic
coordinates and charges. With a comprehensive dataset of 8,503 molecules,
Mol-PECO directly achieves an area-under-the-receiver-operating-characteristic
(AUROC) of 0.813 in 118 odor descriptors, superior to the machine learning of
molecular fingerprints (AUROC of 0.761) and GCN of adjacency matrix (AUROC of
0.678). The learned embeddings by Mol-PECO also capture a meaningful odor space
with global clustering of descriptors and local retrieval of similar odorants.
Our work may promote the understanding and decoding of the olfactory sense and
mechanisms.
| no_new_dataset | 0.711638 |
2305.12434 | Wenxuan Wang | Yuxuan Wan, Wenxuan Wang, Pinjia He, Jiazhen Gu, Haonan Bai, Michael
Lyu | BiasAsker: Measuring the Bias in Conversational AI System | Accepted by FSE 2023 | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Powered by advanced Artificial Intelligence (AI) techniques, conversational
AI systems, such as ChatGPT and digital assistants like Siri, have been widely
deployed in daily life. However, such systems may still produce content
containing biases and stereotypes, causing potential social problems. Due to
the data-driven, black-box nature of modern AI techniques, comprehensively
identifying and measuring biases in conversational systems remains a
challenging task. Particularly, it is hard to generate inputs that can
comprehensively trigger potential bias due to the lack of data containing both
social groups as well as biased properties. In addition, modern conversational
systems can produce diverse responses (e.g., chatting and explanation), which
makes existing bias detection methods simply based on the sentiment and the
toxicity hardly being adopted. In this paper, we propose BiasAsker, an
automated framework to identify and measure social bias in conversational AI
systems. To obtain social groups and biased properties, we construct a
comprehensive social bias dataset, containing a total of 841 groups and 8,110
biased properties. Given the dataset, BiasAsker automatically generates
questions and adopts a novel method based on existence measurement to identify
two types of biases (i.e., absolute bias and related bias) in conversational
systems. Extensive experiments on 8 commercial systems and 2 famous research
models, such as ChatGPT and GPT-3, show that 32.83% of the questions generated
by BiasAsker can trigger biased behaviors in these widely deployed
conversational systems. All the code, data, and experimental results have been
released to facilitate future research.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 11:25:59 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Wan",
"Yuxuan",
""
],
[
"Wang",
"Wenxuan",
""
],
[
"He",
"Pinjia",
""
],
[
"Gu",
"Jiazhen",
""
],
[
"Bai",
"Haonan",
""
],
[
"Lyu",
"Michael",
""
]
] | TITLE: BiasAsker: Measuring the Bias in Conversational AI System
ABSTRACT: Powered by advanced Artificial Intelligence (AI) techniques, conversational
AI systems, such as ChatGPT and digital assistants like Siri, have been widely
deployed in daily life. However, such systems may still produce content
containing biases and stereotypes, causing potential social problems. Due to
the data-driven, black-box nature of modern AI techniques, comprehensively
identifying and measuring biases in conversational systems remains a
challenging task. Particularly, it is hard to generate inputs that can
comprehensively trigger potential bias due to the lack of data containing both
social groups as well as biased properties. In addition, modern conversational
systems can produce diverse responses (e.g., chatting and explanation), which
makes existing bias detection methods simply based on the sentiment and the
toxicity hardly being adopted. In this paper, we propose BiasAsker, an
automated framework to identify and measure social bias in conversational AI
systems. To obtain social groups and biased properties, we construct a
comprehensive social bias dataset, containing a total of 841 groups and 8,110
biased properties. Given the dataset, BiasAsker automatically generates
questions and adopts a novel method based on existence measurement to identify
two types of biases (i.e., absolute bias and related bias) in conversational
systems. Extensive experiments on 8 commercial systems and 2 famous research
models, such as ChatGPT and GPT-3, show that 32.83% of the questions generated
by BiasAsker can trigger biased behaviors in these widely deployed
conversational systems. All the code, data, and experimental results have been
released to facilitate future research.
| new_dataset | 0.716801 |
2305.12450 | Mohan Shi | Mohan Shi, Yuchun Shu, Lingyun Zuo, Qian Chen, Shiliang Zhang, Jie
Zhang, Li-Rong Dai | Semantic VAD: Low-Latency Voice Activity Detection for Speech
Interaction | Accepted by Interspeech2023 | null | null | null | eess.AS cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For speech interaction, voice activity detection (VAD) is often used as a
front-end. However, traditional VAD algorithms usually need to wait for a
continuous tail silence to reach a preset maximum duration before segmentation,
resulting in a large latency that affects user experience. In this paper, we
propose a novel semantic VAD for low-latency segmentation. Different from
existing methods, a frame-level punctuation prediction task is added to the
semantic VAD, and the artificial endpoint is included in the classification
category in addition to the often-used speech presence and absence. To enhance
the semantic information of the model, we also incorporate an automatic speech
recognition (ASR) related semantic loss. Evaluations on an internal dataset
show that the proposed method can reduce the average latency by 53.3% without
significant deterioration of character error rate in the back-end ASR compared
to the traditional VAD approach.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 13:02:53 GMT"
}
] | 2023-05-23T00:00:00 | [
[
"Shi",
"Mohan",
""
],
[
"Shu",
"Yuchun",
""
],
[
"Zuo",
"Lingyun",
""
],
[
"Chen",
"Qian",
""
],
[
"Zhang",
"Shiliang",
""
],
[
"Zhang",
"Jie",
""
],
[
"Dai",
"Li-Rong",
""
]
] | TITLE: Semantic VAD: Low-Latency Voice Activity Detection for Speech
Interaction
ABSTRACT: For speech interaction, voice activity detection (VAD) is often used as a
front-end. However, traditional VAD algorithms usually need to wait for a
continuous tail silence to reach a preset maximum duration before segmentation,
resulting in a large latency that affects user experience. In this paper, we
propose a novel semantic VAD for low-latency segmentation. Different from
existing methods, a frame-level punctuation prediction task is added to the
semantic VAD, and the artificial endpoint is included in the classification
category in addition to the often-used speech presence and absence. To enhance
the semantic information of the model, we also incorporate an automatic speech
recognition (ASR) related semantic loss. Evaluations on an internal dataset
show that the proposed method can reduce the average latency by 53.3% without
significant deterioration of character error rate in the back-end ASR compared
to the traditional VAD approach.
| no_new_dataset | 0.712045 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.