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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2307.02744 | Shuvendu Roy | Shuvendu Roy, Ali Etemad | Active Learning with Contrastive Pre-training for Facial Expression
Recognition | Accepted at 11th International Conference on Affective Computing and
Intelligent Interaction (ACII'2023) | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Deep learning has played a significant role in the success of facial
expression recognition (FER), thanks to large models and vast amounts of
labelled data. However, obtaining labelled data requires a tremendous amount of
human effort, time, and financial resources. Even though some prior works have
focused on reducing the need for large amounts of labelled data using different
unsupervised methods, another promising approach called active learning is
barely explored in the context of FER. This approach involves selecting and
labelling the most representative samples from an unlabelled set to make the
best use of a limited 'labelling budget'. In this paper, we implement and study
8 recent active learning methods on three public FER datasets, FER13, RAF-DB,
and KDEF. Our findings show that existing active learning methods do not
perform well in the context of FER, likely suffering from a phenomenon called
'Cold Start', which occurs when the initial set of labelled samples is not well
representative of the entire dataset. To address this issue, we propose
contrastive self-supervised pre-training, which first learns the underlying
representations based on the entire unlabelled dataset. We then follow this
with the active learning methods and observe that our 2-step approach shows up
to 9.2% improvement over random sampling and up to 6.7% improvement over the
best existing active learning baseline without the pre-training. We will make
the code for this study public upon publication at:
github.com/ShuvenduRoy/ActiveFER.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 03:08:03 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Roy",
"Shuvendu",
""
],
[
"Etemad",
"Ali",
""
]
] | TITLE: Active Learning with Contrastive Pre-training for Facial Expression
Recognition
ABSTRACT: Deep learning has played a significant role in the success of facial
expression recognition (FER), thanks to large models and vast amounts of
labelled data. However, obtaining labelled data requires a tremendous amount of
human effort, time, and financial resources. Even though some prior works have
focused on reducing the need for large amounts of labelled data using different
unsupervised methods, another promising approach called active learning is
barely explored in the context of FER. This approach involves selecting and
labelling the most representative samples from an unlabelled set to make the
best use of a limited 'labelling budget'. In this paper, we implement and study
8 recent active learning methods on three public FER datasets, FER13, RAF-DB,
and KDEF. Our findings show that existing active learning methods do not
perform well in the context of FER, likely suffering from a phenomenon called
'Cold Start', which occurs when the initial set of labelled samples is not well
representative of the entire dataset. To address this issue, we propose
contrastive self-supervised pre-training, which first learns the underlying
representations based on the entire unlabelled dataset. We then follow this
with the active learning methods and observe that our 2-step approach shows up
to 9.2% improvement over random sampling and up to 6.7% improvement over the
best existing active learning baseline without the pre-training. We will make
the code for this study public upon publication at:
github.com/ShuvenduRoy/ActiveFER.
| no_new_dataset | 0.707834 |
2307.02747 | Tiankui Zhang | Yuanpeng Zheng, Tiankui Zhang, Rong Huang and Yapeng Wang | Computing Offloading and Semantic Compression for Intelligent Computing
Tasks in MEC Systems | null | null | 10.1109/WCNC55385.2023.10118995 | null | cs.NI eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates the intelligent computing task-oriented computing
offloading and semantic compression in mobile edge computing (MEC) systems.
With the popularity of intelligent applications in various industries,
terminals increasingly need to offload intelligent computing tasks with complex
demands to MEC servers for computing, which is a great challenge for bandwidth
and computing capacity allocation in MEC systems. Considering the accuracy
requirement of intelligent computing tasks, we formulate an optimization
problem of computing offloading and semantic compression. We jointly optimize
the system utility which are represented as computing accuracy and task delay
respectively to acquire the optimized system utility. To solve the proposed
optimization problem, we decompose it into computing capacity allocation
subproblem and compression offloading subproblem and obtain solutions through
convex optimization and successive convex approximation. After that, the
offloading decisions, computing capacity and compressed ratio are obtained in
closed forms. We design the computing offloading and semantic compression
algorithm for intelligent computing tasks in MEC systems then. Simulation
results represent that our algorithm converges quickly and acquires better
performance and resource utilization efficiency through the trend with total
number of users and computing capacity compared with benchmarks.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 03:15:58 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Zheng",
"Yuanpeng",
""
],
[
"Zhang",
"Tiankui",
""
],
[
"Huang",
"Rong",
""
],
[
"Wang",
"Yapeng",
""
]
] | TITLE: Computing Offloading and Semantic Compression for Intelligent Computing
Tasks in MEC Systems
ABSTRACT: This paper investigates the intelligent computing task-oriented computing
offloading and semantic compression in mobile edge computing (MEC) systems.
With the popularity of intelligent applications in various industries,
terminals increasingly need to offload intelligent computing tasks with complex
demands to MEC servers for computing, which is a great challenge for bandwidth
and computing capacity allocation in MEC systems. Considering the accuracy
requirement of intelligent computing tasks, we formulate an optimization
problem of computing offloading and semantic compression. We jointly optimize
the system utility which are represented as computing accuracy and task delay
respectively to acquire the optimized system utility. To solve the proposed
optimization problem, we decompose it into computing capacity allocation
subproblem and compression offloading subproblem and obtain solutions through
convex optimization and successive convex approximation. After that, the
offloading decisions, computing capacity and compressed ratio are obtained in
closed forms. We design the computing offloading and semantic compression
algorithm for intelligent computing tasks in MEC systems then. Simulation
results represent that our algorithm converges quickly and acquires better
performance and resource utilization efficiency through the trend with total
number of users and computing capacity compared with benchmarks.
| no_new_dataset | 0.709466 |
2307.02748 | Tiankui Zhang | Yuanpeng Zheng, Tiankui Zhang and Jonathan Loo | Dynamic Multi-time Scale User Admission and Resource Allocation for
Semantic Extraction in MEC Systems | null | null | 10.1109/TVT.2023.3290546 | null | cs.NI eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates the semantic extraction task-oriented dynamic
multi-time scale user admission and resourceallocation in mobile edge computing
(MEC) systems. Amid prevalence artifi cial intelligence applications in various
industries,the offloading of semantic extraction tasks which are mainlycomposed
of convolutional neural networks of computer vision isa great challenge for
communication bandwidth and computing capacity allocation in MEC systems.
Considering the stochasticnature of the semantic extraction tasks, we formulate
a stochastic optimization problem by modeling it as the dynamic arrival of
tasks in the temporal domain. We jointly optimize the system revenue and cost
which are represented as user admission in the long term and resource
allocation in the short term respectively. To handle the proposed stochastic
optimization problem, we decompose it into short-time-scale subproblems and a
long-time-scale subproblem by using the Lyapunov optimization technique. After
that, the short-time-scale optimization variables of resource allocation,
including user association, bandwidth allocation, and computing capacity
allocation are obtained in closed form. The user admission optimization on
long-time scales is solved by a heuristic iteration method. Then, the
multi-time scale user admission and resource allocation algorithm is proposed
for dynamic semantic extraction task computing in MEC systems. Simulation
results demonstrate that, compared with the benchmarks, the proposed algorithm
improves the performance of user admission and resource allocation efficiently
and achieves a flexible trade-off between system revenue and cost at multi-time
scales and considering semantic extraction tasks.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 03:16:36 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Zheng",
"Yuanpeng",
""
],
[
"Zhang",
"Tiankui",
""
],
[
"Loo",
"Jonathan",
""
]
] | TITLE: Dynamic Multi-time Scale User Admission and Resource Allocation for
Semantic Extraction in MEC Systems
ABSTRACT: This paper investigates the semantic extraction task-oriented dynamic
multi-time scale user admission and resourceallocation in mobile edge computing
(MEC) systems. Amid prevalence artifi cial intelligence applications in various
industries,the offloading of semantic extraction tasks which are mainlycomposed
of convolutional neural networks of computer vision isa great challenge for
communication bandwidth and computing capacity allocation in MEC systems.
Considering the stochasticnature of the semantic extraction tasks, we formulate
a stochastic optimization problem by modeling it as the dynamic arrival of
tasks in the temporal domain. We jointly optimize the system revenue and cost
which are represented as user admission in the long term and resource
allocation in the short term respectively. To handle the proposed stochastic
optimization problem, we decompose it into short-time-scale subproblems and a
long-time-scale subproblem by using the Lyapunov optimization technique. After
that, the short-time-scale optimization variables of resource allocation,
including user association, bandwidth allocation, and computing capacity
allocation are obtained in closed form. The user admission optimization on
long-time scales is solved by a heuristic iteration method. Then, the
multi-time scale user admission and resource allocation algorithm is proposed
for dynamic semantic extraction task computing in MEC systems. Simulation
results demonstrate that, compared with the benchmarks, the proposed algorithm
improves the performance of user admission and resource allocation efficiently
and achieves a flexible trade-off between system revenue and cost at multi-time
scales and considering semantic extraction tasks.
| no_new_dataset | 0.709447 |
2307.02753 | Jincheng Lu | Jincheng Lu, Xipeng Yang, Jin Ye, Yifu Zhang, Zhikang Zou, Wei Zhang,
Xiao Tan | CityTrack: Improving City-Scale Multi-Camera Multi-Target Tracking by
Location-Aware Tracking and Box-Grained Matching | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-Camera Multi-Target Tracking (MCMT) is a computer vision technique that
involves tracking multiple targets simultaneously across multiple cameras. MCMT
in urban traffic visual analysis faces great challenges due to the complex and
dynamic nature of urban traffic scenes, where multiple cameras with different
views and perspectives are often used to cover a large city-scale area. Targets
in urban traffic scenes often undergo occlusion, illumination changes, and
perspective changes, making it difficult to associate targets across different
cameras accurately. To overcome these challenges, we propose a novel systematic
MCMT framework, called CityTrack. Specifically, we present a Location-Aware
SCMT tracker which integrates various advanced techniques to improve its
effectiveness in the MCMT task and propose a novel Box-Grained Matching (BGM)
method for the ICA module to solve the aforementioned problems. We evaluated
our approach on the public test set of the CityFlowV2 dataset and achieved an
IDF1 of 84.91%, ranking 1st in the 2022 AI CITY CHALLENGE. Our experimental
results demonstrate the effectiveness of our approach in overcoming the
challenges posed by urban traffic scenes.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 03:25:37 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Lu",
"Jincheng",
""
],
[
"Yang",
"Xipeng",
""
],
[
"Ye",
"Jin",
""
],
[
"Zhang",
"Yifu",
""
],
[
"Zou",
"Zhikang",
""
],
[
"Zhang",
"Wei",
""
],
[
"Tan",
"Xiao",
""
]
] | TITLE: CityTrack: Improving City-Scale Multi-Camera Multi-Target Tracking by
Location-Aware Tracking and Box-Grained Matching
ABSTRACT: Multi-Camera Multi-Target Tracking (MCMT) is a computer vision technique that
involves tracking multiple targets simultaneously across multiple cameras. MCMT
in urban traffic visual analysis faces great challenges due to the complex and
dynamic nature of urban traffic scenes, where multiple cameras with different
views and perspectives are often used to cover a large city-scale area. Targets
in urban traffic scenes often undergo occlusion, illumination changes, and
perspective changes, making it difficult to associate targets across different
cameras accurately. To overcome these challenges, we propose a novel systematic
MCMT framework, called CityTrack. Specifically, we present a Location-Aware
SCMT tracker which integrates various advanced techniques to improve its
effectiveness in the MCMT task and propose a novel Box-Grained Matching (BGM)
method for the ICA module to solve the aforementioned problems. We evaluated
our approach on the public test set of the CityFlowV2 dataset and achieved an
IDF1 of 84.91%, ranking 1st in the 2022 AI CITY CHALLENGE. Our experimental
results demonstrate the effectiveness of our approach in overcoming the
challenges posed by urban traffic scenes.
| no_new_dataset | 0.710415 |
2307.02759 | Yuhao Yang | Yuhao Yang, Chao Huang, Lianghao Xia, Chunzhen Huang | Knowledge Graph Self-Supervised Rationalization for Recommendation | Accepted by KDD'23 | null | 10.1145/3580305.3599400 | null | cs.IR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduce a new self-supervised rationalization method,
called KGRec, for knowledge-aware recommender systems. To effectively identify
informative knowledge connections, we propose an attentive knowledge
rationalization mechanism that generates rational scores for knowledge
triplets. With these scores, KGRec integrates generative and contrastive
self-supervised tasks for recommendation through rational masking. To highlight
rationales in the knowledge graph, we design a novel generative task in the
form of masking-reconstructing. By masking important knowledge with high
rational scores, KGRec is trained to rebuild and highlight useful knowledge
connections that serve as rationales. To further rationalize the effect of
collaborative interactions on knowledge graph learning, we introduce a
contrastive learning task that aligns signals from knowledge and user-item
interaction views. To ensure noise-resistant contrasting, potential noisy edges
in both graphs judged by the rational scores are masked. Extensive experiments
on three real-world datasets demonstrate that KGRec outperforms
state-of-the-art methods. We also provide the implementation codes for our
approach at https://github.com/HKUDS/KGRec.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 03:44:40 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Yang",
"Yuhao",
""
],
[
"Huang",
"Chao",
""
],
[
"Xia",
"Lianghao",
""
],
[
"Huang",
"Chunzhen",
""
]
] | TITLE: Knowledge Graph Self-Supervised Rationalization for Recommendation
ABSTRACT: In this paper, we introduce a new self-supervised rationalization method,
called KGRec, for knowledge-aware recommender systems. To effectively identify
informative knowledge connections, we propose an attentive knowledge
rationalization mechanism that generates rational scores for knowledge
triplets. With these scores, KGRec integrates generative and contrastive
self-supervised tasks for recommendation through rational masking. To highlight
rationales in the knowledge graph, we design a novel generative task in the
form of masking-reconstructing. By masking important knowledge with high
rational scores, KGRec is trained to rebuild and highlight useful knowledge
connections that serve as rationales. To further rationalize the effect of
collaborative interactions on knowledge graph learning, we introduce a
contrastive learning task that aligns signals from knowledge and user-item
interaction views. To ensure noise-resistant contrasting, potential noisy edges
in both graphs judged by the rational scores are masked. Extensive experiments
on three real-world datasets demonstrate that KGRec outperforms
state-of-the-art methods. We also provide the implementation codes for our
approach at https://github.com/HKUDS/KGRec.
| no_new_dataset | 0.709573 |
2307.02761 | Haokai Ma | Haokai Ma, Zhuang Qi, Xinxin Dong, Xiangxian Li, Yuze Zheng, Xiangxu
Mengand Lei Meng | Cross-Modal Content Inference and Feature Enrichment for Cold-Start
Recommendation | Accepted by International Joint Conference on Neural Networks 2023
(IJCNN) | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimedia recommendation aims to fuse the multi-modal information of items
for feature enrichment to improve the recommendation performance. However,
existing methods typically introduce multi-modal information based on
collaborative information to improve the overall recommendation precision,
while failing to explore its cold-start recommendation performance. Meanwhile,
these above methods are only applicable when such multi-modal data is
available. To address this problem, this paper proposes a recommendation
framework, named Cross-modal Content Inference and Feature Enrichment
Recommendation (CIERec), which exploits the multi-modal information to improve
its cold-start recommendation performance. Specifically, CIERec first
introduces image annotation as the privileged information to help guide the
mapping of unified features from the visual space to the semantic space in the
training phase. And then CIERec enriches the content representation with the
fusion of collaborative, visual, and cross-modal inferred representations, so
as to improve its cold-start recommendation performance. Experimental results
on two real-world datasets show that the content representations learned by
CIERec are able to achieve superior cold-start recommendation performance over
existing visually-aware recommendation algorithms. More importantly, CIERec can
consistently achieve significant improvements with different conventional
visually-aware backbones, which verifies its universality and effectiveness.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 03:54:19 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Ma",
"Haokai",
""
],
[
"Qi",
"Zhuang",
""
],
[
"Dong",
"Xinxin",
""
],
[
"Li",
"Xiangxian",
""
],
[
"Zheng",
"Yuze",
""
],
[
"Meng",
"Xiangxu Mengand Lei",
""
]
] | TITLE: Cross-Modal Content Inference and Feature Enrichment for Cold-Start
Recommendation
ABSTRACT: Multimedia recommendation aims to fuse the multi-modal information of items
for feature enrichment to improve the recommendation performance. However,
existing methods typically introduce multi-modal information based on
collaborative information to improve the overall recommendation precision,
while failing to explore its cold-start recommendation performance. Meanwhile,
these above methods are only applicable when such multi-modal data is
available. To address this problem, this paper proposes a recommendation
framework, named Cross-modal Content Inference and Feature Enrichment
Recommendation (CIERec), which exploits the multi-modal information to improve
its cold-start recommendation performance. Specifically, CIERec first
introduces image annotation as the privileged information to help guide the
mapping of unified features from the visual space to the semantic space in the
training phase. And then CIERec enriches the content representation with the
fusion of collaborative, visual, and cross-modal inferred representations, so
as to improve its cold-start recommendation performance. Experimental results
on two real-world datasets show that the content representations learned by
CIERec are able to achieve superior cold-start recommendation performance over
existing visually-aware recommendation algorithms. More importantly, CIERec can
consistently achieve significant improvements with different conventional
visually-aware backbones, which verifies its universality and effectiveness.
| no_new_dataset | 0.709856 |
2307.02763 | David Jurgens | David Jurgens, Agrima Seth, Jackson Sargent, Athena Aghighi, Michael
Geraci | Your spouse needs professional help: Determining the Contextual
Appropriateness of Messages through Modeling Social Relationships | ACL 2023, 18 pages, 8 figures, 11 tables | null | null | null | cs.CL cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding interpersonal communication requires, in part, understanding
the social context and norms in which a message is said. However, current
methods for identifying offensive content in such communication largely operate
independent of context, with only a few approaches considering community norms
or prior conversation as context. Here, we introduce a new approach to
identifying inappropriate communication by explicitly modeling the social
relationship between the individuals. We introduce a new dataset of
contextually-situated judgments of appropriateness and show that large language
models can readily incorporate relationship information to accurately identify
appropriateness in a given context. Using data from online conversations and
movie dialogues, we provide insight into how the relationships themselves
function as implicit norms and quantify the degree to which context-sensitivity
is needed in different conversation settings. Further, we also demonstrate that
contextual-appropriateness judgments are predictive of other social factors
expressed in language such as condescension and politeness.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 04:06:05 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Jurgens",
"David",
""
],
[
"Seth",
"Agrima",
""
],
[
"Sargent",
"Jackson",
""
],
[
"Aghighi",
"Athena",
""
],
[
"Geraci",
"Michael",
""
]
] | TITLE: Your spouse needs professional help: Determining the Contextual
Appropriateness of Messages through Modeling Social Relationships
ABSTRACT: Understanding interpersonal communication requires, in part, understanding
the social context and norms in which a message is said. However, current
methods for identifying offensive content in such communication largely operate
independent of context, with only a few approaches considering community norms
or prior conversation as context. Here, we introduce a new approach to
identifying inappropriate communication by explicitly modeling the social
relationship between the individuals. We introduce a new dataset of
contextually-situated judgments of appropriateness and show that large language
models can readily incorporate relationship information to accurately identify
appropriateness in a given context. Using data from online conversations and
movie dialogues, we provide insight into how the relationships themselves
function as implicit norms and quantify the degree to which context-sensitivity
is needed in different conversation settings. Further, we also demonstrate that
contextual-appropriateness judgments are predictive of other social factors
expressed in language such as condescension and politeness.
| new_dataset | 0.716597 |
2307.02768 | Y-Lan Boureau | Mounica Maddela, Megan Ung, Jing Xu, Andrea Madotto, Heather Foran,
Y-Lan Boureau | Training Models to Generate, Recognize, and Reframe Unhelpful Thoughts | ACL 2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many cognitive approaches to well-being, such as recognizing and reframing
unhelpful thoughts, have received considerable empirical support over the past
decades, yet still lack truly widespread adoption in self-help format. A
barrier to that adoption is a lack of adequately specific and diverse dedicated
practice material. This work examines whether current language models can be
leveraged to both produce a virtually unlimited quantity of practice material
illustrating standard unhelpful thought patterns matching specific given
contexts, and generate suitable positive reframing proposals. We propose
PATTERNREFRAME, a novel dataset of about 10k examples of thoughts containing
unhelpful thought patterns conditioned on a given persona, accompanied by about
27k positive reframes. By using this dataset to train and/or evaluate current
models, we show that existing models can already be powerful tools to help
generate an abundance of tailored practice material and hypotheses, with no or
minimal additional model training required.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 04:40:52 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Maddela",
"Mounica",
""
],
[
"Ung",
"Megan",
""
],
[
"Xu",
"Jing",
""
],
[
"Madotto",
"Andrea",
""
],
[
"Foran",
"Heather",
""
],
[
"Boureau",
"Y-Lan",
""
]
] | TITLE: Training Models to Generate, Recognize, and Reframe Unhelpful Thoughts
ABSTRACT: Many cognitive approaches to well-being, such as recognizing and reframing
unhelpful thoughts, have received considerable empirical support over the past
decades, yet still lack truly widespread adoption in self-help format. A
barrier to that adoption is a lack of adequately specific and diverse dedicated
practice material. This work examines whether current language models can be
leveraged to both produce a virtually unlimited quantity of practice material
illustrating standard unhelpful thought patterns matching specific given
contexts, and generate suitable positive reframing proposals. We propose
PATTERNREFRAME, a novel dataset of about 10k examples of thoughts containing
unhelpful thought patterns conditioned on a given persona, accompanied by about
27k positive reframes. By using this dataset to train and/or evaluate current
models, we show that existing models can already be powerful tools to help
generate an abundance of tailored practice material and hypotheses, with no or
minimal additional model training required.
| new_dataset | 0.715579 |
2307.02773 | Shwetank Choudhary | Tuneer Khargonkar, Shwetank Choudhary, Sumit Kumar, Barath Raj KR | SeLiNet: Sentiment enriched Lightweight Network for Emotion Recognition
in Images | Paper submitted in ISCAS 2023 | null | null | null | cs.CV cs.HC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we propose a sentiment-enriched lightweight network SeLiNet
and an end-to-end on-device pipeline for contextual emotion recognition in
images. SeLiNet model consists of body feature extractor, image aesthetics
feature extractor, and learning-based fusion network which jointly estimates
discrete emotion and human sentiments tasks. On the EMOTIC dataset, the
proposed approach achieves an Average Precision (AP) score of 27.17 in
comparison to the baseline AP score of 27.38 while reducing the model size by
>85%. In addition, we report an on-device AP score of 26.42 with reduction in
model size by >93% when compared to the baseline.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 04:49:47 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Khargonkar",
"Tuneer",
""
],
[
"Choudhary",
"Shwetank",
""
],
[
"Kumar",
"Sumit",
""
],
[
"KR",
"Barath Raj",
""
]
] | TITLE: SeLiNet: Sentiment enriched Lightweight Network for Emotion Recognition
in Images
ABSTRACT: In this paper, we propose a sentiment-enriched lightweight network SeLiNet
and an end-to-end on-device pipeline for contextual emotion recognition in
images. SeLiNet model consists of body feature extractor, image aesthetics
feature extractor, and learning-based fusion network which jointly estimates
discrete emotion and human sentiments tasks. On the EMOTIC dataset, the
proposed approach achieves an Average Precision (AP) score of 27.17 in
comparison to the baseline AP score of 27.38 while reducing the model size by
>85%. In addition, we report an on-device AP score of 26.42 with reduction in
model size by >93% when compared to the baseline.
| no_new_dataset | 0.71202 |
2307.02797 | Weihua Li | Mengyan Wang, Yuxuan Hu, Zihan Yuan, Chenting Jiang, Weihua Li,
Shiqing Wu and Quan Bai | BHEISR: Nudging from Bias to Balance -- Promoting Belief Harmony by
Eliminating Ideological Segregation in Knowledge-based Recommendations | 26 pages | null | null | null | cs.IR cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In the realm of personalized recommendation systems, the increasing concern
is the amplification of belief imbalance and user biases, a phenomenon
primarily attributed to the filter bubble. Addressing this critical issue, we
introduce an innovative intermediate agency (BHEISR) between users and existing
recommendation systems to attenuate the negative repercussions of the filter
bubble effect in extant recommendation systems. The main objective is to strike
a belief balance for users while minimizing the detrimental influence caused by
filter bubbles. The BHEISR model amalgamates principles from nudge theory while
upholding democratic and transparent principles. It harnesses user-specific
category information to stimulate curiosity, even in areas users might
initially deem uninteresting. By progressively stimulating interest in novel
categories, the model encourages users to broaden their belief horizons and
explore the information they typically overlook. Our model is time-sensitive
and operates on a user feedback loop. It utilizes the existing recommendation
algorithm of the model and incorporates user feedback from the prior time
frame. This approach endeavors to transcend the constraints of the filter
bubble, enrich recommendation diversity, and strike a belief balance among
users while also catering to user preferences and system-specific business
requirements. To validate the effectiveness and reliability of the BHEISR
model, we conducted a series of comprehensive experiments with real-world
datasets. These experiments compared the performance of the BHEISR model
against several baseline models using nearly 200 filter bubble-impacted users
as test subjects. Our experimental results conclusively illustrate the superior
performance of the BHEISR model in mitigating filter bubbles and balancing user
perspectives.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 06:12:37 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Wang",
"Mengyan",
""
],
[
"Hu",
"Yuxuan",
""
],
[
"Yuan",
"Zihan",
""
],
[
"Jiang",
"Chenting",
""
],
[
"Li",
"Weihua",
""
],
[
"Wu",
"Shiqing",
""
],
[
"Bai",
"Quan",
""
]
] | TITLE: BHEISR: Nudging from Bias to Balance -- Promoting Belief Harmony by
Eliminating Ideological Segregation in Knowledge-based Recommendations
ABSTRACT: In the realm of personalized recommendation systems, the increasing concern
is the amplification of belief imbalance and user biases, a phenomenon
primarily attributed to the filter bubble. Addressing this critical issue, we
introduce an innovative intermediate agency (BHEISR) between users and existing
recommendation systems to attenuate the negative repercussions of the filter
bubble effect in extant recommendation systems. The main objective is to strike
a belief balance for users while minimizing the detrimental influence caused by
filter bubbles. The BHEISR model amalgamates principles from nudge theory while
upholding democratic and transparent principles. It harnesses user-specific
category information to stimulate curiosity, even in areas users might
initially deem uninteresting. By progressively stimulating interest in novel
categories, the model encourages users to broaden their belief horizons and
explore the information they typically overlook. Our model is time-sensitive
and operates on a user feedback loop. It utilizes the existing recommendation
algorithm of the model and incorporates user feedback from the prior time
frame. This approach endeavors to transcend the constraints of the filter
bubble, enrich recommendation diversity, and strike a belief balance among
users while also catering to user preferences and system-specific business
requirements. To validate the effectiveness and reliability of the BHEISR
model, we conducted a series of comprehensive experiments with real-world
datasets. These experiments compared the performance of the BHEISR model
against several baseline models using nearly 200 filter bubble-impacted users
as test subjects. Our experimental results conclusively illustrate the superior
performance of the BHEISR model in mitigating filter bubbles and balancing user
perspectives.
| no_new_dataset | 0.708988 |
2307.02808 | Zicheng Zhang | Zicheng Zhang, Wei Sun, Yingjie Zhou, Haoning Wu, Chunyi Li, Xiongkuo
Min, Xiaohong Liu, Guangtao Zhai, Weisi Lin | Advancing Zero-Shot Digital Human Quality Assessment through
Text-Prompted Evaluation | null | null | null | null | eess.IV cs.CV cs.DB | http://creativecommons.org/licenses/by/4.0/ | Digital humans have witnessed extensive applications in various domains,
necessitating related quality assessment studies. However, there is a lack of
comprehensive digital human quality assessment (DHQA) databases. To address
this gap, we propose SJTU-H3D, a subjective quality assessment database
specifically designed for full-body digital humans. It comprises 40
high-quality reference digital humans and 1,120 labeled distorted counterparts
generated with seven types of distortions. The SJTU-H3D database can serve as a
benchmark for DHQA research, allowing evaluation and refinement of processing
algorithms. Further, we propose a zero-shot DHQA approach that focuses on
no-reference (NR) scenarios to ensure generalization capabilities while
mitigating database bias. Our method leverages semantic and distortion features
extracted from projections, as well as geometry features derived from the mesh
structure of digital humans. Specifically, we employ the Contrastive
Language-Image Pre-training (CLIP) model to measure semantic affinity and
incorporate the Naturalness Image Quality Evaluator (NIQE) model to capture
low-level distortion information. Additionally, we utilize dihedral angles as
geometry descriptors to extract mesh features. By aggregating these measures,
we introduce the Digital Human Quality Index (DHQI), which demonstrates
significant improvements in zero-shot performance. The DHQI can also serve as a
robust baseline for DHQA tasks, facilitating advancements in the field. The
database and the code are available at https://github.com/zzc-1998/SJTU-H3D.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 06:55:30 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Zhang",
"Zicheng",
""
],
[
"Sun",
"Wei",
""
],
[
"Zhou",
"Yingjie",
""
],
[
"Wu",
"Haoning",
""
],
[
"Li",
"Chunyi",
""
],
[
"Min",
"Xiongkuo",
""
],
[
"Liu",
"Xiaohong",
""
],
[
"Zhai",
"Guangtao",
""
],
[
"Lin",
"Weisi",
""
]
] | TITLE: Advancing Zero-Shot Digital Human Quality Assessment through
Text-Prompted Evaluation
ABSTRACT: Digital humans have witnessed extensive applications in various domains,
necessitating related quality assessment studies. However, there is a lack of
comprehensive digital human quality assessment (DHQA) databases. To address
this gap, we propose SJTU-H3D, a subjective quality assessment database
specifically designed for full-body digital humans. It comprises 40
high-quality reference digital humans and 1,120 labeled distorted counterparts
generated with seven types of distortions. The SJTU-H3D database can serve as a
benchmark for DHQA research, allowing evaluation and refinement of processing
algorithms. Further, we propose a zero-shot DHQA approach that focuses on
no-reference (NR) scenarios to ensure generalization capabilities while
mitigating database bias. Our method leverages semantic and distortion features
extracted from projections, as well as geometry features derived from the mesh
structure of digital humans. Specifically, we employ the Contrastive
Language-Image Pre-training (CLIP) model to measure semantic affinity and
incorporate the Naturalness Image Quality Evaluator (NIQE) model to capture
low-level distortion information. Additionally, we utilize dihedral angles as
geometry descriptors to extract mesh features. By aggregating these measures,
we introduce the Digital Human Quality Index (DHQI), which demonstrates
significant improvements in zero-shot performance. The DHQI can also serve as a
robust baseline for DHQA tasks, facilitating advancements in the field. The
database and the code are available at https://github.com/zzc-1998/SJTU-H3D.
| no_new_dataset | 0.538012 |
2307.02819 | Xufeng Caesar Dai | Nathan Koome Murungi, Michael Vinh Pham, Xufeng Dai, Xiaodong Qu | Trends in Machine Learning and Electroencephalogram (EEG): A Review for
Undergraduate Researchers | 14 pages, 1 figure, HCI International 2023 Conference | null | null | null | cs.HC cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper presents a systematic literature review on Brain-Computer
Interfaces (BCIs) in the context of Machine Learning. Our focus is on
Electroencephalography (EEG) research, highlighting the latest trends as of
2023. The objective is to provide undergraduate researchers with an accessible
overview of the BCI field, covering tasks, algorithms, and datasets. By
synthesizing recent findings, our aim is to offer a fundamental understanding
of BCI research, identifying promising avenues for future investigations.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 07:24:38 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Murungi",
"Nathan Koome",
""
],
[
"Pham",
"Michael Vinh",
""
],
[
"Dai",
"Xufeng",
""
],
[
"Qu",
"Xiaodong",
""
]
] | TITLE: Trends in Machine Learning and Electroencephalogram (EEG): A Review for
Undergraduate Researchers
ABSTRACT: This paper presents a systematic literature review on Brain-Computer
Interfaces (BCIs) in the context of Machine Learning. Our focus is on
Electroencephalography (EEG) research, highlighting the latest trends as of
2023. The objective is to provide undergraduate researchers with an accessible
overview of the BCI field, covering tasks, algorithms, and datasets. By
synthesizing recent findings, our aim is to offer a fundamental understanding
of BCI research, identifying promising avenues for future investigations.
| no_new_dataset | 0.712445 |
2307.02828 | Kun He Prof. | Xu Han, Anmin Liu, Chenxuan Yao, Yanbo Fan, Kun He | Sampling-based Fast Gradient Rescaling Method for Highly Transferable
Adversarial Attacks | 10 pages, 6 figures, 7 tables. arXiv admin note: substantial text
overlap with arXiv:2204.02887 | null | null | null | cs.CV cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks are known to be vulnerable to adversarial examples
crafted by adding human-imperceptible perturbations to the benign input. After
achieving nearly 100% attack success rates in white-box setting, more focus is
shifted to black-box attacks, of which the transferability of adversarial
examples has gained significant attention. In either case, the common
gradient-based methods generally use the sign function to generate
perturbations on the gradient update, that offers a roughly correct direction
and has gained great success. But little work pays attention to its possible
limitation. In this work, we observe that the deviation between the original
gradient and the generated noise may lead to inaccurate gradient update
estimation and suboptimal solutions for adversarial transferability. To this
end, we propose a Sampling-based Fast Gradient Rescaling Method (S-FGRM).
Specifically, we use data rescaling to substitute the sign function without
extra computational cost. We further propose a Depth First Sampling method to
eliminate the fluctuation of rescaling and stabilize the gradient update. Our
method could be used in any gradient-based attacks and is extensible to be
integrated with various input transformation or ensemble methods to further
improve the adversarial transferability. Extensive experiments on the standard
ImageNet dataset show that our method could significantly boost the
transferability of gradient-based attacks and outperform the state-of-the-art
baselines.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 07:52:42 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Han",
"Xu",
""
],
[
"Liu",
"Anmin",
""
],
[
"Yao",
"Chenxuan",
""
],
[
"Fan",
"Yanbo",
""
],
[
"He",
"Kun",
""
]
] | TITLE: Sampling-based Fast Gradient Rescaling Method for Highly Transferable
Adversarial Attacks
ABSTRACT: Deep neural networks are known to be vulnerable to adversarial examples
crafted by adding human-imperceptible perturbations to the benign input. After
achieving nearly 100% attack success rates in white-box setting, more focus is
shifted to black-box attacks, of which the transferability of adversarial
examples has gained significant attention. In either case, the common
gradient-based methods generally use the sign function to generate
perturbations on the gradient update, that offers a roughly correct direction
and has gained great success. But little work pays attention to its possible
limitation. In this work, we observe that the deviation between the original
gradient and the generated noise may lead to inaccurate gradient update
estimation and suboptimal solutions for adversarial transferability. To this
end, we propose a Sampling-based Fast Gradient Rescaling Method (S-FGRM).
Specifically, we use data rescaling to substitute the sign function without
extra computational cost. We further propose a Depth First Sampling method to
eliminate the fluctuation of rescaling and stabilize the gradient update. Our
method could be used in any gradient-based attacks and is extensible to be
integrated with various input transformation or ensemble methods to further
improve the adversarial transferability. Extensive experiments on the standard
ImageNet dataset show that our method could significantly boost the
transferability of gradient-based attacks and outperform the state-of-the-art
baselines.
| no_new_dataset | 0.708824 |
2307.02836 | Shiqi Deng | Shiqi Deng and Zhiyu Sun and Ruiyan Zhuang and Jun Gong | Noise-to-Norm Reconstruction for Industrial Anomaly Detection and
Localization | null | null | null | null | cs.CV eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Anomaly detection has a wide range of applications and is especially
important in industrial quality inspection. Currently, many top-performing
anomaly-detection models rely on feature-embedding methods. However, these
methods do not perform well on datasets with large variations in object
locations. Reconstruction-based methods use reconstruction errors to detect
anomalies without considering positional differences between samples. In this
study, a reconstruction-based method using the noise-to-norm paradigm is
proposed, which avoids the invariant reconstruction of anomalous regions. Our
reconstruction network is based on M-net and incorporates multiscale fusion and
residual attention modules to enable end-to-end anomaly detection and
localization. Experiments demonstrate that the method is effective in
reconstructing anomalous regions into normal patterns and achieving accurate
anomaly detection and localization. On the MPDD and VisA datasets, our proposed
method achieved more competitive results than the latest methods, and it set a
new state-of-the-art standard on the MPDD dataset.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 08:06:48 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Deng",
"Shiqi",
""
],
[
"Sun",
"Zhiyu",
""
],
[
"Zhuang",
"Ruiyan",
""
],
[
"Gong",
"Jun",
""
]
] | TITLE: Noise-to-Norm Reconstruction for Industrial Anomaly Detection and
Localization
ABSTRACT: Anomaly detection has a wide range of applications and is especially
important in industrial quality inspection. Currently, many top-performing
anomaly-detection models rely on feature-embedding methods. However, these
methods do not perform well on datasets with large variations in object
locations. Reconstruction-based methods use reconstruction errors to detect
anomalies without considering positional differences between samples. In this
study, a reconstruction-based method using the noise-to-norm paradigm is
proposed, which avoids the invariant reconstruction of anomalous regions. Our
reconstruction network is based on M-net and incorporates multiscale fusion and
residual attention modules to enable end-to-end anomaly detection and
localization. Experiments demonstrate that the method is effective in
reconstructing anomalous regions into normal patterns and achieving accurate
anomaly detection and localization. On the MPDD and VisA datasets, our proposed
method achieved more competitive results than the latest methods, and it set a
new state-of-the-art standard on the MPDD dataset.
| no_new_dataset | 0.70978 |
2307.02865 | Mattia Giovanni Campana | Valerio Arnaboldi, Mattia Giovanni Campana, Franca Delmastro, Elena
Pagani | PLIERS: a Popularity-Based Recommender System for Content Dissemination
in Online Social Networks | Published in SAC '16: Proceedings of the 31st Annual ACM Symposium on
Applied Computing | null | 10.1145/2851613.2851940 | null | cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we propose a novel tag-based recommender system called PLIERS,
which relies on the assumption that users are mainly interested in items and
tags with similar popularity to those they already own. PLIERS is aimed at
reaching a good tradeoff between algorithmic complexity and the level of
personalization of recommended items. To evaluate PLIERS, we performed a set of
experiments on real OSN datasets, demonstrating that it outperforms
state-of-the-art solutions in terms of personalization, relevance, and novelty
of recommendations.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 09:04:58 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Arnaboldi",
"Valerio",
""
],
[
"Campana",
"Mattia Giovanni",
""
],
[
"Delmastro",
"Franca",
""
],
[
"Pagani",
"Elena",
""
]
] | TITLE: PLIERS: a Popularity-Based Recommender System for Content Dissemination
in Online Social Networks
ABSTRACT: In this paper, we propose a novel tag-based recommender system called PLIERS,
which relies on the assumption that users are mainly interested in items and
tags with similar popularity to those they already own. PLIERS is aimed at
reaching a good tradeoff between algorithmic complexity and the level of
personalization of recommended items. To evaluate PLIERS, we performed a set of
experiments on real OSN datasets, demonstrating that it outperforms
state-of-the-art solutions in terms of personalization, relevance, and novelty
of recommendations.
| no_new_dataset | 0.711387 |
2307.02871 | Hanzhang Xue | Hanzhang Xue, Xiaochang Hu, Rui Xie, Hao Fu, Liang Xiao, Yiming Nie,
Bin Dai | Contrastive Label Disambiguation for Self-Supervised Terrain
Traversability Learning in Off-Road Environments | 9 pages, 11 figures | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | Discriminating the traversability of terrains is a crucial task for
autonomous driving in off-road environments. However, it is challenging due to
the diverse, ambiguous, and platform-specific nature of off-road
traversability. In this paper, we propose a novel self-supervised terrain
traversability learning framework, utilizing a contrastive label disambiguation
mechanism. Firstly, weakly labeled training samples with pseudo labels are
automatically generated by projecting actual driving experiences onto the
terrain models constructed in real time. Subsequently, a prototype-based
contrastive representation learning method is designed to learn distinguishable
embeddings, facilitating the self-supervised updating of those pseudo labels.
As the iterative interaction between representation learning and pseudo label
updating, the ambiguities in those pseudo labels are gradually eliminated,
enabling the learning of platform-specific and task-specific traversability
without any human-provided annotations. Experimental results on the RELLIS-3D
dataset and our Gobi Desert driving dataset demonstrate the effectiveness of
the proposed method.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 09:15:38 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Xue",
"Hanzhang",
""
],
[
"Hu",
"Xiaochang",
""
],
[
"Xie",
"Rui",
""
],
[
"Fu",
"Hao",
""
],
[
"Xiao",
"Liang",
""
],
[
"Nie",
"Yiming",
""
],
[
"Dai",
"Bin",
""
]
] | TITLE: Contrastive Label Disambiguation for Self-Supervised Terrain
Traversability Learning in Off-Road Environments
ABSTRACT: Discriminating the traversability of terrains is a crucial task for
autonomous driving in off-road environments. However, it is challenging due to
the diverse, ambiguous, and platform-specific nature of off-road
traversability. In this paper, we propose a novel self-supervised terrain
traversability learning framework, utilizing a contrastive label disambiguation
mechanism. Firstly, weakly labeled training samples with pseudo labels are
automatically generated by projecting actual driving experiences onto the
terrain models constructed in real time. Subsequently, a prototype-based
contrastive representation learning method is designed to learn distinguishable
embeddings, facilitating the self-supervised updating of those pseudo labels.
As the iterative interaction between representation learning and pseudo label
updating, the ambiguities in those pseudo labels are gradually eliminated,
enabling the learning of platform-specific and task-specific traversability
without any human-provided annotations. Experimental results on the RELLIS-3D
dataset and our Gobi Desert driving dataset demonstrate the effectiveness of
the proposed method.
| no_new_dataset | 0.700267 |
2307.02877 | Binbin Xiang | Binbin Xiang, Torben Peters, Theodora Kontogianni, Frawa Vetterli,
Stefano Puliti, Rasmus Astrup, Konrad Schindler | Towards accurate instance segmentation in large-scale LiDAR point clouds | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Panoptic segmentation is the combination of semantic and instance
segmentation: assign the points in a 3D point cloud to semantic categories and
partition them into distinct object instances. It has many obvious applications
for outdoor scene understanding, from city mapping to forest management.
Existing methods struggle to segment nearby instances of the same semantic
category, like adjacent pieces of street furniture or neighbouring trees, which
limits their usability for inventory- or management-type applications that rely
on object instances. This study explores the steps of the panoptic segmentation
pipeline concerned with clustering points into object instances, with the goal
to alleviate that bottleneck. We find that a carefully designed clustering
strategy, which leverages multiple types of learned point embeddings,
significantly improves instance segmentation. Experiments on the NPM3D urban
mobile mapping dataset and the FOR-instance forest dataset demonstrate the
effectiveness and versatility of the proposed strategy.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 09:29:03 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Xiang",
"Binbin",
""
],
[
"Peters",
"Torben",
""
],
[
"Kontogianni",
"Theodora",
""
],
[
"Vetterli",
"Frawa",
""
],
[
"Puliti",
"Stefano",
""
],
[
"Astrup",
"Rasmus",
""
],
[
"Schindler",
"Konrad",
""
]
] | TITLE: Towards accurate instance segmentation in large-scale LiDAR point clouds
ABSTRACT: Panoptic segmentation is the combination of semantic and instance
segmentation: assign the points in a 3D point cloud to semantic categories and
partition them into distinct object instances. It has many obvious applications
for outdoor scene understanding, from city mapping to forest management.
Existing methods struggle to segment nearby instances of the same semantic
category, like adjacent pieces of street furniture or neighbouring trees, which
limits their usability for inventory- or management-type applications that rely
on object instances. This study explores the steps of the panoptic segmentation
pipeline concerned with clustering points into object instances, with the goal
to alleviate that bottleneck. We find that a carefully designed clustering
strategy, which leverages multiple types of learned point embeddings,
significantly improves instance segmentation. Experiments on the NPM3D urban
mobile mapping dataset and the FOR-instance forest dataset demonstrate the
effectiveness and versatility of the proposed strategy.
| no_new_dataset | 0.709655 |
2307.02894 | Georg Rutishauser | Georg Rutishauser, Francesco Conti, Luca Benini | Free Bits: Latency Optimization of Mixed-Precision Quantized Neural
Networks on the Edge | Accepted for publication at the 2023 IEEE International Conference of
Artificial Intelligence Circuits and Systems (AICAS '23) | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mixed-precision quantization, where a deep neural network's layers are
quantized to different precisions, offers the opportunity to optimize the
trade-offs between model size, latency, and statistical accuracy beyond what
can be achieved with homogeneous-bit-width quantization. To navigate the
intractable search space of mixed-precision configurations for a given network,
this paper proposes a hybrid search methodology. It consists of a
hardware-agnostic differentiable search algorithm followed by a hardware-aware
heuristic optimization to find mixed-precision configurations latency-optimized
for a specific hardware target. We evaluate our algorithm on MobileNetV1 and
MobileNetV2 and deploy the resulting networks on a family of multi-core RISC-V
microcontroller platforms with different hardware characteristics. We achieve
up to 28.6% reduction of end-to-end latency compared to an 8-bit model at a
negligible accuracy drop from a full-precision baseline on the 1000-class
ImageNet dataset. We demonstrate speedups relative to an 8-bit baseline, even
on systems with no hardware support for sub-byte arithmetic at negligible
accuracy drop. Furthermore, we show the superiority of our approach with
respect to differentiable search targeting reduced binary operation counts as a
proxy for latency.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 09:57:48 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Rutishauser",
"Georg",
""
],
[
"Conti",
"Francesco",
""
],
[
"Benini",
"Luca",
""
]
] | TITLE: Free Bits: Latency Optimization of Mixed-Precision Quantized Neural
Networks on the Edge
ABSTRACT: Mixed-precision quantization, where a deep neural network's layers are
quantized to different precisions, offers the opportunity to optimize the
trade-offs between model size, latency, and statistical accuracy beyond what
can be achieved with homogeneous-bit-width quantization. To navigate the
intractable search space of mixed-precision configurations for a given network,
this paper proposes a hybrid search methodology. It consists of a
hardware-agnostic differentiable search algorithm followed by a hardware-aware
heuristic optimization to find mixed-precision configurations latency-optimized
for a specific hardware target. We evaluate our algorithm on MobileNetV1 and
MobileNetV2 and deploy the resulting networks on a family of multi-core RISC-V
microcontroller platforms with different hardware characteristics. We achieve
up to 28.6% reduction of end-to-end latency compared to an 8-bit model at a
negligible accuracy drop from a full-precision baseline on the 1000-class
ImageNet dataset. We demonstrate speedups relative to an 8-bit baseline, even
on systems with no hardware support for sub-byte arithmetic at negligible
accuracy drop. Furthermore, we show the superiority of our approach with
respect to differentiable search targeting reduced binary operation counts as a
proxy for latency.
| no_new_dataset | 0.71039 |
2307.02909 | Guinan Li | Guinan Li, Jiajun Deng, Mengzhe Geng, Zengrui Jin, Tianzi Wang, Shujie
Hu, Mingyu Cui, Helen Meng, Xunying Liu | Audio-visual End-to-end Multi-channel Speech Separation, Dereverberation
and Recognition | IEEE/ACM Transactions on Audio, Speech, and Language Processing | null | null | null | eess.AS cs.AI cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate recognition of cocktail party speech containing overlapping
speakers, noise and reverberation remains a highly challenging task to date.
Motivated by the invariance of visual modality to acoustic signal corruption,
an audio-visual multi-channel speech separation, dereverberation and
recognition approach featuring a full incorporation of visual information into
all system components is proposed in this paper. The efficacy of the video
input is consistently demonstrated in mask-based MVDR speech separation,
DNN-WPE or spectral mapping (SpecM) based speech dereverberation front-end and
Conformer ASR back-end. Audio-visual integrated front-end architectures
performing speech separation and dereverberation in a pipelined or joint
fashion via mask-based WPD are investigated. The error cost mismatch between
the speech enhancement front-end and ASR back-end components is minimized by
end-to-end jointly fine-tuning using either the ASR cost function alone, or its
interpolation with the speech enhancement loss. Experiments were conducted on
the mixture overlapped and reverberant speech data constructed using simulation
or replay of the Oxford LRS2 dataset. The proposed audio-visual multi-channel
speech separation, dereverberation and recognition systems consistently
outperformed the comparable audio-only baseline by 9.1% and 6.2% absolute
(41.7% and 36.0% relative) word error rate (WER) reductions. Consistent speech
enhancement improvements were also obtained on PESQ, STOI and SRMR scores.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 10:50:46 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Li",
"Guinan",
""
],
[
"Deng",
"Jiajun",
""
],
[
"Geng",
"Mengzhe",
""
],
[
"Jin",
"Zengrui",
""
],
[
"Wang",
"Tianzi",
""
],
[
"Hu",
"Shujie",
""
],
[
"Cui",
"Mingyu",
""
],
[
"Meng",
"Helen",
""
],
[
"Liu",
"Xunying",
""
]
] | TITLE: Audio-visual End-to-end Multi-channel Speech Separation, Dereverberation
and Recognition
ABSTRACT: Accurate recognition of cocktail party speech containing overlapping
speakers, noise and reverberation remains a highly challenging task to date.
Motivated by the invariance of visual modality to acoustic signal corruption,
an audio-visual multi-channel speech separation, dereverberation and
recognition approach featuring a full incorporation of visual information into
all system components is proposed in this paper. The efficacy of the video
input is consistently demonstrated in mask-based MVDR speech separation,
DNN-WPE or spectral mapping (SpecM) based speech dereverberation front-end and
Conformer ASR back-end. Audio-visual integrated front-end architectures
performing speech separation and dereverberation in a pipelined or joint
fashion via mask-based WPD are investigated. The error cost mismatch between
the speech enhancement front-end and ASR back-end components is minimized by
end-to-end jointly fine-tuning using either the ASR cost function alone, or its
interpolation with the speech enhancement loss. Experiments were conducted on
the mixture overlapped and reverberant speech data constructed using simulation
or replay of the Oxford LRS2 dataset. The proposed audio-visual multi-channel
speech separation, dereverberation and recognition systems consistently
outperformed the comparable audio-only baseline by 9.1% and 6.2% absolute
(41.7% and 36.0% relative) word error rate (WER) reductions. Consistent speech
enhancement improvements were also obtained on PESQ, STOI and SRMR scores.
| no_new_dataset | 0.710025 |
2307.02912 | David Jimenez-Cabello | Mario Almagro, Emilio Almaz\'an, Diego Ortego, David Jim\'enez | LEA: Improving Sentence Similarity Robustness to Typos Using Lexical
Attention Bias | KDD'23 conference (main research track). (*) These authors
contributed equally | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Textual noise, such as typos or abbreviations, is a well-known issue that
penalizes vanilla Transformers for most downstream tasks. We show that this is
also the case for sentence similarity, a fundamental task in multiple domains,
e.g. matching, retrieval or paraphrasing. Sentence similarity can be approached
using cross-encoders, where the two sentences are concatenated in the input
allowing the model to exploit the inter-relations between them. Previous works
addressing the noise issue mainly rely on data augmentation strategies, showing
improved robustness when dealing with corrupted samples that are similar to the
ones used for training. However, all these methods still suffer from the token
distribution shift induced by typos. In this work, we propose to tackle textual
noise by equipping cross-encoders with a novel LExical-aware Attention module
(LEA) that incorporates lexical similarities between words in both sentences.
By using raw text similarities, our approach avoids the tokenization shift
problem obtaining improved robustness. We demonstrate that the attention bias
introduced by LEA helps cross-encoders to tackle complex scenarios with textual
noise, specially in domains with short-text descriptions and limited context.
Experiments using three popular Transformer encoders in five e-commerce
datasets for product matching show that LEA consistently boosts performance
under the presence of noise, while remaining competitive on the original
(clean) splits. We also evaluate our approach in two datasets for textual
entailment and paraphrasing showing that LEA is robust to typos in domains with
longer sentences and more natural context. Additionally, we thoroughly analyze
several design choices in our approach, providing insights about the impact of
the decisions made and fostering future research in cross-encoders dealing with
typos.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 10:53:50 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Almagro",
"Mario",
""
],
[
"Almazán",
"Emilio",
""
],
[
"Ortego",
"Diego",
""
],
[
"Jiménez",
"David",
""
]
] | TITLE: LEA: Improving Sentence Similarity Robustness to Typos Using Lexical
Attention Bias
ABSTRACT: Textual noise, such as typos or abbreviations, is a well-known issue that
penalizes vanilla Transformers for most downstream tasks. We show that this is
also the case for sentence similarity, a fundamental task in multiple domains,
e.g. matching, retrieval or paraphrasing. Sentence similarity can be approached
using cross-encoders, where the two sentences are concatenated in the input
allowing the model to exploit the inter-relations between them. Previous works
addressing the noise issue mainly rely on data augmentation strategies, showing
improved robustness when dealing with corrupted samples that are similar to the
ones used for training. However, all these methods still suffer from the token
distribution shift induced by typos. In this work, we propose to tackle textual
noise by equipping cross-encoders with a novel LExical-aware Attention module
(LEA) that incorporates lexical similarities between words in both sentences.
By using raw text similarities, our approach avoids the tokenization shift
problem obtaining improved robustness. We demonstrate that the attention bias
introduced by LEA helps cross-encoders to tackle complex scenarios with textual
noise, specially in domains with short-text descriptions and limited context.
Experiments using three popular Transformer encoders in five e-commerce
datasets for product matching show that LEA consistently boosts performance
under the presence of noise, while remaining competitive on the original
(clean) splits. We also evaluate our approach in two datasets for textual
entailment and paraphrasing showing that LEA is robust to typos in domains with
longer sentences and more natural context. Additionally, we thoroughly analyze
several design choices in our approach, providing insights about the impact of
the decisions made and fostering future research in cross-encoders dealing with
typos.
| no_new_dataset | 0.710848 |
2307.02935 | Xin Wang | Xin Wang, Tao Tan, Yuan Gao, Luyi Han, Tianyu Zhang, Chunyao Lu,
Regina Beets-Tan, Ruisheng Su, Ritse Mann | DisAsymNet: Disentanglement of Asymmetrical Abnormality on Bilateral
Mammograms using Self-adversarial Learning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Asymmetry is a crucial characteristic of bilateral mammograms (Bi-MG) when
abnormalities are developing. It is widely utilized by radiologists for
diagnosis. The question of 'what the symmetrical Bi-MG would look like when the
asymmetrical abnormalities have been removed ?' has not yet received strong
attention in the development of algorithms on mammograms. Addressing this
question could provide valuable insights into mammographic anatomy and aid in
diagnostic interpretation. Hence, we propose a novel framework, DisAsymNet,
which utilizes asymmetrical abnormality transformer guided self-adversarial
learning for disentangling abnormalities and symmetric Bi-MG. At the same time,
our proposed method is partially guided by randomly synthesized abnormalities.
We conduct experiments on three public and one in-house dataset, and
demonstrate that our method outperforms existing methods in abnormality
classification, segmentation, and localization tasks. Additionally,
reconstructed normal mammograms can provide insights toward better
interpretable visual cues for clinical diagnosis. The code will be accessible
to the public.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 11:52:36 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Wang",
"Xin",
""
],
[
"Tan",
"Tao",
""
],
[
"Gao",
"Yuan",
""
],
[
"Han",
"Luyi",
""
],
[
"Zhang",
"Tianyu",
""
],
[
"Lu",
"Chunyao",
""
],
[
"Beets-Tan",
"Regina",
""
],
[
"Su",
"Ruisheng",
""
],
[
"Mann",
"Ritse",
""
]
] | TITLE: DisAsymNet: Disentanglement of Asymmetrical Abnormality on Bilateral
Mammograms using Self-adversarial Learning
ABSTRACT: Asymmetry is a crucial characteristic of bilateral mammograms (Bi-MG) when
abnormalities are developing. It is widely utilized by radiologists for
diagnosis. The question of 'what the symmetrical Bi-MG would look like when the
asymmetrical abnormalities have been removed ?' has not yet received strong
attention in the development of algorithms on mammograms. Addressing this
question could provide valuable insights into mammographic anatomy and aid in
diagnostic interpretation. Hence, we propose a novel framework, DisAsymNet,
which utilizes asymmetrical abnormality transformer guided self-adversarial
learning for disentangling abnormalities and symmetric Bi-MG. At the same time,
our proposed method is partially guided by randomly synthesized abnormalities.
We conduct experiments on three public and one in-house dataset, and
demonstrate that our method outperforms existing methods in abnormality
classification, segmentation, and localization tasks. Additionally,
reconstructed normal mammograms can provide insights toward better
interpretable visual cues for clinical diagnosis. The code will be accessible
to the public.
| no_new_dataset | 0.705138 |
2307.02946 | Guangyi Zhang | Guangyi Zhang, Nikolaj Tatti, Aristides Gionis | Finding Favourite Tuples on Data Streams with Provably Few Comparisons | To appear in KDD 2023 | null | null | null | cs.DB cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the most fundamental tasks in data science is to assist a user with
unknown preferences in finding high-utility tuples within a large database. To
accurately elicit the unknown user preferences, a widely-adopted way is by
asking the user to compare pairs of tuples. In this paper, we study the problem
of identifying one or more high-utility tuples by adaptively receiving user
input on a minimum number of pairwise comparisons. We devise a single-pass
streaming algorithm, which processes each tuple in the stream at most once,
while ensuring that the memory size and the number of requested comparisons are
in the worst case logarithmic in $n$, where $n$ is the number of all tuples. An
important variant of the problem, which can help to reduce human error in
comparisons, is to allow users to declare ties when confronted with pairs of
tuples of nearly equal utility. We show that the theoretical guarantees of our
method can be maintained for this important problem variant. In addition, we
show how to enhance existing pruning techniques in the literature by leveraging
powerful tools from mathematical programming. Finally, we systematically
evaluate all proposed algorithms over both synthetic and real-life datasets,
examine their scalability, and demonstrate their superior performance over
existing methods.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 12:31:32 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Zhang",
"Guangyi",
""
],
[
"Tatti",
"Nikolaj",
""
],
[
"Gionis",
"Aristides",
""
]
] | TITLE: Finding Favourite Tuples on Data Streams with Provably Few Comparisons
ABSTRACT: One of the most fundamental tasks in data science is to assist a user with
unknown preferences in finding high-utility tuples within a large database. To
accurately elicit the unknown user preferences, a widely-adopted way is by
asking the user to compare pairs of tuples. In this paper, we study the problem
of identifying one or more high-utility tuples by adaptively receiving user
input on a minimum number of pairwise comparisons. We devise a single-pass
streaming algorithm, which processes each tuple in the stream at most once,
while ensuring that the memory size and the number of requested comparisons are
in the worst case logarithmic in $n$, where $n$ is the number of all tuples. An
important variant of the problem, which can help to reduce human error in
comparisons, is to allow users to declare ties when confronted with pairs of
tuples of nearly equal utility. We show that the theoretical guarantees of our
method can be maintained for this important problem variant. In addition, we
show how to enhance existing pruning techniques in the literature by leveraging
powerful tools from mathematical programming. Finally, we systematically
evaluate all proposed algorithms over both synthetic and real-life datasets,
examine their scalability, and demonstrate their superior performance over
existing methods.
| no_new_dataset | 0.707815 |
2307.02971 | Longyue Wang | Bingshuai Liu, Longyue Wang, Chenyang Lyu, Yong Zhang, Jinsong Su,
Shuming Shi, Zhaopeng Tu | On the Cultural Gap in Text-to-Image Generation | Equal contribution: Bingshuai Liu and Longyue Wang. Work done while
Bingshuai Liu and Chengyang Lyu were interning at Tencent AI Lab. Zhaopeng Tu
is the corresponding author | null | null | null | cs.CV cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | One challenge in text-to-image (T2I) generation is the inadvertent reflection
of culture gaps present in the training data, which signifies the disparity in
generated image quality when the cultural elements of the input text are rarely
collected in the training set. Although various T2I models have shown
impressive but arbitrary examples, there is no benchmark to systematically
evaluate a T2I model's ability to generate cross-cultural images. To bridge the
gap, we propose a Challenging Cross-Cultural (C3) benchmark with comprehensive
evaluation criteria, which can assess how well-suited a model is to a target
culture. By analyzing the flawed images generated by the Stable Diffusion model
on the C3 benchmark, we find that the model often fails to generate certain
cultural objects. Accordingly, we propose a novel multi-modal metric that
considers object-text alignment to filter the fine-tuning data in the target
culture, which is used to fine-tune a T2I model to improve cross-cultural
generation. Experimental results show that our multi-modal metric provides
stronger data selection performance on the C3 benchmark than existing metrics,
in which the object-text alignment is crucial. We release the benchmark, data,
code, and generated images to facilitate future research on culturally diverse
T2I generation (https://github.com/longyuewangdcu/C3-Bench).
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 13:17:55 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Liu",
"Bingshuai",
""
],
[
"Wang",
"Longyue",
""
],
[
"Lyu",
"Chenyang",
""
],
[
"Zhang",
"Yong",
""
],
[
"Su",
"Jinsong",
""
],
[
"Shi",
"Shuming",
""
],
[
"Tu",
"Zhaopeng",
""
]
] | TITLE: On the Cultural Gap in Text-to-Image Generation
ABSTRACT: One challenge in text-to-image (T2I) generation is the inadvertent reflection
of culture gaps present in the training data, which signifies the disparity in
generated image quality when the cultural elements of the input text are rarely
collected in the training set. Although various T2I models have shown
impressive but arbitrary examples, there is no benchmark to systematically
evaluate a T2I model's ability to generate cross-cultural images. To bridge the
gap, we propose a Challenging Cross-Cultural (C3) benchmark with comprehensive
evaluation criteria, which can assess how well-suited a model is to a target
culture. By analyzing the flawed images generated by the Stable Diffusion model
on the C3 benchmark, we find that the model often fails to generate certain
cultural objects. Accordingly, we propose a novel multi-modal metric that
considers object-text alignment to filter the fine-tuning data in the target
culture, which is used to fine-tune a T2I model to improve cross-cultural
generation. Experimental results show that our multi-modal metric provides
stronger data selection performance on the C3 benchmark than existing metrics,
in which the object-text alignment is crucial. We release the benchmark, data,
code, and generated images to facilitate future research on culturally diverse
T2I generation (https://github.com/longyuewangdcu/C3-Bench).
| no_new_dataset | 0.706634 |
2307.02974 | Binglu Wang | Yuting Lu, Lingtong Min, Binglu Wang, Le Zheng, Xiaoxu Wang, Yongqiang
Zhao, Teng Long | Cross-Spatial Pixel Integration and Cross-Stage Feature Fusion Based
Transformer Network for Remote Sensing Image Super-Resolution | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Remote sensing image super-resolution (RSISR) plays a vital role in enhancing
spatial detials and improving the quality of satellite imagery. Recently,
Transformer-based models have shown competitive performance in RSISR. To
mitigate the quadratic computational complexity resulting from global
self-attention, various methods constrain attention to a local window,
enhancing its efficiency. Consequently, the receptive fields in a single
attention layer are inadequate, leading to insufficient context modeling.
Furthermore, while most transform-based approaches reuse shallow features
through skip connections, relying solely on these connections treats shallow
and deep features equally, impeding the model's ability to characterize them.
To address these issues, we propose a novel transformer architecture called
Cross-Spatial Pixel Integration and Cross-Stage Feature Fusion Based
Transformer Network (SPIFFNet) for RSISR. Our proposed model effectively
enhances global cognition and understanding of the entire image, facilitating
efficient integration of features cross-stages. The model incorporates
cross-spatial pixel integration attention (CSPIA) to introduce contextual
information into a local window, while cross-stage feature fusion attention
(CSFFA) adaptively fuses features from the previous stage to improve feature
expression in line with the requirements of the current stage. We conducted
comprehensive experiments on multiple benchmark datasets, demonstrating the
superior performance of our proposed SPIFFNet in terms of both quantitative
metrics and visual quality when compared to state-of-the-art methods.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 13:19:06 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Lu",
"Yuting",
""
],
[
"Min",
"Lingtong",
""
],
[
"Wang",
"Binglu",
""
],
[
"Zheng",
"Le",
""
],
[
"Wang",
"Xiaoxu",
""
],
[
"Zhao",
"Yongqiang",
""
],
[
"Long",
"Teng",
""
]
] | TITLE: Cross-Spatial Pixel Integration and Cross-Stage Feature Fusion Based
Transformer Network for Remote Sensing Image Super-Resolution
ABSTRACT: Remote sensing image super-resolution (RSISR) plays a vital role in enhancing
spatial detials and improving the quality of satellite imagery. Recently,
Transformer-based models have shown competitive performance in RSISR. To
mitigate the quadratic computational complexity resulting from global
self-attention, various methods constrain attention to a local window,
enhancing its efficiency. Consequently, the receptive fields in a single
attention layer are inadequate, leading to insufficient context modeling.
Furthermore, while most transform-based approaches reuse shallow features
through skip connections, relying solely on these connections treats shallow
and deep features equally, impeding the model's ability to characterize them.
To address these issues, we propose a novel transformer architecture called
Cross-Spatial Pixel Integration and Cross-Stage Feature Fusion Based
Transformer Network (SPIFFNet) for RSISR. Our proposed model effectively
enhances global cognition and understanding of the entire image, facilitating
efficient integration of features cross-stages. The model incorporates
cross-spatial pixel integration attention (CSPIA) to introduce contextual
information into a local window, while cross-stage feature fusion attention
(CSFFA) adaptively fuses features from the previous stage to improve feature
expression in line with the requirements of the current stage. We conducted
comprehensive experiments on multiple benchmark datasets, demonstrating the
superior performance of our proposed SPIFFNet in terms of both quantitative
metrics and visual quality when compared to state-of-the-art methods.
| no_new_dataset | 0.711268 |
2307.02975 | Mattia Giovanni Campana | Mattia Giovanni Campana, Franca Delmastro, Elena Pagani | Transfer Learning for the Efficient Detection of COVID-19 from
Smartphone Audio Data | null | Pervasive and Mobile Computing, Volume 89, 2023 | 10.1016/j.pmcj.2023.101754 | null | cs.LG cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Disease detection from smartphone data represents an open research challenge
in mobile health (m-health) systems. COVID-19 and its respiratory symptoms are
an important case study in this area and their early detection is a potential
real instrument to counteract the pandemic situation. The efficacy of this
solution mainly depends on the performances of AI algorithms applied to the
collected data and their possible implementation directly on the users' mobile
devices. Considering these issues, and the limited amount of available data, in
this paper we present the experimental evaluation of 3 different deep learning
models, compared also with hand-crafted features, and of two main approaches of
transfer learning in the considered scenario: both feature extraction and
fine-tuning. Specifically, we considered VGGish, YAMNET, and
L\textsuperscript{3}-Net (including 12 different configurations) evaluated
through user-independent experiments on 4 different datasets (13,447 samples in
total). Results clearly show the advantages of L\textsuperscript{3}-Net in all
the experimental settings as it overcomes the other solutions by 12.3\% in
terms of Precision-Recall AUC as features extractor, and by 10\% when the model
is fine-tuned. Moreover, we note that to fine-tune only the fully-connected
layers of the pre-trained models generally leads to worse performances, with an
average drop of 6.6\% with respect to feature extraction. %highlighting the
need for further investigations. Finally, we evaluate the memory footprints of
the different models for their possible applications on commercial mobile
devices.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 13:19:27 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Campana",
"Mattia Giovanni",
""
],
[
"Delmastro",
"Franca",
""
],
[
"Pagani",
"Elena",
""
]
] | TITLE: Transfer Learning for the Efficient Detection of COVID-19 from
Smartphone Audio Data
ABSTRACT: Disease detection from smartphone data represents an open research challenge
in mobile health (m-health) systems. COVID-19 and its respiratory symptoms are
an important case study in this area and their early detection is a potential
real instrument to counteract the pandemic situation. The efficacy of this
solution mainly depends on the performances of AI algorithms applied to the
collected data and their possible implementation directly on the users' mobile
devices. Considering these issues, and the limited amount of available data, in
this paper we present the experimental evaluation of 3 different deep learning
models, compared also with hand-crafted features, and of two main approaches of
transfer learning in the considered scenario: both feature extraction and
fine-tuning. Specifically, we considered VGGish, YAMNET, and
L\textsuperscript{3}-Net (including 12 different configurations) evaluated
through user-independent experiments on 4 different datasets (13,447 samples in
total). Results clearly show the advantages of L\textsuperscript{3}-Net in all
the experimental settings as it overcomes the other solutions by 12.3\% in
terms of Precision-Recall AUC as features extractor, and by 10\% when the model
is fine-tuned. Moreover, we note that to fine-tune only the fully-connected
layers of the pre-trained models generally leads to worse performances, with an
average drop of 6.6\% with respect to feature extraction. %highlighting the
need for further investigations. Finally, we evaluate the memory footprints of
the different models for their possible applications on commercial mobile
devices.
| no_new_dataset | 0.71 |
2307.02984 | Matteo Pennisi | Matteo Pennisi, Federica Proietto Salanitri, Giovanni Bellitto, Simone
Palazzo, Ulas Bagci, Concetto Spampinato | A Privacy-Preserving Walk in the Latent Space of Generative Models for
Medical Applications | Accepted at MICCAI 2023 | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Generative Adversarial Networks (GANs) have demonstrated their ability to
generate synthetic samples that match a target distribution. However, from a
privacy perspective, using GANs as a proxy for data sharing is not a safe
solution, as they tend to embed near-duplicates of real samples in the latent
space. Recent works, inspired by k-anonymity principles, address this issue
through sample aggregation in the latent space, with the drawback of reducing
the dataset by a factor of k. Our work aims to mitigate this problem by
proposing a latent space navigation strategy able to generate diverse synthetic
samples that may support effective training of deep models, while addressing
privacy concerns in a principled way. Our approach leverages an auxiliary
identity classifier as a guide to non-linearly walk between points in the
latent space, minimizing the risk of collision with near-duplicates of real
samples. We empirically demonstrate that, given any random pair of points in
the latent space, our walking strategy is safer than linear interpolation. We
then test our path-finding strategy combined to k-same methods and demonstrate,
on two benchmarks for tuberculosis and diabetic retinopathy classification,
that training a model using samples generated by our approach mitigate drops in
performance, while keeping privacy preservation.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 13:35:48 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Pennisi",
"Matteo",
""
],
[
"Salanitri",
"Federica Proietto",
""
],
[
"Bellitto",
"Giovanni",
""
],
[
"Palazzo",
"Simone",
""
],
[
"Bagci",
"Ulas",
""
],
[
"Spampinato",
"Concetto",
""
]
] | TITLE: A Privacy-Preserving Walk in the Latent Space of Generative Models for
Medical Applications
ABSTRACT: Generative Adversarial Networks (GANs) have demonstrated their ability to
generate synthetic samples that match a target distribution. However, from a
privacy perspective, using GANs as a proxy for data sharing is not a safe
solution, as they tend to embed near-duplicates of real samples in the latent
space. Recent works, inspired by k-anonymity principles, address this issue
through sample aggregation in the latent space, with the drawback of reducing
the dataset by a factor of k. Our work aims to mitigate this problem by
proposing a latent space navigation strategy able to generate diverse synthetic
samples that may support effective training of deep models, while addressing
privacy concerns in a principled way. Our approach leverages an auxiliary
identity classifier as a guide to non-linearly walk between points in the
latent space, minimizing the risk of collision with near-duplicates of real
samples. We empirically demonstrate that, given any random pair of points in
the latent space, our walking strategy is safer than linear interpolation. We
then test our path-finding strategy combined to k-same methods and demonstrate,
on two benchmarks for tuberculosis and diabetic retinopathy classification,
that training a model using samples generated by our approach mitigate drops in
performance, while keeping privacy preservation.
| no_new_dataset | 0.711249 |
2307.02991 | Asma Atamna | Abhijeet Pendyala, Justin Dettmer, Tobias Glasmachers, Asma Atamna | ContainerGym: A Real-World Reinforcement Learning Benchmark for Resource
Allocation | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present ContainerGym, a benchmark for reinforcement learning inspired by a
real-world industrial resource allocation task. The proposed benchmark encodes
a range of challenges commonly encountered in real-world sequential decision
making problems, such as uncertainty. It can be configured to instantiate
problems of varying degrees of difficulty, e.g., in terms of variable
dimensionality. Our benchmark differs from other reinforcement learning
benchmarks, including the ones aiming to encode real-world difficulties, in
that it is directly derived from a real-world industrial problem, which
underwent minimal simplification and streamlining. It is sufficiently versatile
to evaluate reinforcement learning algorithms on any real-world problem that
fits our resource allocation framework. We provide results of standard baseline
methods. Going beyond the usual training reward curves, our results and the
statistical tools used to interpret them allow to highlight interesting
limitations of well-known deep reinforcement learning algorithms, namely PPO,
TRPO and DQN.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 13:44:29 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Pendyala",
"Abhijeet",
""
],
[
"Dettmer",
"Justin",
""
],
[
"Glasmachers",
"Tobias",
""
],
[
"Atamna",
"Asma",
""
]
] | TITLE: ContainerGym: A Real-World Reinforcement Learning Benchmark for Resource
Allocation
ABSTRACT: We present ContainerGym, a benchmark for reinforcement learning inspired by a
real-world industrial resource allocation task. The proposed benchmark encodes
a range of challenges commonly encountered in real-world sequential decision
making problems, such as uncertainty. It can be configured to instantiate
problems of varying degrees of difficulty, e.g., in terms of variable
dimensionality. Our benchmark differs from other reinforcement learning
benchmarks, including the ones aiming to encode real-world difficulties, in
that it is directly derived from a real-world industrial problem, which
underwent minimal simplification and streamlining. It is sufficiently versatile
to evaluate reinforcement learning algorithms on any real-world problem that
fits our resource allocation framework. We provide results of standard baseline
methods. Going beyond the usual training reward curves, our results and the
statistical tools used to interpret them allow to highlight interesting
limitations of well-known deep reinforcement learning algorithms, namely PPO,
TRPO and DQN.
| new_dataset | 0.683023 |
2307.02997 | Xi Jia | Xi Jia, Alexander Thorley, Alberto Gomez, Wenqi Lu, Dipak Kotecha and
Jinming Duan | Fourier-Net+: Leveraging Band-Limited Representation for Efficient 3D
Medical Image Registration | Under review. arXiv admin note: text overlap with arXiv:2211.16342 | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | U-Net style networks are commonly utilized in unsupervised image registration
to predict dense displacement fields, which for high-resolution volumetric
image data is a resource-intensive and time-consuming task. To tackle this
challenge, we first propose Fourier-Net, which replaces the costly U-Net style
expansive path with a parameter-free model-driven decoder. Instead of directly
predicting a full-resolution displacement field, our Fourier-Net learns a
low-dimensional representation of the displacement field in the band-limited
Fourier domain which our model-driven decoder converts to a full-resolution
displacement field in the spatial domain. Expanding upon Fourier-Net, we then
introduce Fourier-Net+, which additionally takes the band-limited spatial
representation of the images as input and further reduces the number of
convolutional layers in the U-Net style network's contracting path. Finally, to
enhance the registration performance, we propose a cascaded version of
Fourier-Net+. We evaluate our proposed methods on three datasets, on which our
proposed Fourier-Net and its variants achieve comparable results with current
state-of-the art methods, while exhibiting faster inference speeds, lower
memory footprint, and fewer multiply-add operations. With such small
computational cost, our Fourier-Net+ enables the efficient training of
large-scale 3D registration on low-VRAM GPUs. Our code is publicly available at
\url{https://github.com/xi-jia/Fourier-Net}.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 13:57:12 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Jia",
"Xi",
""
],
[
"Thorley",
"Alexander",
""
],
[
"Gomez",
"Alberto",
""
],
[
"Lu",
"Wenqi",
""
],
[
"Kotecha",
"Dipak",
""
],
[
"Duan",
"Jinming",
""
]
] | TITLE: Fourier-Net+: Leveraging Band-Limited Representation for Efficient 3D
Medical Image Registration
ABSTRACT: U-Net style networks are commonly utilized in unsupervised image registration
to predict dense displacement fields, which for high-resolution volumetric
image data is a resource-intensive and time-consuming task. To tackle this
challenge, we first propose Fourier-Net, which replaces the costly U-Net style
expansive path with a parameter-free model-driven decoder. Instead of directly
predicting a full-resolution displacement field, our Fourier-Net learns a
low-dimensional representation of the displacement field in the band-limited
Fourier domain which our model-driven decoder converts to a full-resolution
displacement field in the spatial domain. Expanding upon Fourier-Net, we then
introduce Fourier-Net+, which additionally takes the band-limited spatial
representation of the images as input and further reduces the number of
convolutional layers in the U-Net style network's contracting path. Finally, to
enhance the registration performance, we propose a cascaded version of
Fourier-Net+. We evaluate our proposed methods on three datasets, on which our
proposed Fourier-Net and its variants achieve comparable results with current
state-of-the art methods, while exhibiting faster inference speeds, lower
memory footprint, and fewer multiply-add operations. With such small
computational cost, our Fourier-Net+ enables the efficient training of
large-scale 3D registration on low-VRAM GPUs. Our code is publicly available at
\url{https://github.com/xi-jia/Fourier-Net}.
| no_new_dataset | 0.711857 |
2307.03043 | Sandeep Silwal | Ainesh Bakshi, Piotr Indyk, Rajesh Jayaram, Sandeep Silwal, Erik
Waingarten | A Near-Linear Time Algorithm for the Chamfer Distance | null | null | null | null | cs.DS cs.CG cs.GR cs.LG | http://creativecommons.org/licenses/by/4.0/ | For any two point sets $A,B \subset \mathbb{R}^d$ of size up to $n$, the
Chamfer distance from $A$ to $B$ is defined as $\text{CH}(A,B)=\sum_{a \in A}
\min_{b \in B} d_X(a,b)$, where $d_X$ is the underlying distance measure (e.g.,
the Euclidean or Manhattan distance). The Chamfer distance is a popular measure
of dissimilarity between point clouds, used in many machine learning, computer
vision, and graphics applications, and admits a straightforward $O(d n^2)$-time
brute force algorithm. Further, the Chamfer distance is often used as a proxy
for the more computationally demanding Earth-Mover (Optimal Transport)
Distance. However, the \emph{quadratic} dependence on $n$ in the running time
makes the naive approach intractable for large datasets.
We overcome this bottleneck and present the first $(1+\epsilon)$-approximate
algorithm for estimating the Chamfer distance with a near-linear running time.
Specifically, our algorithm runs in time $O(nd \log (n)/\varepsilon^2)$ and is
implementable. Our experiments demonstrate that it is both accurate and fast on
large high-dimensional datasets. We believe that our algorithm will open new
avenues for analyzing large high-dimensional point clouds. We also give
evidence that if the goal is to \emph{report} a $(1+\varepsilon)$-approximate
mapping from $A$ to $B$ (as opposed to just its value), then any sub-quadratic
time algorithm is unlikely to exist.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 15:07:48 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Bakshi",
"Ainesh",
""
],
[
"Indyk",
"Piotr",
""
],
[
"Jayaram",
"Rajesh",
""
],
[
"Silwal",
"Sandeep",
""
],
[
"Waingarten",
"Erik",
""
]
] | TITLE: A Near-Linear Time Algorithm for the Chamfer Distance
ABSTRACT: For any two point sets $A,B \subset \mathbb{R}^d$ of size up to $n$, the
Chamfer distance from $A$ to $B$ is defined as $\text{CH}(A,B)=\sum_{a \in A}
\min_{b \in B} d_X(a,b)$, where $d_X$ is the underlying distance measure (e.g.,
the Euclidean or Manhattan distance). The Chamfer distance is a popular measure
of dissimilarity between point clouds, used in many machine learning, computer
vision, and graphics applications, and admits a straightforward $O(d n^2)$-time
brute force algorithm. Further, the Chamfer distance is often used as a proxy
for the more computationally demanding Earth-Mover (Optimal Transport)
Distance. However, the \emph{quadratic} dependence on $n$ in the running time
makes the naive approach intractable for large datasets.
We overcome this bottleneck and present the first $(1+\epsilon)$-approximate
algorithm for estimating the Chamfer distance with a near-linear running time.
Specifically, our algorithm runs in time $O(nd \log (n)/\varepsilon^2)$ and is
implementable. Our experiments demonstrate that it is both accurate and fast on
large high-dimensional datasets. We believe that our algorithm will open new
avenues for analyzing large high-dimensional point clouds. We also give
evidence that if the goal is to \emph{report} a $(1+\varepsilon)$-approximate
mapping from $A$ to $B$ (as opposed to just its value), then any sub-quadratic
time algorithm is unlikely to exist.
| no_new_dataset | 0.708975 |
2307.03048 | Yan Lin | Yan Lin, Huaiyu Wan, Jilin Hu, Shengnan Guo, Bin Yang, Youfang Lin,
Christian S. Jensen | Origin-Destination Travel Time Oracle for Map-based Services | 15 pages, 12 figures, accepted by SIGMOD International Conference on
Management of Data 2024 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given an origin (O), a destination (D), and a departure time (T), an
Origin-Destination (OD) travel time oracle~(ODT-Oracle) returns an estimate of
the time it takes to travel from O to D when departing at T. ODT-Oracles serve
important purposes in map-based services. To enable the construction of such
oracles, we provide a travel-time estimation (TTE) solution that leverages
historical trajectories to estimate time-varying travel times for OD pairs.
The problem is complicated by the fact that multiple historical trajectories
with different travel times may connect an OD pair, while trajectories may vary
from one another. To solve the problem, it is crucial to remove outlier
trajectories when doing travel time estimation for future queries.
We propose a novel, two-stage framework called Diffusion-based
Origin-destination Travel Time Estimation (DOT), that solves the problem.
First, DOT employs a conditioned Pixelated Trajectories (PiT) denoiser that
enables building a diffusion-based PiT inference process by learning
correlations between OD pairs and historical trajectories. Specifically, given
an OD pair and a departure time, we aim to infer a PiT. Next, DOT encompasses a
Masked Vision Transformer~(MViT) that effectively and efficiently estimates a
travel time based on the inferred PiT. We report on extensive experiments on
two real-world datasets that offer evidence that DOT is capable of
outperforming baseline methods in terms of accuracy, scalability, and
explainability.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 15:14:23 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Lin",
"Yan",
""
],
[
"Wan",
"Huaiyu",
""
],
[
"Hu",
"Jilin",
""
],
[
"Guo",
"Shengnan",
""
],
[
"Yang",
"Bin",
""
],
[
"Lin",
"Youfang",
""
],
[
"Jensen",
"Christian S.",
""
]
] | TITLE: Origin-Destination Travel Time Oracle for Map-based Services
ABSTRACT: Given an origin (O), a destination (D), and a departure time (T), an
Origin-Destination (OD) travel time oracle~(ODT-Oracle) returns an estimate of
the time it takes to travel from O to D when departing at T. ODT-Oracles serve
important purposes in map-based services. To enable the construction of such
oracles, we provide a travel-time estimation (TTE) solution that leverages
historical trajectories to estimate time-varying travel times for OD pairs.
The problem is complicated by the fact that multiple historical trajectories
with different travel times may connect an OD pair, while trajectories may vary
from one another. To solve the problem, it is crucial to remove outlier
trajectories when doing travel time estimation for future queries.
We propose a novel, two-stage framework called Diffusion-based
Origin-destination Travel Time Estimation (DOT), that solves the problem.
First, DOT employs a conditioned Pixelated Trajectories (PiT) denoiser that
enables building a diffusion-based PiT inference process by learning
correlations between OD pairs and historical trajectories. Specifically, given
an OD pair and a departure time, we aim to infer a PiT. Next, DOT encompasses a
Masked Vision Transformer~(MViT) that effectively and efficiently estimates a
travel time based on the inferred PiT. We report on extensive experiments on
two real-world datasets that offer evidence that DOT is capable of
outperforming baseline methods in terms of accuracy, scalability, and
explainability.
| no_new_dataset | 0.710189 |
2307.03054 | Michael \'Alvarez MSc | Carlos Theran, Michael Alvarez, Emmanuel Arzuaga and Heidy Sierra | Multi-source imagery fusion using deep learning in a cloud computing
platform | null | null | null | null | cs.DC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Given the high availability of data collected by different remote sensing
instruments, the data fusion of multi-spectral and hyperspectral images (HSI)
is an important topic in remote sensing. In particular, super-resolution as a
data fusion application using spatial and spectral domains is highly
investigated because its fused images is used to improve the classification and
tracking objects accuracy. On the other hand, the huge amount of data obtained
by remote sensing instruments represent a key concern in terms of data storage,
management and pre-processing. This paper proposes a Big Data Cloud platform
using Hadoop and Spark to store, manages, and process remote sensing data.
Also, a study over the parameter \textit{chunk size} is presented to suggest
the appropriate value for this parameter to download imagery data from Hadoop
into a Spark application, based on the format of our data. We also developed an
alternative approach based on Long Short Term Memory trained with different
patch sizes for super-resolution image. This approach fuse hyperspectral and
multispectral images. As a result, we obtain images with high-spatial and
high-spectral resolution. The experimental results show that for a chunk size
of 64k, an average of 3.5s was required to download data from Hadoop into a
Spark application. The proposed model for super-resolution provides a
structural similarity index of 0.98 and 0.907 for the used dataset.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 15:18:11 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Theran",
"Carlos",
""
],
[
"Alvarez",
"Michael",
""
],
[
"Arzuaga",
"Emmanuel",
""
],
[
"Sierra",
"Heidy",
""
]
] | TITLE: Multi-source imagery fusion using deep learning in a cloud computing
platform
ABSTRACT: Given the high availability of data collected by different remote sensing
instruments, the data fusion of multi-spectral and hyperspectral images (HSI)
is an important topic in remote sensing. In particular, super-resolution as a
data fusion application using spatial and spectral domains is highly
investigated because its fused images is used to improve the classification and
tracking objects accuracy. On the other hand, the huge amount of data obtained
by remote sensing instruments represent a key concern in terms of data storage,
management and pre-processing. This paper proposes a Big Data Cloud platform
using Hadoop and Spark to store, manages, and process remote sensing data.
Also, a study over the parameter \textit{chunk size} is presented to suggest
the appropriate value for this parameter to download imagery data from Hadoop
into a Spark application, based on the format of our data. We also developed an
alternative approach based on Long Short Term Memory trained with different
patch sizes for super-resolution image. This approach fuse hyperspectral and
multispectral images. As a result, we obtain images with high-spatial and
high-spectral resolution. The experimental results show that for a chunk size
of 64k, an average of 3.5s was required to download data from Hadoop into a
Spark application. The proposed model for super-resolution provides a
structural similarity index of 0.98 and 0.907 for the used dataset.
| no_new_dataset | 0.711863 |
2307.03068 | Shadi Sartipi | Shadi Sartipi and Mastaneh Torkamani-Azar and Mujdat Cetin | A Hybrid End-to-End Spatio-Temporal Attention Neural Network with
Graph-Smooth Signals for EEG Emotion Recognition | null | null | null | null | cs.LG eess.SP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recently, physiological data such as electroencephalography (EEG) signals
have attracted significant attention in affective computing. In this context,
the main goal is to design an automated model that can assess emotional states.
Lately, deep neural networks have shown promising performance in emotion
recognition tasks. However, designing a deep architecture that can extract
practical information from raw data is still a challenge. Here, we introduce a
deep neural network that acquires interpretable physiological representations
by a hybrid structure of spatio-temporal encoding and recurrent attention
network blocks. Furthermore, a preprocessing step is applied to the raw data
using graph signal processing tools to perform graph smoothing in the spatial
domain. We demonstrate that our proposed architecture exceeds state-of-the-art
results for emotion classification on the publicly available DEAP dataset. To
explore the generality of the learned model, we also evaluate the performance
of our architecture towards transfer learning (TL) by transferring the model
parameters from a specific source to other target domains. Using DEAP as the
source dataset, we demonstrate the effectiveness of our model in performing
cross-modality TL and improving emotion classification accuracy on DREAMER and
the Emotional English Word (EEWD) datasets, which involve EEG-based emotion
classification tasks with different stimuli.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 15:35:14 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Sartipi",
"Shadi",
""
],
[
"Torkamani-Azar",
"Mastaneh",
""
],
[
"Cetin",
"Mujdat",
""
]
] | TITLE: A Hybrid End-to-End Spatio-Temporal Attention Neural Network with
Graph-Smooth Signals for EEG Emotion Recognition
ABSTRACT: Recently, physiological data such as electroencephalography (EEG) signals
have attracted significant attention in affective computing. In this context,
the main goal is to design an automated model that can assess emotional states.
Lately, deep neural networks have shown promising performance in emotion
recognition tasks. However, designing a deep architecture that can extract
practical information from raw data is still a challenge. Here, we introduce a
deep neural network that acquires interpretable physiological representations
by a hybrid structure of spatio-temporal encoding and recurrent attention
network blocks. Furthermore, a preprocessing step is applied to the raw data
using graph signal processing tools to perform graph smoothing in the spatial
domain. We demonstrate that our proposed architecture exceeds state-of-the-art
results for emotion classification on the publicly available DEAP dataset. To
explore the generality of the learned model, we also evaluate the performance
of our architecture towards transfer learning (TL) by transferring the model
parameters from a specific source to other target domains. Using DEAP as the
source dataset, we demonstrate the effectiveness of our model in performing
cross-modality TL and improving emotion classification accuracy on DREAMER and
the Emotional English Word (EEWD) datasets, which involve EEG-based emotion
classification tasks with different stimuli.
| no_new_dataset | 0.709019 |
2307.03083 | Abhay Goyal | Abhay Goyal, Nimay Parekh, Lam Yin Cheung, Koustuv Saha, Frederick L
Altice, Robin O'hanlon, Roger Ho Chun Man, Christian Poellabauer, Honoria
Guarino, Pedro Mateu Gelabert, Navin Kumar | Predicting Opioid Use Outcomes in Minoritized Communities | null | null | null | null | cs.CE | http://creativecommons.org/licenses/by/4.0/ | Machine learning algorithms can sometimes exacerbate health disparities based
on ethnicity, gender, and other factors. There has been limited work at
exploring potential biases within algorithms deployed on a small scale, and/or
within minoritized communities. Understanding the nature of potential biases
may improve the prediction of various health outcomes. As a case study, we used
data from a sample of 539 young adults from minoritized communities who engaged
in nonmedical use of prescription opioids and/or heroin. We addressed the
indicated issues through the following contributions: 1) Using machine learning
techniques, we predicted a range of opioid use outcomes for participants in our
dataset; 2) We assessed if algorithms trained only on a majority sub-sample
(e.g., Non-Hispanic/Latino, male), could accurately predict opioid use outcomes
for a minoritized sub-sample (e.g., Latino, female). Results indicated that
models trained on a random sample of our data could predict a range of opioid
use outcomes with high precision. However, we noted a decrease in precision
when we trained our models on data from a majority sub-sample, and tested these
models on a minoritized sub-sample. We posit that a range of cultural factors
and systemic forms of discrimination are not captured by data from majority
sub-samples. Broadly, for predictions to be valid, models should be trained on
data that includes adequate representation of the groups of people about whom
predictions will be made. Stakeholders may utilize our findings to mitigate
biases in models for predicting opioid use outcomes within minoritized
communities.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 15:50:58 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Goyal",
"Abhay",
""
],
[
"Parekh",
"Nimay",
""
],
[
"Cheung",
"Lam Yin",
""
],
[
"Saha",
"Koustuv",
""
],
[
"Altice",
"Frederick L",
""
],
[
"O'hanlon",
"Robin",
""
],
[
"Man",
"Roger Ho Chun",
""
],
[
"Poellabauer",
"Christian",
""
],
[
"Guarino",
"Honoria",
""
],
[
"Gelabert",
"Pedro Mateu",
""
],
[
"Kumar",
"Navin",
""
]
] | TITLE: Predicting Opioid Use Outcomes in Minoritized Communities
ABSTRACT: Machine learning algorithms can sometimes exacerbate health disparities based
on ethnicity, gender, and other factors. There has been limited work at
exploring potential biases within algorithms deployed on a small scale, and/or
within minoritized communities. Understanding the nature of potential biases
may improve the prediction of various health outcomes. As a case study, we used
data from a sample of 539 young adults from minoritized communities who engaged
in nonmedical use of prescription opioids and/or heroin. We addressed the
indicated issues through the following contributions: 1) Using machine learning
techniques, we predicted a range of opioid use outcomes for participants in our
dataset; 2) We assessed if algorithms trained only on a majority sub-sample
(e.g., Non-Hispanic/Latino, male), could accurately predict opioid use outcomes
for a minoritized sub-sample (e.g., Latino, female). Results indicated that
models trained on a random sample of our data could predict a range of opioid
use outcomes with high precision. However, we noted a decrease in precision
when we trained our models on data from a majority sub-sample, and tested these
models on a minoritized sub-sample. We posit that a range of cultural factors
and systemic forms of discrimination are not captured by data from majority
sub-samples. Broadly, for predictions to be valid, models should be trained on
data that includes adequate representation of the groups of people about whom
predictions will be made. Stakeholders may utilize our findings to mitigate
biases in models for predicting opioid use outcomes within minoritized
communities.
| no_new_dataset | 0.705405 |
2307.03101 | Jie Zhang | Jie Zhang, Masanori Suganuma, Takayuki Okatani | Contextual Affinity Distillation for Image Anomaly Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Previous works on unsupervised industrial anomaly detection mainly focus on
local structural anomalies such as cracks and color contamination. While
achieving significantly high detection performance on this kind of anomaly,
they are faced with logical anomalies that violate the long-range dependencies
such as a normal object placed in the wrong position. In this paper, based on
previous knowledge distillation works, we propose to use two students (local
and global) to better mimic the teacher's behavior. The local student, which is
used in previous studies mainly focuses on structural anomaly detection while
the global student pays attention to logical anomalies. To further encourage
the global student's learning to capture long-range dependencies, we design the
global context condensing block (GCCB) and propose a contextual affinity loss
for the student training and anomaly scoring. Experimental results show the
proposed method doesn't need cumbersome training techniques and achieves a new
state-of-the-art performance on the MVTec LOCO AD dataset.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 16:18:14 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Zhang",
"Jie",
""
],
[
"Suganuma",
"Masanori",
""
],
[
"Okatani",
"Takayuki",
""
]
] | TITLE: Contextual Affinity Distillation for Image Anomaly Detection
ABSTRACT: Previous works on unsupervised industrial anomaly detection mainly focus on
local structural anomalies such as cracks and color contamination. While
achieving significantly high detection performance on this kind of anomaly,
they are faced with logical anomalies that violate the long-range dependencies
such as a normal object placed in the wrong position. In this paper, based on
previous knowledge distillation works, we propose to use two students (local
and global) to better mimic the teacher's behavior. The local student, which is
used in previous studies mainly focuses on structural anomaly detection while
the global student pays attention to logical anomalies. To further encourage
the global student's learning to capture long-range dependencies, we design the
global context condensing block (GCCB) and propose a contextual affinity loss
for the student training and anomaly scoring. Experimental results show the
proposed method doesn't need cumbersome training techniques and achieves a new
state-of-the-art performance on the MVTec LOCO AD dataset.
| no_new_dataset | 0.711243 |
2307.03110 | Arjun Sridhar | Bhavna Gopal, Arjun Sridhar, Tunhou Zhang and Yiran Chen | LISSNAS: Locality-based Iterative Search Space Shrinkage for Neural
Architecture Search | null | IJCAI 2023 | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Search spaces hallmark the advancement of Neural Architecture Search (NAS).
Large and complex search spaces with versatile building operators and
structures provide more opportunities to brew promising architectures, yet pose
severe challenges on efficient exploration and exploitation. Subsequently,
several search space shrinkage methods optimize by selecting a single
sub-region that contains some well-performing networks. Small performance and
efficiency gains are observed with these methods but such techniques leave room
for significantly improved search performance and are ineffective at retaining
architectural diversity. We propose LISSNAS, an automated algorithm that
shrinks a large space into a diverse, small search space with SOTA search
performance. Our approach leverages locality, the relationship between
structural and performance similarity, to efficiently extract many pockets of
well-performing networks. We showcase our method on an array of search spaces
spanning various sizes and datasets. We accentuate the effectiveness of our
shrunk spaces when used in one-shot search by achieving the best Top-1 accuracy
in two different search spaces. Our method achieves a SOTA Top-1 accuracy of
77.6\% in ImageNet under mobile constraints, best-in-class Kendal-Tau,
architectural diversity, and search space size.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 16:28:51 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Gopal",
"Bhavna",
""
],
[
"Sridhar",
"Arjun",
""
],
[
"Zhang",
"Tunhou",
""
],
[
"Chen",
"Yiran",
""
]
] | TITLE: LISSNAS: Locality-based Iterative Search Space Shrinkage for Neural
Architecture Search
ABSTRACT: Search spaces hallmark the advancement of Neural Architecture Search (NAS).
Large and complex search spaces with versatile building operators and
structures provide more opportunities to brew promising architectures, yet pose
severe challenges on efficient exploration and exploitation. Subsequently,
several search space shrinkage methods optimize by selecting a single
sub-region that contains some well-performing networks. Small performance and
efficiency gains are observed with these methods but such techniques leave room
for significantly improved search performance and are ineffective at retaining
architectural diversity. We propose LISSNAS, an automated algorithm that
shrinks a large space into a diverse, small search space with SOTA search
performance. Our approach leverages locality, the relationship between
structural and performance similarity, to efficiently extract many pockets of
well-performing networks. We showcase our method on an array of search spaces
spanning various sizes and datasets. We accentuate the effectiveness of our
shrunk spaces when used in one-shot search by achieving the best Top-1 accuracy
in two different search spaces. Our method achieves a SOTA Top-1 accuracy of
77.6\% in ImageNet under mobile constraints, best-in-class Kendal-Tau,
architectural diversity, and search space size.
| no_new_dataset | 0.711819 |
2307.03113 | Michael Mior | Michael J. Mior | JSONoid: Monoid-based Enrichment for Configurable and Scalable
Data-Driven Schema Discovery | null | null | null | null | cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Schema discovery is an important aspect to working with data in formats such
as JSON. Unlike relational databases, JSON data sets often do not have
associated structural information. Consumers of such datasets are often left to
browse through data in an attempt to observe commonalities in structure across
documents to construct suitable code for data processing. However, this process
is time-consuming and error-prone. Existing distributed approaches to mining
schemas present a significant usability advantage as they provide useful
metadata for large data sources. However, depending on the data source, ad hoc
queries for estimating other properties to help with crafting an efficient data
pipeline can be expensive. We propose JSONoid, a distributed schema discovery
process augmented with additional metadata in the form of monoid data
structures that are easily maintainable in a distributed setting. JSONoid
subsumes several existing approaches to distributed schema discovery with
similar performance. Our approach also adds significant useful additional
information about data values to discovered schemas with linear scalability.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 16:30:38 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Mior",
"Michael J.",
""
]
] | TITLE: JSONoid: Monoid-based Enrichment for Configurable and Scalable
Data-Driven Schema Discovery
ABSTRACT: Schema discovery is an important aspect to working with data in formats such
as JSON. Unlike relational databases, JSON data sets often do not have
associated structural information. Consumers of such datasets are often left to
browse through data in an attempt to observe commonalities in structure across
documents to construct suitable code for data processing. However, this process
is time-consuming and error-prone. Existing distributed approaches to mining
schemas present a significant usability advantage as they provide useful
metadata for large data sources. However, depending on the data source, ad hoc
queries for estimating other properties to help with crafting an efficient data
pipeline can be expensive. We propose JSONoid, a distributed schema discovery
process augmented with additional metadata in the form of monoid data
structures that are easily maintainable in a distributed setting. JSONoid
subsumes several existing approaches to distributed schema discovery with
similar performance. Our approach also adds significant useful additional
information about data values to discovered schemas with linear scalability.
| no_new_dataset | 0.710396 |
2307.03115 | Zijun Yao | Zijun Yao, Yantao Liu, Xin Lv, Shulin Cao, Jifan Yu, Lei Hou, Juanzi
Li | KoRC: Knowledge oriented Reading Comprehension Benchmark for Deep Text
Understanding | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep text understanding, which requires the connections between a given
document and prior knowledge beyond its text, has been highlighted by many
benchmarks in recent years. However, these benchmarks have encountered two
major limitations. On the one hand, most of them require human annotation of
knowledge, which leads to limited knowledge coverage. On the other hand, they
usually use choices or spans in the texts as the answers, which results in
narrow answer space. To overcome these limitations, we build a new challenging
benchmark named KoRc in this paper. Compared with previous benchmarks, KoRC has
two advantages, i.e., broad knowledge coverage and flexible answer format.
Specifically, we utilize massive knowledge bases to guide annotators or large
language models (LLMs) to construct knowledgable questions. Moreover, we use
labels in knowledge bases rather than spans or choices as the final answers. We
test state-of-the-art models on KoRC and the experimental results show that the
strongest baseline only achieves 68.3% and 30.0% F1 measure in the
in-distribution and out-of-distribution test set, respectively. These results
indicate that deep text understanding is still an unsolved challenge. The
benchmark dataset, leaderboard, and baseline methods are released in
https://github.com/THU-KEG/KoRC.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 16:35:25 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Yao",
"Zijun",
""
],
[
"Liu",
"Yantao",
""
],
[
"Lv",
"Xin",
""
],
[
"Cao",
"Shulin",
""
],
[
"Yu",
"Jifan",
""
],
[
"Hou",
"Lei",
""
],
[
"Li",
"Juanzi",
""
]
] | TITLE: KoRC: Knowledge oriented Reading Comprehension Benchmark for Deep Text
Understanding
ABSTRACT: Deep text understanding, which requires the connections between a given
document and prior knowledge beyond its text, has been highlighted by many
benchmarks in recent years. However, these benchmarks have encountered two
major limitations. On the one hand, most of them require human annotation of
knowledge, which leads to limited knowledge coverage. On the other hand, they
usually use choices or spans in the texts as the answers, which results in
narrow answer space. To overcome these limitations, we build a new challenging
benchmark named KoRc in this paper. Compared with previous benchmarks, KoRC has
two advantages, i.e., broad knowledge coverage and flexible answer format.
Specifically, we utilize massive knowledge bases to guide annotators or large
language models (LLMs) to construct knowledgable questions. Moreover, we use
labels in knowledge bases rather than spans or choices as the final answers. We
test state-of-the-art models on KoRC and the experimental results show that the
strongest baseline only achieves 68.3% and 30.0% F1 measure in the
in-distribution and out-of-distribution test set, respectively. These results
indicate that deep text understanding is still an unsolved challenge. The
benchmark dataset, leaderboard, and baseline methods are released in
https://github.com/THU-KEG/KoRC.
| new_dataset | 0.715194 |
2307.03133 | Yongcan Yu | Yongcan Yu, Lijun Sheng, Ran He, Jian Liang | Benchmarking Test-Time Adaptation against Distribution Shifts in Image
Classification | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Test-time adaptation (TTA) is a technique aimed at enhancing the
generalization performance of models by leveraging unlabeled samples solely
during prediction. Given the need for robustness in neural network systems when
faced with distribution shifts, numerous TTA methods have recently been
proposed. However, evaluating these methods is often done under different
settings, such as varying distribution shifts, backbones, and designing
scenarios, leading to a lack of consistent and fair benchmarks to validate
their effectiveness. To address this issue, we present a benchmark that
systematically evaluates 13 prominent TTA methods and their variants on five
widely used image classification datasets: CIFAR-10-C, CIFAR-100-C, ImageNet-C,
DomainNet, and Office-Home. These methods encompass a wide range of adaptation
scenarios (e.g. online adaptation v.s. offline adaptation, instance adaptation
v.s. batch adaptation v.s. domain adaptation). Furthermore, we explore the
compatibility of different TTA methods with diverse network backbones. To
implement this benchmark, we have developed a unified framework in PyTorch,
which allows for consistent evaluation and comparison of the TTA methods across
the different datasets and network architectures. By establishing this
benchmark, we aim to provide researchers and practitioners with a reliable
means of assessing and comparing the effectiveness of TTA methods in improving
model robustness and generalization performance. Our code is available at
https://github.com/yuyongcan/Benchmark-TTA.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 16:59:53 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Yu",
"Yongcan",
""
],
[
"Sheng",
"Lijun",
""
],
[
"He",
"Ran",
""
],
[
"Liang",
"Jian",
""
]
] | TITLE: Benchmarking Test-Time Adaptation against Distribution Shifts in Image
Classification
ABSTRACT: Test-time adaptation (TTA) is a technique aimed at enhancing the
generalization performance of models by leveraging unlabeled samples solely
during prediction. Given the need for robustness in neural network systems when
faced with distribution shifts, numerous TTA methods have recently been
proposed. However, evaluating these methods is often done under different
settings, such as varying distribution shifts, backbones, and designing
scenarios, leading to a lack of consistent and fair benchmarks to validate
their effectiveness. To address this issue, we present a benchmark that
systematically evaluates 13 prominent TTA methods and their variants on five
widely used image classification datasets: CIFAR-10-C, CIFAR-100-C, ImageNet-C,
DomainNet, and Office-Home. These methods encompass a wide range of adaptation
scenarios (e.g. online adaptation v.s. offline adaptation, instance adaptation
v.s. batch adaptation v.s. domain adaptation). Furthermore, we explore the
compatibility of different TTA methods with diverse network backbones. To
implement this benchmark, we have developed a unified framework in PyTorch,
which allows for consistent evaluation and comparison of the TTA methods across
the different datasets and network architectures. By establishing this
benchmark, we aim to provide researchers and practitioners with a reliable
means of assessing and comparing the effectiveness of TTA methods in improving
model robustness and generalization performance. Our code is available at
https://github.com/yuyongcan/Benchmark-TTA.
| no_new_dataset | 0.693434 |
2307.03153 | Kate Sanders | Kate Sanders, David Etter, Reno Kriz, Benjamin Van Durme | MultiVENT: Multilingual Videos of Events with Aligned Natural Text | null | null | null | null | cs.IR cs.CV cs.MM | http://creativecommons.org/licenses/by/4.0/ | Everyday news coverage has shifted from traditional broadcasts towards a wide
range of presentation formats such as first-hand, unedited video footage.
Datasets that reflect the diverse array of multimodal, multilingual news
sources available online could be used to teach models to benefit from this
shift, but existing news video datasets focus on traditional news broadcasts
produced for English-speaking audiences. We address this limitation by
constructing MultiVENT, a dataset of multilingual, event-centric videos
grounded in text documents across five target languages. MultiVENT includes
both news broadcast videos and non-professional event footage, which we use to
analyze the state of online news videos and how they can be leveraged to build
robust, factually accurate models. Finally, we provide a model for complex,
multilingual video retrieval to serve as a baseline for information retrieval
using MultiVENT.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 17:29:34 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Sanders",
"Kate",
""
],
[
"Etter",
"David",
""
],
[
"Kriz",
"Reno",
""
],
[
"Van Durme",
"Benjamin",
""
]
] | TITLE: MultiVENT: Multilingual Videos of Events with Aligned Natural Text
ABSTRACT: Everyday news coverage has shifted from traditional broadcasts towards a wide
range of presentation formats such as first-hand, unedited video footage.
Datasets that reflect the diverse array of multimodal, multilingual news
sources available online could be used to teach models to benefit from this
shift, but existing news video datasets focus on traditional news broadcasts
produced for English-speaking audiences. We address this limitation by
constructing MultiVENT, a dataset of multilingual, event-centric videos
grounded in text documents across five target languages. MultiVENT includes
both news broadcast videos and non-professional event footage, which we use to
analyze the state of online news videos and how they can be leveraged to build
robust, factually accurate models. Finally, we provide a model for complex,
multilingual video retrieval to serve as a baseline for information retrieval
using MultiVENT.
| new_dataset | 0.715001 |
2307.03171 | Rahul Mihir Patel | Rahul Patel, Elias B. Khalil | LEO: Learning Efficient Orderings for Multiobjective Binary Decision
Diagrams | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Approaches based on Binary decision diagrams (BDDs) have recently achieved
state-of-the-art results for multiobjective integer programming problems. The
variable ordering used in constructing BDDs can have a significant impact on
their size and on the quality of bounds derived from relaxed or restricted BDDs
for single-objective optimization problems. We first showcase a similar impact
of variable ordering on the Pareto frontier (PF) enumeration time for the
multiobjective knapsack problem, suggesting the need for deriving variable
ordering methods that improve the scalability of the multiobjective BDD
approach. To that end, we derive a novel parameter configuration space based on
variable scoring functions which are linear in a small set of interpretable and
easy-to-compute variable features. We show how the configuration space can be
efficiently explored using black-box optimization, circumventing the curse of
dimensionality (in the number of variables and objectives), and finding good
orderings that reduce the PF enumeration time. However, black-box optimization
approaches incur a computational overhead that outweighs the reduction in time
due to good variable ordering. To alleviate this issue, we propose LEO, a
supervised learning approach for finding efficient variable orderings that
reduce the enumeration time. Experiments on benchmark sets from the knapsack
problem with 3-7 objectives and up to 80 variables show that LEO is ~30-300%
and ~10-200% faster at PF enumeration than common ordering strategies and
algorithm configuration. Our code and instances are available at
https://github.com/khalil-research/leo.
| [
{
"version": "v1",
"created": "Thu, 6 Jul 2023 17:52:29 GMT"
}
] | 2023-07-07T00:00:00 | [
[
"Patel",
"Rahul",
""
],
[
"Khalil",
"Elias B.",
""
]
] | TITLE: LEO: Learning Efficient Orderings for Multiobjective Binary Decision
Diagrams
ABSTRACT: Approaches based on Binary decision diagrams (BDDs) have recently achieved
state-of-the-art results for multiobjective integer programming problems. The
variable ordering used in constructing BDDs can have a significant impact on
their size and on the quality of bounds derived from relaxed or restricted BDDs
for single-objective optimization problems. We first showcase a similar impact
of variable ordering on the Pareto frontier (PF) enumeration time for the
multiobjective knapsack problem, suggesting the need for deriving variable
ordering methods that improve the scalability of the multiobjective BDD
approach. To that end, we derive a novel parameter configuration space based on
variable scoring functions which are linear in a small set of interpretable and
easy-to-compute variable features. We show how the configuration space can be
efficiently explored using black-box optimization, circumventing the curse of
dimensionality (in the number of variables and objectives), and finding good
orderings that reduce the PF enumeration time. However, black-box optimization
approaches incur a computational overhead that outweighs the reduction in time
due to good variable ordering. To alleviate this issue, we propose LEO, a
supervised learning approach for finding efficient variable orderings that
reduce the enumeration time. Experiments on benchmark sets from the knapsack
problem with 3-7 objectives and up to 80 variables show that LEO is ~30-300%
and ~10-200% faster at PF enumeration than common ordering strategies and
algorithm configuration. Our code and instances are available at
https://github.com/khalil-research/leo.
| no_new_dataset | 0.711813 |
1705.03387 | Hyeungill Lee | Hyeungill Lee, Sungyeob Han, Jungwoo Lee | Generative Adversarial Trainer: Defense to Adversarial Perturbations
with GAN | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We propose a novel technique to make neural network robust to adversarial
examples using a generative adversarial network. We alternately train both
classifier and generator networks. The generator network generates an
adversarial perturbation that can easily fool the classifier network by using a
gradient of each image. Simultaneously, the classifier network is trained to
classify correctly both original and adversarial images generated by the
generator. These procedures help the classifier network to become more robust
to adversarial perturbations. Furthermore, our adversarial training framework
efficiently reduces overfitting and outperforms other regularization methods
such as Dropout. We applied our method to supervised learning for CIFAR
datasets, and experimantal results show that our method significantly lowers
the generalization error of the network. To the best of our knowledge, this is
the first method which uses GAN to improve supervised learning.
| [
{
"version": "v1",
"created": "Tue, 9 May 2017 15:30:58 GMT"
},
{
"version": "v2",
"created": "Fri, 26 May 2017 21:44:32 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Jul 2023 06:49:26 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Lee",
"Hyeungill",
""
],
[
"Han",
"Sungyeob",
""
],
[
"Lee",
"Jungwoo",
""
]
] | TITLE: Generative Adversarial Trainer: Defense to Adversarial Perturbations
with GAN
ABSTRACT: We propose a novel technique to make neural network robust to adversarial
examples using a generative adversarial network. We alternately train both
classifier and generator networks. The generator network generates an
adversarial perturbation that can easily fool the classifier network by using a
gradient of each image. Simultaneously, the classifier network is trained to
classify correctly both original and adversarial images generated by the
generator. These procedures help the classifier network to become more robust
to adversarial perturbations. Furthermore, our adversarial training framework
efficiently reduces overfitting and outperforms other regularization methods
such as Dropout. We applied our method to supervised learning for CIFAR
datasets, and experimantal results show that our method significantly lowers
the generalization error of the network. To the best of our knowledge, this is
the first method which uses GAN to improve supervised learning.
| no_new_dataset | 0.711425 |
1901.03427 | Kurmanbek Kaiyrbekov | Kurmanbek Kaiyrbekov, Metin Sezgin | Stroke-based sketched symbol reconstruction and segmentation | null | IEEE Computer Graphics and Applications, vol. 40, no. 1, pp.
112-126, 1 Jan.-Feb. 2020 | 10.1109/MCG.2019.2943333 | null | cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hand-drawn objects usually consist of multiple semantically meaningful parts.
For example, a stick figure consists of a head, a torso, and pairs of legs and
arms. Efficient and accurate identification of these subparts promises to
significantly improve algorithms for stylization, deformation, morphing and
animation of 2D drawings. In this paper, we propose a neural network model that
segments symbols into stroke-level components. Our segmentation framework has
two main elements: a fixed feature extractor and a Multilayer Perceptron (MLP)
network that identifies a component based on the feature. As the feature
extractor we utilize an encoder of a stroke-rnn, which is our newly proposed
generative Variational Auto-Encoder (VAE) model that reconstructs symbols on a
stroke by stroke basis. Experiments show that a single encoder could be reused
for segmenting multiple categories of sketched symbols with negligible effects
on segmentation accuracies. Our segmentation scores surpass existing
methodologies on an available small state of the art dataset. Moreover,
extensive evaluations on our newly annotated big dataset demonstrate that our
framework obtains significantly better accuracies as compared to baseline
models. We release the dataset to the community.
| [
{
"version": "v1",
"created": "Thu, 10 Jan 2019 23:04:46 GMT"
},
{
"version": "v2",
"created": "Sat, 19 Jan 2019 07:32:09 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Kaiyrbekov",
"Kurmanbek",
""
],
[
"Sezgin",
"Metin",
""
]
] | TITLE: Stroke-based sketched symbol reconstruction and segmentation
ABSTRACT: Hand-drawn objects usually consist of multiple semantically meaningful parts.
For example, a stick figure consists of a head, a torso, and pairs of legs and
arms. Efficient and accurate identification of these subparts promises to
significantly improve algorithms for stylization, deformation, morphing and
animation of 2D drawings. In this paper, we propose a neural network model that
segments symbols into stroke-level components. Our segmentation framework has
two main elements: a fixed feature extractor and a Multilayer Perceptron (MLP)
network that identifies a component based on the feature. As the feature
extractor we utilize an encoder of a stroke-rnn, which is our newly proposed
generative Variational Auto-Encoder (VAE) model that reconstructs symbols on a
stroke by stroke basis. Experiments show that a single encoder could be reused
for segmenting multiple categories of sketched symbols with negligible effects
on segmentation accuracies. Our segmentation scores surpass existing
methodologies on an available small state of the art dataset. Moreover,
extensive evaluations on our newly annotated big dataset demonstrate that our
framework obtains significantly better accuracies as compared to baseline
models. We release the dataset to the community.
| new_dataset | 0.715014 |
2003.09053 | Dmitry Petrov | Marios Loizou, Siddhant Garg, Dmitry Petrov, Melinos Averkiou,
Evangelos Kalogerakis | Cross-Shape Attention for Part Segmentation of 3D Point Clouds | null | null | null | null | cs.CV cs.GR cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a deep learning method that propagates point-wise feature
representations across shapes within a collection for the purpose of 3D shape
segmentation. We propose a cross-shape attention mechanism to enable
interactions between a shape's point-wise features and those of other shapes.
The mechanism assesses both the degree of interaction between points and also
mediates feature propagation across shapes, improving the accuracy and
consistency of the resulting point-wise feature representations for shape
segmentation. Our method also proposes a shape retrieval measure to select
suitable shapes for cross-shape attention operations for each test shape. Our
experiments demonstrate that our approach yields state-of-the-art results in
the popular PartNet dataset.
| [
{
"version": "v1",
"created": "Fri, 20 Mar 2020 00:23:10 GMT"
},
{
"version": "v2",
"created": "Wed, 1 Apr 2020 15:53:11 GMT"
},
{
"version": "v3",
"created": "Mon, 6 Apr 2020 18:09:01 GMT"
},
{
"version": "v4",
"created": "Sat, 6 Aug 2022 16:25:37 GMT"
},
{
"version": "v5",
"created": "Thu, 25 May 2023 21:40:43 GMT"
},
{
"version": "v6",
"created": "Wed, 5 Jul 2023 16:26:28 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Loizou",
"Marios",
""
],
[
"Garg",
"Siddhant",
""
],
[
"Petrov",
"Dmitry",
""
],
[
"Averkiou",
"Melinos",
""
],
[
"Kalogerakis",
"Evangelos",
""
]
] | TITLE: Cross-Shape Attention for Part Segmentation of 3D Point Clouds
ABSTRACT: We present a deep learning method that propagates point-wise feature
representations across shapes within a collection for the purpose of 3D shape
segmentation. We propose a cross-shape attention mechanism to enable
interactions between a shape's point-wise features and those of other shapes.
The mechanism assesses both the degree of interaction between points and also
mediates feature propagation across shapes, improving the accuracy and
consistency of the resulting point-wise feature representations for shape
segmentation. Our method also proposes a shape retrieval measure to select
suitable shapes for cross-shape attention operations for each test shape. Our
experiments demonstrate that our approach yields state-of-the-art results in
the popular PartNet dataset.
| no_new_dataset | 0.710823 |
2011.04106 | Jieming Zhu | Jieming Zhu, Jinyang Liu, Weiqi Li, Jincai Lai, Xiuqiang He, Liang
Chen, Zibin Zheng | Ensemble Knowledge Distillation for CTR Prediction | Published in CIKM'2020 | null | null | null | cs.LG cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, deep learning-based models have been widely studied for
click-through rate (CTR) prediction and lead to improved prediction accuracy in
many industrial applications. However, current research focuses primarily on
building complex network architectures to better capture sophisticated feature
interactions and dynamic user behaviors. The increased model complexity may
slow down online inference and hinder its adoption in real-time applications.
Instead, our work targets at a new model training strategy based on knowledge
distillation (KD). KD is a teacher-student learning framework to transfer
knowledge learned from a teacher model to a student model. The KD strategy not
only allows us to simplify the student model as a vanilla DNN model but also
achieves significant accuracy improvements over the state-of-the-art teacher
models. The benefits thus motivate us to further explore the use of a powerful
ensemble of teachers for more accurate student model training. We also propose
some novel techniques to facilitate ensembled CTR prediction, including teacher
gating and early stopping by distillation loss. We conduct comprehensive
experiments against 12 existing models and across three industrial datasets.
Both offline and online A/B testing results show the effectiveness of our
KD-based training strategy.
| [
{
"version": "v1",
"created": "Sun, 8 Nov 2020 23:37:58 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 03:27:45 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Zhu",
"Jieming",
""
],
[
"Liu",
"Jinyang",
""
],
[
"Li",
"Weiqi",
""
],
[
"Lai",
"Jincai",
""
],
[
"He",
"Xiuqiang",
""
],
[
"Chen",
"Liang",
""
],
[
"Zheng",
"Zibin",
""
]
] | TITLE: Ensemble Knowledge Distillation for CTR Prediction
ABSTRACT: Recently, deep learning-based models have been widely studied for
click-through rate (CTR) prediction and lead to improved prediction accuracy in
many industrial applications. However, current research focuses primarily on
building complex network architectures to better capture sophisticated feature
interactions and dynamic user behaviors. The increased model complexity may
slow down online inference and hinder its adoption in real-time applications.
Instead, our work targets at a new model training strategy based on knowledge
distillation (KD). KD is a teacher-student learning framework to transfer
knowledge learned from a teacher model to a student model. The KD strategy not
only allows us to simplify the student model as a vanilla DNN model but also
achieves significant accuracy improvements over the state-of-the-art teacher
models. The benefits thus motivate us to further explore the use of a powerful
ensemble of teachers for more accurate student model training. We also propose
some novel techniques to facilitate ensembled CTR prediction, including teacher
gating and early stopping by distillation loss. We conduct comprehensive
experiments against 12 existing models and across three industrial datasets.
Both offline and online A/B testing results show the effectiveness of our
KD-based training strategy.
| no_new_dataset | 0.709252 |
2012.13392 | Chen Chen | Ce Zheng and Wenhan Wu and Chen Chen and Taojiannan Yang and Sijie Zhu
and Ju Shen and Nasser Kehtarnavaz and Mubarak Shah | Deep Learning-Based Human Pose Estimation: A Survey | null | null | null | null | cs.CV cs.GR cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human pose estimation aims to locate the human body parts and build human
body representation (e.g., body skeleton) from input data such as images and
videos. It has drawn increasing attention during the past decade and has been
utilized in a wide range of applications including human-computer interaction,
motion analysis, augmented reality, and virtual reality. Although the recently
developed deep learning-based solutions have achieved high performance in human
pose estimation, there still remain challenges due to insufficient training
data, depth ambiguities, and occlusion. The goal of this survey paper is to
provide a comprehensive review of recent deep learning-based solutions for both
2D and 3D pose estimation via a systematic analysis and comparison of these
solutions based on their input data and inference procedures. More than 250
research papers since 2014 are covered in this survey. Furthermore, 2D and 3D
human pose estimation datasets and evaluation metrics are included.
Quantitative performance comparisons of the reviewed methods on popular
datasets are summarized and discussed. Finally, the challenges involved,
applications, and future research directions are concluded. A regularly updated
project page is provided: \url{https://github.com/zczcwh/DL-HPE}
| [
{
"version": "v1",
"created": "Thu, 24 Dec 2020 18:49:06 GMT"
},
{
"version": "v2",
"created": "Sun, 27 Dec 2020 21:50:33 GMT"
},
{
"version": "v3",
"created": "Sat, 2 Jan 2021 19:24:54 GMT"
},
{
"version": "v4",
"created": "Sun, 23 Jan 2022 18:52:19 GMT"
},
{
"version": "v5",
"created": "Mon, 3 Jul 2023 21:21:19 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Zheng",
"Ce",
""
],
[
"Wu",
"Wenhan",
""
],
[
"Chen",
"Chen",
""
],
[
"Yang",
"Taojiannan",
""
],
[
"Zhu",
"Sijie",
""
],
[
"Shen",
"Ju",
""
],
[
"Kehtarnavaz",
"Nasser",
""
],
[
"Shah",
"Mubarak",
""
]
] | TITLE: Deep Learning-Based Human Pose Estimation: A Survey
ABSTRACT: Human pose estimation aims to locate the human body parts and build human
body representation (e.g., body skeleton) from input data such as images and
videos. It has drawn increasing attention during the past decade and has been
utilized in a wide range of applications including human-computer interaction,
motion analysis, augmented reality, and virtual reality. Although the recently
developed deep learning-based solutions have achieved high performance in human
pose estimation, there still remain challenges due to insufficient training
data, depth ambiguities, and occlusion. The goal of this survey paper is to
provide a comprehensive review of recent deep learning-based solutions for both
2D and 3D pose estimation via a systematic analysis and comparison of these
solutions based on their input data and inference procedures. More than 250
research papers since 2014 are covered in this survey. Furthermore, 2D and 3D
human pose estimation datasets and evaluation metrics are included.
Quantitative performance comparisons of the reviewed methods on popular
datasets are summarized and discussed. Finally, the challenges involved,
applications, and future research directions are concluded. A regularly updated
project page is provided: \url{https://github.com/zczcwh/DL-HPE}
| no_new_dataset | 0.709239 |
2101.11183 | Haipeng Li | Haipeng Li, Shuaicheng Liu, Jue Wang | DeepOIS: Gyroscope-Guided Deep Optical Image Stabilizer Compensation | null | IEEE Transactions on Circuits and Systems for Video Technology (
Volume: 32, Issue: 5, May 2022) | 10.1109/TCSVT.2021.3103281 | 21690602 | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Mobile captured images can be aligned using their gyroscope sensors. Optical
image stabilizer (OIS) terminates this possibility by adjusting the images
during the capturing. In this work, we propose a deep network that compensates
the motions caused by the OIS, such that the gyroscopes can be used for image
alignment on the OIS cameras. To achieve this, first, we record both videos and
gyroscopes with an OIS camera as training data. Then, we convert gyroscope
readings into motion fields. Second, we propose a Fundamental Mixtures motion
model for rolling shutter cameras, where an array of rotations within a frame
are extracted as the ground-truth guidance. Third, we train a convolutional
neural network with gyroscope motions as input to compensate for the OIS
motion. Once finished, the compensation network can be applied for other
scenes, where the image alignment is purely based on gyroscopes with no need
for images contents, delivering strong robustness. Experiments show that our
results are comparable with that of non-OIS cameras, and outperform image-based
alignment results with a relatively large margin. Code and dataset are
available at https://github.com/lhaippp/DeepOIS
| [
{
"version": "v1",
"created": "Wed, 27 Jan 2021 03:23:46 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 07:30:21 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Li",
"Haipeng",
""
],
[
"Liu",
"Shuaicheng",
""
],
[
"Wang",
"Jue",
""
]
] | TITLE: DeepOIS: Gyroscope-Guided Deep Optical Image Stabilizer Compensation
ABSTRACT: Mobile captured images can be aligned using their gyroscope sensors. Optical
image stabilizer (OIS) terminates this possibility by adjusting the images
during the capturing. In this work, we propose a deep network that compensates
the motions caused by the OIS, such that the gyroscopes can be used for image
alignment on the OIS cameras. To achieve this, first, we record both videos and
gyroscopes with an OIS camera as training data. Then, we convert gyroscope
readings into motion fields. Second, we propose a Fundamental Mixtures motion
model for rolling shutter cameras, where an array of rotations within a frame
are extracted as the ground-truth guidance. Third, we train a convolutional
neural network with gyroscope motions as input to compensate for the OIS
motion. Once finished, the compensation network can be applied for other
scenes, where the image alignment is purely based on gyroscopes with no need
for images contents, delivering strong robustness. Experiments show that our
results are comparable with that of non-OIS cameras, and outperform image-based
alignment results with a relatively large margin. Code and dataset are
available at https://github.com/lhaippp/DeepOIS
| no_new_dataset | 0.712664 |
2103.07050 | Chenhao Xu | Chenhao Xu, Jiaqi Ge, Yong Li, Yao Deng, Longxiang Gao, Mengshi Zhang,
Yong Xiang, Xi Zheng | SCEI: A Smart-Contract Driven Edge Intelligence Framework for IoT
Systems | null | null | 10.1109/TMC.2023.3290925 | null | cs.LG cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated learning (FL) enables collaborative training of a shared model on
edge devices while maintaining data privacy. FL is effective when dealing with
independent and identically distributed (iid) datasets, but struggles with
non-iid datasets. Various personalized approaches have been proposed, but such
approaches fail to handle underlying shifts in data distribution, such as data
distribution skew commonly observed in real-world scenarios (e.g., driver
behavior in smart transportation systems changing across time and location).
Additionally, trust concerns among unacquainted devices and security concerns
with the centralized aggregator pose additional challenges. To address these
challenges, this paper presents a dynamically optimized personal deep learning
scheme based on blockchain and federated learning. Specifically, the innovative
smart contract implemented in the blockchain allows distributed edge devices to
reach a consensus on the optimal weights of personalized models. Experimental
evaluations using multiple models and real-world datasets demonstrate that the
proposed scheme achieves higher accuracy and faster convergence compared to
traditional federated and personalized learning approaches.
| [
{
"version": "v1",
"created": "Fri, 12 Mar 2021 02:57:05 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 04:15:37 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Xu",
"Chenhao",
""
],
[
"Ge",
"Jiaqi",
""
],
[
"Li",
"Yong",
""
],
[
"Deng",
"Yao",
""
],
[
"Gao",
"Longxiang",
""
],
[
"Zhang",
"Mengshi",
""
],
[
"Xiang",
"Yong",
""
],
[
"Zheng",
"Xi",
""
]
] | TITLE: SCEI: A Smart-Contract Driven Edge Intelligence Framework for IoT
Systems
ABSTRACT: Federated learning (FL) enables collaborative training of a shared model on
edge devices while maintaining data privacy. FL is effective when dealing with
independent and identically distributed (iid) datasets, but struggles with
non-iid datasets. Various personalized approaches have been proposed, but such
approaches fail to handle underlying shifts in data distribution, such as data
distribution skew commonly observed in real-world scenarios (e.g., driver
behavior in smart transportation systems changing across time and location).
Additionally, trust concerns among unacquainted devices and security concerns
with the centralized aggregator pose additional challenges. To address these
challenges, this paper presents a dynamically optimized personal deep learning
scheme based on blockchain and federated learning. Specifically, the innovative
smart contract implemented in the blockchain allows distributed edge devices to
reach a consensus on the optimal weights of personalized models. Experimental
evaluations using multiple models and real-world datasets demonstrate that the
proposed scheme achieves higher accuracy and faster convergence compared to
traditional federated and personalized learning approaches.
| no_new_dataset | 0.709806 |
2104.12363 | Takuya Kanazawa | Takuya Kanazawa | One-parameter family of acquisition functions for efficient global
optimization | 13 pages, 6 figures. Accepted to IJCNN 2022 | 2022 International Joint Conference on Neural Networks (IJCNN),
Padua, Italy, 2022, pp. 1-8 | 10.1109/IJCNN55064.2022.9892219 | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bayesian optimization (BO) with Gaussian processes is a powerful methodology
to optimize an expensive black-box function with as few function evaluations as
possible. The expected improvement (EI) and probability of improvement (PI) are
among the most widely used schemes for BO. There is a plethora of other schemes
that outperform EI and PI, but most of them are numerically far more expensive
than EI and PI. In this work, we propose a new one-parameter family of
acquisition functions for BO that unifies EI and PI. The proposed method is
numerically inexpensive, is easy to implement, can be easily parallelized, and
on benchmark tasks shows a performance superior to EI and GP-UCB. Its
generalization to BO with Student-t processes is also presented.
| [
{
"version": "v1",
"created": "Mon, 26 Apr 2021 06:41:30 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Kanazawa",
"Takuya",
""
]
] | TITLE: One-parameter family of acquisition functions for efficient global
optimization
ABSTRACT: Bayesian optimization (BO) with Gaussian processes is a powerful methodology
to optimize an expensive black-box function with as few function evaluations as
possible. The expected improvement (EI) and probability of improvement (PI) are
among the most widely used schemes for BO. There is a plethora of other schemes
that outperform EI and PI, but most of them are numerically far more expensive
than EI and PI. In this work, we propose a new one-parameter family of
acquisition functions for BO that unifies EI and PI. The proposed method is
numerically inexpensive, is easy to implement, can be easily parallelized, and
on benchmark tasks shows a performance superior to EI and GP-UCB. Its
generalization to BO with Student-t processes is also presented.
| no_new_dataset | 0.709044 |
2105.00717 | Alon Shoshan | Alon Shoshan, Nadav Bhonker, Igor Kviatkovsky, Matan Fintz, Gerard
Medioni | Synthetic Data for Model Selection | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Recent breakthroughs in synthetic data generation approaches made it possible
to produce highly photorealistic images which are hardly distinguishable from
real ones. Furthermore, synthetic generation pipelines have the potential to
generate an unlimited number of images. The combination of high photorealism
and scale turn synthetic data into a promising candidate for improving various
machine learning (ML) pipelines. Thus far, a large body of research in this
field has focused on using synthetic images for training, by augmenting and
enlarging training data. In contrast to using synthetic data for training, in
this work we explore whether synthetic data can be beneficial for model
selection. Considering the task of image classification, we demonstrate that
when data is scarce, synthetic data can be used to replace the held out
validation set, thus allowing to train on a larger dataset. We also introduce a
novel method to calibrate the synthetic error estimation to fit that of the
real domain. We show that such calibration significantly improves the
usefulness of synthetic data for model selection.
| [
{
"version": "v1",
"created": "Mon, 3 May 2021 09:52:03 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 15:59:52 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Shoshan",
"Alon",
""
],
[
"Bhonker",
"Nadav",
""
],
[
"Kviatkovsky",
"Igor",
""
],
[
"Fintz",
"Matan",
""
],
[
"Medioni",
"Gerard",
""
]
] | TITLE: Synthetic Data for Model Selection
ABSTRACT: Recent breakthroughs in synthetic data generation approaches made it possible
to produce highly photorealistic images which are hardly distinguishable from
real ones. Furthermore, synthetic generation pipelines have the potential to
generate an unlimited number of images. The combination of high photorealism
and scale turn synthetic data into a promising candidate for improving various
machine learning (ML) pipelines. Thus far, a large body of research in this
field has focused on using synthetic images for training, by augmenting and
enlarging training data. In contrast to using synthetic data for training, in
this work we explore whether synthetic data can be beneficial for model
selection. Considering the task of image classification, we demonstrate that
when data is scarce, synthetic data can be used to replace the held out
validation set, thus allowing to train on a larger dataset. We also introduce a
novel method to calibrate the synthetic error estimation to fit that of the
real domain. We show that such calibration significantly improves the
usefulness of synthetic data for model selection.
| no_new_dataset | 0.7134 |
2106.04140 | Byeonggeun Kim | Byeonggeun Kim, Simyung Chang, Jinkyu Lee, Dooyong Sung | Broadcasted Residual Learning for Efficient Keyword Spotting | Proceedings of INTERSPEECH 2021 | null | null | null | cs.SD cs.LG eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Keyword spotting is an important research field because it plays a key role
in device wake-up and user interaction on smart devices. However, it is
challenging to minimize errors while operating efficiently in devices with
limited resources such as mobile phones. We present a broadcasted residual
learning method to achieve high accuracy with small model size and
computational load. Our method configures most of the residual functions as 1D
temporal convolution while still allows 2D convolution together using a
broadcasted-residual connection that expands temporal output to
frequency-temporal dimension. This residual mapping enables the network to
effectively represent useful audio features with much less computation than
conventional convolutional neural networks. We also propose a novel network
architecture, Broadcasting-residual network (BC-ResNet), based on broadcasted
residual learning and describe how to scale up the model according to the
target device's resources. BC-ResNets achieve state-of-the-art 98.0% and 98.7%
top-1 accuracy on Google speech command datasets v1 and v2, respectively, and
consistently outperform previous approaches, using fewer computations and
parameters. Code is available at
https://github.com/Qualcomm-AI-research/bcresnet.
| [
{
"version": "v1",
"created": "Tue, 8 Jun 2021 06:55:39 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jun 2021 02:56:06 GMT"
},
{
"version": "v3",
"created": "Thu, 27 Oct 2022 12:50:28 GMT"
},
{
"version": "v4",
"created": "Wed, 5 Jul 2023 15:18:54 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Kim",
"Byeonggeun",
""
],
[
"Chang",
"Simyung",
""
],
[
"Lee",
"Jinkyu",
""
],
[
"Sung",
"Dooyong",
""
]
] | TITLE: Broadcasted Residual Learning for Efficient Keyword Spotting
ABSTRACT: Keyword spotting is an important research field because it plays a key role
in device wake-up and user interaction on smart devices. However, it is
challenging to minimize errors while operating efficiently in devices with
limited resources such as mobile phones. We present a broadcasted residual
learning method to achieve high accuracy with small model size and
computational load. Our method configures most of the residual functions as 1D
temporal convolution while still allows 2D convolution together using a
broadcasted-residual connection that expands temporal output to
frequency-temporal dimension. This residual mapping enables the network to
effectively represent useful audio features with much less computation than
conventional convolutional neural networks. We also propose a novel network
architecture, Broadcasting-residual network (BC-ResNet), based on broadcasted
residual learning and describe how to scale up the model according to the
target device's resources. BC-ResNets achieve state-of-the-art 98.0% and 98.7%
top-1 accuracy on Google speech command datasets v1 and v2, respectively, and
consistently outperform previous approaches, using fewer computations and
parameters. Code is available at
https://github.com/Qualcomm-AI-research/bcresnet.
| no_new_dataset | 0.709252 |
2106.06847 | Jiezhang Cao | Jiezhang Cao, Yawei Li, Kai Zhang, Luc Van Gool | Video Super-Resolution Transformer | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video super-resolution (VSR), with the aim to restore a high-resolution video
from its corresponding low-resolution version, is a spatial-temporal sequence
prediction problem. Recently, Transformer has been gaining popularity due to
its parallel computing ability for sequence-to-sequence modeling. Thus, it
seems to be straightforward to apply the vision Transformer to solve VSR.
However, the typical block design of Transformer with a fully connected
self-attention layer and a token-wise feed-forward layer does not fit well for
VSR due to the following two reasons. First, the fully connected self-attention
layer neglects to exploit the data locality because this layer relies on linear
layers to compute attention maps. Second, the token-wise feed-forward layer
lacks the feature alignment which is important for VSR since this layer
independently processes each of the input token embeddings without any
interaction among them. In this paper, we make the first attempt to adapt
Transformer for VSR. Specifically, to tackle the first issue, we present a
spatial-temporal convolutional self-attention layer with a theoretical
understanding to exploit the locality information. For the second issue, we
design a bidirectional optical flow-based feed-forward layer to discover the
correlations across different video frames and also align features. Extensive
experiments on several benchmark datasets demonstrate the effectiveness of our
proposed method. The code will be available at
https://github.com/caojiezhang/VSR-Transformer.
| [
{
"version": "v1",
"created": "Sat, 12 Jun 2021 20:00:32 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Mar 2023 09:39:43 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Jul 2023 15:30:58 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Cao",
"Jiezhang",
""
],
[
"Li",
"Yawei",
""
],
[
"Zhang",
"Kai",
""
],
[
"Van Gool",
"Luc",
""
]
] | TITLE: Video Super-Resolution Transformer
ABSTRACT: Video super-resolution (VSR), with the aim to restore a high-resolution video
from its corresponding low-resolution version, is a spatial-temporal sequence
prediction problem. Recently, Transformer has been gaining popularity due to
its parallel computing ability for sequence-to-sequence modeling. Thus, it
seems to be straightforward to apply the vision Transformer to solve VSR.
However, the typical block design of Transformer with a fully connected
self-attention layer and a token-wise feed-forward layer does not fit well for
VSR due to the following two reasons. First, the fully connected self-attention
layer neglects to exploit the data locality because this layer relies on linear
layers to compute attention maps. Second, the token-wise feed-forward layer
lacks the feature alignment which is important for VSR since this layer
independently processes each of the input token embeddings without any
interaction among them. In this paper, we make the first attempt to adapt
Transformer for VSR. Specifically, to tackle the first issue, we present a
spatial-temporal convolutional self-attention layer with a theoretical
understanding to exploit the locality information. For the second issue, we
design a bidirectional optical flow-based feed-forward layer to discover the
correlations across different video frames and also align features. Extensive
experiments on several benchmark datasets demonstrate the effectiveness of our
proposed method. The code will be available at
https://github.com/caojiezhang/VSR-Transformer.
| no_new_dataset | 0.710409 |
2111.15430 | Bingyuan Liu | Bingyuan Liu, Ismail Ben Ayed, Adrian Galdran, Jose Dolz | The Devil is in the Margin: Margin-based Label Smoothing for Network
Calibration | CVPR 2022. Code: https://github.com/by-liu/MbLS | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In spite of the dominant performances of deep neural networks, recent works
have shown that they are poorly calibrated, resulting in over-confident
predictions. Miscalibration can be exacerbated by overfitting due to the
minimization of the cross-entropy during training, as it promotes the predicted
softmax probabilities to match the one-hot label assignments. This yields a
pre-softmax activation of the correct class that is significantly larger than
the remaining activations. Recent evidence from the literature suggests that
loss functions that embed implicit or explicit maximization of the entropy of
predictions yield state-of-the-art calibration performances. We provide a
unifying constrained-optimization perspective of current state-of-the-art
calibration losses. Specifically, these losses could be viewed as
approximations of a linear penalty (or a Lagrangian) imposing equality
constraints on logit distances. This points to an important limitation of such
underlying equality constraints, whose ensuing gradients constantly push
towards a non-informative solution, which might prevent from reaching the best
compromise between the discriminative performance and calibration of the model
during gradient-based optimization. Following our observations, we propose a
simple and flexible generalization based on inequality constraints, which
imposes a controllable margin on logit distances. Comprehensive experiments on
a variety of image classification, semantic segmentation and NLP benchmarks
demonstrate that our method sets novel state-of-the-art results on these tasks
in terms of network calibration, without affecting the discriminative
performance. The code is available at https://github.com/by-liu/MbLS .
| [
{
"version": "v1",
"created": "Tue, 30 Nov 2021 14:21:47 GMT"
},
{
"version": "v2",
"created": "Thu, 3 Mar 2022 02:39:36 GMT"
},
{
"version": "v3",
"created": "Tue, 29 Mar 2022 19:31:53 GMT"
},
{
"version": "v4",
"created": "Wed, 5 Jul 2023 07:10:35 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Liu",
"Bingyuan",
""
],
[
"Ayed",
"Ismail Ben",
""
],
[
"Galdran",
"Adrian",
""
],
[
"Dolz",
"Jose",
""
]
] | TITLE: The Devil is in the Margin: Margin-based Label Smoothing for Network
Calibration
ABSTRACT: In spite of the dominant performances of deep neural networks, recent works
have shown that they are poorly calibrated, resulting in over-confident
predictions. Miscalibration can be exacerbated by overfitting due to the
minimization of the cross-entropy during training, as it promotes the predicted
softmax probabilities to match the one-hot label assignments. This yields a
pre-softmax activation of the correct class that is significantly larger than
the remaining activations. Recent evidence from the literature suggests that
loss functions that embed implicit or explicit maximization of the entropy of
predictions yield state-of-the-art calibration performances. We provide a
unifying constrained-optimization perspective of current state-of-the-art
calibration losses. Specifically, these losses could be viewed as
approximations of a linear penalty (or a Lagrangian) imposing equality
constraints on logit distances. This points to an important limitation of such
underlying equality constraints, whose ensuing gradients constantly push
towards a non-informative solution, which might prevent from reaching the best
compromise between the discriminative performance and calibration of the model
during gradient-based optimization. Following our observations, we propose a
simple and flexible generalization based on inequality constraints, which
imposes a controllable margin on logit distances. Comprehensive experiments on
a variety of image classification, semantic segmentation and NLP benchmarks
demonstrate that our method sets novel state-of-the-art results on these tasks
in terms of network calibration, without affecting the discriminative
performance. The code is available at https://github.com/by-liu/MbLS .
| no_new_dataset | 0.709982 |
2112.01914 | Zheyuan Zhou | Zheyuan Zhou, Liang Du, Xiaoqing Ye, Zhikang Zou, Xiao Tan, Li Zhang,
Xiangyang Xue, Jianfeng Feng | SGM3D: Stereo Guided Monocular 3D Object Detection | 8 pages, 5 figures | null | 10.1109/LRA.2022.3191849 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Monocular 3D object detection aims to predict the object location, dimension
and orientation in 3D space alongside the object category given only a
monocular image. It poses a great challenge due to its ill-posed property which
is critically lack of depth information in the 2D image plane. While there
exist approaches leveraging off-the-shelve depth estimation or relying on LiDAR
sensors to mitigate this problem, the dependence on the additional depth model
or expensive equipment severely limits their scalability to generic 3D
perception. In this paper, we propose a stereo-guided monocular 3D object
detection framework, dubbed SGM3D, adapting the robust 3D features learned from
stereo inputs to enhance the feature for monocular detection. We innovatively
present a multi-granularity domain adaptation (MG-DA) mechanism to exploit the
network's ability to generate stereo-mimicking features given only on monocular
cues. Coarse BEV feature-level, as well as the fine anchor-level domain
adaptation, are both leveraged for guidance in the monocular domain.In
addition, we introduce an IoU matching-based alignment (IoU-MA) method for
object-level domain adaptation between the stereo and monocular predictions to
alleviate the mismatches while adopting the MG-DA. Extensive experiments
demonstrate state-of-the-art results on KITTI and Lyft datasets.
| [
{
"version": "v1",
"created": "Fri, 3 Dec 2021 13:57:14 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Feb 2022 16:43:36 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Zhou",
"Zheyuan",
""
],
[
"Du",
"Liang",
""
],
[
"Ye",
"Xiaoqing",
""
],
[
"Zou",
"Zhikang",
""
],
[
"Tan",
"Xiao",
""
],
[
"Zhang",
"Li",
""
],
[
"Xue",
"Xiangyang",
""
],
[
"Feng",
"Jianfeng",
""
]
] | TITLE: SGM3D: Stereo Guided Monocular 3D Object Detection
ABSTRACT: Monocular 3D object detection aims to predict the object location, dimension
and orientation in 3D space alongside the object category given only a
monocular image. It poses a great challenge due to its ill-posed property which
is critically lack of depth information in the 2D image plane. While there
exist approaches leveraging off-the-shelve depth estimation or relying on LiDAR
sensors to mitigate this problem, the dependence on the additional depth model
or expensive equipment severely limits their scalability to generic 3D
perception. In this paper, we propose a stereo-guided monocular 3D object
detection framework, dubbed SGM3D, adapting the robust 3D features learned from
stereo inputs to enhance the feature for monocular detection. We innovatively
present a multi-granularity domain adaptation (MG-DA) mechanism to exploit the
network's ability to generate stereo-mimicking features given only on monocular
cues. Coarse BEV feature-level, as well as the fine anchor-level domain
adaptation, are both leveraged for guidance in the monocular domain.In
addition, we introduce an IoU matching-based alignment (IoU-MA) method for
object-level domain adaptation between the stereo and monocular predictions to
alleviate the mismatches while adopting the MG-DA. Extensive experiments
demonstrate state-of-the-art results on KITTI and Lyft datasets.
| no_new_dataset | 0.708187 |
2112.08645 | Zachariah Carmichael | Zachariah Carmichael, Tim Moon, Sam Ade Jacobs | Learning Interpretable Models Through Multi-Objective Neural
Architecture Search | International Conference on Automated Machine Learning (AutoML)
Workshop | null | null | null | cs.LG cs.AI cs.NE | http://creativecommons.org/licenses/by/4.0/ | Monumental advances in deep learning have led to unprecedented achievements
across various domains. While the performance of deep neural networks is
indubitable, the architectural design and interpretability of such models are
nontrivial. Research has been introduced to automate the design of neural
network architectures through neural architecture search (NAS). Recent progress
has made these methods more pragmatic by exploiting distributed computation and
novel optimization algorithms. However, there is little work in optimizing
architectures for interpretability. To this end, we propose a multi-objective
distributed NAS framework that optimizes for both task performance and
"introspectability," a surrogate metric for aspects of interpretability. We
leverage the non-dominated sorting genetic algorithm (NSGA-II) and explainable
AI (XAI) techniques to reward architectures that can be better comprehended by
domain experts. The framework is evaluated on several image classification
datasets. We demonstrate that jointly optimizing for task error and
introspectability leads to more disentangled and debuggable architectures that
perform within tolerable error.
| [
{
"version": "v1",
"created": "Thu, 16 Dec 2021 05:50:55 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Mar 2022 00:27:24 GMT"
},
{
"version": "v3",
"created": "Sun, 13 Nov 2022 02:36:10 GMT"
},
{
"version": "v4",
"created": "Tue, 4 Jul 2023 16:31:25 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Carmichael",
"Zachariah",
""
],
[
"Moon",
"Tim",
""
],
[
"Jacobs",
"Sam Ade",
""
]
] | TITLE: Learning Interpretable Models Through Multi-Objective Neural
Architecture Search
ABSTRACT: Monumental advances in deep learning have led to unprecedented achievements
across various domains. While the performance of deep neural networks is
indubitable, the architectural design and interpretability of such models are
nontrivial. Research has been introduced to automate the design of neural
network architectures through neural architecture search (NAS). Recent progress
has made these methods more pragmatic by exploiting distributed computation and
novel optimization algorithms. However, there is little work in optimizing
architectures for interpretability. To this end, we propose a multi-objective
distributed NAS framework that optimizes for both task performance and
"introspectability," a surrogate metric for aspects of interpretability. We
leverage the non-dominated sorting genetic algorithm (NSGA-II) and explainable
AI (XAI) techniques to reward architectures that can be better comprehended by
domain experts. The framework is evaluated on several image classification
datasets. We demonstrate that jointly optimizing for task error and
introspectability leads to more disentangled and debuggable architectures that
perform within tolerable error.
| no_new_dataset | 0.707379 |
2201.05121 | Yunfan Ye | Yunfan Ye, Renjiao Yi, Zhiping Cai, Kai Xu | STEdge: Self-training Edge Detection with Multi-layer Teaching and
Regularization | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning-based edge detection has hereunto been strongly supervised with
pixel-wise annotations which are tedious to obtain manually. We study the
problem of self-training edge detection, leveraging the untapped wealth of
large-scale unlabeled image datasets. We design a self-supervised framework
with multi-layer regularization and self-teaching. In particular, we impose a
consistency regularization which enforces the outputs from each of the multiple
layers to be consistent for the input image and its perturbed counterpart. We
adopt L0-smoothing as the 'perturbation' to encourage edge prediction lying on
salient boundaries following the cluster assumption in self-supervised
learning. Meanwhile, the network is trained with multi-layer supervision by
pseudo labels which are initialized with Canny edges and then iteratively
refined by the network as the training proceeds. The regularization and
self-teaching together attain a good balance of precision and recall, leading
to a significant performance boost over supervised methods, with lightweight
refinement on the target dataset. Furthermore, our method demonstrates strong
cross-dataset generality. For example, it attains 4.8% improvement for ODS and
5.8% for OIS when tested on the unseen BIPED dataset, compared to the
state-of-the-art methods.
| [
{
"version": "v1",
"created": "Thu, 13 Jan 2022 18:26:36 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 17:52:07 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Ye",
"Yunfan",
""
],
[
"Yi",
"Renjiao",
""
],
[
"Cai",
"Zhiping",
""
],
[
"Xu",
"Kai",
""
]
] | TITLE: STEdge: Self-training Edge Detection with Multi-layer Teaching and
Regularization
ABSTRACT: Learning-based edge detection has hereunto been strongly supervised with
pixel-wise annotations which are tedious to obtain manually. We study the
problem of self-training edge detection, leveraging the untapped wealth of
large-scale unlabeled image datasets. We design a self-supervised framework
with multi-layer regularization and self-teaching. In particular, we impose a
consistency regularization which enforces the outputs from each of the multiple
layers to be consistent for the input image and its perturbed counterpart. We
adopt L0-smoothing as the 'perturbation' to encourage edge prediction lying on
salient boundaries following the cluster assumption in self-supervised
learning. Meanwhile, the network is trained with multi-layer supervision by
pseudo labels which are initialized with Canny edges and then iteratively
refined by the network as the training proceeds. The regularization and
self-teaching together attain a good balance of precision and recall, leading
to a significant performance boost over supervised methods, with lightweight
refinement on the target dataset. Furthermore, our method demonstrates strong
cross-dataset generality. For example, it attains 4.8% improvement for ODS and
5.8% for OIS when tested on the unseen BIPED dataset, compared to the
state-of-the-art methods.
| no_new_dataset | 0.709069 |
2201.12926 | Ekin Aky\"urek | Ekin Aky\"urek and Jacob Andreas | Compositionality as Lexical Symmetry | ACL2023 Final Version | null | null | null | cs.CL cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In tasks like semantic parsing, instruction following, and question
answering, standard deep networks fail to generalize compositionally from small
datasets. Many existing approaches overcome this limitation with model
architectures that enforce a compositional process of sentence interpretation.
In this paper, we present a domain-general and model-agnostic formulation of
compositionality as a constraint on symmetries of data distributions rather
than models. Informally, we prove that whenever a task can be solved by a
compositional model, there is a corresponding data augmentation scheme -- a
procedure for transforming examples into other well formed examples -- that
imparts compositional inductive bias on any model trained to solve the same
task. We describe a procedure called LEXSYM that discovers these
transformations automatically, then applies them to training data for ordinary
neural sequence models. Unlike existing compositional data augmentation
procedures, LEXSYM can be deployed agnostically across text, structured data,
and even images. It matches or surpasses state-of-the-art, task-specific models
on COGS semantic parsing, SCAN and ALCHEMY instruction following, and
CLEVR-COGENT visual question answering datasets.
| [
{
"version": "v1",
"created": "Sun, 30 Jan 2022 21:44:46 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 17:59:33 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Akyürek",
"Ekin",
""
],
[
"Andreas",
"Jacob",
""
]
] | TITLE: Compositionality as Lexical Symmetry
ABSTRACT: In tasks like semantic parsing, instruction following, and question
answering, standard deep networks fail to generalize compositionally from small
datasets. Many existing approaches overcome this limitation with model
architectures that enforce a compositional process of sentence interpretation.
In this paper, we present a domain-general and model-agnostic formulation of
compositionality as a constraint on symmetries of data distributions rather
than models. Informally, we prove that whenever a task can be solved by a
compositional model, there is a corresponding data augmentation scheme -- a
procedure for transforming examples into other well formed examples -- that
imparts compositional inductive bias on any model trained to solve the same
task. We describe a procedure called LEXSYM that discovers these
transformations automatically, then applies them to training data for ordinary
neural sequence models. Unlike existing compositional data augmentation
procedures, LEXSYM can be deployed agnostically across text, structured data,
and even images. It matches or surpasses state-of-the-art, task-specific models
on COGS semantic parsing, SCAN and ALCHEMY instruction following, and
CLEVR-COGENT visual question answering datasets.
| no_new_dataset | 0.709265 |
2204.12900 | Xin Zhang | Xin Zhang and Xiaohua Xie and Jianhuang Lai and Wei-Shi Zheng | Cross-Camera Trajectories Help Person Retrieval in a Camera Network | IEEE Transactions on Image Processing (TIP), 2023 | null | 10.1109/TIP.2023.3290515 | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We are concerned with retrieving a query person from multiple videos captured
by a non-overlapping camera network. Existing methods often rely on purely
visual matching or consider temporal constraints but ignore the spatial
information of the camera network. To address this issue, we propose a
pedestrian retrieval framework based on cross-camera trajectory generation,
which integrates both temporal and spatial information. To obtain pedestrian
trajectories, we propose a novel cross-camera spatio-temporal model that
integrates pedestrians' walking habits and the path layout between cameras to
form a joint probability distribution. Such a spatio-temporal model among a
camera network can be specified using sparsely sampled pedestrian data. Based
on the spatio-temporal model, cross-camera trajectories can be extracted by the
conditional random field model and further optimized by restricted non-negative
matrix factorization. Finally, a trajectory re-ranking technique is proposed to
improve the pedestrian retrieval results. To verify the effectiveness of our
method, we construct the first cross-camera pedestrian trajectory dataset, the
Person Trajectory Dataset, in real surveillance scenarios. Extensive
experiments verify the effectiveness and robustness of the proposed method.
| [
{
"version": "v1",
"created": "Wed, 27 Apr 2022 13:10:48 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Mar 2023 10:41:04 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Jul 2023 02:20:38 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Zhang",
"Xin",
""
],
[
"Xie",
"Xiaohua",
""
],
[
"Lai",
"Jianhuang",
""
],
[
"Zheng",
"Wei-Shi",
""
]
] | TITLE: Cross-Camera Trajectories Help Person Retrieval in a Camera Network
ABSTRACT: We are concerned with retrieving a query person from multiple videos captured
by a non-overlapping camera network. Existing methods often rely on purely
visual matching or consider temporal constraints but ignore the spatial
information of the camera network. To address this issue, we propose a
pedestrian retrieval framework based on cross-camera trajectory generation,
which integrates both temporal and spatial information. To obtain pedestrian
trajectories, we propose a novel cross-camera spatio-temporal model that
integrates pedestrians' walking habits and the path layout between cameras to
form a joint probability distribution. Such a spatio-temporal model among a
camera network can be specified using sparsely sampled pedestrian data. Based
on the spatio-temporal model, cross-camera trajectories can be extracted by the
conditional random field model and further optimized by restricted non-negative
matrix factorization. Finally, a trajectory re-ranking technique is proposed to
improve the pedestrian retrieval results. To verify the effectiveness of our
method, we construct the first cross-camera pedestrian trajectory dataset, the
Person Trajectory Dataset, in real surveillance scenarios. Extensive
experiments verify the effectiveness and robustness of the proposed method.
| new_dataset | 0.716001 |
2205.11018 | Lalita Kumari | Lalita Kumari, Sukhdeep Singh, Vaibhav Varish Singh Rathore and Anuj
Sharma | A Comprehensive Handwritten Paragraph Text Recognition System:
LexiconNet | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this study, we have presented an efficient procedure using two
state-of-the-art approaches from the literature of handwritten text recognition
as Vertical Attention Network and Word Beam Search. The attention module is
responsible for internal line segmentation that consequently processes a page
in a line-by-line manner. At the decoding step, we have added a connectionist
temporal classification-based word beam search decoder as a post-processing
step. In this study, an end-to-end paragraph recognition system is presented
with a lexicon decoder as a post-processing step. Our procedure reports
state-of-the-art results on standard datasets. The reported character error
rate is 3.24% on the IAM dataset with 27.19% improvement, 1.13% on RIMES with
40.83% improvement and 2.43% on the READ-16 dataset with 32.31% improvement
from existing literature and the word error rate is 8.29% on IAM dataset with
43.02% improvement, 2.94% on RIMES dataset with 56.25% improvement and 7.35% on
READ-2016 dataset with 47.27% improvement from the existing results. The
character error rate and word error rate reported in this work surpass the
results reported in the literature.
| [
{
"version": "v1",
"created": "Mon, 23 May 2022 03:35:45 GMT"
},
{
"version": "v2",
"created": "Sat, 29 Oct 2022 03:16:59 GMT"
},
{
"version": "v3",
"created": "Wed, 5 Jul 2023 06:08:29 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Kumari",
"Lalita",
""
],
[
"Singh",
"Sukhdeep",
""
],
[
"Rathore",
"Vaibhav Varish Singh",
""
],
[
"Sharma",
"Anuj",
""
]
] | TITLE: A Comprehensive Handwritten Paragraph Text Recognition System:
LexiconNet
ABSTRACT: In this study, we have presented an efficient procedure using two
state-of-the-art approaches from the literature of handwritten text recognition
as Vertical Attention Network and Word Beam Search. The attention module is
responsible for internal line segmentation that consequently processes a page
in a line-by-line manner. At the decoding step, we have added a connectionist
temporal classification-based word beam search decoder as a post-processing
step. In this study, an end-to-end paragraph recognition system is presented
with a lexicon decoder as a post-processing step. Our procedure reports
state-of-the-art results on standard datasets. The reported character error
rate is 3.24% on the IAM dataset with 27.19% improvement, 1.13% on RIMES with
40.83% improvement and 2.43% on the READ-16 dataset with 32.31% improvement
from existing literature and the word error rate is 8.29% on IAM dataset with
43.02% improvement, 2.94% on RIMES dataset with 56.25% improvement and 7.35% on
READ-2016 dataset with 47.27% improvement from the existing results. The
character error rate and word error rate reported in this work surpass the
results reported in the literature.
| no_new_dataset | 0.713251 |
2206.12336 | Chen Ling | Chen Ling, Carl Yang, and Liang Zhao | Deep Generation of Heterogeneous Networks | 10 pages, accepted by 2021 IEEE International Conference on Data
Mining (ICDM). IEEE, 2021 | null | 10.1109/ICDM51629.2021.00049 | null | cs.SI cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | Heterogeneous graphs are ubiquitous data structures that can inherently
capture multi-type and multi-modal interactions between objects. In recent
years, research on encoding heterogeneous graph into latent representations
have enjoyed a rapid increase. However, its reverse process, namely how to
construct heterogeneous graphs from underlying representations and
distributions have not been well explored due to several challenges in 1)
modeling the local heterogeneous semantic distribution; 2) preserving the
graph-structured distributions over the local semantics; and 3) characterizing
the global heterogeneous graph distributions. To address these challenges, we
propose a novel framework for heterogeneous graph generation (HGEN) that
jointly captures the semantic, structural, and global distributions of
heterogeneous graphs. Specifically, we propose a heterogeneous walk generator
that hierarchically generates meta-paths and their path instances. In addition,
a novel heterogeneous graph assembler is developed that can sample and combine
the generated meta-path instances (e.g., walks) into heterogeneous graphs in a
stratified manner. Theoretical analysis on the preservation of heterogeneous
graph patterns by the proposed generation process has been performed. Extensive
experiments on multiple real-world and synthetic heterogeneous graph datasets
demonstrate the effectiveness of the proposed HGEN in generating realistic
heterogeneous graphs.
| [
{
"version": "v1",
"created": "Fri, 24 Jun 2022 15:14:18 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 14:59:15 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Ling",
"Chen",
""
],
[
"Yang",
"Carl",
""
],
[
"Zhao",
"Liang",
""
]
] | TITLE: Deep Generation of Heterogeneous Networks
ABSTRACT: Heterogeneous graphs are ubiquitous data structures that can inherently
capture multi-type and multi-modal interactions between objects. In recent
years, research on encoding heterogeneous graph into latent representations
have enjoyed a rapid increase. However, its reverse process, namely how to
construct heterogeneous graphs from underlying representations and
distributions have not been well explored due to several challenges in 1)
modeling the local heterogeneous semantic distribution; 2) preserving the
graph-structured distributions over the local semantics; and 3) characterizing
the global heterogeneous graph distributions. To address these challenges, we
propose a novel framework for heterogeneous graph generation (HGEN) that
jointly captures the semantic, structural, and global distributions of
heterogeneous graphs. Specifically, we propose a heterogeneous walk generator
that hierarchically generates meta-paths and their path instances. In addition,
a novel heterogeneous graph assembler is developed that can sample and combine
the generated meta-path instances (e.g., walks) into heterogeneous graphs in a
stratified manner. Theoretical analysis on the preservation of heterogeneous
graph patterns by the proposed generation process has been performed. Extensive
experiments on multiple real-world and synthetic heterogeneous graph datasets
demonstrate the effectiveness of the proposed HGEN in generating realistic
heterogeneous graphs.
| no_new_dataset | 0.707582 |
2207.06079 | Awet Haileslassie Gebrehiwot | Awet Haileslassie Gebrehiwot, Patrik Vacek, David Hurych, Karel
Zimmermann, Patrick Perez, Tom\'a\v{s} Svoboda | Teachers in concordance for pseudo-labeling of 3D sequential data | This work has been submitted to the IEEE for publication | in IEEE Robotics and Automation Letters, vol. 8, no. 2, pp.
536-543, Feb. 2023 | 10.1109/LRA.2022.3226029 | null | cs.CV cs.RO | http://creativecommons.org/licenses/by-sa/4.0/ | Automatic pseudo-labeling is a powerful tool to tap into large amounts of
sequential unlabeled data. It is specially appealing in safety-critical
applications of autonomous driving, where performance requirements are extreme,
datasets are large, and manual labeling is very challenging. We propose to
leverage sequences of point clouds to boost the pseudolabeling technique in a
teacher-student setup via training multiple teachers, each with access to
different temporal information. This set of teachers, dubbed Concordance,
provides higher quality pseudo-labels for student training than standard
methods. The output of multiple teachers is combined via a novel pseudo label
confidence-guided criterion. Our experimental evaluation focuses on the 3D
point cloud domain and urban driving scenarios. We show the performance of our
method applied to 3D semantic segmentation and 3D object detection on three
benchmark datasets. Our approach, which uses only 20% manual labels,
outperforms some fully supervised methods. A notable performance boost is
achieved for classes rarely appearing in training data.
| [
{
"version": "v1",
"created": "Wed, 13 Jul 2022 09:40:22 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 04:31:58 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Gebrehiwot",
"Awet Haileslassie",
""
],
[
"Vacek",
"Patrik",
""
],
[
"Hurych",
"David",
""
],
[
"Zimmermann",
"Karel",
""
],
[
"Perez",
"Patrick",
""
],
[
"Svoboda",
"Tomáš",
""
]
] | TITLE: Teachers in concordance for pseudo-labeling of 3D sequential data
ABSTRACT: Automatic pseudo-labeling is a powerful tool to tap into large amounts of
sequential unlabeled data. It is specially appealing in safety-critical
applications of autonomous driving, where performance requirements are extreme,
datasets are large, and manual labeling is very challenging. We propose to
leverage sequences of point clouds to boost the pseudolabeling technique in a
teacher-student setup via training multiple teachers, each with access to
different temporal information. This set of teachers, dubbed Concordance,
provides higher quality pseudo-labels for student training than standard
methods. The output of multiple teachers is combined via a novel pseudo label
confidence-guided criterion. Our experimental evaluation focuses on the 3D
point cloud domain and urban driving scenarios. We show the performance of our
method applied to 3D semantic segmentation and 3D object detection on three
benchmark datasets. Our approach, which uses only 20% manual labels,
outperforms some fully supervised methods. A notable performance boost is
achieved for classes rarely appearing in training data.
| no_new_dataset | 0.710465 |
2207.06825 | David Bruggemann | David Bruggemann, Christos Sakaridis, Prune Truong, Luc Van Gool | Refign: Align and Refine for Adaptation of Semantic Segmentation to
Adverse Conditions | IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Due to the scarcity of dense pixel-level semantic annotations for images
recorded in adverse visual conditions, there has been a keen interest in
unsupervised domain adaptation (UDA) for the semantic segmentation of such
images. UDA adapts models trained on normal conditions to the target
adverse-condition domains. Meanwhile, multiple datasets with driving scenes
provide corresponding images of the same scenes across multiple conditions,
which can serve as a form of weak supervision for domain adaptation. We propose
Refign, a generic extension to self-training-based UDA methods which leverages
these cross-domain correspondences. Refign consists of two steps: (1) aligning
the normal-condition image to the corresponding adverse-condition image using
an uncertainty-aware dense matching network, and (2) refining the adverse
prediction with the normal prediction using an adaptive label correction
mechanism. We design custom modules to streamline both steps and set the new
state of the art for domain-adaptive semantic segmentation on several
adverse-condition benchmarks, including ACDC and Dark Zurich. The approach
introduces no extra training parameters, minimal computational overhead --
during training only -- and can be used as a drop-in extension to improve any
given self-training-based UDA method. Code is available at
https://github.com/brdav/refign.
| [
{
"version": "v1",
"created": "Thu, 14 Jul 2022 11:30:38 GMT"
},
{
"version": "v2",
"created": "Tue, 20 Sep 2022 12:15:30 GMT"
},
{
"version": "v3",
"created": "Mon, 3 Jul 2023 19:10:55 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Bruggemann",
"David",
""
],
[
"Sakaridis",
"Christos",
""
],
[
"Truong",
"Prune",
""
],
[
"Van Gool",
"Luc",
""
]
] | TITLE: Refign: Align and Refine for Adaptation of Semantic Segmentation to
Adverse Conditions
ABSTRACT: Due to the scarcity of dense pixel-level semantic annotations for images
recorded in adverse visual conditions, there has been a keen interest in
unsupervised domain adaptation (UDA) for the semantic segmentation of such
images. UDA adapts models trained on normal conditions to the target
adverse-condition domains. Meanwhile, multiple datasets with driving scenes
provide corresponding images of the same scenes across multiple conditions,
which can serve as a form of weak supervision for domain adaptation. We propose
Refign, a generic extension to self-training-based UDA methods which leverages
these cross-domain correspondences. Refign consists of two steps: (1) aligning
the normal-condition image to the corresponding adverse-condition image using
an uncertainty-aware dense matching network, and (2) refining the adverse
prediction with the normal prediction using an adaptive label correction
mechanism. We design custom modules to streamline both steps and set the new
state of the art for domain-adaptive semantic segmentation on several
adverse-condition benchmarks, including ACDC and Dark Zurich. The approach
introduces no extra training parameters, minimal computational overhead --
during training only -- and can be used as a drop-in extension to improve any
given self-training-based UDA method. Code is available at
https://github.com/brdav/refign.
| no_new_dataset | 0.710239 |
2207.07268 | Youpeng Zhao | Youpeng Zhao, Huadong Tang, Yingying Jiang, Yong A and Qiang Wu | Lightweight Vision Transformer with Cross Feature Attention | Technical Report. A shorter version has been accepted to ICIP 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in vision transformers (ViTs) have achieved great performance
in visual recognition tasks. Convolutional neural networks (CNNs) exploit
spatial inductive bias to learn visual representations, but these networks are
spatially local. ViTs can learn global representations with their
self-attention mechanism, but they are usually heavy-weight and unsuitable for
mobile devices. In this paper, we propose cross feature attention (XFA) to
bring down computation cost for transformers, and combine efficient mobile CNNs
to form a novel efficient light-weight CNN-ViT hybrid model, XFormer, which can
serve as a general-purpose backbone to learn both global and local
representation. Experimental results show that XFormer outperforms numerous CNN
and ViT-based models across different tasks and datasets. On ImageNet1K
dataset, XFormer achieves top-1 accuracy of 78.5% with 5.5 million parameters,
which is 2.2% and 6.3% more accurate than EfficientNet-B0 (CNN-based) and DeiT
(ViT-based) for similar number of parameters. Our model also performs well when
transferring to object detection and semantic segmentation tasks. On MS COCO
dataset, XFormer exceeds MobileNetV2 by 10.5 AP (22.7 -> 33.2 AP) in YOLOv3
framework with only 6.3M parameters and 3.8G FLOPs. On Cityscapes dataset, with
only a simple all-MLP decoder, XFormer achieves mIoU of 78.5 and FPS of 15.3,
surpassing state-of-the-art lightweight segmentation networks.
| [
{
"version": "v1",
"created": "Fri, 15 Jul 2022 03:27:13 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 16:11:41 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Zhao",
"Youpeng",
""
],
[
"Tang",
"Huadong",
""
],
[
"Jiang",
"Yingying",
""
],
[
"A",
"Yong",
""
],
[
"Wu",
"Qiang",
""
]
] | TITLE: Lightweight Vision Transformer with Cross Feature Attention
ABSTRACT: Recent advances in vision transformers (ViTs) have achieved great performance
in visual recognition tasks. Convolutional neural networks (CNNs) exploit
spatial inductive bias to learn visual representations, but these networks are
spatially local. ViTs can learn global representations with their
self-attention mechanism, but they are usually heavy-weight and unsuitable for
mobile devices. In this paper, we propose cross feature attention (XFA) to
bring down computation cost for transformers, and combine efficient mobile CNNs
to form a novel efficient light-weight CNN-ViT hybrid model, XFormer, which can
serve as a general-purpose backbone to learn both global and local
representation. Experimental results show that XFormer outperforms numerous CNN
and ViT-based models across different tasks and datasets. On ImageNet1K
dataset, XFormer achieves top-1 accuracy of 78.5% with 5.5 million parameters,
which is 2.2% and 6.3% more accurate than EfficientNet-B0 (CNN-based) and DeiT
(ViT-based) for similar number of parameters. Our model also performs well when
transferring to object detection and semantic segmentation tasks. On MS COCO
dataset, XFormer exceeds MobileNetV2 by 10.5 AP (22.7 -> 33.2 AP) in YOLOv3
framework with only 6.3M parameters and 3.8G FLOPs. On Cityscapes dataset, with
only a simple all-MLP decoder, XFormer achieves mIoU of 78.5 and FPS of 15.3,
surpassing state-of-the-art lightweight segmentation networks.
| no_new_dataset | 0.711193 |
2207.13730 | Takuya Kanazawa | Takuya Kanazawa, Haiyan Wang, Chetan Gupta | Distributional Actor-Critic Ensemble for Uncertainty-Aware Continuous
Control | 10 pages, 6 figures. Accepted to International Joint Conference on
Neural Networks (IJCNN 2022), July 18-23, Padua, Italy | 2022 International Joint Conference on Neural Networks (IJCNN),
Padua, Italy, 2022, pp. 1-10 | 10.1109/IJCNN55064.2022.9892771 | null | cs.LG cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Uncertainty quantification is one of the central challenges for machine
learning in real-world applications. In reinforcement learning, an agent
confronts two kinds of uncertainty, called epistemic uncertainty and aleatoric
uncertainty. Disentangling and evaluating these uncertainties simultaneously
stands a chance of improving the agent's final performance, accelerating
training, and facilitating quality assurance after deployment. In this work, we
propose an uncertainty-aware reinforcement learning algorithm for continuous
control tasks that extends the Deep Deterministic Policy Gradient algorithm
(DDPG). It exploits epistemic uncertainty to accelerate exploration and
aleatoric uncertainty to learn a risk-sensitive policy. We conduct numerical
experiments showing that our variant of DDPG outperforms vanilla DDPG without
uncertainty estimation in benchmark tasks on robotic control and power-grid
optimization.
| [
{
"version": "v1",
"created": "Wed, 27 Jul 2022 18:11:04 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Kanazawa",
"Takuya",
""
],
[
"Wang",
"Haiyan",
""
],
[
"Gupta",
"Chetan",
""
]
] | TITLE: Distributional Actor-Critic Ensemble for Uncertainty-Aware Continuous
Control
ABSTRACT: Uncertainty quantification is one of the central challenges for machine
learning in real-world applications. In reinforcement learning, an agent
confronts two kinds of uncertainty, called epistemic uncertainty and aleatoric
uncertainty. Disentangling and evaluating these uncertainties simultaneously
stands a chance of improving the agent's final performance, accelerating
training, and facilitating quality assurance after deployment. In this work, we
propose an uncertainty-aware reinforcement learning algorithm for continuous
control tasks that extends the Deep Deterministic Policy Gradient algorithm
(DDPG). It exploits epistemic uncertainty to accelerate exploration and
aleatoric uncertainty to learn a risk-sensitive policy. We conduct numerical
experiments showing that our variant of DDPG outperforms vanilla DDPG without
uncertainty estimation in benchmark tasks on robotic control and power-grid
optimization.
| no_new_dataset | 0.709422 |
2208.06339 | Casper Gyurik | Casper Gyurik, Vedran Dunjko | On establishing learning separations between classical and quantum
machine learning with classical data | this article is superseded by arXiv:2306.16028 | null | null | null | quant-ph cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite years of effort, the quantum machine learning community has only been
able to show quantum learning advantages for certain contrived
cryptography-inspired datasets in the case of classical data. In this note, we
discuss the challenges of finding learning problems that quantum learning
algorithms can learn much faster than any classical learning algorithm, and we
study how to identify such learning problems. Specifically, we reflect on the
main concepts in computational learning theory pertaining to this question, and
we discuss how subtle changes in definitions can mean conceptually
significantly different tasks, which can either lead to a separation or no
separation at all. Moreover, we study existing learning problems with a
provable quantum speedup to distill sets of more general and sufficient
conditions (i.e., ``checklists'') for a learning problem to exhibit a
separation between classical and quantum learners. These checklists are
intended to streamline one's approach to proving quantum speedups for learning
problems, or to elucidate bottlenecks. Finally, to illustrate its application,
we analyze examples of potential separations (i.e., when the learning problem
is build from computational separations, or when the data comes from a quantum
experiment) through the lens of our approach.
| [
{
"version": "v1",
"created": "Fri, 12 Aug 2022 16:00:30 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 12:05:50 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Gyurik",
"Casper",
""
],
[
"Dunjko",
"Vedran",
""
]
] | TITLE: On establishing learning separations between classical and quantum
machine learning with classical data
ABSTRACT: Despite years of effort, the quantum machine learning community has only been
able to show quantum learning advantages for certain contrived
cryptography-inspired datasets in the case of classical data. In this note, we
discuss the challenges of finding learning problems that quantum learning
algorithms can learn much faster than any classical learning algorithm, and we
study how to identify such learning problems. Specifically, we reflect on the
main concepts in computational learning theory pertaining to this question, and
we discuss how subtle changes in definitions can mean conceptually
significantly different tasks, which can either lead to a separation or no
separation at all. Moreover, we study existing learning problems with a
provable quantum speedup to distill sets of more general and sufficient
conditions (i.e., ``checklists'') for a learning problem to exhibit a
separation between classical and quantum learners. These checklists are
intended to streamline one's approach to proving quantum speedups for learning
problems, or to elucidate bottlenecks. Finally, to illustrate its application,
we analyze examples of potential separations (i.e., when the learning problem
is build from computational separations, or when the data comes from a quantum
experiment) through the lens of our approach.
| no_new_dataset | 0.708584 |
2208.13648 | Jingru Li | Jingru Li, Sheng Zhou, Liangcheng Li, Haishuai Wang, Zhi Yu, Jiajun Bu | Dynamic Data-Free Knowledge Distillation by Easy-to-Hard Learning
Strategy | Accepted by Information Sciences, Proof version provided | null | 10.1016/j.ins.2023.119202 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data-free knowledge distillation (DFKD) is a widely-used strategy for
Knowledge Distillation (KD) whose training data is not available. It trains a
lightweight student model with the aid of a large pretrained teacher model
without any access to training data. However, existing DFKD methods suffer from
inadequate and unstable training process, as they do not adjust the generation
target dynamically based on the status of the student model during learning. To
address this limitation, we propose a novel DFKD method called CuDFKD. It
teaches students by a dynamic strategy that gradually generates easy-to-hard
pseudo samples, mirroring how humans learn. Besides, CuDFKD adapts the
generation target dynamically according to the status of student model.
Moreover, We provide a theoretical analysis of the majorization minimization
(MM) algorithm and explain the convergence of CuDFKD. To measure the robustness
and fidelity of DFKD methods, we propose two more metrics, and experiments
shows CuDFKD has comparable performance to state-of-the-art (SOTA) DFKD methods
on all datasets. Experiments also present that our CuDFKD has the fastest
convergence and best robustness over other SOTA DFKD methods.
| [
{
"version": "v1",
"created": "Mon, 29 Aug 2022 14:51:57 GMT"
},
{
"version": "v2",
"created": "Mon, 16 Jan 2023 07:47:39 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Jul 2023 01:57:51 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Li",
"Jingru",
""
],
[
"Zhou",
"Sheng",
""
],
[
"Li",
"Liangcheng",
""
],
[
"Wang",
"Haishuai",
""
],
[
"Yu",
"Zhi",
""
],
[
"Bu",
"Jiajun",
""
]
] | TITLE: Dynamic Data-Free Knowledge Distillation by Easy-to-Hard Learning
Strategy
ABSTRACT: Data-free knowledge distillation (DFKD) is a widely-used strategy for
Knowledge Distillation (KD) whose training data is not available. It trains a
lightweight student model with the aid of a large pretrained teacher model
without any access to training data. However, existing DFKD methods suffer from
inadequate and unstable training process, as they do not adjust the generation
target dynamically based on the status of the student model during learning. To
address this limitation, we propose a novel DFKD method called CuDFKD. It
teaches students by a dynamic strategy that gradually generates easy-to-hard
pseudo samples, mirroring how humans learn. Besides, CuDFKD adapts the
generation target dynamically according to the status of student model.
Moreover, We provide a theoretical analysis of the majorization minimization
(MM) algorithm and explain the convergence of CuDFKD. To measure the robustness
and fidelity of DFKD methods, we propose two more metrics, and experiments
shows CuDFKD has comparable performance to state-of-the-art (SOTA) DFKD methods
on all datasets. Experiments also present that our CuDFKD has the fastest
convergence and best robustness over other SOTA DFKD methods.
| no_new_dataset | 0.709466 |
2209.01857 | Christoph Jansen PhD | Christoph Jansen (1), Malte Nalenz (1), Georg Schollmeyer (1), Thomas
Augustin (1) ((1) Ludwig-Maximilians-Universit\"at Munich) | Statistical Comparisons of Classifiers by Generalized Stochastic
Dominance | Accepted for publication in: Journal of Machine Learning Research
(JMLR) | null | null | null | stat.ML cs.LG stat.ME | http://creativecommons.org/licenses/by/4.0/ | Although being a crucial question for the development of machine learning
algorithms, there is still no consensus on how to compare classifiers over
multiple data sets with respect to several criteria. Every comparison framework
is confronted with (at least) three fundamental challenges: the multiplicity of
quality criteria, the multiplicity of data sets and the randomness of the
selection of data sets. In this paper, we add a fresh view to the vivid debate
by adopting recent developments in decision theory. Based on so-called
preference systems, our framework ranks classifiers by a generalized concept of
stochastic dominance, which powerfully circumvents the cumbersome, and often
even self-contradictory, reliance on aggregates. Moreover, we show that
generalized stochastic dominance can be operationalized by solving
easy-to-handle linear programs and moreover statistically tested employing an
adapted two-sample observation-randomization test. This yields indeed a
powerful framework for the statistical comparison of classifiers over multiple
data sets with respect to multiple quality criteria simultaneously. We
illustrate and investigate our framework in a simulation study and with a set
of standard benchmark data sets.
| [
{
"version": "v1",
"created": "Mon, 5 Sep 2022 09:28:15 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 13:56:24 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Jansen",
"Christoph",
"",
"Ludwig-Maximilians-Universität Munich"
],
[
"Nalenz",
"Malte",
"",
"Ludwig-Maximilians-Universität Munich"
],
[
"Schollmeyer",
"Georg",
"",
"Ludwig-Maximilians-Universität Munich"
],
[
"Augustin",
"Thomas",
"",
"Ludwig-Maximilians-Universität Munich"
]
] | TITLE: Statistical Comparisons of Classifiers by Generalized Stochastic
Dominance
ABSTRACT: Although being a crucial question for the development of machine learning
algorithms, there is still no consensus on how to compare classifiers over
multiple data sets with respect to several criteria. Every comparison framework
is confronted with (at least) three fundamental challenges: the multiplicity of
quality criteria, the multiplicity of data sets and the randomness of the
selection of data sets. In this paper, we add a fresh view to the vivid debate
by adopting recent developments in decision theory. Based on so-called
preference systems, our framework ranks classifiers by a generalized concept of
stochastic dominance, which powerfully circumvents the cumbersome, and often
even self-contradictory, reliance on aggregates. Moreover, we show that
generalized stochastic dominance can be operationalized by solving
easy-to-handle linear programs and moreover statistically tested employing an
adapted two-sample observation-randomization test. This yields indeed a
powerful framework for the statistical comparison of classifiers over multiple
data sets with respect to multiple quality criteria simultaneously. We
illustrate and investigate our framework in a simulation study and with a set
of standard benchmark data sets.
| no_new_dataset | 0.709057 |
2209.06606 | Gr\'egoire Petit | Gr\'egoire Petit, Adrian Popescu, Eden Belouadah, David Picard,
Bertrand Delezoide | PlaStIL: Plastic and Stable Memory-Free Class-Incremental Learning | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Plasticity and stability are needed in class-incremental learning in order to
learn from new data while preserving past knowledge. Due to catastrophic
forgetting, finding a compromise between these two properties is particularly
challenging when no memory buffer is available. Mainstream methods need to
store two deep models since they integrate new classes using fine-tuning with
knowledge distillation from the previous incremental state. We propose a method
which has similar number of parameters but distributes them differently in
order to find a better balance between plasticity and stability. Following an
approach already deployed by transfer-based incremental methods, we freeze the
feature extractor after the initial state. Classes in the oldest incremental
states are trained with this frozen extractor to ensure stability. Recent
classes are predicted using partially fine-tuned models in order to introduce
plasticity. Our proposed plasticity layer can be incorporated to any
transfer-based method designed for exemplar-free incremental learning, and we
apply it to two such methods. Evaluation is done with three large-scale
datasets. Results show that performance gains are obtained in all tested
configurations compared to existing methods.
| [
{
"version": "v1",
"created": "Wed, 14 Sep 2022 12:53:00 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 09:48:35 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Petit",
"Grégoire",
""
],
[
"Popescu",
"Adrian",
""
],
[
"Belouadah",
"Eden",
""
],
[
"Picard",
"David",
""
],
[
"Delezoide",
"Bertrand",
""
]
] | TITLE: PlaStIL: Plastic and Stable Memory-Free Class-Incremental Learning
ABSTRACT: Plasticity and stability are needed in class-incremental learning in order to
learn from new data while preserving past knowledge. Due to catastrophic
forgetting, finding a compromise between these two properties is particularly
challenging when no memory buffer is available. Mainstream methods need to
store two deep models since they integrate new classes using fine-tuning with
knowledge distillation from the previous incremental state. We propose a method
which has similar number of parameters but distributes them differently in
order to find a better balance between plasticity and stability. Following an
approach already deployed by transfer-based incremental methods, we freeze the
feature extractor after the initial state. Classes in the oldest incremental
states are trained with this frozen extractor to ensure stability. Recent
classes are predicted using partially fine-tuned models in order to introduce
plasticity. Our proposed plasticity layer can be incorporated to any
transfer-based method designed for exemplar-free incremental learning, and we
apply it to two such methods. Evaluation is done with three large-scale
datasets. Results show that performance gains are obtained in all tested
configurations compared to existing methods.
| no_new_dataset | 0.70978 |
2209.08418 | Takuya Kanazawa | Takuya Kanazawa and Chetan Gupta | Sample-based Uncertainty Quantification with a Single Deterministic
Neural Network | 16 pages, 17 figures, 2 tables. Accepted by the 14th International
Conference on Neural Computation Theory and Applications (NCTA 2022) held as
part of IJCCI 2022, October 24-26, 2022, Valletta, Malta | Proceedings of the 14th International Joint Conference on
Computational Intelligence - Volume 1: NCTA, 292-304, 2022 , Valletta, Malta | 10.5220/0011546800003332 | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Development of an accurate, flexible, and numerically efficient uncertainty
quantification (UQ) method is one of fundamental challenges in machine
learning. Previously, a UQ method called DISCO Nets has been proposed
(Bouchacourt et al., 2016), which trains a neural network by minimizing the
energy score. In this method, a random noise vector in
$\mathbb{R}^{10\text{--}100}$ is concatenated with the original input vector in
order to produce a diverse ensemble forecast despite using a single neural
network. While this method has shown promising performance on a hand pose
estimation task in computer vision, it remained unexplored whether this method
works as nicely for regression on tabular data, and how it competes with more
recent advanced UQ methods such as NGBoost. In this paper, we propose an
improved neural architecture of DISCO Nets that admits faster and more stable
training while only using a compact noise vector of dimension $\sim
\mathcal{O}(1)$. We benchmark this approach on miscellaneous real-world tabular
datasets and confirm that it is competitive with or even superior to standard
UQ baselines. Moreover we observe that it exhibits better point forecast
performance than a neural network of the same size trained with the
conventional mean squared error. As another advantage of the proposed method,
we show that local feature importance computation methods such as SHAP can be
easily applied to any subregion of the predictive distribution. A new
elementary proof for the validity of using the energy score to learn predictive
distributions is also provided.
| [
{
"version": "v1",
"created": "Sat, 17 Sep 2022 22:37:39 GMT"
},
{
"version": "v2",
"created": "Thu, 3 Nov 2022 23:01:50 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Kanazawa",
"Takuya",
""
],
[
"Gupta",
"Chetan",
""
]
] | TITLE: Sample-based Uncertainty Quantification with a Single Deterministic
Neural Network
ABSTRACT: Development of an accurate, flexible, and numerically efficient uncertainty
quantification (UQ) method is one of fundamental challenges in machine
learning. Previously, a UQ method called DISCO Nets has been proposed
(Bouchacourt et al., 2016), which trains a neural network by minimizing the
energy score. In this method, a random noise vector in
$\mathbb{R}^{10\text{--}100}$ is concatenated with the original input vector in
order to produce a diverse ensemble forecast despite using a single neural
network. While this method has shown promising performance on a hand pose
estimation task in computer vision, it remained unexplored whether this method
works as nicely for regression on tabular data, and how it competes with more
recent advanced UQ methods such as NGBoost. In this paper, we propose an
improved neural architecture of DISCO Nets that admits faster and more stable
training while only using a compact noise vector of dimension $\sim
\mathcal{O}(1)$. We benchmark this approach on miscellaneous real-world tabular
datasets and confirm that it is competitive with or even superior to standard
UQ baselines. Moreover we observe that it exhibits better point forecast
performance than a neural network of the same size trained with the
conventional mean squared error. As another advantage of the proposed method,
we show that local feature importance computation methods such as SHAP can be
easily applied to any subregion of the predictive distribution. A new
elementary proof for the validity of using the energy score to learn predictive
distributions is also provided.
| no_new_dataset | 0.707821 |
2209.08691 | Francesco Ragusa | Francesco Ragusa and Antonino Furnari and Giovanni Maria Farinella | MECCANO: A Multimodal Egocentric Dataset for Humans Behavior
Understanding in the Industrial-like Domain | arXiv admin note: text overlap with arXiv:2010.05654 | Computer Vision and Image Understanding 2023 | 10.1016/S1077-3142(23)00144-3 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Wearable cameras allow to acquire images and videos from the user's
perspective. These data can be processed to understand humans behavior. Despite
human behavior analysis has been thoroughly investigated in third person
vision, it is still understudied in egocentric settings and in particular in
industrial scenarios. To encourage research in this field, we present MECCANO,
a multimodal dataset of egocentric videos to study humans behavior
understanding in industrial-like settings. The multimodality is characterized
by the presence of gaze signals, depth maps and RGB videos acquired
simultaneously with a custom headset. The dataset has been explicitly labeled
for fundamental tasks in the context of human behavior understanding from a
first person view, such as recognizing and anticipating human-object
interactions. With the MECCANO dataset, we explored five different tasks
including 1) Action Recognition, 2) Active Objects Detection and Recognition,
3) Egocentric Human-Objects Interaction Detection, 4) Action Anticipation and
5) Next-Active Objects Detection. We propose a benchmark aimed to study human
behavior in the considered industrial-like scenario which demonstrates that the
investigated tasks and the considered scenario are challenging for
state-of-the-art algorithms. To support research in this field, we publicy
release the dataset at https://iplab.dmi.unict.it/MECCANO/.
| [
{
"version": "v1",
"created": "Mon, 19 Sep 2022 00:52:42 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Ragusa",
"Francesco",
""
],
[
"Furnari",
"Antonino",
""
],
[
"Farinella",
"Giovanni Maria",
""
]
] | TITLE: MECCANO: A Multimodal Egocentric Dataset for Humans Behavior
Understanding in the Industrial-like Domain
ABSTRACT: Wearable cameras allow to acquire images and videos from the user's
perspective. These data can be processed to understand humans behavior. Despite
human behavior analysis has been thoroughly investigated in third person
vision, it is still understudied in egocentric settings and in particular in
industrial scenarios. To encourage research in this field, we present MECCANO,
a multimodal dataset of egocentric videos to study humans behavior
understanding in industrial-like settings. The multimodality is characterized
by the presence of gaze signals, depth maps and RGB videos acquired
simultaneously with a custom headset. The dataset has been explicitly labeled
for fundamental tasks in the context of human behavior understanding from a
first person view, such as recognizing and anticipating human-object
interactions. With the MECCANO dataset, we explored five different tasks
including 1) Action Recognition, 2) Active Objects Detection and Recognition,
3) Egocentric Human-Objects Interaction Detection, 4) Action Anticipation and
5) Next-Active Objects Detection. We propose a benchmark aimed to study human
behavior in the considered industrial-like scenario which demonstrates that the
investigated tasks and the considered scenario are challenging for
state-of-the-art algorithms. To support research in this field, we publicy
release the dataset at https://iplab.dmi.unict.it/MECCANO/.
| new_dataset | 0.71484 |
2210.08376 | Navidreza Asadi | Navidreza Asadi, Maziar Goudarzi | Variant Parallelism: Lightweight Deep Convolutional Models for
Distributed Inference on IoT Devices | 8 pages, 6 figures, 7 tables | null | 10.1109/JIOT.2023.3285877 | null | cs.DC cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Two major techniques are commonly used to meet real-time inference
limitations when distributing models across resource-constrained IoT devices:
(1) model parallelism (MP) and (2) class parallelism (CP). In MP, transmitting
bulky intermediate data (orders of magnitude larger than input) between devices
imposes huge communication overhead. Although CP solves this problem, it has
limitations on the number of sub-models. In addition, both solutions are fault
intolerant, an issue when deployed on edge devices. We propose variant
parallelism (VP), an ensemble-based deep learning distribution method where
different variants of a main model are generated and can be deployed on
separate machines. We design a family of lighter models around the original
model, and train them simultaneously to improve accuracy over single models.
Our experimental results on six common mid-sized object recognition datasets
demonstrate that our models can have 5.8-7.1x fewer parameters, 4.3-31x fewer
multiply-accumulations (MACs), and 2.5-13.2x less response time on atomic
inputs compared to MobileNetV2 while achieving comparable or higher accuracy.
Our technique easily generates several variants of the base architecture. Each
variant returns only 2k outputs 1 <= k <= (#classes/2), representing Top-k
classes, instead of tons of floating point values required in MP. Since each
variant provides a full-class prediction, our approach maintains higher
availability compared with MP and CP in presence of failure.
| [
{
"version": "v1",
"created": "Sat, 15 Oct 2022 20:52:28 GMT"
},
{
"version": "v2",
"created": "Sun, 11 Jun 2023 21:10:25 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Asadi",
"Navidreza",
""
],
[
"Goudarzi",
"Maziar",
""
]
] | TITLE: Variant Parallelism: Lightweight Deep Convolutional Models for
Distributed Inference on IoT Devices
ABSTRACT: Two major techniques are commonly used to meet real-time inference
limitations when distributing models across resource-constrained IoT devices:
(1) model parallelism (MP) and (2) class parallelism (CP). In MP, transmitting
bulky intermediate data (orders of magnitude larger than input) between devices
imposes huge communication overhead. Although CP solves this problem, it has
limitations on the number of sub-models. In addition, both solutions are fault
intolerant, an issue when deployed on edge devices. We propose variant
parallelism (VP), an ensemble-based deep learning distribution method where
different variants of a main model are generated and can be deployed on
separate machines. We design a family of lighter models around the original
model, and train them simultaneously to improve accuracy over single models.
Our experimental results on six common mid-sized object recognition datasets
demonstrate that our models can have 5.8-7.1x fewer parameters, 4.3-31x fewer
multiply-accumulations (MACs), and 2.5-13.2x less response time on atomic
inputs compared to MobileNetV2 while achieving comparable or higher accuracy.
Our technique easily generates several variants of the base architecture. Each
variant returns only 2k outputs 1 <= k <= (#classes/2), representing Top-k
classes, instead of tons of floating point values required in MP. Since each
variant provides a full-class prediction, our approach maintains higher
availability compared with MP and CP in presence of failure.
| no_new_dataset | 0.71027 |
2210.09643 | Yidong Ouyang | Yidong Ouyang, Liyan Xie, Guang Cheng | Improving Adversarial Robustness by Contrastive Guided Diffusion Process | null | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Synthetic data generation has become an emerging tool to help improve the
adversarial robustness in classification tasks since robust learning requires a
significantly larger amount of training samples compared with standard
classification tasks. Among various deep generative models, the diffusion model
has been shown to produce high-quality synthetic images and has achieved good
performance in improving the adversarial robustness. However, diffusion-type
methods are typically slow in data generation as compared with other generative
models. Although different acceleration techniques have been proposed recently,
it is also of great importance to study how to improve the sample efficiency of
generated data for the downstream task. In this paper, we first analyze the
optimality condition of synthetic distribution for achieving non-trivial robust
accuracy. We show that enhancing the distinguishability among the generated
data is critical for improving adversarial robustness. Thus, we propose the
Contrastive-Guided Diffusion Process (Contrastive-DP), which adopts the
contrastive loss to guide the diffusion model in data generation. We verify our
theoretical results using simulations and demonstrate the good performance of
Contrastive-DP on image datasets.
| [
{
"version": "v1",
"created": "Tue, 18 Oct 2022 07:20:53 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 12:08:26 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Ouyang",
"Yidong",
""
],
[
"Xie",
"Liyan",
""
],
[
"Cheng",
"Guang",
""
]
] | TITLE: Improving Adversarial Robustness by Contrastive Guided Diffusion Process
ABSTRACT: Synthetic data generation has become an emerging tool to help improve the
adversarial robustness in classification tasks since robust learning requires a
significantly larger amount of training samples compared with standard
classification tasks. Among various deep generative models, the diffusion model
has been shown to produce high-quality synthetic images and has achieved good
performance in improving the adversarial robustness. However, diffusion-type
methods are typically slow in data generation as compared with other generative
models. Although different acceleration techniques have been proposed recently,
it is also of great importance to study how to improve the sample efficiency of
generated data for the downstream task. In this paper, we first analyze the
optimality condition of synthetic distribution for achieving non-trivial robust
accuracy. We show that enhancing the distinguishability among the generated
data is critical for improving adversarial robustness. Thus, we propose the
Contrastive-Guided Diffusion Process (Contrastive-DP), which adopts the
contrastive loss to guide the diffusion model in data generation. We verify our
theoretical results using simulations and demonstrate the good performance of
Contrastive-DP on image datasets.
| no_new_dataset | 0.712057 |
2210.14816 | Gerben Izaak Beintema | Gerben I. Beintema, Maarten Schoukens, Roland T\'oth | Deep Subspace Encoders for Nonlinear System Identification | Accepted in Automatica | null | null | null | eess.SY cs.LG cs.SY | http://creativecommons.org/licenses/by/4.0/ | Using Artificial Neural Networks (ANN) for nonlinear system identification
has proven to be a promising approach, but despite of all recent research
efforts, many practical and theoretical problems still remain open.
Specifically, noise handling and models, issues of consistency and reliable
estimation under minimisation of the prediction error are the most severe
problems. The latter comes with numerous practical challenges such as explosion
of the computational cost in terms of the number of data samples and the
occurrence of instabilities during optimization. In this paper, we aim to
overcome these issues by proposing a method which uses a truncated prediction
loss and a subspace encoder for state estimation. The truncated prediction loss
is computed by selecting multiple truncated subsections from the time series
and computing the average prediction loss. To obtain a computationally
efficient estimation method that minimizes the truncated prediction loss, a
subspace encoder represented by an artificial neural network is introduced.
This encoder aims to approximate the state reconstructability map of the
estimated model to provide an initial state for each truncated subsection given
past inputs and outputs. By theoretical analysis, we show that, under mild
conditions, the proposed method is locally consistent, increases optimization
stability, and achieves increased data efficiency by allowing for overlap
between the subsections. Lastly, we provide practical insights and user
guidelines employing a numerical example and state-of-the-art benchmark
results.
| [
{
"version": "v1",
"created": "Wed, 26 Oct 2022 16:04:38 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 13:52:23 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Beintema",
"Gerben I.",
""
],
[
"Schoukens",
"Maarten",
""
],
[
"Tóth",
"Roland",
""
]
] | TITLE: Deep Subspace Encoders for Nonlinear System Identification
ABSTRACT: Using Artificial Neural Networks (ANN) for nonlinear system identification
has proven to be a promising approach, but despite of all recent research
efforts, many practical and theoretical problems still remain open.
Specifically, noise handling and models, issues of consistency and reliable
estimation under minimisation of the prediction error are the most severe
problems. The latter comes with numerous practical challenges such as explosion
of the computational cost in terms of the number of data samples and the
occurrence of instabilities during optimization. In this paper, we aim to
overcome these issues by proposing a method which uses a truncated prediction
loss and a subspace encoder for state estimation. The truncated prediction loss
is computed by selecting multiple truncated subsections from the time series
and computing the average prediction loss. To obtain a computationally
efficient estimation method that minimizes the truncated prediction loss, a
subspace encoder represented by an artificial neural network is introduced.
This encoder aims to approximate the state reconstructability map of the
estimated model to provide an initial state for each truncated subsection given
past inputs and outputs. By theoretical analysis, we show that, under mild
conditions, the proposed method is locally consistent, increases optimization
stability, and achieves increased data efficiency by allowing for overlap
between the subsections. Lastly, we provide practical insights and user
guidelines employing a numerical example and state-of-the-art benchmark
results.
| no_new_dataset | 0.709019 |
2211.01407 | Ilia Sucholutsky | Ilia Sucholutsky and Ruairidh M. Battleday and Katherine M. Collins
and Raja Marjieh and Joshua C. Peterson and Pulkit Singh and Umang Bhatt and
Nori Jacoby and Adrian Weller and Thomas L. Griffiths | On the Informativeness of Supervision Signals | Proceedings of UAI 2023 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Supervised learning typically focuses on learning transferable
representations from training examples annotated by humans. While rich
annotations (like soft labels) carry more information than sparse annotations
(like hard labels), they are also more expensive to collect. For example, while
hard labels only provide information about the closest class an object belongs
to (e.g., "this is a dog"), soft labels provide information about the object's
relationship with multiple classes (e.g., "this is most likely a dog, but it
could also be a wolf or a coyote"). We use information theory to compare how a
number of commonly-used supervision signals contribute to
representation-learning performance, as well as how their capacity is affected
by factors such as the number of labels, classes, dimensions, and noise. Our
framework provides theoretical justification for using hard labels in the
big-data regime, but richer supervision signals for few-shot learning and
out-of-distribution generalization. We validate these results empirically in a
series of experiments with over 1 million crowdsourced image annotations and
conduct a cost-benefit analysis to establish a tradeoff curve that enables
users to optimize the cost of supervising representation learning on their own
datasets.
| [
{
"version": "v1",
"created": "Wed, 2 Nov 2022 18:02:31 GMT"
},
{
"version": "v2",
"created": "Thu, 22 Jun 2023 19:04:25 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Jul 2023 15:25:05 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Sucholutsky",
"Ilia",
""
],
[
"Battleday",
"Ruairidh M.",
""
],
[
"Collins",
"Katherine M.",
""
],
[
"Marjieh",
"Raja",
""
],
[
"Peterson",
"Joshua C.",
""
],
[
"Singh",
"Pulkit",
""
],
[
"Bhatt",
"Umang",
""
],
[
"Jacoby",
"Nori",
""
],
[
"Weller",
"Adrian",
""
],
[
"Griffiths",
"Thomas L.",
""
]
] | TITLE: On the Informativeness of Supervision Signals
ABSTRACT: Supervised learning typically focuses on learning transferable
representations from training examples annotated by humans. While rich
annotations (like soft labels) carry more information than sparse annotations
(like hard labels), they are also more expensive to collect. For example, while
hard labels only provide information about the closest class an object belongs
to (e.g., "this is a dog"), soft labels provide information about the object's
relationship with multiple classes (e.g., "this is most likely a dog, but it
could also be a wolf or a coyote"). We use information theory to compare how a
number of commonly-used supervision signals contribute to
representation-learning performance, as well as how their capacity is affected
by factors such as the number of labels, classes, dimensions, and noise. Our
framework provides theoretical justification for using hard labels in the
big-data regime, but richer supervision signals for few-shot learning and
out-of-distribution generalization. We validate these results empirically in a
series of experiments with over 1 million crowdsourced image annotations and
conduct a cost-benefit analysis to establish a tradeoff curve that enables
users to optimize the cost of supervising representation learning on their own
datasets.
| no_new_dataset | 0.709818 |
2211.07398 | Yuxiang Wei | Yuxiang Wei, Tengfei Xue, Yogesh Rathi, Nikos Makris, Fan Zhang,
Lauren J. O'Donnell | Age Prediction Performance Varies Across Deep, Superficial, and
Cerebellar White Matter Connections | 5 pages, 1 figure | null | null | null | q-bio.NC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The brain's white matter (WM) undergoes developmental and degenerative
processes during the human lifespan. To investigate the relationship between WM
anatomical regions and age, we study diffusion magnetic resonance imaging
tractography that is finely parcellated into fiber clusters in the deep,
superficial, and cerebellar WM. We propose a deep-learning-based age prediction
model that leverages large convolutional kernels and inverted bottlenecks. We
improve performance using novel discrete multi-faceted mix data augmentation
and a novel prior-knowledge-based loss function that encourages age predictions
in the expected range. We study a dataset of 965 healthy young adults (22-37
years) derived from the Human Connectome Project (HCP). Experimental results
demonstrate that the proposed model achieves a mean absolute error of 2.59
years and outperforms compared methods. We find that the deep WM is the most
informative for age prediction in this cohort, while the superficial WM is the
least informative. Overall, the most predictive WM tracts are the
thalamo-frontal tract from the deep WM and the intracerebellar input and
Purkinje tract from the cerebellar WM.
| [
{
"version": "v1",
"created": "Fri, 11 Nov 2022 15:23:09 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 13:08:31 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Wei",
"Yuxiang",
""
],
[
"Xue",
"Tengfei",
""
],
[
"Rathi",
"Yogesh",
""
],
[
"Makris",
"Nikos",
""
],
[
"Zhang",
"Fan",
""
],
[
"O'Donnell",
"Lauren J.",
""
]
] | TITLE: Age Prediction Performance Varies Across Deep, Superficial, and
Cerebellar White Matter Connections
ABSTRACT: The brain's white matter (WM) undergoes developmental and degenerative
processes during the human lifespan. To investigate the relationship between WM
anatomical regions and age, we study diffusion magnetic resonance imaging
tractography that is finely parcellated into fiber clusters in the deep,
superficial, and cerebellar WM. We propose a deep-learning-based age prediction
model that leverages large convolutional kernels and inverted bottlenecks. We
improve performance using novel discrete multi-faceted mix data augmentation
and a novel prior-knowledge-based loss function that encourages age predictions
in the expected range. We study a dataset of 965 healthy young adults (22-37
years) derived from the Human Connectome Project (HCP). Experimental results
demonstrate that the proposed model achieves a mean absolute error of 2.59
years and outperforms compared methods. We find that the deep WM is the most
informative for age prediction in this cohort, while the superficial WM is the
least informative. Overall, the most predictive WM tracts are the
thalamo-frontal tract from the deep WM and the intracerebellar input and
Purkinje tract from the cerebellar WM.
| no_new_dataset | 0.710791 |
2211.17226 | Keith Mills | Keith G. Mills, Fred X. Han, Jialin Zhang, Fabian Chudak, Ali Safari
Mamaghani, Mohammad Salameh, Wei Lu, Shangling Jui, Di Niu | GENNAPE: Towards Generalized Neural Architecture Performance Estimators | AAAI 2023 Oral Presentation; includes supplementary materials with
more details on introduced benchmarks; 14 Pages, 6 Figures, 10 Tables | null | 10.1609/aaai.v37i8.26102 | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Predicting neural architecture performance is a challenging task and is
crucial to neural architecture design and search. Existing approaches either
rely on neural performance predictors which are limited to modeling
architectures in a predefined design space involving specific sets of operators
and connection rules, and cannot generalize to unseen architectures, or resort
to zero-cost proxies which are not always accurate. In this paper, we propose
GENNAPE, a Generalized Neural Architecture Performance Estimator, which is
pretrained on open neural architecture benchmarks, and aims to generalize to
completely unseen architectures through combined innovations in network
representation, contrastive pretraining, and fuzzy clustering-based predictor
ensemble. Specifically, GENNAPE represents a given neural network as a
Computation Graph (CG) of atomic operations which can model an arbitrary
architecture. It first learns a graph encoder via Contrastive Learning to
encourage network separation by topological features, and then trains multiple
predictor heads, which are soft-aggregated according to the fuzzy membership of
a neural network. Experiments show that GENNAPE pretrained on NAS-Bench-101 can
achieve superior transferability to 5 different public neural network
benchmarks, including NAS-Bench-201, NAS-Bench-301, MobileNet and ResNet
families under no or minimum fine-tuning. We further introduce 3 challenging
newly labelled neural network benchmarks: HiAML, Inception and Two-Path, which
can concentrate in narrow accuracy ranges. Extensive experiments show that
GENNAPE can correctly discern high-performance architectures in these families.
Finally, when paired with a search algorithm, GENNAPE can find architectures
that improve accuracy while reducing FLOPs on three families.
| [
{
"version": "v1",
"created": "Wed, 30 Nov 2022 18:27:41 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 20:01:14 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Mills",
"Keith G.",
""
],
[
"Han",
"Fred X.",
""
],
[
"Zhang",
"Jialin",
""
],
[
"Chudak",
"Fabian",
""
],
[
"Mamaghani",
"Ali Safari",
""
],
[
"Salameh",
"Mohammad",
""
],
[
"Lu",
"Wei",
""
],
[
"Jui",
"Shangling",
""
],
[
"Niu",
"Di",
""
]
] | TITLE: GENNAPE: Towards Generalized Neural Architecture Performance Estimators
ABSTRACT: Predicting neural architecture performance is a challenging task and is
crucial to neural architecture design and search. Existing approaches either
rely on neural performance predictors which are limited to modeling
architectures in a predefined design space involving specific sets of operators
and connection rules, and cannot generalize to unseen architectures, or resort
to zero-cost proxies which are not always accurate. In this paper, we propose
GENNAPE, a Generalized Neural Architecture Performance Estimator, which is
pretrained on open neural architecture benchmarks, and aims to generalize to
completely unseen architectures through combined innovations in network
representation, contrastive pretraining, and fuzzy clustering-based predictor
ensemble. Specifically, GENNAPE represents a given neural network as a
Computation Graph (CG) of atomic operations which can model an arbitrary
architecture. It first learns a graph encoder via Contrastive Learning to
encourage network separation by topological features, and then trains multiple
predictor heads, which are soft-aggregated according to the fuzzy membership of
a neural network. Experiments show that GENNAPE pretrained on NAS-Bench-101 can
achieve superior transferability to 5 different public neural network
benchmarks, including NAS-Bench-201, NAS-Bench-301, MobileNet and ResNet
families under no or minimum fine-tuning. We further introduce 3 challenging
newly labelled neural network benchmarks: HiAML, Inception and Two-Path, which
can concentrate in narrow accuracy ranges. Extensive experiments show that
GENNAPE can correctly discern high-performance architectures in these families.
Finally, when paired with a search algorithm, GENNAPE can find architectures
that improve accuracy while reducing FLOPs on three families.
| no_new_dataset | 0.707354 |
2211.17228 | Keith Mills | Keith G. Mills, Di Niu, Mohammad Salameh, Weichen Qiu, Fred X. Han,
Puyuan Liu, Jialin Zhang, Wei Lu, Shangling Jui | AIO-P: Expanding Neural Performance Predictors Beyond Image
Classification | AAAI 2023 Oral Presentation; version includes supplementary material;
16 Pages, 4 Figures, 22 Tables | null | 10.1609/aaai.v37i8.26101 | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Evaluating neural network performance is critical to deep neural network
design but a costly procedure. Neural predictors provide an efficient solution
by treating architectures as samples and learning to estimate their performance
on a given task. However, existing predictors are task-dependent, predominantly
estimating neural network performance on image classification benchmarks. They
are also search-space dependent; each predictor is designed to make predictions
for a specific architecture search space with predefined topologies and set of
operations. In this paper, we propose a novel All-in-One Predictor (AIO-P),
which aims to pretrain neural predictors on architecture examples from
multiple, separate computer vision (CV) task domains and multiple architecture
spaces, and then transfer to unseen downstream CV tasks or neural
architectures. We describe our proposed techniques for general graph
representation, efficient predictor pretraining and knowledge infusion
techniques, as well as methods to transfer to downstream tasks/spaces.
Extensive experimental results show that AIO-P can achieve Mean Absolute Error
(MAE) and Spearman's Rank Correlation (SRCC) below 1% and above 0.5,
respectively, on a breadth of target downstream CV tasks with or without
fine-tuning, outperforming a number of baselines. Moreover, AIO-P can directly
transfer to new architectures not seen during training, accurately rank them
and serve as an effective performance estimator when paired with an algorithm
designed to preserve performance while reducing FLOPs.
| [
{
"version": "v1",
"created": "Wed, 30 Nov 2022 18:30:41 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Apr 2023 20:07:09 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Mills",
"Keith G.",
""
],
[
"Niu",
"Di",
""
],
[
"Salameh",
"Mohammad",
""
],
[
"Qiu",
"Weichen",
""
],
[
"Han",
"Fred X.",
""
],
[
"Liu",
"Puyuan",
""
],
[
"Zhang",
"Jialin",
""
],
[
"Lu",
"Wei",
""
],
[
"Jui",
"Shangling",
""
]
] | TITLE: AIO-P: Expanding Neural Performance Predictors Beyond Image
Classification
ABSTRACT: Evaluating neural network performance is critical to deep neural network
design but a costly procedure. Neural predictors provide an efficient solution
by treating architectures as samples and learning to estimate their performance
on a given task. However, existing predictors are task-dependent, predominantly
estimating neural network performance on image classification benchmarks. They
are also search-space dependent; each predictor is designed to make predictions
for a specific architecture search space with predefined topologies and set of
operations. In this paper, we propose a novel All-in-One Predictor (AIO-P),
which aims to pretrain neural predictors on architecture examples from
multiple, separate computer vision (CV) task domains and multiple architecture
spaces, and then transfer to unseen downstream CV tasks or neural
architectures. We describe our proposed techniques for general graph
representation, efficient predictor pretraining and knowledge infusion
techniques, as well as methods to transfer to downstream tasks/spaces.
Extensive experimental results show that AIO-P can achieve Mean Absolute Error
(MAE) and Spearman's Rank Correlation (SRCC) below 1% and above 0.5,
respectively, on a breadth of target downstream CV tasks with or without
fine-tuning, outperforming a number of baselines. Moreover, AIO-P can directly
transfer to new architectures not seen during training, accurately rank them
and serve as an effective performance estimator when paired with an algorithm
designed to preserve performance while reducing FLOPs.
| no_new_dataset | 0.708799 |
2212.02090 | Sangwoo Mo | Junhyun Nam, Sangwoo Mo, Jaeho Lee, Jinwoo Shin | Breaking the Spurious Causality of Conditional Generation via Fairness
Intervention with Corrective Sampling | TMLR 2023 | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To capture the relationship between samples and labels, conditional
generative models often inherit spurious correlations from the training
dataset. This can result in label-conditional distributions that are imbalanced
with respect to another latent attribute. To mitigate this issue, which we call
spurious causality of conditional generation, we propose a general two-step
strategy. (a) Fairness Intervention (FI): emphasize the minority samples that
are hard to generate due to the spurious correlation in the training dataset.
(b) Corrective Sampling (CS): explicitly filter the generated samples and
ensure that they follow the desired latent attribute distribution. We have
designed the fairness intervention to work for various degrees of supervision
on the spurious attribute, including unsupervised, weakly-supervised, and
semi-supervised scenarios. Our experimental results demonstrate that FICS can
effectively resolve spurious causality of conditional generation across various
datasets.
| [
{
"version": "v1",
"created": "Mon, 5 Dec 2022 08:09:33 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 09:17:49 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Nam",
"Junhyun",
""
],
[
"Mo",
"Sangwoo",
""
],
[
"Lee",
"Jaeho",
""
],
[
"Shin",
"Jinwoo",
""
]
] | TITLE: Breaking the Spurious Causality of Conditional Generation via Fairness
Intervention with Corrective Sampling
ABSTRACT: To capture the relationship between samples and labels, conditional
generative models often inherit spurious correlations from the training
dataset. This can result in label-conditional distributions that are imbalanced
with respect to another latent attribute. To mitigate this issue, which we call
spurious causality of conditional generation, we propose a general two-step
strategy. (a) Fairness Intervention (FI): emphasize the minority samples that
are hard to generate due to the spurious correlation in the training dataset.
(b) Corrective Sampling (CS): explicitly filter the generated samples and
ensure that they follow the desired latent attribute distribution. We have
designed the fairness intervention to work for various degrees of supervision
on the spurious attribute, including unsupervised, weakly-supervised, and
semi-supervised scenarios. Our experimental results demonstrate that FICS can
effectively resolve spurious causality of conditional generation across various
datasets.
| no_new_dataset | 0.712851 |
2212.05136 | Kevin Joo | Hyekang Kevin Joo, Khoa Vo, Kashu Yamazaki, Ngan Le | CLIP-TSA: CLIP-Assisted Temporal Self-Attention for Weakly-Supervised
Video Anomaly Detection | Published at the 30th IEEE International Conference on Image
Processing (IEEE ICIP 2023) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video anomaly detection (VAD) -- commonly formulated as a multiple-instance
learning problem in a weakly-supervised manner due to its labor-intensive
nature -- is a challenging problem in video surveillance where the frames of
anomaly need to be localized in an untrimmed video. In this paper, we first
propose to utilize the ViT-encoded visual features from CLIP, in contrast with
the conventional C3D or I3D features in the domain, to efficiently extract
discriminative representations in the novel technique. We then model temporal
dependencies and nominate the snippets of interest by leveraging our proposed
Temporal Self-Attention (TSA). The ablation study confirms the effectiveness of
TSA and ViT feature. The extensive experiments show that our proposed CLIP-TSA
outperforms the existing state-of-the-art (SOTA) methods by a large margin on
three commonly-used benchmark datasets in the VAD problem (UCF-Crime,
ShanghaiTech Campus, and XD-Violence). Our source code is available at
https://github.com/joos2010kj/CLIP-TSA.
| [
{
"version": "v1",
"created": "Fri, 9 Dec 2022 22:28:24 GMT"
},
{
"version": "v2",
"created": "Fri, 5 May 2023 19:50:05 GMT"
},
{
"version": "v3",
"created": "Mon, 3 Jul 2023 23:03:22 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Joo",
"Hyekang Kevin",
""
],
[
"Vo",
"Khoa",
""
],
[
"Yamazaki",
"Kashu",
""
],
[
"Le",
"Ngan",
""
]
] | TITLE: CLIP-TSA: CLIP-Assisted Temporal Self-Attention for Weakly-Supervised
Video Anomaly Detection
ABSTRACT: Video anomaly detection (VAD) -- commonly formulated as a multiple-instance
learning problem in a weakly-supervised manner due to its labor-intensive
nature -- is a challenging problem in video surveillance where the frames of
anomaly need to be localized in an untrimmed video. In this paper, we first
propose to utilize the ViT-encoded visual features from CLIP, in contrast with
the conventional C3D or I3D features in the domain, to efficiently extract
discriminative representations in the novel technique. We then model temporal
dependencies and nominate the snippets of interest by leveraging our proposed
Temporal Self-Attention (TSA). The ablation study confirms the effectiveness of
TSA and ViT feature. The extensive experiments show that our proposed CLIP-TSA
outperforms the existing state-of-the-art (SOTA) methods by a large margin on
three commonly-used benchmark datasets in the VAD problem (UCF-Crime,
ShanghaiTech Campus, and XD-Violence). Our source code is available at
https://github.com/joos2010kj/CLIP-TSA.
| no_new_dataset | 0.710597 |
2212.05961 | Zhengqing Yuan | Zhengqing Yuan, Xiaolong Zhang, Yue Wang, Xuecong Hou, Huiwen Xue,
Zhuanzhe Zhao and Yongming Liu | RPN: A Word Vector Level Data Augmentation Algorithm in Deep Learning
for Language Understanding | Accepted by 2023 IEEE International Conference on Systems, Man, and
Cybernetics (IEEE SMC) | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data augmentation is a widely used technique in machine learning to improve
model performance. However, existing data augmentation techniques in natural
language understanding (NLU) may not fully capture the complexity of natural
language variations, and they can be challenging to apply to large datasets.
This paper proposes the Random Position Noise (RPN) algorithm, a novel data
augmentation technique that operates at the word vector level. RPN modifies the
word embeddings of the original text by introducing noise based on the existing
values of selected word vectors, allowing for more fine-grained modifications
and better capturing natural language variations. Unlike traditional data
augmentation methods, RPN does not require gradients in the computational graph
during virtual sample updates, making it simpler to apply to large datasets.
Experimental results demonstrate that RPN consistently outperforms existing
data augmentation techniques across various NLU tasks, including sentiment
analysis, natural language inference, and paraphrase detection. Moreover, RPN
performs well in low-resource settings and is applicable to any model featuring
a word embeddings layer. The proposed RPN algorithm is a promising approach for
enhancing NLU performance and addressing the challenges associated with
traditional data augmentation techniques in large-scale NLU tasks. Our
experimental results demonstrated that the RPN algorithm achieved
state-of-the-art performance in all seven NLU tasks, thereby highlighting its
effectiveness and potential for real-world NLU applications.
| [
{
"version": "v1",
"created": "Mon, 12 Dec 2022 15:16:12 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Apr 2023 06:49:53 GMT"
},
{
"version": "v3",
"created": "Wed, 5 Jul 2023 01:18:04 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Yuan",
"Zhengqing",
""
],
[
"Zhang",
"Xiaolong",
""
],
[
"Wang",
"Yue",
""
],
[
"Hou",
"Xuecong",
""
],
[
"Xue",
"Huiwen",
""
],
[
"Zhao",
"Zhuanzhe",
""
],
[
"Liu",
"Yongming",
""
]
] | TITLE: RPN: A Word Vector Level Data Augmentation Algorithm in Deep Learning
for Language Understanding
ABSTRACT: Data augmentation is a widely used technique in machine learning to improve
model performance. However, existing data augmentation techniques in natural
language understanding (NLU) may not fully capture the complexity of natural
language variations, and they can be challenging to apply to large datasets.
This paper proposes the Random Position Noise (RPN) algorithm, a novel data
augmentation technique that operates at the word vector level. RPN modifies the
word embeddings of the original text by introducing noise based on the existing
values of selected word vectors, allowing for more fine-grained modifications
and better capturing natural language variations. Unlike traditional data
augmentation methods, RPN does not require gradients in the computational graph
during virtual sample updates, making it simpler to apply to large datasets.
Experimental results demonstrate that RPN consistently outperforms existing
data augmentation techniques across various NLU tasks, including sentiment
analysis, natural language inference, and paraphrase detection. Moreover, RPN
performs well in low-resource settings and is applicable to any model featuring
a word embeddings layer. The proposed RPN algorithm is a promising approach for
enhancing NLU performance and addressing the challenges associated with
traditional data augmentation techniques in large-scale NLU tasks. Our
experimental results demonstrated that the RPN algorithm achieved
state-of-the-art performance in all seven NLU tasks, thereby highlighting its
effectiveness and potential for real-world NLU applications.
| no_new_dataset | 0.710189 |
2212.08846 | Junyan Cao | Junyan Cao, Yan Hong, Li Niu | Painterly Image Harmonization in Dual Domains | Accepted by AAAI2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image harmonization aims to produce visually harmonious composite images by
adjusting the foreground appearance to be compatible with the background. When
the composite image has photographic foreground and painterly background, the
task is called painterly image harmonization. There are only few works on this
task, which are either time-consuming or weak in generating well-harmonized
results. In this work, we propose a novel painterly harmonization network
consisting of a dual-domain generator and a dual-domain discriminator, which
harmonizes the composite image in both spatial domain and frequency domain. The
dual-domain generator performs harmonization by using AdaIN modules in the
spatial domain and our proposed ResFFT modules in the frequency domain. The
dual-domain discriminator attempts to distinguish the inharmonious patches
based on the spatial feature and frequency feature of each patch, which can
enhance the ability of generator in an adversarial manner. Extensive
experiments on the benchmark dataset show the effectiveness of our method. Our
code and model are available at
https://github.com/bcmi/PHDNet-Painterly-Image-Harmonization.
| [
{
"version": "v1",
"created": "Sat, 17 Dec 2022 11:00:34 GMT"
},
{
"version": "v2",
"created": "Tue, 20 Dec 2022 07:51:12 GMT"
},
{
"version": "v3",
"created": "Tue, 28 Feb 2023 02:30:31 GMT"
},
{
"version": "v4",
"created": "Tue, 4 Jul 2023 07:31:18 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Cao",
"Junyan",
""
],
[
"Hong",
"Yan",
""
],
[
"Niu",
"Li",
""
]
] | TITLE: Painterly Image Harmonization in Dual Domains
ABSTRACT: Image harmonization aims to produce visually harmonious composite images by
adjusting the foreground appearance to be compatible with the background. When
the composite image has photographic foreground and painterly background, the
task is called painterly image harmonization. There are only few works on this
task, which are either time-consuming or weak in generating well-harmonized
results. In this work, we propose a novel painterly harmonization network
consisting of a dual-domain generator and a dual-domain discriminator, which
harmonizes the composite image in both spatial domain and frequency domain. The
dual-domain generator performs harmonization by using AdaIN modules in the
spatial domain and our proposed ResFFT modules in the frequency domain. The
dual-domain discriminator attempts to distinguish the inharmonious patches
based on the spatial feature and frequency feature of each patch, which can
enhance the ability of generator in an adversarial manner. Extensive
experiments on the benchmark dataset show the effectiveness of our method. Our
code and model are available at
https://github.com/bcmi/PHDNet-Painterly-Image-Harmonization.
| no_new_dataset | 0.712595 |
2212.10549 | Rohan Pandey | Rohan Pandey, Rulin Shao, Paul Pu Liang, Ruslan Salakhutdinov,
Louis-Philippe Morency | Cross-modal Attention Congruence Regularization for Vision-Language
Relation Alignment | ACL 2023 | null | null | null | cs.CL cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite recent progress towards scaling up multimodal vision-language models,
these models are still known to struggle on compositional generalization
benchmarks such as Winoground. We find that a critical component lacking from
current vision-language models is relation-level alignment: the ability to
match directional semantic relations in text (e.g., "mug in grass") with
spatial relationships in the image (e.g., the position of the mug relative to
the grass). To tackle this problem, we show that relation alignment can be
enforced by encouraging the directed language attention from 'mug' to 'grass'
(capturing the semantic relation 'in') to match the directed visual attention
from the mug to the grass. Tokens and their corresponding objects are softly
identified using the cross-modal attention. We prove that this notion of soft
relation alignment is equivalent to enforcing congruence between vision and
language attention matrices under a 'change of basis' provided by the
cross-modal attention matrix. Intuitively, our approach projects visual
attention into the language attention space to calculate its divergence from
the actual language attention, and vice versa. We apply our Cross-modal
Attention Congruence Regularization (CACR) loss to UNITER and improve on the
state-of-the-art approach to Winoground.
| [
{
"version": "v1",
"created": "Tue, 20 Dec 2022 18:53:14 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 13:18:29 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Pandey",
"Rohan",
""
],
[
"Shao",
"Rulin",
""
],
[
"Liang",
"Paul Pu",
""
],
[
"Salakhutdinov",
"Ruslan",
""
],
[
"Morency",
"Louis-Philippe",
""
]
] | TITLE: Cross-modal Attention Congruence Regularization for Vision-Language
Relation Alignment
ABSTRACT: Despite recent progress towards scaling up multimodal vision-language models,
these models are still known to struggle on compositional generalization
benchmarks such as Winoground. We find that a critical component lacking from
current vision-language models is relation-level alignment: the ability to
match directional semantic relations in text (e.g., "mug in grass") with
spatial relationships in the image (e.g., the position of the mug relative to
the grass). To tackle this problem, we show that relation alignment can be
enforced by encouraging the directed language attention from 'mug' to 'grass'
(capturing the semantic relation 'in') to match the directed visual attention
from the mug to the grass. Tokens and their corresponding objects are softly
identified using the cross-modal attention. We prove that this notion of soft
relation alignment is equivalent to enforcing congruence between vision and
language attention matrices under a 'change of basis' provided by the
cross-modal attention matrix. Intuitively, our approach projects visual
attention into the language attention space to calculate its divergence from
the actual language attention, and vice versa. We apply our Cross-modal
Attention Congruence Regularization (CACR) loss to UNITER and improve on the
state-of-the-art approach to Winoground.
| no_new_dataset | 0.710459 |
2301.00363 | Leikun Yin | Leikun Yin, Rahul Ghosh, Chenxi Lin, David Hale, Christoph Weigl,
James Obarowski, Junxiong Zhou, Jessica Till, Xiaowei Jia, Troy Mao, Vipin
Kumar, Zhenong Jin | Mapping smallholder cashew plantations to inform sustainable tree crop
expansion in Benin | null | Remote Sensing of Environment, 295, p.113695 (2023) | 10.1016/j.rse.2023.113695 | null | cs.CV cs.LG stat.AP | http://creativecommons.org/licenses/by/4.0/ | Cashews are grown by over 3 million smallholders in more than 40 countries
worldwide as a principal source of income. As the third largest cashew producer
in Africa, Benin has nearly 200,000 smallholder cashew growers contributing 15%
of the country's national export earnings. However, a lack of information on
where and how cashew trees grow across the country hinders decision-making that
could support increased cashew production and poverty alleviation. By
leveraging 2.4-m Planet Basemaps and 0.5-m aerial imagery, newly developed deep
learning algorithms, and large-scale ground truth datasets, we successfully
produced the first national map of cashew in Benin and characterized the
expansion of cashew plantations between 2015 and 2021. In particular, we
developed a SpatioTemporal Classification with Attention (STCA) model to map
the distribution of cashew plantations, which can fully capture texture
information from discriminative time steps during a growing season. We further
developed a Clustering Augmented Self-supervised Temporal Classification
(CASTC) model to distinguish high-density versus low-density cashew plantations
by automatic feature extraction and optimized clustering. Results show that the
STCA model has an overall accuracy over 85% and the CASTC model achieved an
overall accuracy of 76%. We found that the cashew area in Benin almost doubled
from 2015 to 2021 with 60% of new plantation development coming from cropland
or fallow land, while encroachment of cashew plantations into protected areas
has increased by 55%. Only half of cashew plantations were high-density in
2021, suggesting high potential for intensification. Our study illustrates the
power of combining high-resolution remote sensing imagery and state-of-the-art
deep learning algorithms to better understand tree crops in the heterogeneous
smallholder landscape.
| [
{
"version": "v1",
"created": "Sun, 1 Jan 2023 07:18:47 GMT"
},
{
"version": "v2",
"created": "Sun, 15 Jan 2023 18:04:42 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Yin",
"Leikun",
""
],
[
"Ghosh",
"Rahul",
""
],
[
"Lin",
"Chenxi",
""
],
[
"Hale",
"David",
""
],
[
"Weigl",
"Christoph",
""
],
[
"Obarowski",
"James",
""
],
[
"Zhou",
"Junxiong",
""
],
[
"Till",
"Jessica",
""
],
[
"Jia",
"Xiaowei",
""
],
[
"Mao",
"Troy",
""
],
[
"Kumar",
"Vipin",
""
],
[
"Jin",
"Zhenong",
""
]
] | TITLE: Mapping smallholder cashew plantations to inform sustainable tree crop
expansion in Benin
ABSTRACT: Cashews are grown by over 3 million smallholders in more than 40 countries
worldwide as a principal source of income. As the third largest cashew producer
in Africa, Benin has nearly 200,000 smallholder cashew growers contributing 15%
of the country's national export earnings. However, a lack of information on
where and how cashew trees grow across the country hinders decision-making that
could support increased cashew production and poverty alleviation. By
leveraging 2.4-m Planet Basemaps and 0.5-m aerial imagery, newly developed deep
learning algorithms, and large-scale ground truth datasets, we successfully
produced the first national map of cashew in Benin and characterized the
expansion of cashew plantations between 2015 and 2021. In particular, we
developed a SpatioTemporal Classification with Attention (STCA) model to map
the distribution of cashew plantations, which can fully capture texture
information from discriminative time steps during a growing season. We further
developed a Clustering Augmented Self-supervised Temporal Classification
(CASTC) model to distinguish high-density versus low-density cashew plantations
by automatic feature extraction and optimized clustering. Results show that the
STCA model has an overall accuracy over 85% and the CASTC model achieved an
overall accuracy of 76%. We found that the cashew area in Benin almost doubled
from 2015 to 2021 with 60% of new plantation development coming from cropland
or fallow land, while encroachment of cashew plantations into protected areas
has increased by 55%. Only half of cashew plantations were high-density in
2021, suggesting high potential for intensification. Our study illustrates the
power of combining high-resolution remote sensing imagery and state-of-the-art
deep learning algorithms to better understand tree crops in the heterogeneous
smallholder landscape.
| no_new_dataset | 0.70878 |
2301.12005 | Seungyeon Kim | Seungyeon Kim, Ankit Singh Rawat, Manzil Zaheer, Sadeep Jayasumana,
Veeranjaneyulu Sadhanala, Wittawat Jitkrittum, Aditya Krishna Menon, Rob
Fergus, Sanjiv Kumar | EmbedDistill: A Geometric Knowledge Distillation for Information
Retrieval | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large neural models (such as Transformers) achieve state-of-the-art
performance for information retrieval (IR). In this paper, we aim to improve
distillation methods that pave the way for the resource-efficient deployment of
such models in practice. Inspired by our theoretical analysis of the
teacher-student generalization gap for IR models, we propose a novel
distillation approach that leverages the relative geometry among queries and
documents learned by the large teacher model. Unlike existing teacher
score-based distillation methods, our proposed approach employs embedding
matching tasks to provide a stronger signal to align the representations of the
teacher and student models. In addition, it utilizes query generation to
explore the data manifold to reduce the discrepancies between the student and
the teacher where training data is sparse. Furthermore, our analysis also
motivates novel asymmetric architectures for student models which realizes
better embedding alignment without increasing online inference cost. On
standard benchmarks like MSMARCO, we show that our approach successfully
distills from both dual-encoder (DE) and cross-encoder (CE) teacher models to
1/10th size asymmetric students that can retain 95-97% of the teacher
performance.
| [
{
"version": "v1",
"created": "Fri, 27 Jan 2023 22:04:37 GMT"
},
{
"version": "v2",
"created": "Mon, 3 Jul 2023 20:02:28 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Kim",
"Seungyeon",
""
],
[
"Rawat",
"Ankit Singh",
""
],
[
"Zaheer",
"Manzil",
""
],
[
"Jayasumana",
"Sadeep",
""
],
[
"Sadhanala",
"Veeranjaneyulu",
""
],
[
"Jitkrittum",
"Wittawat",
""
],
[
"Menon",
"Aditya Krishna",
""
],
[
"Fergus",
"Rob",
""
],
[
"Kumar",
"Sanjiv",
""
]
] | TITLE: EmbedDistill: A Geometric Knowledge Distillation for Information
Retrieval
ABSTRACT: Large neural models (such as Transformers) achieve state-of-the-art
performance for information retrieval (IR). In this paper, we aim to improve
distillation methods that pave the way for the resource-efficient deployment of
such models in practice. Inspired by our theoretical analysis of the
teacher-student generalization gap for IR models, we propose a novel
distillation approach that leverages the relative geometry among queries and
documents learned by the large teacher model. Unlike existing teacher
score-based distillation methods, our proposed approach employs embedding
matching tasks to provide a stronger signal to align the representations of the
teacher and student models. In addition, it utilizes query generation to
explore the data manifold to reduce the discrepancies between the student and
the teacher where training data is sparse. Furthermore, our analysis also
motivates novel asymmetric architectures for student models which realizes
better embedding alignment without increasing online inference cost. On
standard benchmarks like MSMARCO, we show that our approach successfully
distills from both dual-encoder (DE) and cross-encoder (CE) teacher models to
1/10th size asymmetric students that can retain 95-97% of the teacher
performance.
| no_new_dataset | 0.709856 |
2301.12040 | Minghao Xu | Minghao Xu, Xinyu Yuan, Santiago Miret, Jian Tang | ProtST: Multi-Modality Learning of Protein Sequences and Biomedical
Texts | Accpeted by ICML 2023 (Oral), code and data released | null | null | null | q-bio.BM cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current protein language models (PLMs) learn protein representations mainly
based on their sequences, thereby well capturing co-evolutionary information,
but they are unable to explicitly acquire protein functions, which is the end
goal of protein representation learning. Fortunately, for many proteins, their
textual property descriptions are available, where their various functions are
also described. Motivated by this fact, we first build the ProtDescribe dataset
to augment protein sequences with text descriptions of their functions and
other important properties. Based on this dataset, we propose the ProtST
framework to enhance Protein Sequence pre-training and understanding by
biomedical Texts. During pre-training, we design three types of tasks, i.e.,
unimodal mask prediction, multimodal representation alignment and multimodal
mask prediction, to enhance a PLM with protein property information with
different granularities and, at the same time, preserve the PLM's original
representation power. On downstream tasks, ProtST enables both supervised
learning and zero-shot prediction. We verify the superiority of ProtST-induced
PLMs over previous ones on diverse representation learning benchmarks. Under
the zero-shot setting, we show the effectiveness of ProtST on zero-shot protein
classification, and ProtST also enables functional protein retrieval from a
large-scale database without any function annotation.
| [
{
"version": "v1",
"created": "Sat, 28 Jan 2023 00:58:48 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 03:17:48 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Xu",
"Minghao",
""
],
[
"Yuan",
"Xinyu",
""
],
[
"Miret",
"Santiago",
""
],
[
"Tang",
"Jian",
""
]
] | TITLE: ProtST: Multi-Modality Learning of Protein Sequences and Biomedical
Texts
ABSTRACT: Current protein language models (PLMs) learn protein representations mainly
based on their sequences, thereby well capturing co-evolutionary information,
but they are unable to explicitly acquire protein functions, which is the end
goal of protein representation learning. Fortunately, for many proteins, their
textual property descriptions are available, where their various functions are
also described. Motivated by this fact, we first build the ProtDescribe dataset
to augment protein sequences with text descriptions of their functions and
other important properties. Based on this dataset, we propose the ProtST
framework to enhance Protein Sequence pre-training and understanding by
biomedical Texts. During pre-training, we design three types of tasks, i.e.,
unimodal mask prediction, multimodal representation alignment and multimodal
mask prediction, to enhance a PLM with protein property information with
different granularities and, at the same time, preserve the PLM's original
representation power. On downstream tasks, ProtST enables both supervised
learning and zero-shot prediction. We verify the superiority of ProtST-induced
PLMs over previous ones on diverse representation learning benchmarks. Under
the zero-shot setting, we show the effectiveness of ProtST on zero-shot protein
classification, and ProtST also enables functional protein retrieval from a
large-scale database without any function annotation.
| new_dataset | 0.714871 |
2302.05118 | Christian Tomani | Christian Tomani, Futa Waseda, Yuesong Shen and Daniel Cremers | Beyond In-Domain Scenarios: Robust Density-Aware Calibration | In Proceedings of the International Conference on Machine Learning
(ICML), 2023. Code available at
https://github.com/futakw/DensityAwareCalibration | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Calibrating deep learning models to yield uncertainty-aware predictions is
crucial as deep neural networks get increasingly deployed in safety-critical
applications. While existing post-hoc calibration methods achieve impressive
results on in-domain test datasets, they are limited by their inability to
yield reliable uncertainty estimates in domain-shift and out-of-domain (OOD)
scenarios. We aim to bridge this gap by proposing DAC, an accuracy-preserving
as well as Density-Aware Calibration method based on k-nearest-neighbors (KNN).
In contrast to existing post-hoc methods, we utilize hidden layers of
classifiers as a source for uncertainty-related information and study their
importance. We show that DAC is a generic method that can readily be combined
with state-of-the-art post-hoc methods. DAC boosts the robustness of
calibration performance in domain-shift and OOD, while maintaining excellent
in-domain predictive uncertainty estimates. We demonstrate that DAC leads to
consistently better calibration across a large number of model architectures,
datasets, and metrics. Additionally, we show that DAC improves calibration
substantially on recent large-scale neural networks pre-trained on vast amounts
of data.
| [
{
"version": "v1",
"created": "Fri, 10 Feb 2023 08:48:32 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 16:46:16 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Tomani",
"Christian",
""
],
[
"Waseda",
"Futa",
""
],
[
"Shen",
"Yuesong",
""
],
[
"Cremers",
"Daniel",
""
]
] | TITLE: Beyond In-Domain Scenarios: Robust Density-Aware Calibration
ABSTRACT: Calibrating deep learning models to yield uncertainty-aware predictions is
crucial as deep neural networks get increasingly deployed in safety-critical
applications. While existing post-hoc calibration methods achieve impressive
results on in-domain test datasets, they are limited by their inability to
yield reliable uncertainty estimates in domain-shift and out-of-domain (OOD)
scenarios. We aim to bridge this gap by proposing DAC, an accuracy-preserving
as well as Density-Aware Calibration method based on k-nearest-neighbors (KNN).
In contrast to existing post-hoc methods, we utilize hidden layers of
classifiers as a source for uncertainty-related information and study their
importance. We show that DAC is a generic method that can readily be combined
with state-of-the-art post-hoc methods. DAC boosts the robustness of
calibration performance in domain-shift and OOD, while maintaining excellent
in-domain predictive uncertainty estimates. We demonstrate that DAC leads to
consistently better calibration across a large number of model architectures,
datasets, and metrics. Additionally, we show that DAC improves calibration
substantially on recent large-scale neural networks pre-trained on vast amounts
of data.
| no_new_dataset | 0.710176 |
2302.10429 | Li Shen | Yan Sun, Li Shen, Tiansheng Huang, Liang Ding, Dacheng Tao | FedSpeed: Larger Local Interval, Less Communication Round, and Higher
Generalization Accuracy | ICLR 2023 | null | null | null | cs.LG cs.DC math.OC | http://creativecommons.org/licenses/by/4.0/ | Federated learning is an emerging distributed machine learning framework
which jointly trains a global model via a large number of local devices with
data privacy protections. Its performance suffers from the non-vanishing biases
introduced by the local inconsistent optimal and the rugged client-drifts by
the local over-fitting. In this paper, we propose a novel and practical method,
FedSpeed, to alleviate the negative impacts posed by these problems.
Concretely, FedSpeed applies the prox-correction term on the current local
updates to efficiently reduce the biases introduced by the prox-term, a
necessary regularizer to maintain the strong local consistency. Furthermore,
FedSpeed merges the vanilla stochastic gradient with a perturbation computed
from an extra gradient ascent step in the neighborhood, thereby alleviating the
issue of local over-fitting. Our theoretical analysis indicates that the
convergence rate is related to both the communication rounds $T$ and local
intervals $K$ with a upper bound $\small \mathcal{O}(1/T)$ if setting a proper
local interval. Moreover, we conduct extensive experiments on the real-world
dataset to demonstrate the efficiency of our proposed FedSpeed, which performs
significantly faster and achieves the state-of-the-art (SOTA) performance on
the general FL experimental settings than several baselines. Our code is
available at \url{https://github.com/woodenchild95/FL-Simulator.git}.
| [
{
"version": "v1",
"created": "Tue, 21 Feb 2023 03:55:29 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 09:45:47 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Sun",
"Yan",
""
],
[
"Shen",
"Li",
""
],
[
"Huang",
"Tiansheng",
""
],
[
"Ding",
"Liang",
""
],
[
"Tao",
"Dacheng",
""
]
] | TITLE: FedSpeed: Larger Local Interval, Less Communication Round, and Higher
Generalization Accuracy
ABSTRACT: Federated learning is an emerging distributed machine learning framework
which jointly trains a global model via a large number of local devices with
data privacy protections. Its performance suffers from the non-vanishing biases
introduced by the local inconsistent optimal and the rugged client-drifts by
the local over-fitting. In this paper, we propose a novel and practical method,
FedSpeed, to alleviate the negative impacts posed by these problems.
Concretely, FedSpeed applies the prox-correction term on the current local
updates to efficiently reduce the biases introduced by the prox-term, a
necessary regularizer to maintain the strong local consistency. Furthermore,
FedSpeed merges the vanilla stochastic gradient with a perturbation computed
from an extra gradient ascent step in the neighborhood, thereby alleviating the
issue of local over-fitting. Our theoretical analysis indicates that the
convergence rate is related to both the communication rounds $T$ and local
intervals $K$ with a upper bound $\small \mathcal{O}(1/T)$ if setting a proper
local interval. Moreover, we conduct extensive experiments on the real-world
dataset to demonstrate the efficiency of our proposed FedSpeed, which performs
significantly faster and achieves the state-of-the-art (SOTA) performance on
the general FL experimental settings than several baselines. Our code is
available at \url{https://github.com/woodenchild95/FL-Simulator.git}.
| no_new_dataset | 0.709453 |
2303.00865 | Ramin Nakhli | Ramin Nakhli, Puria Azadi Moghadam, Haoyang Mi, Hossein Farahani,
Alexander Baras, Blake Gilks, Ali Bashashati | AMIGO: Sparse Multi-Modal Graph Transformer with Shared-Context
Processing for Representation Learning of Giga-pixel Images | Accepted at CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Processing giga-pixel whole slide histopathology images (WSI) is a
computationally expensive task. Multiple instance learning (MIL) has become the
conventional approach to process WSIs, in which these images are split into
smaller patches for further processing. However, MIL-based techniques ignore
explicit information about the individual cells within a patch. In this paper,
by defining the novel concept of shared-context processing, we designed a
multi-modal Graph Transformer (AMIGO) that uses the celluar graph within the
tissue to provide a single representation for a patient while taking advantage
of the hierarchical structure of the tissue, enabling a dynamic focus between
cell-level and tissue-level information. We benchmarked the performance of our
model against multiple state-of-the-art methods in survival prediction and
showed that ours can significantly outperform all of them including
hierarchical Vision Transformer (ViT). More importantly, we show that our model
is strongly robust to missing information to an extent that it can achieve the
same performance with as low as 20% of the data. Finally, in two different
cancer datasets, we demonstrated that our model was able to stratify the
patients into low-risk and high-risk groups while other state-of-the-art
methods failed to achieve this goal. We also publish a large dataset of
immunohistochemistry images (InUIT) containing 1,600 tissue microarray (TMA)
cores from 188 patients along with their survival information, making it one of
the largest publicly available datasets in this context.
| [
{
"version": "v1",
"created": "Wed, 1 Mar 2023 23:37:45 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 13:25:47 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Nakhli",
"Ramin",
""
],
[
"Moghadam",
"Puria Azadi",
""
],
[
"Mi",
"Haoyang",
""
],
[
"Farahani",
"Hossein",
""
],
[
"Baras",
"Alexander",
""
],
[
"Gilks",
"Blake",
""
],
[
"Bashashati",
"Ali",
""
]
] | TITLE: AMIGO: Sparse Multi-Modal Graph Transformer with Shared-Context
Processing for Representation Learning of Giga-pixel Images
ABSTRACT: Processing giga-pixel whole slide histopathology images (WSI) is a
computationally expensive task. Multiple instance learning (MIL) has become the
conventional approach to process WSIs, in which these images are split into
smaller patches for further processing. However, MIL-based techniques ignore
explicit information about the individual cells within a patch. In this paper,
by defining the novel concept of shared-context processing, we designed a
multi-modal Graph Transformer (AMIGO) that uses the celluar graph within the
tissue to provide a single representation for a patient while taking advantage
of the hierarchical structure of the tissue, enabling a dynamic focus between
cell-level and tissue-level information. We benchmarked the performance of our
model against multiple state-of-the-art methods in survival prediction and
showed that ours can significantly outperform all of them including
hierarchical Vision Transformer (ViT). More importantly, we show that our model
is strongly robust to missing information to an extent that it can achieve the
same performance with as low as 20% of the data. Finally, in two different
cancer datasets, we demonstrated that our model was able to stratify the
patients into low-risk and high-risk groups while other state-of-the-art
methods failed to achieve this goal. We also publish a large dataset of
immunohistochemistry images (InUIT) containing 1,600 tissue microarray (TMA)
cores from 188 patients along with their survival information, making it one of
the largest publicly available datasets in this context.
| new_dataset | 0.701764 |
2303.08325 | Zikang Xu | Zikang Xu, Shang Zhao, Quan Quan, Qingsong Yao, and S. Kevin Zhou | FairAdaBN: Mitigating unfairness with adaptive batch normalization and
its application to dermatological disease classification | Accepted by MICCAI 2023 | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | Deep learning is becoming increasingly ubiquitous in medical research and
applications while involving sensitive information and even critical diagnosis
decisions. Researchers observe a significant performance disparity among
subgroups with different demographic attributes, which is called model
unfairness, and put lots of effort into carefully designing elegant
architectures to address unfairness, which poses heavy training burden, brings
poor generalization, and reveals the trade-off between model performance and
fairness. To tackle these issues, we propose FairAdaBN by making batch
normalization adaptive to sensitive attribute. This simple but effective design
can be adopted to several classification backbones that are originally unaware
of fairness. Additionally, we derive a novel loss function that restrains
statistical parity between subgroups on mini-batches, encouraging the model to
converge with considerable fairness. In order to evaluate the trade-off between
model performance and fairness, we propose a new metric, named
Fairness-Accuracy Trade-off Efficiency (FATE), to compute normalized fairness
improvement over accuracy drop. Experiments on two dermatological datasets show
that our proposed method outperforms other methods on fairness criteria and
FATE.
| [
{
"version": "v1",
"created": "Wed, 15 Mar 2023 02:22:07 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 05:17:09 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Xu",
"Zikang",
""
],
[
"Zhao",
"Shang",
""
],
[
"Quan",
"Quan",
""
],
[
"Yao",
"Qingsong",
""
],
[
"Zhou",
"S. Kevin",
""
]
] | TITLE: FairAdaBN: Mitigating unfairness with adaptive batch normalization and
its application to dermatological disease classification
ABSTRACT: Deep learning is becoming increasingly ubiquitous in medical research and
applications while involving sensitive information and even critical diagnosis
decisions. Researchers observe a significant performance disparity among
subgroups with different demographic attributes, which is called model
unfairness, and put lots of effort into carefully designing elegant
architectures to address unfairness, which poses heavy training burden, brings
poor generalization, and reveals the trade-off between model performance and
fairness. To tackle these issues, we propose FairAdaBN by making batch
normalization adaptive to sensitive attribute. This simple but effective design
can be adopted to several classification backbones that are originally unaware
of fairness. Additionally, we derive a novel loss function that restrains
statistical parity between subgroups on mini-batches, encouraging the model to
converge with considerable fairness. In order to evaluate the trade-off between
model performance and fairness, we propose a new metric, named
Fairness-Accuracy Trade-off Efficiency (FATE), to compute normalized fairness
improvement over accuracy drop. Experiments on two dermatological datasets show
that our proposed method outperforms other methods on fairness criteria and
FATE.
| no_new_dataset | 0.707582 |
2303.12776 | Shilong Zhang | Shilong Zhang, Xinjiang Wang, Jiaqi Wang, Jiangmiao Pang, Chengqi Lyu,
Wenwei Zhang, Ping Luo, Kai Chen | Dense Distinct Query for End-to-End Object Detection | Accepted to CVPR2023. Code has been released at
https://github.com/jshilong/DDQ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One-to-one label assignment in object detection has successfully obviated the
need for non-maximum suppression (NMS) as postprocessing and makes the pipeline
end-to-end. However, it triggers a new dilemma as the widely used sparse
queries cannot guarantee a high recall, while dense queries inevitably bring
more similar queries and encounter optimization difficulties. As both sparse
and dense queries are problematic, then what are the expected queries in
end-to-end object detection? This paper shows that the solution should be Dense
Distinct Queries (DDQ). Concretely, we first lay dense queries like traditional
detectors and then select distinct ones for one-to-one assignments. DDQ blends
the advantages of traditional and recent end-to-end detectors and significantly
improves the performance of various detectors including FCN, R-CNN, and DETRs.
Most impressively, DDQ-DETR achieves 52.1 AP on MS-COCO dataset within 12
epochs using a ResNet-50 backbone, outperforming all existing detectors in the
same setting. DDQ also shares the benefit of end-to-end detectors in crowded
scenes and achieves 93.8 AP on CrowdHuman. We hope DDQ can inspire researchers
to consider the complementarity between traditional methods and end-to-end
detectors. The source code can be found at
\url{https://github.com/jshilong/DDQ}.
| [
{
"version": "v1",
"created": "Wed, 22 Mar 2023 17:42:22 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 13:36:43 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Zhang",
"Shilong",
""
],
[
"Wang",
"Xinjiang",
""
],
[
"Wang",
"Jiaqi",
""
],
[
"Pang",
"Jiangmiao",
""
],
[
"Lyu",
"Chengqi",
""
],
[
"Zhang",
"Wenwei",
""
],
[
"Luo",
"Ping",
""
],
[
"Chen",
"Kai",
""
]
] | TITLE: Dense Distinct Query for End-to-End Object Detection
ABSTRACT: One-to-one label assignment in object detection has successfully obviated the
need for non-maximum suppression (NMS) as postprocessing and makes the pipeline
end-to-end. However, it triggers a new dilemma as the widely used sparse
queries cannot guarantee a high recall, while dense queries inevitably bring
more similar queries and encounter optimization difficulties. As both sparse
and dense queries are problematic, then what are the expected queries in
end-to-end object detection? This paper shows that the solution should be Dense
Distinct Queries (DDQ). Concretely, we first lay dense queries like traditional
detectors and then select distinct ones for one-to-one assignments. DDQ blends
the advantages of traditional and recent end-to-end detectors and significantly
improves the performance of various detectors including FCN, R-CNN, and DETRs.
Most impressively, DDQ-DETR achieves 52.1 AP on MS-COCO dataset within 12
epochs using a ResNet-50 backbone, outperforming all existing detectors in the
same setting. DDQ also shares the benefit of end-to-end detectors in crowded
scenes and achieves 93.8 AP on CrowdHuman. We hope DDQ can inspire researchers
to consider the complementarity between traditional methods and end-to-end
detectors. The source code can be found at
\url{https://github.com/jshilong/DDQ}.
| no_new_dataset | 0.709221 |
2304.03682 | Shadman Rohan | Shadman Rohan, Mojammel Hossain, Mohammad Mamun Or Rashid, Nabeel
Mohammed | BenCoref: A Multi-Domain Dataset of Nominal Phrases and Pronominal
Reference Annotations | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Coreference Resolution is a well studied problem in NLP. While widely studied
for English and other resource-rich languages, research on coreference
resolution in Bengali largely remains unexplored due to the absence of relevant
datasets. Bengali, being a low-resource language, exhibits greater
morphological richness compared to English. In this article, we introduce a new
dataset, BenCoref, comprising coreference annotations for Bengali texts
gathered from four distinct domains. This relatively small dataset contains
5200 mention annotations forming 502 mention clusters within 48,569 tokens. We
describe the process of creating this dataset and report performance of
multiple models trained using BenCoref. We expect that our work provides some
valuable insights on the variations in coreference phenomena across several
domains in Bengali and encourages the development of additional resources for
Bengali. Furthermore, we found poor crosslingual performance at zero-shot
setting from English, highlighting the need for more language-specific
resources for this task.
| [
{
"version": "v1",
"created": "Fri, 7 Apr 2023 15:08:46 GMT"
},
{
"version": "v2",
"created": "Tue, 30 May 2023 13:42:48 GMT"
},
{
"version": "v3",
"created": "Mon, 3 Jul 2023 18:33:23 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Rohan",
"Shadman",
""
],
[
"Hossain",
"Mojammel",
""
],
[
"Rashid",
"Mohammad Mamun Or",
""
],
[
"Mohammed",
"Nabeel",
""
]
] | TITLE: BenCoref: A Multi-Domain Dataset of Nominal Phrases and Pronominal
Reference Annotations
ABSTRACT: Coreference Resolution is a well studied problem in NLP. While widely studied
for English and other resource-rich languages, research on coreference
resolution in Bengali largely remains unexplored due to the absence of relevant
datasets. Bengali, being a low-resource language, exhibits greater
morphological richness compared to English. In this article, we introduce a new
dataset, BenCoref, comprising coreference annotations for Bengali texts
gathered from four distinct domains. This relatively small dataset contains
5200 mention annotations forming 502 mention clusters within 48,569 tokens. We
describe the process of creating this dataset and report performance of
multiple models trained using BenCoref. We expect that our work provides some
valuable insights on the variations in coreference phenomena across several
domains in Bengali and encourages the development of additional resources for
Bengali. Furthermore, we found poor crosslingual performance at zero-shot
setting from English, highlighting the need for more language-specific
resources for this task.
| new_dataset | 0.715859 |
2304.04106 | Kun Han | Kun Han, Yifeng Xiong, Chenyu You, Pooya Khosravi, Shanlin Sun,
Xiangyi Yan, James Duncan, Xiaohui Xie | MedGen3D: A Deep Generative Framework for Paired 3D Image and Mask
Generation | Accepted by MICCAI 2023. Project Page:
https://krishan999.github.io/MedGen3D/ | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | Acquiring and annotating sufficient labeled data is crucial in developing
accurate and robust learning-based models, but obtaining such data can be
challenging in many medical image segmentation tasks. One promising solution is
to synthesize realistic data with ground-truth mask annotations. However, no
prior studies have explored generating complete 3D volumetric images with
masks. In this paper, we present MedGen3D, a deep generative framework that can
generate paired 3D medical images and masks. First, we represent the 3D medical
data as 2D sequences and propose the Multi-Condition Diffusion Probabilistic
Model (MC-DPM) to generate multi-label mask sequences adhering to anatomical
geometry. Then, we use an image sequence generator and semantic diffusion
refiner conditioned on the generated mask sequences to produce realistic 3D
medical images that align with the generated masks. Our proposed framework
guarantees accurate alignment between synthetic images and segmentation maps.
Experiments on 3D thoracic CT and brain MRI datasets show that our synthetic
data is both diverse and faithful to the original data, and demonstrate the
benefits for downstream segmentation tasks. We anticipate that MedGen3D's
ability to synthesize paired 3D medical images and masks will prove valuable in
training deep learning models for medical imaging tasks.
| [
{
"version": "v1",
"created": "Sat, 8 Apr 2023 21:43:26 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 23:06:50 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Han",
"Kun",
""
],
[
"Xiong",
"Yifeng",
""
],
[
"You",
"Chenyu",
""
],
[
"Khosravi",
"Pooya",
""
],
[
"Sun",
"Shanlin",
""
],
[
"Yan",
"Xiangyi",
""
],
[
"Duncan",
"James",
""
],
[
"Xie",
"Xiaohui",
""
]
] | TITLE: MedGen3D: A Deep Generative Framework for Paired 3D Image and Mask
Generation
ABSTRACT: Acquiring and annotating sufficient labeled data is crucial in developing
accurate and robust learning-based models, but obtaining such data can be
challenging in many medical image segmentation tasks. One promising solution is
to synthesize realistic data with ground-truth mask annotations. However, no
prior studies have explored generating complete 3D volumetric images with
masks. In this paper, we present MedGen3D, a deep generative framework that can
generate paired 3D medical images and masks. First, we represent the 3D medical
data as 2D sequences and propose the Multi-Condition Diffusion Probabilistic
Model (MC-DPM) to generate multi-label mask sequences adhering to anatomical
geometry. Then, we use an image sequence generator and semantic diffusion
refiner conditioned on the generated mask sequences to produce realistic 3D
medical images that align with the generated masks. Our proposed framework
guarantees accurate alignment between synthetic images and segmentation maps.
Experiments on 3D thoracic CT and brain MRI datasets show that our synthetic
data is both diverse and faithful to the original data, and demonstrate the
benefits for downstream segmentation tasks. We anticipate that MedGen3D's
ability to synthesize paired 3D medical images and masks will prove valuable in
training deep learning models for medical imaging tasks.
| no_new_dataset | 0.709189 |
2304.13009 | Lucas Prado Osco | Lucas Prado Osco, Eduardo Lopes de Lemos, Wesley Nunes Gon\c{c}alves,
Ana Paula Marques Ramos and Jos\'e Marcato Junior | The Potential of Visual ChatGPT For Remote Sensing | null | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in Natural Language Processing (NLP), particularly in
Large Language Models (LLMs), associated with deep learning-based computer
vision techniques, have shown substantial potential for automating a variety of
tasks. One notable model is Visual ChatGPT, which combines ChatGPT's LLM
capabilities with visual computation to enable effective image analysis. The
model's ability to process images based on textual inputs can revolutionize
diverse fields. However, its application in the remote sensing domain remains
unexplored. This is the first paper to examine the potential of Visual ChatGPT,
a cutting-edge LLM founded on the GPT architecture, to tackle the aspects of
image processing related to the remote sensing domain. Among its current
capabilities, Visual ChatGPT can generate textual descriptions of images,
perform canny edge and straight line detection, and conduct image segmentation.
These offer valuable insights into image content and facilitate the
interpretation and extraction of information. By exploring the applicability of
these techniques within publicly available datasets of satellite images, we
demonstrate the current model's limitations in dealing with remote sensing
images, highlighting its challenges and future prospects. Although still in
early development, we believe that the combination of LLMs and visual models
holds a significant potential to transform remote sensing image processing,
creating accessible and practical application opportunities in the field.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 17:29:47 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 14:09:09 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Osco",
"Lucas Prado",
""
],
[
"de Lemos",
"Eduardo Lopes",
""
],
[
"Gonçalves",
"Wesley Nunes",
""
],
[
"Ramos",
"Ana Paula Marques",
""
],
[
"Junior",
"José Marcato",
""
]
] | TITLE: The Potential of Visual ChatGPT For Remote Sensing
ABSTRACT: Recent advancements in Natural Language Processing (NLP), particularly in
Large Language Models (LLMs), associated with deep learning-based computer
vision techniques, have shown substantial potential for automating a variety of
tasks. One notable model is Visual ChatGPT, which combines ChatGPT's LLM
capabilities with visual computation to enable effective image analysis. The
model's ability to process images based on textual inputs can revolutionize
diverse fields. However, its application in the remote sensing domain remains
unexplored. This is the first paper to examine the potential of Visual ChatGPT,
a cutting-edge LLM founded on the GPT architecture, to tackle the aspects of
image processing related to the remote sensing domain. Among its current
capabilities, Visual ChatGPT can generate textual descriptions of images,
perform canny edge and straight line detection, and conduct image segmentation.
These offer valuable insights into image content and facilitate the
interpretation and extraction of information. By exploring the applicability of
these techniques within publicly available datasets of satellite images, we
demonstrate the current model's limitations in dealing with remote sensing
images, highlighting its challenges and future prospects. Although still in
early development, we believe that the combination of LLMs and visual models
holds a significant potential to transform remote sensing image processing,
creating accessible and practical application opportunities in the field.
| no_new_dataset | 0.708023 |
2305.00417 | Sonali C S | C. S. Sonali, Chinmayi B S, Ahana Balasubramanian | Transformer-based Sequence Labeling for Audio Classification based on
MFCCs | Error in the explanation as well inadequate results and conclusion | null | null | null | cs.SD cs.CV eess.AS | http://creativecommons.org/licenses/by/4.0/ | Audio classification is vital in areas such as speech and music recognition.
Feature extraction from the audio signal, such as Mel-Spectrograms and MFCCs,
is a critical step in audio classification. These features are transformed into
spectrograms for classification. Researchers have explored various techniques,
including traditional machine and deep learning methods to classify
spectrograms, but these can be computationally expensive. To simplify this
process, a more straightforward approach inspired by sequence classification in
NLP can be used. This paper proposes a Transformer-encoder-based model for
audio classification using MFCCs. The model was benchmarked against the ESC-50,
Speech Commands v0.02 and UrbanSound8k datasets and has shown strong
performance, with the highest accuracy of 95.2% obtained upon training the
model on the UrbanSound8k dataset. The model consisted of a mere 127,544 total
parameters, making it light-weight yet highly efficient at the audio
classification task.
| [
{
"version": "v1",
"created": "Sun, 30 Apr 2023 07:25:43 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 05:50:38 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Sonali",
"C. S.",
""
],
[
"S",
"Chinmayi B",
""
],
[
"Balasubramanian",
"Ahana",
""
]
] | TITLE: Transformer-based Sequence Labeling for Audio Classification based on
MFCCs
ABSTRACT: Audio classification is vital in areas such as speech and music recognition.
Feature extraction from the audio signal, such as Mel-Spectrograms and MFCCs,
is a critical step in audio classification. These features are transformed into
spectrograms for classification. Researchers have explored various techniques,
including traditional machine and deep learning methods to classify
spectrograms, but these can be computationally expensive. To simplify this
process, a more straightforward approach inspired by sequence classification in
NLP can be used. This paper proposes a Transformer-encoder-based model for
audio classification using MFCCs. The model was benchmarked against the ESC-50,
Speech Commands v0.02 and UrbanSound8k datasets and has shown strong
performance, with the highest accuracy of 95.2% obtained upon training the
model on the UrbanSound8k dataset. The model consisted of a mere 127,544 total
parameters, making it light-weight yet highly efficient at the audio
classification task.
| no_new_dataset | 0.710189 |
2305.02301 | Cheng-Yu Hsieh | Cheng-Yu Hsieh, Chun-Liang Li, Chih-Kuan Yeh, Hootan Nakhost, Yasuhisa
Fujii, Alexander Ratner, Ranjay Krishna, Chen-Yu Lee, Tomas Pfister | Distilling Step-by-Step! Outperforming Larger Language Models with Less
Training Data and Smaller Model Sizes | Accepted to Findings of ACL 2023 | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deploying large language models (LLMs) is challenging because they are memory
inefficient and compute-intensive for practical applications. In reaction,
researchers train smaller task-specific models by either finetuning with human
labels or distilling using LLM-generated labels. However, finetuning and
distillation require large amounts of training data to achieve comparable
performance to LLMs. We introduce Distilling step-by-step, a new mechanism that
(a) trains smaller models that outperform LLMs, and (b) achieves so by
leveraging less training data needed by finetuning or distillation. Our method
extracts LLM rationales as additional supervision for training small models
within a multi-task framework. We present three findings across 4 NLP
benchmarks: First, compared to both finetuning and distillation, our mechanism
achieves better performance with much fewer labeled/unlabeled training
examples. Second, compared to few-shot prompted LLMs, we achieve better
performance using substantially smaller model sizes. Third, we reduce both the
model size and the amount of data required to outperform LLMs; our finetuned
770M T5 model outperforms the few-shot prompted 540B PaLM model using only 80%
of available data on a benchmark, whereas standard finetuning the same T5 model
struggles to match even by using 100% of the dataset. We release the code at:
https://github.com/google-research/distilling-step-by-step .
| [
{
"version": "v1",
"created": "Wed, 3 May 2023 17:50:56 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 16:59:31 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Hsieh",
"Cheng-Yu",
""
],
[
"Li",
"Chun-Liang",
""
],
[
"Yeh",
"Chih-Kuan",
""
],
[
"Nakhost",
"Hootan",
""
],
[
"Fujii",
"Yasuhisa",
""
],
[
"Ratner",
"Alexander",
""
],
[
"Krishna",
"Ranjay",
""
],
[
"Lee",
"Chen-Yu",
""
],
[
"Pfister",
"Tomas",
""
]
] | TITLE: Distilling Step-by-Step! Outperforming Larger Language Models with Less
Training Data and Smaller Model Sizes
ABSTRACT: Deploying large language models (LLMs) is challenging because they are memory
inefficient and compute-intensive for practical applications. In reaction,
researchers train smaller task-specific models by either finetuning with human
labels or distilling using LLM-generated labels. However, finetuning and
distillation require large amounts of training data to achieve comparable
performance to LLMs. We introduce Distilling step-by-step, a new mechanism that
(a) trains smaller models that outperform LLMs, and (b) achieves so by
leveraging less training data needed by finetuning or distillation. Our method
extracts LLM rationales as additional supervision for training small models
within a multi-task framework. We present three findings across 4 NLP
benchmarks: First, compared to both finetuning and distillation, our mechanism
achieves better performance with much fewer labeled/unlabeled training
examples. Second, compared to few-shot prompted LLMs, we achieve better
performance using substantially smaller model sizes. Third, we reduce both the
model size and the amount of data required to outperform LLMs; our finetuned
770M T5 model outperforms the few-shot prompted 540B PaLM model using only 80%
of available data on a benchmark, whereas standard finetuning the same T5 model
struggles to match even by using 100% of the dataset. We release the code at:
https://github.com/google-research/distilling-step-by-step .
| no_new_dataset | 0.711437 |
2305.02562 | Anderson de Andrade | Anderson de Andrade, Alon Harell, Yalda Foroutan, Ivan V. Baji\'c | Conditional and Residual Methods in Scalable Coding for Humans and
Machines | IEEE ICME Workshop on Coding for Machines, Brisbane, Australia, 2023 | null | null | null | eess.IV cs.IT cs.LG math.IT | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present methods for conditional and residual coding in the context of
scalable coding for humans and machines. Our focus is on optimizing the
rate-distortion performance of the reconstruction task using the information
available in the computer vision task. We include an information analysis of
both approaches to provide baselines and also propose an entropy model suitable
for conditional coding with increased modelling capacity and similar
tractability as previous work. We apply these methods to image reconstruction,
using, in one instance, representations created for semantic segmentation on
the Cityscapes dataset, and in another instance, representations created for
object detection on the COCO dataset. In both experiments, we obtain similar
performance between the conditional and residual methods, with the resulting
rate-distortion curves contained within our baselines.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 05:32:44 GMT"
},
{
"version": "v2",
"created": "Tue, 4 Jul 2023 23:27:16 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"de Andrade",
"Anderson",
""
],
[
"Harell",
"Alon",
""
],
[
"Foroutan",
"Yalda",
""
],
[
"Bajić",
"Ivan V.",
""
]
] | TITLE: Conditional and Residual Methods in Scalable Coding for Humans and
Machines
ABSTRACT: We present methods for conditional and residual coding in the context of
scalable coding for humans and machines. Our focus is on optimizing the
rate-distortion performance of the reconstruction task using the information
available in the computer vision task. We include an information analysis of
both approaches to provide baselines and also propose an entropy model suitable
for conditional coding with increased modelling capacity and similar
tractability as previous work. We apply these methods to image reconstruction,
using, in one instance, representations created for semantic segmentation on
the Cityscapes dataset, and in another instance, representations created for
object detection on the COCO dataset. In both experiments, we obtain similar
performance between the conditional and residual methods, with the resulting
rate-distortion curves contained within our baselines.
| no_new_dataset | 0.713806 |
2305.07611 | Songning Lai | Songning Lai, Xifeng Hu, Haoxuan Xu, Zhaoxia Ren and Zhi Liu | Multimodal Sentiment Analysis: A Survey | It needs to be returned for major modifications | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal sentiment analysis has become an important research area in the
field of artificial intelligence. With the latest advances in deep learning,
this technology has reached new heights. It has great potential for both
application and research, making it a popular research topic. This review
provides an overview of the definition, background, and development of
multimodal sentiment analysis. It also covers recent datasets and advanced
models, emphasizing the challenges and future prospects of this technology.
Finally, it looks ahead to future research directions. It should be noted that
this review provides constructive suggestions for promising research directions
and building better performing multimodal sentiment analysis models, which can
help researchers in this field.
| [
{
"version": "v1",
"created": "Fri, 12 May 2023 16:56:13 GMT"
},
{
"version": "v2",
"created": "Wed, 21 Jun 2023 11:35:00 GMT"
},
{
"version": "v3",
"created": "Mon, 3 Jul 2023 19:05:30 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Lai",
"Songning",
""
],
[
"Hu",
"Xifeng",
""
],
[
"Xu",
"Haoxuan",
""
],
[
"Ren",
"Zhaoxia",
""
],
[
"Liu",
"Zhi",
""
]
] | TITLE: Multimodal Sentiment Analysis: A Survey
ABSTRACT: Multimodal sentiment analysis has become an important research area in the
field of artificial intelligence. With the latest advances in deep learning,
this technology has reached new heights. It has great potential for both
application and research, making it a popular research topic. This review
provides an overview of the definition, background, and development of
multimodal sentiment analysis. It also covers recent datasets and advanced
models, emphasizing the challenges and future prospects of this technology.
Finally, it looks ahead to future research directions. It should be noted that
this review provides constructive suggestions for promising research directions
and building better performing multimodal sentiment analysis models, which can
help researchers in this field.
| no_new_dataset | 0.709849 |
2305.11991 | Ren\'e Peinl | Ren\'e Peinl and Johannes Wirth | Evaluation of medium-large Language Models at zero-shot closed book
generative question answering | null | Under review in ARIA 2023 | null | null | cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | Large language models (LLMs) have garnered significant attention, but the
definition of "large" lacks clarity. This paper focuses on medium-sized
language models (MLMs), defined as having at least six billion parameters but
less than 100 billion. The study evaluates MLMs regarding zero-shot generative
question answering, which requires models to provide elaborate answers without
external document retrieval. The paper introduces an own test dataset and
presents results from human evaluation. Results show that combining the best
answers from different MLMs yielded an overall correct answer rate of 82.7%
which is better than the 60.9% of ChatGPT. The best MLM achieved 71.8% and has
33B parameters, which highlights the importance of using appropriate training
data for fine-tuning rather than solely relying on the number of parameters.
More fine-grained feedback should be used to further improve the quality of
answers. The open source community is quickly closing the gap to the best
commercial models.
| [
{
"version": "v1",
"created": "Fri, 19 May 2023 20:33:19 GMT"
},
{
"version": "v2",
"created": "Mon, 3 Jul 2023 18:26:23 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Peinl",
"René",
""
],
[
"Wirth",
"Johannes",
""
]
] | TITLE: Evaluation of medium-large Language Models at zero-shot closed book
generative question answering
ABSTRACT: Large language models (LLMs) have garnered significant attention, but the
definition of "large" lacks clarity. This paper focuses on medium-sized
language models (MLMs), defined as having at least six billion parameters but
less than 100 billion. The study evaluates MLMs regarding zero-shot generative
question answering, which requires models to provide elaborate answers without
external document retrieval. The paper introduces an own test dataset and
presents results from human evaluation. Results show that combining the best
answers from different MLMs yielded an overall correct answer rate of 82.7%
which is better than the 60.9% of ChatGPT. The best MLM achieved 71.8% and has
33B parameters, which highlights the importance of using appropriate training
data for fine-tuning rather than solely relying on the number of parameters.
More fine-grained feedback should be used to further improve the quality of
answers. The open source community is quickly closing the gap to the best
commercial models.
| new_dataset | 0.713594 |
2305.12199 | Xuan-Quy Dao | Xuan-Quy Dao, Ngoc-Bich Le, The-Duy Vo, Xuan-Dung Phan, Bac-Bien Ngo,
Van-Tien Nguyen, Thi-My-Thanh Nguyen, and Hong-Phuoc Nguyen | VNHSGE: VietNamese High School Graduation Examination Dataset for Large
Language Models | 74 pages, 44 figures | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The VNHSGE (VietNamese High School Graduation Examination) dataset, developed
exclusively for evaluating large language models (LLMs), is introduced in this
article. The dataset, which covers nine subjects, was generated from the
Vietnamese National High School Graduation Examination and comparable tests.
300 literary essays have been included, and there are over 19,000
multiple-choice questions on a range of topics. The dataset assesses LLMs in
multitasking situations such as question answering, text generation, reading
comprehension, visual question answering, and more by including both textual
data and accompanying images. Using ChatGPT and BingChat, we evaluated LLMs on
the VNHSGE dataset and contrasted their performance with that of Vietnamese
students to see how well they performed. The results show that ChatGPT and
BingChat both perform at a human level in a number of areas, including
literature, English, history, geography, and civics education. They still have
space to grow, though, especially in the areas of mathematics, physics,
chemistry, and biology. The VNHSGE dataset seeks to provide an adequate
benchmark for assessing the abilities of LLMs with its wide-ranging coverage
and variety of activities. We intend to promote future developments in the
creation of LLMs by making this dataset available to the scientific community,
especially in resolving LLMs' limits in disciplines involving mathematics and
the natural sciences.
| [
{
"version": "v1",
"created": "Sat, 20 May 2023 14:13:08 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Dao",
"Xuan-Quy",
""
],
[
"Le",
"Ngoc-Bich",
""
],
[
"Vo",
"The-Duy",
""
],
[
"Phan",
"Xuan-Dung",
""
],
[
"Ngo",
"Bac-Bien",
""
],
[
"Nguyen",
"Van-Tien",
""
],
[
"Nguyen",
"Thi-My-Thanh",
""
],
[
"Nguyen",
"Hong-Phuoc",
""
]
] | TITLE: VNHSGE: VietNamese High School Graduation Examination Dataset for Large
Language Models
ABSTRACT: The VNHSGE (VietNamese High School Graduation Examination) dataset, developed
exclusively for evaluating large language models (LLMs), is introduced in this
article. The dataset, which covers nine subjects, was generated from the
Vietnamese National High School Graduation Examination and comparable tests.
300 literary essays have been included, and there are over 19,000
multiple-choice questions on a range of topics. The dataset assesses LLMs in
multitasking situations such as question answering, text generation, reading
comprehension, visual question answering, and more by including both textual
data and accompanying images. Using ChatGPT and BingChat, we evaluated LLMs on
the VNHSGE dataset and contrasted their performance with that of Vietnamese
students to see how well they performed. The results show that ChatGPT and
BingChat both perform at a human level in a number of areas, including
literature, English, history, geography, and civics education. They still have
space to grow, though, especially in the areas of mathematics, physics,
chemistry, and biology. The VNHSGE dataset seeks to provide an adequate
benchmark for assessing the abilities of LLMs with its wide-ranging coverage
and variety of activities. We intend to promote future developments in the
creation of LLMs by making this dataset available to the scientific community,
especially in resolving LLMs' limits in disciplines involving mathematics and
the natural sciences.
| new_dataset | 0.721157 |
2305.14705 | Sheng Shen | Sheng Shen, Le Hou, Yanqi Zhou, Nan Du, Shayne Longpre, Jason Wei,
Hyung Won Chung, Barret Zoph, William Fedus, Xinyun Chen, Tu Vu, Yuexin Wu,
Wuyang Chen, Albert Webson, Yunxuan Li, Vincent Zhao, Hongkun Yu, Kurt
Keutzer, Trevor Darrell, Denny Zhou | Mixture-of-Experts Meets Instruction Tuning:A Winning Combination for
Large Language Models | Preprint | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Sparse Mixture-of-Experts (MoE) is a neural architecture design that can be
utilized to add learnable parameters to Large Language Models (LLMs) without
increasing inference cost. Instruction tuning is a technique for training LLMs
to follow instructions. We advocate combining these two approaches, as we find
that MoE models benefit more from instruction tuning than dense models. In
particular, we conduct empirical studies across three experimental setups: (i)
Direct finetuning on individual downstream tasks devoid of instruction tuning;
(ii) Instructiontuning followed by in-context few-shot or zero-shot
generalization on downstream tasks; and (iii) Instruction tuning supplemented
by further finetuning on individual downstream tasks. In the first scenario,
MoE models overall underperform dense models of identical computational
capacity. This narrative, however, dramatically changes with the introduction
of instruction tuning (second and third scenario), used independently or in
conjunction with task-specific finetuning. Our most powerful model,
FLAN-MOE-32B, surpasses the performance of FLAN-PALM-62B on four benchmark
tasks, while using only a third of the FLOPs. The advancements embodied
byFLAN-MOE inspire a reevaluation of the design principles of large-scale,
high-performance language models in the framework of task-agnostic learning.
| [
{
"version": "v1",
"created": "Wed, 24 May 2023 04:22:26 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Jul 2023 17:24:43 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Shen",
"Sheng",
""
],
[
"Hou",
"Le",
""
],
[
"Zhou",
"Yanqi",
""
],
[
"Du",
"Nan",
""
],
[
"Longpre",
"Shayne",
""
],
[
"Wei",
"Jason",
""
],
[
"Chung",
"Hyung Won",
""
],
[
"Zoph",
"Barret",
""
],
[
"Fedus",
"William",
""
],
[
"Chen",
"Xinyun",
""
],
[
"Vu",
"Tu",
""
],
[
"Wu",
"Yuexin",
""
],
[
"Chen",
"Wuyang",
""
],
[
"Webson",
"Albert",
""
],
[
"Li",
"Yunxuan",
""
],
[
"Zhao",
"Vincent",
""
],
[
"Yu",
"Hongkun",
""
],
[
"Keutzer",
"Kurt",
""
],
[
"Darrell",
"Trevor",
""
],
[
"Zhou",
"Denny",
""
]
] | TITLE: Mixture-of-Experts Meets Instruction Tuning:A Winning Combination for
Large Language Models
ABSTRACT: Sparse Mixture-of-Experts (MoE) is a neural architecture design that can be
utilized to add learnable parameters to Large Language Models (LLMs) without
increasing inference cost. Instruction tuning is a technique for training LLMs
to follow instructions. We advocate combining these two approaches, as we find
that MoE models benefit more from instruction tuning than dense models. In
particular, we conduct empirical studies across three experimental setups: (i)
Direct finetuning on individual downstream tasks devoid of instruction tuning;
(ii) Instructiontuning followed by in-context few-shot or zero-shot
generalization on downstream tasks; and (iii) Instruction tuning supplemented
by further finetuning on individual downstream tasks. In the first scenario,
MoE models overall underperform dense models of identical computational
capacity. This narrative, however, dramatically changes with the introduction
of instruction tuning (second and third scenario), used independently or in
conjunction with task-specific finetuning. Our most powerful model,
FLAN-MOE-32B, surpasses the performance of FLAN-PALM-62B on four benchmark
tasks, while using only a third of the FLOPs. The advancements embodied
byFLAN-MOE inspire a reevaluation of the design principles of large-scale,
high-performance language models in the framework of task-agnostic learning.
| no_new_dataset | 0.70825 |
2305.16243 | Ehsan Doostmohammadi | Ehsan Doostmohammadi, Tobias Norlund, Marco Kuhlmann, Richard
Johansson | Surface-Based Retrieval Reduces Perplexity of Retrieval-Augmented
Language Models | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Augmenting language models with a retrieval mechanism has been shown to
significantly improve their performance while keeping the number of parameters
low. Retrieval-augmented models commonly rely on a semantic retrieval mechanism
based on the similarity between dense representations of the query chunk and
potential neighbors. In this paper, we study the state-of-the-art Retro model
and observe that its performance gain is better explained by surface-level
similarities, such as token overlap. Inspired by this, we replace the semantic
retrieval in Retro with a surface-level method based on BM25, obtaining a
significant reduction in perplexity. As full BM25 retrieval can be
computationally costly for large datasets, we also apply it in a re-ranking
scenario, gaining part of the perplexity reduction with minimal computational
overhead.
| [
{
"version": "v1",
"created": "Thu, 25 May 2023 16:56:26 GMT"
},
{
"version": "v2",
"created": "Sat, 10 Jun 2023 11:08:06 GMT"
},
{
"version": "v3",
"created": "Tue, 4 Jul 2023 07:59:15 GMT"
}
] | 2023-07-06T00:00:00 | [
[
"Doostmohammadi",
"Ehsan",
""
],
[
"Norlund",
"Tobias",
""
],
[
"Kuhlmann",
"Marco",
""
],
[
"Johansson",
"Richard",
""
]
] | TITLE: Surface-Based Retrieval Reduces Perplexity of Retrieval-Augmented
Language Models
ABSTRACT: Augmenting language models with a retrieval mechanism has been shown to
significantly improve their performance while keeping the number of parameters
low. Retrieval-augmented models commonly rely on a semantic retrieval mechanism
based on the similarity between dense representations of the query chunk and
potential neighbors. In this paper, we study the state-of-the-art Retro model
and observe that its performance gain is better explained by surface-level
similarities, such as token overlap. Inspired by this, we replace the semantic
retrieval in Retro with a surface-level method based on BM25, obtaining a
significant reduction in perplexity. As full BM25 retrieval can be
computationally costly for large datasets, we also apply it in a re-ranking
scenario, gaining part of the perplexity reduction with minimal computational
overhead.
| no_new_dataset | 0.710415 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.