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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2302.13004 | Yaqi Liu | Yaqi Liu, Binbin Lv, Xin Jin, Xiaoyu Chen, and Xiaokun Zhang | TBFormer: Two-Branch Transformer for Image Forgery Localization | 5 pages, 3 figures | null | 10.1109/LSP.2023.3279018 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image forgery localization aims to identify forged regions by capturing
subtle traces from high-quality discriminative features. In this paper, we
propose a Transformer-style network with two feature extraction branches for
image forgery localization, and it is named as Two-Branch Transformer
(TBFormer). Firstly, two feature extraction branches are elaborately designed,
taking advantage of the discriminative stacked Transformer layers, for both RGB
and noise domain features. Secondly, an Attention-aware Hierarchical-feature
Fusion Module (AHFM) is proposed to effectively fuse hierarchical features from
two different domains. Although the two feature extraction branches have the
same architecture, their features have significant differences since they are
extracted from different domains. We adopt position attention to embed them
into a unified feature domain for hierarchical feature investigation. Finally,
a Transformer decoder is constructed for feature reconstruction to generate the
predicted mask. Extensive experiments on publicly available datasets
demonstrate the effectiveness of the proposed model.
| [
{
"version": "v1",
"created": "Sat, 25 Feb 2023 06:45:01 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Liu",
"Yaqi",
""
],
[
"Lv",
"Binbin",
""
],
[
"Jin",
"Xin",
""
],
[
"Chen",
"Xiaoyu",
""
],
[
"Zhang",
"Xiaokun",
""
]
] | TITLE: TBFormer: Two-Branch Transformer for Image Forgery Localization
ABSTRACT: Image forgery localization aims to identify forged regions by capturing
subtle traces from high-quality discriminative features. In this paper, we
propose a Transformer-style network with two feature extraction branches for
image forgery localization, and it is named as Two-Branch Transformer
(TBFormer). Firstly, two feature extraction branches are elaborately designed,
taking advantage of the discriminative stacked Transformer layers, for both RGB
and noise domain features. Secondly, an Attention-aware Hierarchical-feature
Fusion Module (AHFM) is proposed to effectively fuse hierarchical features from
two different domains. Although the two feature extraction branches have the
same architecture, their features have significant differences since they are
extracted from different domains. We adopt position attention to embed them
into a unified feature domain for hierarchical feature investigation. Finally,
a Transformer decoder is constructed for feature reconstruction to generate the
predicted mask. Extensive experiments on publicly available datasets
demonstrate the effectiveness of the proposed model.
|
2303.01801 | Ivan Maksymov | Ivan S. Maksymov and Andrey Pototsky | Reservoir computing based on solitary-like waves dynamics of film flows:
a proof of concept | null | null | 10.1209/0295-5075/acd471 | null | physics.flu-dyn cs.LG cs.NE nlin.PS | http://creativecommons.org/licenses/by/4.0/ | Several theoretical works have shown that solitons -- waves that
self-maintain constant shape and velocity as they propagate -- can be used as a
physical computational reservoir, a concept where machine learning algorithms
designed for digital computers are replaced by analog physical systems that
exhibit nonlinear dynamical behaviour. Here we propose and experimentally
validate a novel reservoir computing (RC) system that for the first time
employs solitary-like (SL) waves propagating on the surface of a liquid film
flowing over an inclined surface. We demonstrate the ability of the SL wave RC
system (SLRC) to forecast chaotic time series and to successfully pass
essential benchmark tests, including a memory capacity test and a Mackey-Glass
model test.
| [
{
"version": "v1",
"created": "Fri, 3 Mar 2023 09:17:53 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Maksymov",
"Ivan S.",
""
],
[
"Pototsky",
"Andrey",
""
]
] | TITLE: Reservoir computing based on solitary-like waves dynamics of film flows:
a proof of concept
ABSTRACT: Several theoretical works have shown that solitons -- waves that
self-maintain constant shape and velocity as they propagate -- can be used as a
physical computational reservoir, a concept where machine learning algorithms
designed for digital computers are replaced by analog physical systems that
exhibit nonlinear dynamical behaviour. Here we propose and experimentally
validate a novel reservoir computing (RC) system that for the first time
employs solitary-like (SL) waves propagating on the surface of a liquid film
flowing over an inclined surface. We demonstrate the ability of the SL wave RC
system (SLRC) to forecast chaotic time series and to successfully pass
essential benchmark tests, including a memory capacity test and a Mackey-Glass
model test.
|
2303.03315 | Antoine Gu\'edon | Antoine Gu\'edon, Tom Monnier, Pascal Monasse and Vincent Lepetit | MACARONS: Mapping And Coverage Anticipation with RGB Online
Self-Supervision | To appear at CVPR 2023. Project Webpage:
https://imagine.enpc.fr/~guedona/MACARONS/ | null | null | null | cs.CV cs.AI cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a method that simultaneously learns to explore new large
environments and to reconstruct them in 3D from color images only. This is
closely related to the Next Best View problem (NBV), where one has to identify
where to move the camera next to improve the coverage of an unknown scene.
However, most of the current NBV methods rely on depth sensors, need 3D
supervision and/or do not scale to large scenes. Our method requires only a
color camera and no 3D supervision. It simultaneously learns in a
self-supervised fashion to predict a "volume occupancy field" from color images
and, from this field, to predict the NBV. Thanks to this approach, our method
performs well on new scenes as it is not biased towards any training 3D data.
We demonstrate this on a recent dataset made of various 3D scenes and show it
performs even better than recent methods requiring a depth sensor, which is not
a realistic assumption for outdoor scenes captured with a flying drone.
| [
{
"version": "v1",
"created": "Mon, 6 Mar 2023 17:38:03 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 16:16:16 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Guédon",
"Antoine",
""
],
[
"Monnier",
"Tom",
""
],
[
"Monasse",
"Pascal",
""
],
[
"Lepetit",
"Vincent",
""
]
] | TITLE: MACARONS: Mapping And Coverage Anticipation with RGB Online
Self-Supervision
ABSTRACT: We introduce a method that simultaneously learns to explore new large
environments and to reconstruct them in 3D from color images only. This is
closely related to the Next Best View problem (NBV), where one has to identify
where to move the camera next to improve the coverage of an unknown scene.
However, most of the current NBV methods rely on depth sensors, need 3D
supervision and/or do not scale to large scenes. Our method requires only a
color camera and no 3D supervision. It simultaneously learns in a
self-supervised fashion to predict a "volume occupancy field" from color images
and, from this field, to predict the NBV. Thanks to this approach, our method
performs well on new scenes as it is not biased towards any training 3D data.
We demonstrate this on a recent dataset made of various 3D scenes and show it
performs even better than recent methods requiring a depth sensor, which is not
a realistic assumption for outdoor scenes captured with a flying drone.
|
2303.07284 | Bo He | Bo He, Jun Wang, Jielin Qiu, Trung Bui, Abhinav Shrivastava, Zhaowen
Wang | Align and Attend: Multimodal Summarization with Dual Contrastive Losses | Accepted at CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The goal of multimodal summarization is to extract the most important
information from different modalities to form output summaries. Unlike the
unimodal summarization, the multimodal summarization task explicitly leverages
cross-modal information to help generate more reliable and high-quality
summaries. However, existing methods fail to leverage the temporal
correspondence between different modalities and ignore the intrinsic
correlation between different samples. To address this issue, we introduce
Align and Attend Multimodal Summarization (A2Summ), a unified multimodal
transformer-based model which can effectively align and attend the multimodal
input. In addition, we propose two novel contrastive losses to model both
inter-sample and intra-sample correlations. Extensive experiments on two
standard video summarization datasets (TVSum and SumMe) and two multimodal
summarization datasets (Daily Mail and CNN) demonstrate the superiority of
A2Summ, achieving state-of-the-art performances on all datasets. Moreover, we
collected a large-scale multimodal summarization dataset BLiSS, which contains
livestream videos and transcribed texts with annotated summaries. Our code and
dataset are publicly available at ~\url{https://boheumd.github.io/A2Summ/}.
| [
{
"version": "v1",
"created": "Mon, 13 Mar 2023 17:01:42 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 01:11:09 GMT"
},
{
"version": "v3",
"created": "Mon, 12 Jun 2023 18:13:44 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"He",
"Bo",
""
],
[
"Wang",
"Jun",
""
],
[
"Qiu",
"Jielin",
""
],
[
"Bui",
"Trung",
""
],
[
"Shrivastava",
"Abhinav",
""
],
[
"Wang",
"Zhaowen",
""
]
] | TITLE: Align and Attend: Multimodal Summarization with Dual Contrastive Losses
ABSTRACT: The goal of multimodal summarization is to extract the most important
information from different modalities to form output summaries. Unlike the
unimodal summarization, the multimodal summarization task explicitly leverages
cross-modal information to help generate more reliable and high-quality
summaries. However, existing methods fail to leverage the temporal
correspondence between different modalities and ignore the intrinsic
correlation between different samples. To address this issue, we introduce
Align and Attend Multimodal Summarization (A2Summ), a unified multimodal
transformer-based model which can effectively align and attend the multimodal
input. In addition, we propose two novel contrastive losses to model both
inter-sample and intra-sample correlations. Extensive experiments on two
standard video summarization datasets (TVSum and SumMe) and two multimodal
summarization datasets (Daily Mail and CNN) demonstrate the superiority of
A2Summ, achieving state-of-the-art performances on all datasets. Moreover, we
collected a large-scale multimodal summarization dataset BLiSS, which contains
livestream videos and transcribed texts with annotated summaries. Our code and
dataset are publicly available at ~\url{https://boheumd.github.io/A2Summ/}.
|
2303.15693 | Masataka Kawai | Masataka Kawai, Noriaki Ota, Shinsuke Yamaoka | Large-scale pretraining on pathological images for fine-tuning of small
pathological benchmarks | 20 pages, 6 figures | null | null | null | cs.CV cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pretraining a deep learning model on large image datasets is a standard step
before fine-tuning the model on small targeted datasets. The large dataset is
usually general images (e.g. imagenet2012) while the small dataset can be
specialized datasets that have different distributions from the large dataset.
However, this 'large-to-small' strategy is not well-validated when the large
dataset is specialized and has a similar distribution to small datasets. We
newly compiled three hematoxylin and eosin-stained image datasets, one large
(PTCGA200) and two magnification-adjusted small datasets (PCam200 and
segPANDA200). Major deep learning models were trained with supervised and
self-supervised learning methods and fine-tuned on the small datasets for tumor
classification and tissue segmentation benchmarks. ResNet50 pretrained with
MoCov2, SimCLR, and BYOL on PTCGA200 was better than imagenet2012 pretraining
when fine-tuned on PTCGA200 (accuracy of 83.94%, 86.41%, 84.91%, and 82.72%,
respectively). ResNet50 pre-trained on PTCGA200 with MoCov2 exceeded the
COCOtrain2017-pretrained baseline and was the best in ResNet50 for the tissue
segmentation benchmark (mIoU of 63.53% and 63.22%). We found re-training
imagenet-pretrained models (ResNet50, BiT-M-R50x1, and ViT-S/16) on PTCGA200
improved downstream benchmarks.
| [
{
"version": "v1",
"created": "Tue, 28 Mar 2023 02:46:10 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 02:31:10 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Kawai",
"Masataka",
""
],
[
"Ota",
"Noriaki",
""
],
[
"Yamaoka",
"Shinsuke",
""
]
] | TITLE: Large-scale pretraining on pathological images for fine-tuning of small
pathological benchmarks
ABSTRACT: Pretraining a deep learning model on large image datasets is a standard step
before fine-tuning the model on small targeted datasets. The large dataset is
usually general images (e.g. imagenet2012) while the small dataset can be
specialized datasets that have different distributions from the large dataset.
However, this 'large-to-small' strategy is not well-validated when the large
dataset is specialized and has a similar distribution to small datasets. We
newly compiled three hematoxylin and eosin-stained image datasets, one large
(PTCGA200) and two magnification-adjusted small datasets (PCam200 and
segPANDA200). Major deep learning models were trained with supervised and
self-supervised learning methods and fine-tuned on the small datasets for tumor
classification and tissue segmentation benchmarks. ResNet50 pretrained with
MoCov2, SimCLR, and BYOL on PTCGA200 was better than imagenet2012 pretraining
when fine-tuned on PTCGA200 (accuracy of 83.94%, 86.41%, 84.91%, and 82.72%,
respectively). ResNet50 pre-trained on PTCGA200 with MoCov2 exceeded the
COCOtrain2017-pretrained baseline and was the best in ResNet50 for the tissue
segmentation benchmark (mIoU of 63.53% and 63.22%). We found re-training
imagenet-pretrained models (ResNet50, BiT-M-R50x1, and ViT-S/16) on PTCGA200
improved downstream benchmarks.
|
2303.16778 | Nazmus Sakib | Nazmus Sakib, G. M. Shahariar, Md. Mohsinul Kabir, Md. Kamrul Hasan
and Hasan Mahmud | Assorted, Archetypal and Annotated Two Million (3A2M) Cooking Recipes
Dataset based on Active Learning | null | International Conference on Machine Intelligence and Emerging
Technologies. MIET 2022. Lecture Notes of the Institute for Computer
Sciences, Social Informatics and Telecommunications Engineering, vol 491, pp
188-203, Springer, Cham | 10.1007/978-3-031-34622-4_15 | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Cooking recipes allow individuals to exchange culinary ideas and provide food
preparation instructions. Due to a lack of adequate labeled data, categorizing
raw recipes found online to the appropriate food genres is a challenging task
in this domain. Utilizing the knowledge of domain experts to categorize recipes
could be a solution. In this study, we present a novel dataset of two million
culinary recipes labeled in respective categories leveraging the knowledge of
food experts and an active learning technique. To construct the dataset, we
collect the recipes from the RecipeNLG dataset. Then, we employ three human
experts whose trustworthiness score is higher than 86.667% to categorize 300K
recipe by their Named Entity Recognition (NER) and assign it to one of the nine
categories: bakery, drinks, non-veg, vegetables, fast food, cereals, meals,
sides and fusion. Finally, we categorize the remaining 1900K recipes using
Active Learning method with a blend of Query-by-Committee and Human In The Loop
(HITL) approaches. There are more than two million recipes in our dataset, each
of which is categorized and has a confidence score linked with it. For the 9
genres, the Fleiss Kappa score of this massive dataset is roughly 0.56026. We
believe that the research community can use this dataset to perform various
machine learning tasks such as recipe genre classification, recipe generation
of a specific genre, new recipe creation, etc. The dataset can also be used to
train and evaluate the performance of various NLP tasks such as named entity
recognition, part-of-speech tagging, semantic role labeling, and so on. The
dataset will be available upon publication: https://tinyurl.com/3zu4778y.
| [
{
"version": "v1",
"created": "Mon, 27 Mar 2023 07:53:18 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Sakib",
"Nazmus",
""
],
[
"Shahariar",
"G. M.",
""
],
[
"Kabir",
"Md. Mohsinul",
""
],
[
"Hasan",
"Md. Kamrul",
""
],
[
"Mahmud",
"Hasan",
""
]
] | TITLE: Assorted, Archetypal and Annotated Two Million (3A2M) Cooking Recipes
Dataset based on Active Learning
ABSTRACT: Cooking recipes allow individuals to exchange culinary ideas and provide food
preparation instructions. Due to a lack of adequate labeled data, categorizing
raw recipes found online to the appropriate food genres is a challenging task
in this domain. Utilizing the knowledge of domain experts to categorize recipes
could be a solution. In this study, we present a novel dataset of two million
culinary recipes labeled in respective categories leveraging the knowledge of
food experts and an active learning technique. To construct the dataset, we
collect the recipes from the RecipeNLG dataset. Then, we employ three human
experts whose trustworthiness score is higher than 86.667% to categorize 300K
recipe by their Named Entity Recognition (NER) and assign it to one of the nine
categories: bakery, drinks, non-veg, vegetables, fast food, cereals, meals,
sides and fusion. Finally, we categorize the remaining 1900K recipes using
Active Learning method with a blend of Query-by-Committee and Human In The Loop
(HITL) approaches. There are more than two million recipes in our dataset, each
of which is categorized and has a confidence score linked with it. For the 9
genres, the Fleiss Kappa score of this massive dataset is roughly 0.56026. We
believe that the research community can use this dataset to perform various
machine learning tasks such as recipe genre classification, recipe generation
of a specific genre, new recipe creation, etc. The dataset can also be used to
train and evaluate the performance of various NLP tasks such as named entity
recognition, part-of-speech tagging, semantic role labeling, and so on. The
dataset will be available upon publication: https://tinyurl.com/3zu4778y.
|
2304.02721 | Daniel Campos | Daniel Campos, ChengXiang Zhai | To Asymmetry and Beyond: Structured Pruning of Sequence to Sequence
Models for Improved Inference Efficiency | SustaiNLP2023 @ ACL 2023,9 pages, 6 figures, 33 tables | null | null | null | cs.CL cs.AI | http://creativecommons.org/publicdomain/zero/1.0/ | Sequence-to-sequence language models can be used to produce abstractive
summaries which are coherent, relevant, and concise. Still, model sizes can
make deployment in latency-sensitive or web-scale implementations difficult.
This paper studies the relationship between model size, structured pruning,
inference efficiency, and summarization accuracy on widely used summarization
datasets. We show that model accuracy is tied to the encoder size while
inference efficiency is connected to the decoder. Using asymmetric pruning can
lead to nearly 3x improvement in inference latency with ~1 point loss in
Rouge-2. Moreover, we find both the average degradation and the role of
asymmetry to be consistent across model sizes and variations in datasets.
| [
{
"version": "v1",
"created": "Wed, 5 Apr 2023 19:44:20 GMT"
},
{
"version": "v2",
"created": "Sat, 3 Jun 2023 02:01:21 GMT"
},
{
"version": "v3",
"created": "Mon, 12 Jun 2023 21:13:14 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Campos",
"Daniel",
""
],
[
"Zhai",
"ChengXiang",
""
]
] | TITLE: To Asymmetry and Beyond: Structured Pruning of Sequence to Sequence
Models for Improved Inference Efficiency
ABSTRACT: Sequence-to-sequence language models can be used to produce abstractive
summaries which are coherent, relevant, and concise. Still, model sizes can
make deployment in latency-sensitive or web-scale implementations difficult.
This paper studies the relationship between model size, structured pruning,
inference efficiency, and summarization accuracy on widely used summarization
datasets. We show that model accuracy is tied to the encoder size while
inference efficiency is connected to the decoder. Using asymmetric pruning can
lead to nearly 3x improvement in inference latency with ~1 point loss in
Rouge-2. Moreover, we find both the average degradation and the role of
asymmetry to be consistent across model sizes and variations in datasets.
|
2304.03279 | Alexander Pan | Alexander Pan, Jun Shern Chan, Andy Zou, Nathaniel Li, Steven Basart,
Thomas Woodside, Jonathan Ng, Hanlin Zhang, Scott Emmons, Dan Hendrycks | Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards
and Ethical Behavior in the MACHIAVELLI Benchmark | ICML 2023 Oral (camera-ready); 31 pages, 5 figures | null | null | null | cs.LG cs.AI cs.CL cs.CY | http://creativecommons.org/licenses/by/4.0/ | Artificial agents have traditionally been trained to maximize reward, which
may incentivize power-seeking and deception, analogous to how next-token
prediction in language models (LMs) may incentivize toxicity. So do agents
naturally learn to be Machiavellian? And how do we measure these behaviors in
general-purpose models such as GPT-4? Towards answering these questions, we
introduce MACHIAVELLI, a benchmark of 134 Choose-Your-Own-Adventure games
containing over half a million rich, diverse scenarios that center on social
decision-making. Scenario labeling is automated with LMs, which are more
performant than human annotators. We mathematize dozens of harmful behaviors
and use our annotations to evaluate agents' tendencies to be power-seeking,
cause disutility, and commit ethical violations. We observe some tension
between maximizing reward and behaving ethically. To improve this trade-off, we
investigate LM-based methods to steer agents' towards less harmful behaviors.
Our results show that agents can both act competently and morally, so concrete
progress can currently be made in machine ethics--designing agents that are
Pareto improvements in both safety and capabilities.
| [
{
"version": "v1",
"created": "Thu, 6 Apr 2023 17:59:03 GMT"
},
{
"version": "v2",
"created": "Mon, 1 May 2023 22:58:44 GMT"
},
{
"version": "v3",
"created": "Thu, 8 Jun 2023 02:04:23 GMT"
},
{
"version": "v4",
"created": "Tue, 13 Jun 2023 01:01:42 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Pan",
"Alexander",
""
],
[
"Chan",
"Jun Shern",
""
],
[
"Zou",
"Andy",
""
],
[
"Li",
"Nathaniel",
""
],
[
"Basart",
"Steven",
""
],
[
"Woodside",
"Thomas",
""
],
[
"Ng",
"Jonathan",
""
],
[
"Zhang",
"Hanlin",
""
],
[
"Emmons",
"Scott",
""
],
[
"Hendrycks",
"Dan",
""
]
] | TITLE: Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards
and Ethical Behavior in the MACHIAVELLI Benchmark
ABSTRACT: Artificial agents have traditionally been trained to maximize reward, which
may incentivize power-seeking and deception, analogous to how next-token
prediction in language models (LMs) may incentivize toxicity. So do agents
naturally learn to be Machiavellian? And how do we measure these behaviors in
general-purpose models such as GPT-4? Towards answering these questions, we
introduce MACHIAVELLI, a benchmark of 134 Choose-Your-Own-Adventure games
containing over half a million rich, diverse scenarios that center on social
decision-making. Scenario labeling is automated with LMs, which are more
performant than human annotators. We mathematize dozens of harmful behaviors
and use our annotations to evaluate agents' tendencies to be power-seeking,
cause disutility, and commit ethical violations. We observe some tension
between maximizing reward and behaving ethically. To improve this trade-off, we
investigate LM-based methods to steer agents' towards less harmful behaviors.
Our results show that agents can both act competently and morally, so concrete
progress can currently be made in machine ethics--designing agents that are
Pareto improvements in both safety and capabilities.
|
2304.05047 | Abdullah-Al-Zubaer Imran | Attiano Purpura-Pontoniere, Demetri Terzopoulos, Adam Wang,
Abdullah-Al-Zubaer Imran | Semi-Supervised Relational Contrastive Learning | 10 pages, 5 figures, 2 tables | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Disease diagnosis from medical images via supervised learning is usually
dependent on tedious, error-prone, and costly image labeling by medical
experts. Alternatively, semi-supervised learning and self-supervised learning
offer effectiveness through the acquisition of valuable insights from readily
available unlabeled images. We present Semi-Supervised Relational Contrastive
Learning (SRCL), a novel semi-supervised learning model that leverages
self-supervised contrastive loss and sample relation consistency for the more
meaningful and effective exploitation of unlabeled data. Our experimentation
with the SRCL model explores both pre-train/fine-tune and joint learning of the
pretext (contrastive learning) and downstream (diagnostic classification)
tasks. We validate against the ISIC 2018 Challenge benchmark skin lesion
classification dataset and demonstrate the effectiveness of our semi-supervised
method on varying amounts of labeled data.
| [
{
"version": "v1",
"created": "Tue, 11 Apr 2023 08:14:30 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 06:34:37 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Purpura-Pontoniere",
"Attiano",
""
],
[
"Terzopoulos",
"Demetri",
""
],
[
"Wang",
"Adam",
""
],
[
"Imran",
"Abdullah-Al-Zubaer",
""
]
] | TITLE: Semi-Supervised Relational Contrastive Learning
ABSTRACT: Disease diagnosis from medical images via supervised learning is usually
dependent on tedious, error-prone, and costly image labeling by medical
experts. Alternatively, semi-supervised learning and self-supervised learning
offer effectiveness through the acquisition of valuable insights from readily
available unlabeled images. We present Semi-Supervised Relational Contrastive
Learning (SRCL), a novel semi-supervised learning model that leverages
self-supervised contrastive loss and sample relation consistency for the more
meaningful and effective exploitation of unlabeled data. Our experimentation
with the SRCL model explores both pre-train/fine-tune and joint learning of the
pretext (contrastive learning) and downstream (diagnostic classification)
tasks. We validate against the ISIC 2018 Challenge benchmark skin lesion
classification dataset and demonstrate the effectiveness of our semi-supervised
method on varying amounts of labeled data.
|
2304.09252 | Md Hasibul Amin | Md Hasibul Amin, Mohammed E. Elbtity and Ramtin Zand | IMAC-Sim: A Circuit-level Simulator For In-Memory Analog Computing
Architectures | null | Proceedings of the Great Lakes Symposium on VLSI 2023 (GLSVLSI
'23), Association for Computing Machinery, New York, NY, USA, 659-664 | 10.1145/3583781.3590264 | null | cs.ET cs.AR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the increased attention to memristive-based in-memory analog computing
(IMAC) architectures as an alternative for energy-hungry computer systems for
machine learning applications, a tool that enables exploring their device- and
circuit-level design space can significantly boost the research and development
in this area. Thus, in this paper, we develop IMAC-Sim, a circuit-level
simulator for the design space exploration of IMAC architectures. IMAC-Sim is a
Python-based simulation framework, which creates the SPICE netlist of the IMAC
circuit based on various device- and circuit-level hyperparameters selected by
the user, and automatically evaluates the accuracy, power consumption, and
latency of the developed circuit using a user-specified dataset. Moreover,
IMAC-Sim simulates the interconnect parasitic resistance and capacitance in the
IMAC architectures and is also equipped with horizontal and vertical
partitioning techniques to surmount these reliability challenges. IMAC-Sim is a
flexible tool that supports a broad range of device- and circuit-level
hyperparameters. In this paper, we perform controlled experiments to exhibit
some of the important capabilities of the IMAC-Sim, while the entirety of its
features is available for researchers via an open-source tool.
| [
{
"version": "v1",
"created": "Tue, 18 Apr 2023 19:22:34 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Amin",
"Md Hasibul",
""
],
[
"Elbtity",
"Mohammed E.",
""
],
[
"Zand",
"Ramtin",
""
]
] | TITLE: IMAC-Sim: A Circuit-level Simulator For In-Memory Analog Computing
Architectures
ABSTRACT: With the increased attention to memristive-based in-memory analog computing
(IMAC) architectures as an alternative for energy-hungry computer systems for
machine learning applications, a tool that enables exploring their device- and
circuit-level design space can significantly boost the research and development
in this area. Thus, in this paper, we develop IMAC-Sim, a circuit-level
simulator for the design space exploration of IMAC architectures. IMAC-Sim is a
Python-based simulation framework, which creates the SPICE netlist of the IMAC
circuit based on various device- and circuit-level hyperparameters selected by
the user, and automatically evaluates the accuracy, power consumption, and
latency of the developed circuit using a user-specified dataset. Moreover,
IMAC-Sim simulates the interconnect parasitic resistance and capacitance in the
IMAC architectures and is also equipped with horizontal and vertical
partitioning techniques to surmount these reliability challenges. IMAC-Sim is a
flexible tool that supports a broad range of device- and circuit-level
hyperparameters. In this paper, we perform controlled experiments to exhibit
some of the important capabilities of the IMAC-Sim, while the entirety of its
features is available for researchers via an open-source tool.
|
2305.02549 | Chen-Yu Lee | Chen-Yu Lee, Chun-Liang Li, Hao Zhang, Timothy Dozat, Vincent Perot,
Guolong Su, Xiang Zhang, Kihyuk Sohn, Nikolai Glushnev, Renshen Wang, Joshua
Ainslie, Shangbang Long, Siyang Qin, Yasuhisa Fujii, Nan Hua, Tomas Pfister | FormNetV2: Multimodal Graph Contrastive Learning for Form Document
Information Extraction | Accepted to ACL 2023 | null | null | null | cs.CL cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | The recent advent of self-supervised pre-training techniques has led to a
surge in the use of multimodal learning in form document understanding.
However, existing approaches that extend the mask language modeling to other
modalities require careful multi-task tuning, complex reconstruction target
designs, or additional pre-training data. In FormNetV2, we introduce a
centralized multimodal graph contrastive learning strategy to unify
self-supervised pre-training for all modalities in one loss. The graph
contrastive objective maximizes the agreement of multimodal representations,
providing a natural interplay for all modalities without special customization.
In addition, we extract image features within the bounding box that joins a
pair of tokens connected by a graph edge, capturing more targeted visual cues
without loading a sophisticated and separately pre-trained image embedder.
FormNetV2 establishes new state-of-the-art performance on FUNSD, CORD, SROIE
and Payment benchmarks with a more compact model size.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 05:02:04 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 04:27:14 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Lee",
"Chen-Yu",
""
],
[
"Li",
"Chun-Liang",
""
],
[
"Zhang",
"Hao",
""
],
[
"Dozat",
"Timothy",
""
],
[
"Perot",
"Vincent",
""
],
[
"Su",
"Guolong",
""
],
[
"Zhang",
"Xiang",
""
],
[
"Sohn",
"Kihyuk",
""
],
[
"Glushnev",
"Nikolai",
""
],
[
"Wang",
"Renshen",
""
],
[
"Ainslie",
"Joshua",
""
],
[
"Long",
"Shangbang",
""
],
[
"Qin",
"Siyang",
""
],
[
"Fujii",
"Yasuhisa",
""
],
[
"Hua",
"Nan",
""
],
[
"Pfister",
"Tomas",
""
]
] | TITLE: FormNetV2: Multimodal Graph Contrastive Learning for Form Document
Information Extraction
ABSTRACT: The recent advent of self-supervised pre-training techniques has led to a
surge in the use of multimodal learning in form document understanding.
However, existing approaches that extend the mask language modeling to other
modalities require careful multi-task tuning, complex reconstruction target
designs, or additional pre-training data. In FormNetV2, we introduce a
centralized multimodal graph contrastive learning strategy to unify
self-supervised pre-training for all modalities in one loss. The graph
contrastive objective maximizes the agreement of multimodal representations,
providing a natural interplay for all modalities without special customization.
In addition, we extract image features within the bounding box that joins a
pair of tokens connected by a graph edge, capturing more targeted visual cues
without loading a sophisticated and separately pre-trained image embedder.
FormNetV2 establishes new state-of-the-art performance on FUNSD, CORD, SROIE
and Payment benchmarks with a more compact model size.
|
2305.03956 | Sanghyun Kim | Sanghyun Kim and Jiwon Seo | Machine-Learning-Based Classification of GPS Signal Reception Conditions
Using a Dual-Polarized Antenna in Urban Areas | Submitted to IEEE ION PLANS 2023 | null | 10.1109/PLANS53410.2023.10140036 | null | cs.LG eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In urban areas, dense buildings frequently block and reflect global
positioning system (GPS) signals, resulting in the reception of a few visible
satellites with many multipath signals. This is a significant problem that
results in unreliable positioning in urban areas. If a signal reception
condition from a certain satellite can be detected, the positioning performance
can be improved by excluding or de-weighting the multipath contaminated
satellite signal. Thus, we developed a machine-learning-based method of
classifying GPS signal reception conditions using a dual-polarized antenna. We
employed a decision tree algorithm for classification using three features, one
of which can be obtained only from a dual-polarized antenna. A machine-learning
model was trained using GPS signals collected from various locations. When the
features extracted from the GPS raw signal are input, the generated
machine-learning model outputs one of the three signal reception conditions:
non-line-of-sight (NLOS) only, line-of-sight (LOS) only, or LOS+NLOS. Multiple
testing datasets were used to analyze the classification accuracy, which was
then compared with an existing method using dual single-polarized antennas.
Consequently, when the testing dataset was collected at different locations
from the training dataset, a classification accuracy of 64.47% was obtained,
which was slightly higher than the accuracy of the existing method using dual
single-polarized antennas. Therefore, the dual-polarized antenna solution is
more beneficial than the dual single-polarized antenna solution because it has
a more compact form factor and its performance is similar to that of the other
solution.
| [
{
"version": "v1",
"created": "Sat, 6 May 2023 06:45:47 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Kim",
"Sanghyun",
""
],
[
"Seo",
"Jiwon",
""
]
] | TITLE: Machine-Learning-Based Classification of GPS Signal Reception Conditions
Using a Dual-Polarized Antenna in Urban Areas
ABSTRACT: In urban areas, dense buildings frequently block and reflect global
positioning system (GPS) signals, resulting in the reception of a few visible
satellites with many multipath signals. This is a significant problem that
results in unreliable positioning in urban areas. If a signal reception
condition from a certain satellite can be detected, the positioning performance
can be improved by excluding or de-weighting the multipath contaminated
satellite signal. Thus, we developed a machine-learning-based method of
classifying GPS signal reception conditions using a dual-polarized antenna. We
employed a decision tree algorithm for classification using three features, one
of which can be obtained only from a dual-polarized antenna. A machine-learning
model was trained using GPS signals collected from various locations. When the
features extracted from the GPS raw signal are input, the generated
machine-learning model outputs one of the three signal reception conditions:
non-line-of-sight (NLOS) only, line-of-sight (LOS) only, or LOS+NLOS. Multiple
testing datasets were used to analyze the classification accuracy, which was
then compared with an existing method using dual single-polarized antennas.
Consequently, when the testing dataset was collected at different locations
from the training dataset, a classification accuracy of 64.47% was obtained,
which was slightly higher than the accuracy of the existing method using dual
single-polarized antennas. Therefore, the dual-polarized antenna solution is
more beneficial than the dual single-polarized antenna solution because it has
a more compact form factor and its performance is similar to that of the other
solution.
|
2305.04790 | Tao Gong | Tao Gong, Chengqi Lyu, Shilong Zhang, Yudong Wang, Miao Zheng, Qian
Zhao, Kuikun Liu, Wenwei Zhang, Ping Luo, Kai Chen | MultiModal-GPT: A Vision and Language Model for Dialogue with Humans | 10 pages, 8 figures | null | null | null | cs.CV cs.CL | http://creativecommons.org/licenses/by/4.0/ | We present a vision and language model named MultiModal-GPT to conduct
multi-round dialogue with humans. MultiModal-GPT can follow various
instructions from humans, such as generating a detailed caption, counting the
number of interested objects, and answering general questions from users.
MultiModal-GPT is parameter-efficiently fine-tuned from OpenFlamingo, with
Low-rank Adapter (LoRA) added both in the cross-attention part and the
self-attention part of the language model. We first construct instruction
templates with vision and language data for multi-modality instruction tuning
to make the model understand and follow human instructions. We find the quality
of training data is vital for the dialogue performance, where few data
containing short answers can lead the model to respond shortly to any
instructions. To further enhance the ability to chat with humans of the
MultiModal-GPT, we utilize language-only instruction-following data to train
the MultiModal-GPT jointly. The joint training of language-only and
visual-language instructions with the \emph{same} instruction template
effectively improves dialogue performance. Various demos show the ability of
continuous dialogue of MultiModal-GPT with humans. Code, dataset, and demo are
at https://github.com/open-mmlab/Multimodal-GPT
| [
{
"version": "v1",
"created": "Mon, 8 May 2023 15:45:42 GMT"
},
{
"version": "v2",
"created": "Tue, 9 May 2023 11:41:53 GMT"
},
{
"version": "v3",
"created": "Tue, 13 Jun 2023 13:31:12 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Gong",
"Tao",
""
],
[
"Lyu",
"Chengqi",
""
],
[
"Zhang",
"Shilong",
""
],
[
"Wang",
"Yudong",
""
],
[
"Zheng",
"Miao",
""
],
[
"Zhao",
"Qian",
""
],
[
"Liu",
"Kuikun",
""
],
[
"Zhang",
"Wenwei",
""
],
[
"Luo",
"Ping",
""
],
[
"Chen",
"Kai",
""
]
] | TITLE: MultiModal-GPT: A Vision and Language Model for Dialogue with Humans
ABSTRACT: We present a vision and language model named MultiModal-GPT to conduct
multi-round dialogue with humans. MultiModal-GPT can follow various
instructions from humans, such as generating a detailed caption, counting the
number of interested objects, and answering general questions from users.
MultiModal-GPT is parameter-efficiently fine-tuned from OpenFlamingo, with
Low-rank Adapter (LoRA) added both in the cross-attention part and the
self-attention part of the language model. We first construct instruction
templates with vision and language data for multi-modality instruction tuning
to make the model understand and follow human instructions. We find the quality
of training data is vital for the dialogue performance, where few data
containing short answers can lead the model to respond shortly to any
instructions. To further enhance the ability to chat with humans of the
MultiModal-GPT, we utilize language-only instruction-following data to train
the MultiModal-GPT jointly. The joint training of language-only and
visual-language instructions with the \emph{same} instruction template
effectively improves dialogue performance. Various demos show the ability of
continuous dialogue of MultiModal-GPT with humans. Code, dataset, and demo are
at https://github.com/open-mmlab/Multimodal-GPT
|
2305.05374 | Yuxiang Zhao | Yuxiang Zhao, Zhuomin Chai, Yibo Lin, Runsheng Wang, Ru Huang | HybridNet: Dual-Branch Fusion of Geometrical and Topological Views for
VLSI Congestion Prediction | null | 2023 IEEE International Symposium of EDA | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate early congestion prediction can prevent unpleasant surprises at the
routing stage, playing a crucial character in assisting designers to iterate
faster in VLSI design cycles. In this paper, we introduce a novel strategy to
fully incorporate topological and geometrical features of circuits by making
several key designs in our network architecture. To be more specific, we
construct two individual graphs (geometry-graph, topology-graph) with distinct
edge construction schemes according to their unique properties. We then propose
a dual-branch network with different encoder layers in each pathway and
aggregate representations with a sophisticated fusion strategy. Our network,
named HybridNet, not only provides a simple yet effective way to capture the
geometric interactions of cells, but also preserves the original topological
relationships in the netlist. Experimental results on the ISPD2015 benchmarks
show that we achieve an improvement of 10.9% compared to previous methods.
| [
{
"version": "v1",
"created": "Sun, 7 May 2023 15:02:02 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 02:50:12 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Zhao",
"Yuxiang",
""
],
[
"Chai",
"Zhuomin",
""
],
[
"Lin",
"Yibo",
""
],
[
"Wang",
"Runsheng",
""
],
[
"Huang",
"Ru",
""
]
] | TITLE: HybridNet: Dual-Branch Fusion of Geometrical and Topological Views for
VLSI Congestion Prediction
ABSTRACT: Accurate early congestion prediction can prevent unpleasant surprises at the
routing stage, playing a crucial character in assisting designers to iterate
faster in VLSI design cycles. In this paper, we introduce a novel strategy to
fully incorporate topological and geometrical features of circuits by making
several key designs in our network architecture. To be more specific, we
construct two individual graphs (geometry-graph, topology-graph) with distinct
edge construction schemes according to their unique properties. We then propose
a dual-branch network with different encoder layers in each pathway and
aggregate representations with a sophisticated fusion strategy. Our network,
named HybridNet, not only provides a simple yet effective way to capture the
geometric interactions of cells, but also preserves the original topological
relationships in the netlist. Experimental results on the ISPD2015 benchmarks
show that we achieve an improvement of 10.9% compared to previous methods.
|
2305.09996 | YeCong Wan | Ye-Cong Wan, Ming-Wen Shao, Yuan-Shuo Cheng, Yue-Xian Liu, Zhi-Yuan
Bao | Restoring Images Captured in Arbitrary Hybrid Adverse Weather Conditions
in One Go | In submission | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Adverse conditions typically suffer from stochastic hybrid weather
degradations (e.g., rainy and hazy night), while existing image restoration
algorithms envisage that weather degradations occur independently, thus may
fail to handle real-world complicated scenarios. Besides, supervised training
is not feasible due to the lack of a comprehensive paired dataset to
characterize hybrid conditions. To this end, we have advanced the
aforementioned limitations with two tactics: framework and data. First, we
present a novel unified framework, dubbed RAHC, to Restore Arbitrary Hybrid
adverse weather Conditions in one go. Specifically, our RAHC leverages a
multi-head aggregation architecture to learn multiple degradation
representation subspaces and then constrains the network to flexibly handle
multiple hybrid adverse weather in a unified paradigm through a discrimination
mechanism in the output space. Furthermore, we devise a reconstruction vectors
aided scheme to provide auxiliary visual content cues for reconstruction, thus
can comfortably cope with hybrid scenarios with insufficient remaining image
constituents. Second, we construct a new dataset, termed HAC, for learning and
benchmarking arbitrary Hybrid Adverse Conditions restoration. HAC contains 31
scenarios composed of an arbitrary combination of five common weather, with a
total of ~316K adverse-weather/clean pairs. Extensive experiments yield
superior results and establish new state-of-the-art results on both HAC and
conventional datasets.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 06:42:42 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 05:57:07 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Wan",
"Ye-Cong",
""
],
[
"Shao",
"Ming-Wen",
""
],
[
"Cheng",
"Yuan-Shuo",
""
],
[
"Liu",
"Yue-Xian",
""
],
[
"Bao",
"Zhi-Yuan",
""
]
] | TITLE: Restoring Images Captured in Arbitrary Hybrid Adverse Weather Conditions
in One Go
ABSTRACT: Adverse conditions typically suffer from stochastic hybrid weather
degradations (e.g., rainy and hazy night), while existing image restoration
algorithms envisage that weather degradations occur independently, thus may
fail to handle real-world complicated scenarios. Besides, supervised training
is not feasible due to the lack of a comprehensive paired dataset to
characterize hybrid conditions. To this end, we have advanced the
aforementioned limitations with two tactics: framework and data. First, we
present a novel unified framework, dubbed RAHC, to Restore Arbitrary Hybrid
adverse weather Conditions in one go. Specifically, our RAHC leverages a
multi-head aggregation architecture to learn multiple degradation
representation subspaces and then constrains the network to flexibly handle
multiple hybrid adverse weather in a unified paradigm through a discrimination
mechanism in the output space. Furthermore, we devise a reconstruction vectors
aided scheme to provide auxiliary visual content cues for reconstruction, thus
can comfortably cope with hybrid scenarios with insufficient remaining image
constituents. Second, we construct a new dataset, termed HAC, for learning and
benchmarking arbitrary Hybrid Adverse Conditions restoration. HAC contains 31
scenarios composed of an arbitrary combination of five common weather, with a
total of ~316K adverse-weather/clean pairs. Extensive experiments yield
superior results and establish new state-of-the-art results on both HAC and
conventional datasets.
|
2305.12331 | Shubo Lv | Shubo Lv, Xiong Wang, Sining Sun, Long Ma, Lei Xie | DCCRN-KWS: an audio bias based model for noise robust small-footprint
keyword spotting | Accepted by INTERSPEECH2023 | null | null | null | eess.AS cs.SD | http://creativecommons.org/licenses/by/4.0/ | Real-world complex acoustic environments especially the ones with a low
signal-to-noise ratio (SNR) will bring tremendous challenges to a keyword
spotting (KWS) system. Inspired by the recent advances of neural speech
enhancement and context bias in speech recognition, we propose a robust audio
context bias based DCCRN-KWS model to address this challenge. We form the whole
architecture as a multi-task learning framework for both denosing and keyword
spotting, where the DCCRN encoder is connected with the KWS model. Helped with
the denoising task, we further introduce an audio context bias module to
leverage the real keyword samples and bias the network to better iscriminate
keywords in noisy conditions. Feature merge and complex context linear modules
are also introduced to strength such discrimination and to effectively leverage
contextual information respectively. Experiments on the internal challenging
dataset and the HIMIYA public dataset show that our DCCRN-KWS system is
superior in performance, while ablation study demonstrates the good design of
the whole model.
| [
{
"version": "v1",
"created": "Sun, 21 May 2023 03:36:56 GMT"
},
{
"version": "v2",
"created": "Tue, 23 May 2023 03:21:37 GMT"
},
{
"version": "v3",
"created": "Tue, 13 Jun 2023 01:07:16 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Lv",
"Shubo",
""
],
[
"Wang",
"Xiong",
""
],
[
"Sun",
"Sining",
""
],
[
"Ma",
"Long",
""
],
[
"Xie",
"Lei",
""
]
] | TITLE: DCCRN-KWS: an audio bias based model for noise robust small-footprint
keyword spotting
ABSTRACT: Real-world complex acoustic environments especially the ones with a low
signal-to-noise ratio (SNR) will bring tremendous challenges to a keyword
spotting (KWS) system. Inspired by the recent advances of neural speech
enhancement and context bias in speech recognition, we propose a robust audio
context bias based DCCRN-KWS model to address this challenge. We form the whole
architecture as a multi-task learning framework for both denosing and keyword
spotting, where the DCCRN encoder is connected with the KWS model. Helped with
the denoising task, we further introduce an audio context bias module to
leverage the real keyword samples and bias the network to better iscriminate
keywords in noisy conditions. Feature merge and complex context linear modules
are also introduced to strength such discrimination and to effectively leverage
contextual information respectively. Experiments on the internal challenging
dataset and the HIMIYA public dataset show that our DCCRN-KWS system is
superior in performance, while ablation study demonstrates the good design of
the whole model.
|
2305.14839 | Yunshui Li | Yunshui Li, Binyuan Hui, ZhiChao Yin, Min Yang, Fei Huang and Yongbin
Li | PaCE: Unified Multi-modal Dialogue Pre-training with Progressive and
Compositional Experts | ACL 2023 | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Perceiving multi-modal information and fulfilling dialogues with humans is a
long-term goal of artificial intelligence. Pre-training is commonly regarded as
an effective approach for multi-modal dialogue. However, due to the limited
availability of multi-modal dialogue data, there is still scarce research on
multi-modal dialogue pre-training. Yet another intriguing challenge emerges
from the encompassing nature of multi-modal dialogue, which involves various
modalities and tasks. Moreover, new forms of tasks may arise at unpredictable
points in the future. Hence, it is essential for designed multi-modal dialogue
models to possess sufficient flexibility to adapt to such scenarios. This paper
proposes \textbf{PaCE}, a unified, structured, compositional multi-modal
dialogue pre-training framework. It utilizes a combination of several
fundamental experts to accommodate multiple dialogue-related tasks and can be
pre-trained using limited dialogue and extensive non-dialogue multi-modal data.
Furthermore, we propose a progressive training method where old experts from
the past can assist new experts, facilitating the expansion of their
capabilities. Experimental results demonstrate that PaCE achieves
state-of-the-art results on eight multi-modal dialog benchmarks.
| [
{
"version": "v1",
"created": "Wed, 24 May 2023 07:43:29 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 06:31:46 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Li",
"Yunshui",
""
],
[
"Hui",
"Binyuan",
""
],
[
"Yin",
"ZhiChao",
""
],
[
"Yang",
"Min",
""
],
[
"Huang",
"Fei",
""
],
[
"Li",
"Yongbin",
""
]
] | TITLE: PaCE: Unified Multi-modal Dialogue Pre-training with Progressive and
Compositional Experts
ABSTRACT: Perceiving multi-modal information and fulfilling dialogues with humans is a
long-term goal of artificial intelligence. Pre-training is commonly regarded as
an effective approach for multi-modal dialogue. However, due to the limited
availability of multi-modal dialogue data, there is still scarce research on
multi-modal dialogue pre-training. Yet another intriguing challenge emerges
from the encompassing nature of multi-modal dialogue, which involves various
modalities and tasks. Moreover, new forms of tasks may arise at unpredictable
points in the future. Hence, it is essential for designed multi-modal dialogue
models to possess sufficient flexibility to adapt to such scenarios. This paper
proposes \textbf{PaCE}, a unified, structured, compositional multi-modal
dialogue pre-training framework. It utilizes a combination of several
fundamental experts to accommodate multiple dialogue-related tasks and can be
pre-trained using limited dialogue and extensive non-dialogue multi-modal data.
Furthermore, we propose a progressive training method where old experts from
the past can assist new experts, facilitating the expansion of their
capabilities. Experimental results demonstrate that PaCE achieves
state-of-the-art results on eight multi-modal dialog benchmarks.
|
2305.19395 | Yuchen Zhuang | Yuchen Zhuang, Yue Yu, Lingkai Kong, Xiang Chen, Chao Zhang | DyGen: Learning from Noisy Labels via Dynamics-Enhanced Generative
Modeling | Accepted by KDD 2023 research track | null | 10.1145/3580305.3599318 | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning from noisy labels is a challenge that arises in many real-world
applications where training data can contain incorrect or corrupted labels.
When fine-tuning language models with noisy labels, models can easily overfit
the label noise, leading to decreased performance. Most existing methods for
learning from noisy labels use static input features for denoising, but these
methods are limited by the information they can provide on true label
distributions and can result in biased or incorrect predictions. In this work,
we propose the Dynamics-Enhanced Generative Model (DyGen), which uses dynamic
patterns in the embedding space during the fine-tuning process of language
models to improve noisy label predictions. DyGen uses the variational
auto-encoding framework to infer the posterior distributions of true labels
from noisy labels and training dynamics. Additionally, a co-regularization
mechanism is used to minimize the impact of potentially noisy labels and
priors. DyGen demonstrates an average accuracy improvement of 3.10% on two
synthetic noise datasets and 1.48% on three real-world noise datasets compared
to the previous state-of-the-art. Extensive experiments and analyses show the
effectiveness of each component in DyGen. Our code is available for
reproducibility on GitHub.
| [
{
"version": "v1",
"created": "Tue, 30 May 2023 20:19:41 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 05:14:45 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Zhuang",
"Yuchen",
""
],
[
"Yu",
"Yue",
""
],
[
"Kong",
"Lingkai",
""
],
[
"Chen",
"Xiang",
""
],
[
"Zhang",
"Chao",
""
]
] | TITLE: DyGen: Learning from Noisy Labels via Dynamics-Enhanced Generative
Modeling
ABSTRACT: Learning from noisy labels is a challenge that arises in many real-world
applications where training data can contain incorrect or corrupted labels.
When fine-tuning language models with noisy labels, models can easily overfit
the label noise, leading to decreased performance. Most existing methods for
learning from noisy labels use static input features for denoising, but these
methods are limited by the information they can provide on true label
distributions and can result in biased or incorrect predictions. In this work,
we propose the Dynamics-Enhanced Generative Model (DyGen), which uses dynamic
patterns in the embedding space during the fine-tuning process of language
models to improve noisy label predictions. DyGen uses the variational
auto-encoding framework to infer the posterior distributions of true labels
from noisy labels and training dynamics. Additionally, a co-regularization
mechanism is used to minimize the impact of potentially noisy labels and
priors. DyGen demonstrates an average accuracy improvement of 3.10% on two
synthetic noise datasets and 1.48% on three real-world noise datasets compared
to the previous state-of-the-art. Extensive experiments and analyses show the
effectiveness of each component in DyGen. Our code is available for
reproducibility on GitHub.
|
2305.20011 | Virginia Aglietti | Virginia Aglietti, Alan Malek, Ira Ktena, Silvia Chiappa | Constrained Causal Bayesian Optimization | null | International Conference on Machine Learning, 2023 | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose constrained causal Bayesian optimization (cCBO), an approach for
finding interventions in a known causal graph that optimize a target variable
under some constraints. cCBO first reduces the search space by exploiting the
graph structure and, if available, an observational dataset; and then solves
the restricted optimization problem by modelling target and constraint
quantities using Gaussian processes and by sequentially selecting interventions
via a constrained expected improvement acquisition function. We propose
different surrogate models that enable to integrate observational and
interventional data while capturing correlation among effects with increasing
levels of sophistication. We evaluate cCBO on artificial and real-world causal
graphs showing successful trade off between fast convergence and percentage of
feasible interventions.
| [
{
"version": "v1",
"created": "Wed, 31 May 2023 16:34:58 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Aglietti",
"Virginia",
""
],
[
"Malek",
"Alan",
""
],
[
"Ktena",
"Ira",
""
],
[
"Chiappa",
"Silvia",
""
]
] | TITLE: Constrained Causal Bayesian Optimization
ABSTRACT: We propose constrained causal Bayesian optimization (cCBO), an approach for
finding interventions in a known causal graph that optimize a target variable
under some constraints. cCBO first reduces the search space by exploiting the
graph structure and, if available, an observational dataset; and then solves
the restricted optimization problem by modelling target and constraint
quantities using Gaussian processes and by sequentially selecting interventions
via a constrained expected improvement acquisition function. We propose
different surrogate models that enable to integrate observational and
interventional data while capturing correlation among effects with increasing
levels of sophistication. We evaluate cCBO on artificial and real-world causal
graphs showing successful trade off between fast convergence and percentage of
feasible interventions.
|
2306.04192 | Shiqian Zhao | Shiqian Zhao, Kangjie Chen, Meng Hao, Jian Zhang, Guowen Xu, Hongwei
Li, Tianwei Zhang | Extracting Cloud-based Model with Prior Knowledge | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine Learning-as-a-Service, a pay-as-you-go business pattern, is widely
accepted by third-party users and developers. However, the open inference APIs
may be utilized by malicious customers to conduct model extraction attacks,
i.e., attackers can replicate a cloud-based black-box model merely via querying
malicious examples. Existing model extraction attacks mainly depend on the
posterior knowledge (i.e., predictions of query samples) from Oracle. Thus,
they either require high query overhead to simulate the decision boundary, or
suffer from generalization errors and overfitting problems due to query budget
limitations. To mitigate it, this work proposes an efficient model extraction
attack based on prior knowledge for the first time. The insight is that prior
knowledge of unlabeled proxy datasets is conducive to the search for the
decision boundary (e.g., informative samples). Specifically, we leverage
self-supervised learning including autoencoder and contrastive learning to
pre-compile the prior knowledge of the proxy dataset into the feature extractor
of the substitute model. Then we adopt entropy to measure and sample the most
informative examples to query the target model. Our design leverages both prior
and posterior knowledge to extract the model and thus eliminates
generalizability errors and overfitting problems. We conduct extensive
experiments on open APIs like Traffic Recognition, Flower Recognition,
Moderation Recognition, and NSFW Recognition from real-world platforms, Azure
and Clarifai. The experimental results demonstrate the effectiveness and
efficiency of our attack. For example, our attack achieves 95.1% fidelity with
merely 1.8K queries (cost 2.16$) on the NSFW Recognition API. Also, the
adversarial examples generated with our substitute model have better
transferability than others, which reveals that our scheme is more conducive to
downstream attacks.
| [
{
"version": "v1",
"created": "Wed, 7 Jun 2023 07:00:02 GMT"
},
{
"version": "v2",
"created": "Thu, 8 Jun 2023 04:27:17 GMT"
},
{
"version": "v3",
"created": "Fri, 9 Jun 2023 12:35:18 GMT"
},
{
"version": "v4",
"created": "Tue, 13 Jun 2023 07:55:15 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Zhao",
"Shiqian",
""
],
[
"Chen",
"Kangjie",
""
],
[
"Hao",
"Meng",
""
],
[
"Zhang",
"Jian",
""
],
[
"Xu",
"Guowen",
""
],
[
"Li",
"Hongwei",
""
],
[
"Zhang",
"Tianwei",
""
]
] | TITLE: Extracting Cloud-based Model with Prior Knowledge
ABSTRACT: Machine Learning-as-a-Service, a pay-as-you-go business pattern, is widely
accepted by third-party users and developers. However, the open inference APIs
may be utilized by malicious customers to conduct model extraction attacks,
i.e., attackers can replicate a cloud-based black-box model merely via querying
malicious examples. Existing model extraction attacks mainly depend on the
posterior knowledge (i.e., predictions of query samples) from Oracle. Thus,
they either require high query overhead to simulate the decision boundary, or
suffer from generalization errors and overfitting problems due to query budget
limitations. To mitigate it, this work proposes an efficient model extraction
attack based on prior knowledge for the first time. The insight is that prior
knowledge of unlabeled proxy datasets is conducive to the search for the
decision boundary (e.g., informative samples). Specifically, we leverage
self-supervised learning including autoencoder and contrastive learning to
pre-compile the prior knowledge of the proxy dataset into the feature extractor
of the substitute model. Then we adopt entropy to measure and sample the most
informative examples to query the target model. Our design leverages both prior
and posterior knowledge to extract the model and thus eliminates
generalizability errors and overfitting problems. We conduct extensive
experiments on open APIs like Traffic Recognition, Flower Recognition,
Moderation Recognition, and NSFW Recognition from real-world platforms, Azure
and Clarifai. The experimental results demonstrate the effectiveness and
efficiency of our attack. For example, our attack achieves 95.1% fidelity with
merely 1.8K queries (cost 2.16$) on the NSFW Recognition API. Also, the
adversarial examples generated with our substitute model have better
transferability than others, which reveals that our scheme is more conducive to
downstream attacks.
|
2306.04455 | Rolf Jagerman | Zhen Qin, Rolf Jagerman, Rama Pasumarthi, Honglei Zhuang, He Zhang,
Aijun Bai, Kai Hui, Le Yan, Xuanhui Wang | RD-Suite: A Benchmark for Ranking Distillation | 15 pages, 2 figures. arXiv admin note: text overlap with
arXiv:2011.04006 by other authors | null | null | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | The distillation of ranking models has become an important topic in both
academia and industry. In recent years, several advanced methods have been
proposed to tackle this problem, often leveraging ranking information from
teacher rankers that is absent in traditional classification settings. To date,
there is no well-established consensus on how to evaluate this class of models.
Moreover, inconsistent benchmarking on a wide range of tasks and datasets make
it difficult to assess or invigorate advances in this field. This paper first
examines representative prior arts on ranking distillation, and raises three
questions to be answered around methodology and reproducibility. To that end,
we propose a systematic and unified benchmark, Ranking Distillation Suite
(RD-Suite), which is a suite of tasks with 4 large real-world datasets,
encompassing two major modalities (textual and numeric) and two applications
(standard distillation and distillation transfer). RD-Suite consists of
benchmark results that challenge some of the common wisdom in the field, and
the release of datasets with teacher scores and evaluation scripts for future
research. RD-Suite paves the way towards better understanding of ranking
distillation, facilities more research in this direction, and presents new
challenges.
| [
{
"version": "v1",
"created": "Wed, 7 Jun 2023 14:29:12 GMT"
},
{
"version": "v2",
"created": "Mon, 12 Jun 2023 09:53:56 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Qin",
"Zhen",
""
],
[
"Jagerman",
"Rolf",
""
],
[
"Pasumarthi",
"Rama",
""
],
[
"Zhuang",
"Honglei",
""
],
[
"Zhang",
"He",
""
],
[
"Bai",
"Aijun",
""
],
[
"Hui",
"Kai",
""
],
[
"Yan",
"Le",
""
],
[
"Wang",
"Xuanhui",
""
]
] | TITLE: RD-Suite: A Benchmark for Ranking Distillation
ABSTRACT: The distillation of ranking models has become an important topic in both
academia and industry. In recent years, several advanced methods have been
proposed to tackle this problem, often leveraging ranking information from
teacher rankers that is absent in traditional classification settings. To date,
there is no well-established consensus on how to evaluate this class of models.
Moreover, inconsistent benchmarking on a wide range of tasks and datasets make
it difficult to assess or invigorate advances in this field. This paper first
examines representative prior arts on ranking distillation, and raises three
questions to be answered around methodology and reproducibility. To that end,
we propose a systematic and unified benchmark, Ranking Distillation Suite
(RD-Suite), which is a suite of tasks with 4 large real-world datasets,
encompassing two major modalities (textual and numeric) and two applications
(standard distillation and distillation transfer). RD-Suite consists of
benchmark results that challenge some of the common wisdom in the field, and
the release of datasets with teacher scores and evaluation scripts for future
research. RD-Suite paves the way towards better understanding of ranking
distillation, facilities more research in this direction, and presents new
challenges.
|
2306.04911 | Jungwuk Park | Jungwuk Park, Dong-Jun Han, Soyeong Kim, Jaekyun Moon | Test-Time Style Shifting: Handling Arbitrary Styles in Domain
Generalization | ICML 2023 camera-ready version | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In domain generalization (DG), the target domain is unknown when the model is
being trained, and the trained model should successfully work on an arbitrary
(and possibly unseen) target domain during inference. This is a difficult
problem, and despite active studies in recent years, it remains a great
challenge. In this paper, we take a simple yet effective approach to tackle
this issue. We propose test-time style shifting, which shifts the style of the
test sample (that has a large style gap with the source domains) to the nearest
source domain that the model is already familiar with, before making the
prediction. This strategy enables the model to handle any target domains with
arbitrary style statistics, without additional model update at test-time.
Additionally, we propose style balancing, which provides a great platform for
maximizing the advantage of test-time style shifting by handling the
DG-specific imbalance issues. The proposed ideas are easy to implement and
successfully work in conjunction with various other DG schemes. Experimental
results on different datasets show the effectiveness of our methods.
| [
{
"version": "v1",
"created": "Thu, 8 Jun 2023 03:26:16 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 00:37:33 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Park",
"Jungwuk",
""
],
[
"Han",
"Dong-Jun",
""
],
[
"Kim",
"Soyeong",
""
],
[
"Moon",
"Jaekyun",
""
]
] | TITLE: Test-Time Style Shifting: Handling Arbitrary Styles in Domain
Generalization
ABSTRACT: In domain generalization (DG), the target domain is unknown when the model is
being trained, and the trained model should successfully work on an arbitrary
(and possibly unseen) target domain during inference. This is a difficult
problem, and despite active studies in recent years, it remains a great
challenge. In this paper, we take a simple yet effective approach to tackle
this issue. We propose test-time style shifting, which shifts the style of the
test sample (that has a large style gap with the source domains) to the nearest
source domain that the model is already familiar with, before making the
prediction. This strategy enables the model to handle any target domains with
arbitrary style statistics, without additional model update at test-time.
Additionally, we propose style balancing, which provides a great platform for
maximizing the advantage of test-time style shifting by handling the
DG-specific imbalance issues. The proposed ideas are easy to implement and
successfully work in conjunction with various other DG schemes. Experimental
results on different datasets show the effectiveness of our methods.
|
2306.05083 | Cheng-Han Chiang | Cheng-Han Chiang, Yung-Sung Chuang, James Glass, Hung-yi Lee | Revealing the Blind Spot of Sentence Encoder Evaluation by HEROS | ACL 2023 repl4nlp (representation learning for NLP) workshop poster
paper. Dataset at https://huggingface.co/datasets/dcml0714/Heros | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Existing sentence textual similarity benchmark datasets only use a single
number to summarize how similar the sentence encoder's decision is to humans'.
However, it is unclear what kind of sentence pairs a sentence encoder (SE)
would consider similar. Moreover, existing SE benchmarks mainly consider
sentence pairs with low lexical overlap, so it is unclear how the SEs behave
when two sentences have high lexical overlap. We introduce a high-quality SE
diagnostic dataset, HEROS. HEROS is constructed by transforming an original
sentence into a new sentence based on certain rules to form a \textit{minimal
pair}, and the minimal pair has high lexical overlaps. The rules include
replacing a word with a synonym, an antonym, a typo, a random word, and
converting the original sentence into its negation. Different rules yield
different subsets of HEROS. By systematically comparing the performance of over
60 supervised and unsupervised SEs on HEROS, we reveal that most unsupervised
sentence encoders are insensitive to negation. We find the datasets used to
train the SE are the main determinants of what kind of sentence pairs an SE
considers similar. We also show that even if two SEs have similar performance
on STS benchmarks, they can have very different behavior on HEROS. Our result
reveals the blind spot of traditional STS benchmarks when evaluating SEs.
| [
{
"version": "v1",
"created": "Thu, 8 Jun 2023 10:24:02 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 04:56:06 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Chiang",
"Cheng-Han",
""
],
[
"Chuang",
"Yung-Sung",
""
],
[
"Glass",
"James",
""
],
[
"Lee",
"Hung-yi",
""
]
] | TITLE: Revealing the Blind Spot of Sentence Encoder Evaluation by HEROS
ABSTRACT: Existing sentence textual similarity benchmark datasets only use a single
number to summarize how similar the sentence encoder's decision is to humans'.
However, it is unclear what kind of sentence pairs a sentence encoder (SE)
would consider similar. Moreover, existing SE benchmarks mainly consider
sentence pairs with low lexical overlap, so it is unclear how the SEs behave
when two sentences have high lexical overlap. We introduce a high-quality SE
diagnostic dataset, HEROS. HEROS is constructed by transforming an original
sentence into a new sentence based on certain rules to form a \textit{minimal
pair}, and the minimal pair has high lexical overlaps. The rules include
replacing a word with a synonym, an antonym, a typo, a random word, and
converting the original sentence into its negation. Different rules yield
different subsets of HEROS. By systematically comparing the performance of over
60 supervised and unsupervised SEs on HEROS, we reveal that most unsupervised
sentence encoders are insensitive to negation. We find the datasets used to
train the SE are the main determinants of what kind of sentence pairs an SE
considers similar. We also show that even if two SEs have similar performance
on STS benchmarks, they can have very different behavior on HEROS. Our result
reveals the blind spot of traditional STS benchmarks when evaluating SEs.
|
2306.05785 | Anshul Nasery | Anshul Nasery, Hardik Shah, Arun Sai Suggala, Prateek Jain | End-to-End Neural Network Compression via $\frac{\ell_1}{\ell_2}$
Regularized Latency Surrogates | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Neural network (NN) compression via techniques such as pruning, quantization
requires setting compression hyperparameters (e.g., number of channels to be
pruned, bitwidths for quantization) for each layer either manually or via
neural architecture search (NAS) which can be computationally expensive. We
address this problem by providing an end-to-end technique that optimizes for
model's Floating Point Operations (FLOPs) or for on-device latency via a novel
$\frac{\ell_1}{\ell_2}$ latency surrogate. Our algorithm is versatile and can
be used with many popular compression methods including pruning, low-rank
factorization, and quantization. Crucially, it is fast and runs in almost the
same amount of time as single model training; which is a significant training
speed-up over standard NAS methods. For BERT compression on GLUE fine-tuning
tasks, we achieve $50\%$ reduction in FLOPs with only $1\%$ drop in
performance. For compressing MobileNetV3 on ImageNet-1K, we achieve $15\%$
reduction in FLOPs, and $11\%$ reduction in on-device latency without drop in
accuracy, while still requiring $3\times$ less training compute than SOTA
compression techniques. Finally, for transfer learning on smaller datasets, our
technique identifies $1.2\times$-$1.4\times$ cheaper architectures than
standard MobileNetV3, EfficientNet suite of architectures at almost the same
training cost and accuracy.
| [
{
"version": "v1",
"created": "Fri, 9 Jun 2023 09:57:17 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 06:41:47 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Nasery",
"Anshul",
""
],
[
"Shah",
"Hardik",
""
],
[
"Suggala",
"Arun Sai",
""
],
[
"Jain",
"Prateek",
""
]
] | TITLE: End-to-End Neural Network Compression via $\frac{\ell_1}{\ell_2}$
Regularized Latency Surrogates
ABSTRACT: Neural network (NN) compression via techniques such as pruning, quantization
requires setting compression hyperparameters (e.g., number of channels to be
pruned, bitwidths for quantization) for each layer either manually or via
neural architecture search (NAS) which can be computationally expensive. We
address this problem by providing an end-to-end technique that optimizes for
model's Floating Point Operations (FLOPs) or for on-device latency via a novel
$\frac{\ell_1}{\ell_2}$ latency surrogate. Our algorithm is versatile and can
be used with many popular compression methods including pruning, low-rank
factorization, and quantization. Crucially, it is fast and runs in almost the
same amount of time as single model training; which is a significant training
speed-up over standard NAS methods. For BERT compression on GLUE fine-tuning
tasks, we achieve $50\%$ reduction in FLOPs with only $1\%$ drop in
performance. For compressing MobileNetV3 on ImageNet-1K, we achieve $15\%$
reduction in FLOPs, and $11\%$ reduction in on-device latency without drop in
accuracy, while still requiring $3\times$ less training compute than SOTA
compression techniques. Finally, for transfer learning on smaller datasets, our
technique identifies $1.2\times$-$1.4\times$ cheaper architectures than
standard MobileNetV3, EfficientNet suite of architectures at almost the same
training cost and accuracy.
|
2306.06193 | Dan Ley | Dan Ley, Leonard Tang, Matthew Nazari, Hongjin Lin, Suraj Srinivas,
Himabindu Lakkaraju | Consistent Explanations in the Face of Model Indeterminacy via
Ensembling | null | null | null | null | cs.LG cs.AI cs.CY | http://creativecommons.org/licenses/by/4.0/ | This work addresses the challenge of providing consistent explanations for
predictive models in the presence of model indeterminacy, which arises due to
the existence of multiple (nearly) equally well-performing models for a given
dataset and task. Despite their similar performance, such models often exhibit
inconsistent or even contradictory explanations for their predictions, posing
challenges to end users who rely on these models to make critical decisions.
Recognizing this issue, we introduce ensemble methods as an approach to enhance
the consistency of the explanations provided in these scenarios. Leveraging
insights from recent work on neural network loss landscapes and mode
connectivity, we devise ensemble strategies to efficiently explore the
underspecification set -- the set of models with performance variations
resulting solely from changes in the random seed during training. Experiments
on five benchmark financial datasets reveal that ensembling can yield
significant improvements when it comes to explanation similarity, and
demonstrate the potential of existing ensemble methods to explore the
underspecification set efficiently. Our findings highlight the importance of
considering model indeterminacy when interpreting explanations and showcase the
effectiveness of ensembles in enhancing the reliability of explanations in
machine learning.
| [
{
"version": "v1",
"created": "Fri, 9 Jun 2023 18:45:43 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 02:04:21 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Ley",
"Dan",
""
],
[
"Tang",
"Leonard",
""
],
[
"Nazari",
"Matthew",
""
],
[
"Lin",
"Hongjin",
""
],
[
"Srinivas",
"Suraj",
""
],
[
"Lakkaraju",
"Himabindu",
""
]
] | TITLE: Consistent Explanations in the Face of Model Indeterminacy via
Ensembling
ABSTRACT: This work addresses the challenge of providing consistent explanations for
predictive models in the presence of model indeterminacy, which arises due to
the existence of multiple (nearly) equally well-performing models for a given
dataset and task. Despite their similar performance, such models often exhibit
inconsistent or even contradictory explanations for their predictions, posing
challenges to end users who rely on these models to make critical decisions.
Recognizing this issue, we introduce ensemble methods as an approach to enhance
the consistency of the explanations provided in these scenarios. Leveraging
insights from recent work on neural network loss landscapes and mode
connectivity, we devise ensemble strategies to efficiently explore the
underspecification set -- the set of models with performance variations
resulting solely from changes in the random seed during training. Experiments
on five benchmark financial datasets reveal that ensembling can yield
significant improvements when it comes to explanation similarity, and
demonstrate the potential of existing ensemble methods to explore the
underspecification set efficiently. Our findings highlight the importance of
considering model indeterminacy when interpreting explanations and showcase the
effectiveness of ensembles in enhancing the reliability of explanations in
machine learning.
|
2306.06265 | Donghao Li | Donghao Li, Ruiquan Huang, Cong Shen, Jing Yang | Near-optimal Conservative Exploration in Reinforcement Learning under
Episode-wise Constraints | Accepted by ICML2023 | null | null | null | cs.LG cs.IT math.IT stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates conservative exploration in reinforcement learning
where the performance of the learning agent is guaranteed to be above a certain
threshold throughout the learning process. It focuses on the tabular episodic
Markov Decision Process (MDP) setting that has finite states and actions. With
the knowledge of an existing safe baseline policy, an algorithm termed as
StepMix is proposed to balance the exploitation and exploration while ensuring
that the conservative constraint is never violated in each episode with high
probability. StepMix features a unique design of a mixture policy that
adaptively and smoothly interpolates between the baseline policy and the
optimistic policy. Theoretical analysis shows that StepMix achieves
near-optimal regret order as in the constraint-free setting, indicating that
obeying the stringent episode-wise conservative constraint does not compromise
the learning performance. Besides, a randomization-based EpsMix algorithm is
also proposed and shown to achieve the same performance as StepMix. The
algorithm design and theoretical analysis are further extended to the setting
where the baseline policy is not given a priori but must be learned from an
offline dataset, and it is proved that similar conservative guarantee and
regret can be achieved if the offline dataset is sufficiently large. Experiment
results corroborate the theoretical analysis and demonstrate the effectiveness
of the proposed conservative exploration strategies.
| [
{
"version": "v1",
"created": "Fri, 9 Jun 2023 21:26:57 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Li",
"Donghao",
""
],
[
"Huang",
"Ruiquan",
""
],
[
"Shen",
"Cong",
""
],
[
"Yang",
"Jing",
""
]
] | TITLE: Near-optimal Conservative Exploration in Reinforcement Learning under
Episode-wise Constraints
ABSTRACT: This paper investigates conservative exploration in reinforcement learning
where the performance of the learning agent is guaranteed to be above a certain
threshold throughout the learning process. It focuses on the tabular episodic
Markov Decision Process (MDP) setting that has finite states and actions. With
the knowledge of an existing safe baseline policy, an algorithm termed as
StepMix is proposed to balance the exploitation and exploration while ensuring
that the conservative constraint is never violated in each episode with high
probability. StepMix features a unique design of a mixture policy that
adaptively and smoothly interpolates between the baseline policy and the
optimistic policy. Theoretical analysis shows that StepMix achieves
near-optimal regret order as in the constraint-free setting, indicating that
obeying the stringent episode-wise conservative constraint does not compromise
the learning performance. Besides, a randomization-based EpsMix algorithm is
also proposed and shown to achieve the same performance as StepMix. The
algorithm design and theoretical analysis are further extended to the setting
where the baseline policy is not given a priori but must be learned from an
offline dataset, and it is proved that similar conservative guarantee and
regret can be achieved if the offline dataset is sufficiently large. Experiment
results corroborate the theoretical analysis and demonstrate the effectiveness
of the proposed conservative exploration strategies.
|
2306.06362 | Xiaqing Pan | Xiaqing Pan, Nicholas Charron, Yongqian Yang, Scott Peters, Thomas
Whelan, Chen Kong, Omkar Parkhi, Richard Newcombe, Carl Yuheng Ren | Aria Digital Twin: A New Benchmark Dataset for Egocentric 3D Machine
Perception | null | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We introduce the Aria Digital Twin (ADT) - an egocentric dataset captured
using Aria glasses with extensive object, environment, and human level ground
truth. This ADT release contains 200 sequences of real-world activities
conducted by Aria wearers in two real indoor scenes with 398 object instances
(324 stationary and 74 dynamic). Each sequence consists of: a) raw data of two
monochrome camera streams, one RGB camera stream, two IMU streams; b) complete
sensor calibration; c) ground truth data including continuous
6-degree-of-freedom (6DoF) poses of the Aria devices, object 6DoF poses, 3D eye
gaze vectors, 3D human poses, 2D image segmentations, image depth maps; and d)
photo-realistic synthetic renderings. To the best of our knowledge, there is no
existing egocentric dataset with a level of accuracy, photo-realism and
comprehensiveness comparable to ADT. By contributing ADT to the research
community, our mission is to set a new standard for evaluation in the
egocentric machine perception domain, which includes very challenging research
problems such as 3D object detection and tracking, scene reconstruction and
understanding, sim-to-real learning, human pose prediction - while also
inspiring new machine perception tasks for augmented reality (AR) applications.
To kick start exploration of the ADT research use cases, we evaluated several
existing state-of-the-art methods for object detection, segmentation and image
translation tasks that demonstrate the usefulness of ADT as a benchmarking
dataset.
| [
{
"version": "v1",
"created": "Sat, 10 Jun 2023 06:46:32 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 06:38:47 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Pan",
"Xiaqing",
""
],
[
"Charron",
"Nicholas",
""
],
[
"Yang",
"Yongqian",
""
],
[
"Peters",
"Scott",
""
],
[
"Whelan",
"Thomas",
""
],
[
"Kong",
"Chen",
""
],
[
"Parkhi",
"Omkar",
""
],
[
"Newcombe",
"Richard",
""
],
[
"Ren",
"Carl Yuheng",
""
]
] | TITLE: Aria Digital Twin: A New Benchmark Dataset for Egocentric 3D Machine
Perception
ABSTRACT: We introduce the Aria Digital Twin (ADT) - an egocentric dataset captured
using Aria glasses with extensive object, environment, and human level ground
truth. This ADT release contains 200 sequences of real-world activities
conducted by Aria wearers in two real indoor scenes with 398 object instances
(324 stationary and 74 dynamic). Each sequence consists of: a) raw data of two
monochrome camera streams, one RGB camera stream, two IMU streams; b) complete
sensor calibration; c) ground truth data including continuous
6-degree-of-freedom (6DoF) poses of the Aria devices, object 6DoF poses, 3D eye
gaze vectors, 3D human poses, 2D image segmentations, image depth maps; and d)
photo-realistic synthetic renderings. To the best of our knowledge, there is no
existing egocentric dataset with a level of accuracy, photo-realism and
comprehensiveness comparable to ADT. By contributing ADT to the research
community, our mission is to set a new standard for evaluation in the
egocentric machine perception domain, which includes very challenging research
problems such as 3D object detection and tracking, scene reconstruction and
understanding, sim-to-real learning, human pose prediction - while also
inspiring new machine perception tasks for augmented reality (AR) applications.
To kick start exploration of the ADT research use cases, we evaluated several
existing state-of-the-art methods for object detection, segmentation and image
translation tasks that demonstrate the usefulness of ADT as a benchmarking
dataset.
|
2306.06819 | Anderson Avila | Anderson R. Avila, Mehdi Rezagholizadeh, Chao Xing | Multimodal Audio-textual Architecture for Robust Spoken Language
Understanding | null | null | null | null | cs.CL cs.LG eess.AS | http://creativecommons.org/licenses/by/4.0/ | Recent voice assistants are usually based on the cascade spoken language
understanding (SLU) solution, which consists of an automatic speech recognition
(ASR) engine and a natural language understanding (NLU) system. Because such
approach relies on the ASR output, it often suffers from the so-called ASR
error propagation. In this work, we investigate impacts of this ASR error
propagation on state-of-the-art NLU systems based on pre-trained language
models (PLM), such as BERT and RoBERTa. Moreover, a multimodal language
understanding (MLU) module is proposed to mitigate SLU performance degradation
caused by errors present in the ASR transcript. The MLU benefits from
self-supervised features learned from both audio and text modalities,
specifically Wav2Vec for speech and Bert/RoBERTa for language. Our MLU combines
an encoder network to embed the audio signal and a text encoder to process text
transcripts followed by a late fusion layer to fuse audio and text logits. We
found that the proposed MLU showed to be robust towards poor quality ASR
transcripts, while the performance of BERT and RoBERTa are severely
compromised. Our model is evaluated on five tasks from three SLU datasets and
robustness is tested using ASR transcripts from three ASR engines. Results show
that the proposed approach effectively mitigates the ASR error propagation
problem, surpassing the PLM models' performance across all datasets for the
academic ASR engine.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 01:55:53 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 15:41:11 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Avila",
"Anderson R.",
""
],
[
"Rezagholizadeh",
"Mehdi",
""
],
[
"Xing",
"Chao",
""
]
] | TITLE: Multimodal Audio-textual Architecture for Robust Spoken Language
Understanding
ABSTRACT: Recent voice assistants are usually based on the cascade spoken language
understanding (SLU) solution, which consists of an automatic speech recognition
(ASR) engine and a natural language understanding (NLU) system. Because such
approach relies on the ASR output, it often suffers from the so-called ASR
error propagation. In this work, we investigate impacts of this ASR error
propagation on state-of-the-art NLU systems based on pre-trained language
models (PLM), such as BERT and RoBERTa. Moreover, a multimodal language
understanding (MLU) module is proposed to mitigate SLU performance degradation
caused by errors present in the ASR transcript. The MLU benefits from
self-supervised features learned from both audio and text modalities,
specifically Wav2Vec for speech and Bert/RoBERTa for language. Our MLU combines
an encoder network to embed the audio signal and a text encoder to process text
transcripts followed by a late fusion layer to fuse audio and text logits. We
found that the proposed MLU showed to be robust towards poor quality ASR
transcripts, while the performance of BERT and RoBERTa are severely
compromised. Our model is evaluated on five tasks from three SLU datasets and
robustness is tested using ASR transcripts from three ASR engines. Results show
that the proposed approach effectively mitigates the ASR error propagation
problem, surpassing the PLM models' performance across all datasets for the
academic ASR engine.
|
2306.07201 | Guian Fang | Ziyang Ma, Mengsha Liu, Guian Fang, Ying Shen | LTCR: Long-Text Chinese Rumor Detection Dataset | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | False information can spread quickly on social media, negatively influencing
the citizens' behaviors and responses to social events. To better detect all of
the fake news, especially long texts which are harder to find completely, a
Long-Text Chinese Rumor detection dataset named LTCR is proposed. The LTCR
dataset provides a valuable resource for accurately detecting misinformation,
especially in the context of complex fake news related to COVID-19. The dataset
consists of 1,729 and 500 pieces of real and fake news, respectively. The
average lengths of real and fake news are approximately 230 and 152 characters.
We also propose \method, Salience-aware Fake News Detection Model, which
achieves the highest accuracy (95.85%), fake news recall (90.91%) and F-score
(90.60%) on the dataset. (https://github.com/Enderfga/DoubleCheck)
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 16:03:36 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 08:08:18 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Ma",
"Ziyang",
""
],
[
"Liu",
"Mengsha",
""
],
[
"Fang",
"Guian",
""
],
[
"Shen",
"Ying",
""
]
] | TITLE: LTCR: Long-Text Chinese Rumor Detection Dataset
ABSTRACT: False information can spread quickly on social media, negatively influencing
the citizens' behaviors and responses to social events. To better detect all of
the fake news, especially long texts which are harder to find completely, a
Long-Text Chinese Rumor detection dataset named LTCR is proposed. The LTCR
dataset provides a valuable resource for accurately detecting misinformation,
especially in the context of complex fake news related to COVID-19. The dataset
consists of 1,729 and 500 pieces of real and fake news, respectively. The
average lengths of real and fake news are approximately 230 and 152 characters.
We also propose \method, Salience-aware Fake News Detection Model, which
achieves the highest accuracy (95.85%), fake news recall (90.91%) and F-score
(90.60%) on the dataset. (https://github.com/Enderfga/DoubleCheck)
|
2306.07265 | Shilong Liu | Tianhe Ren, Shilong Liu, Feng Li, Hao Zhang, Ailing Zeng, Jie Yang,
Xingyu Liao, Ding Jia, Hongyang Li, He Cao, Jianan Wang, Zhaoyang Zeng,
Xianbiao Qi, Yuhui Yuan, Jianwei Yang, Lei Zhang | detrex: Benchmarking Detection Transformers | project link: https://github.com/IDEA-Research/detrex | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The DEtection TRansformer (DETR) algorithm has received considerable
attention in the research community and is gradually emerging as a mainstream
approach for object detection and other perception tasks. However, the current
field lacks a unified and comprehensive benchmark specifically tailored for
DETR-based models. To address this issue, we develop a unified, highly modular,
and lightweight codebase called detrex, which supports a majority of the
mainstream DETR-based instance recognition algorithms, covering various
fundamental tasks, including object detection, segmentation, and pose
estimation. We conduct extensive experiments under detrex and perform a
comprehensive benchmark for DETR-based models. Moreover, we enhance the
performance of detection transformers through the refinement of training
hyper-parameters, providing strong baselines for supported algorithms.We hope
that detrex could offer research communities a standardized and unified
platform to evaluate and compare different DETR-based models while fostering a
deeper understanding and driving advancements in DETR-based instance
recognition. Our code is available at https://github.com/IDEA-Research/detrex.
The project is currently being actively developed. We encourage the community
to use detrex codebase for further development and contributions.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 17:52:11 GMT"
},
{
"version": "v2",
"created": "Tue, 13 Jun 2023 17:53:15 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Ren",
"Tianhe",
""
],
[
"Liu",
"Shilong",
""
],
[
"Li",
"Feng",
""
],
[
"Zhang",
"Hao",
""
],
[
"Zeng",
"Ailing",
""
],
[
"Yang",
"Jie",
""
],
[
"Liao",
"Xingyu",
""
],
[
"Jia",
"Ding",
""
],
[
"Li",
"Hongyang",
""
],
[
"Cao",
"He",
""
],
[
"Wang",
"Jianan",
""
],
[
"Zeng",
"Zhaoyang",
""
],
[
"Qi",
"Xianbiao",
""
],
[
"Yuan",
"Yuhui",
""
],
[
"Yang",
"Jianwei",
""
],
[
"Zhang",
"Lei",
""
]
] | TITLE: detrex: Benchmarking Detection Transformers
ABSTRACT: The DEtection TRansformer (DETR) algorithm has received considerable
attention in the research community and is gradually emerging as a mainstream
approach for object detection and other perception tasks. However, the current
field lacks a unified and comprehensive benchmark specifically tailored for
DETR-based models. To address this issue, we develop a unified, highly modular,
and lightweight codebase called detrex, which supports a majority of the
mainstream DETR-based instance recognition algorithms, covering various
fundamental tasks, including object detection, segmentation, and pose
estimation. We conduct extensive experiments under detrex and perform a
comprehensive benchmark for DETR-based models. Moreover, we enhance the
performance of detection transformers through the refinement of training
hyper-parameters, providing strong baselines for supported algorithms.We hope
that detrex could offer research communities a standardized and unified
platform to evaluate and compare different DETR-based models while fostering a
deeper understanding and driving advancements in DETR-based instance
recognition. Our code is available at https://github.com/IDEA-Research/detrex.
The project is currently being actively developed. We encourage the community
to use detrex codebase for further development and contributions.
|
2306.07290 | Bogdan Mazoure | Bogdan Mazoure, Walter Talbott, Miguel Angel Bautista, Devon Hjelm,
Alexander Toshev, Josh Susskind | Value function estimation using conditional diffusion models for control | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | A fairly reliable trend in deep reinforcement learning is that the
performance scales with the number of parameters, provided a complimentary
scaling in amount of training data. As the appetite for large models increases,
it is imperative to address, sooner than later, the potential problem of
running out of high-quality demonstrations. In this case, instead of collecting
only new data via costly human demonstrations or risking a simulation-to-real
transfer with uncertain effects, it would be beneficial to leverage vast
amounts of readily-available low-quality data. Since classical control
algorithms such as behavior cloning or temporal difference learning cannot be
used on reward-free or action-free data out-of-the-box, this solution warrants
novel training paradigms for continuous control. We propose a simple algorithm
called Diffused Value Function (DVF), which learns a joint multi-step model of
the environment-robot interaction dynamics using a diffusion model. This model
can be efficiently learned from state sequences (i.e., without access to reward
functions nor actions), and subsequently used to estimate the value of each
action out-of-the-box. We show how DVF can be used to efficiently capture the
state visitation measure for multiple controllers, and show promising
qualitative and quantitative results on challenging robotics benchmarks.
| [
{
"version": "v1",
"created": "Fri, 9 Jun 2023 18:40:55 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Mazoure",
"Bogdan",
""
],
[
"Talbott",
"Walter",
""
],
[
"Bautista",
"Miguel Angel",
""
],
[
"Hjelm",
"Devon",
""
],
[
"Toshev",
"Alexander",
""
],
[
"Susskind",
"Josh",
""
]
] | TITLE: Value function estimation using conditional diffusion models for control
ABSTRACT: A fairly reliable trend in deep reinforcement learning is that the
performance scales with the number of parameters, provided a complimentary
scaling in amount of training data. As the appetite for large models increases,
it is imperative to address, sooner than later, the potential problem of
running out of high-quality demonstrations. In this case, instead of collecting
only new data via costly human demonstrations or risking a simulation-to-real
transfer with uncertain effects, it would be beneficial to leverage vast
amounts of readily-available low-quality data. Since classical control
algorithms such as behavior cloning or temporal difference learning cannot be
used on reward-free or action-free data out-of-the-box, this solution warrants
novel training paradigms for continuous control. We propose a simple algorithm
called Diffused Value Function (DVF), which learns a joint multi-step model of
the environment-robot interaction dynamics using a diffusion model. This model
can be efficiently learned from state sequences (i.e., without access to reward
functions nor actions), and subsequently used to estimate the value of each
action out-of-the-box. We show how DVF can be used to efficiently capture the
state visitation measure for multiple controllers, and show promising
qualitative and quantitative results on challenging robotics benchmarks.
|
2306.07296 | Andri Pranolo | Andri Pranolo, Yingchi Mao, Aji Prasetya Wibawa, Agung Bella Putra
Utama, Felix Andika Dwiyanto | Optimized Three Deep Learning Models Based-PSO Hyperparameters for
Beijing PM2.5 Prediction | Volume 5 (1): 53-66 | Knowledge Engineering and Data Science, 2022, Vol 5 No 1, pp:
53-66 | 10.17977/um018v5i12022p53-66 | null | cs.LG cs.AI cs.NE | http://creativecommons.org/licenses/by-sa/4.0/ | Deep learning is a machine learning approach that produces excellent
performance in various applications, including natural language processing,
image identification, and forecasting. Deep learning network performance
depends on the hyperparameter settings. This research attempts to optimize the
deep learning architecture of Long short term memory (LSTM), Convolutional
neural network (CNN), and Multilayer perceptron (MLP) for forecasting tasks
using Particle swarm optimization (PSO), a swarm intelligence-based
metaheuristic optimization methodology: Proposed M-1 (PSO-LSTM), M-2 (PSO-CNN),
and M-3 (PSO-MLP). Beijing PM2.5 datasets was analyzed to measure the
performance of the proposed models. PM2.5 as a target variable was affected by
dew point, pressure, temperature, cumulated wind speed, hours of snow, and
hours of rain. The deep learning network inputs consist of three different
scenarios: daily, weekly, and monthly. The results show that the proposed M-1
with three hidden layers produces the best results of RMSE and MAPE compared to
the proposed M-2, M-3, and all the baselines. A recommendation for air
pollution management could be generated by using these optimized models
| [
{
"version": "v1",
"created": "Sat, 10 Jun 2023 16:06:44 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Pranolo",
"Andri",
""
],
[
"Mao",
"Yingchi",
""
],
[
"Wibawa",
"Aji Prasetya",
""
],
[
"Utama",
"Agung Bella Putra",
""
],
[
"Dwiyanto",
"Felix Andika",
""
]
] | TITLE: Optimized Three Deep Learning Models Based-PSO Hyperparameters for
Beijing PM2.5 Prediction
ABSTRACT: Deep learning is a machine learning approach that produces excellent
performance in various applications, including natural language processing,
image identification, and forecasting. Deep learning network performance
depends on the hyperparameter settings. This research attempts to optimize the
deep learning architecture of Long short term memory (LSTM), Convolutional
neural network (CNN), and Multilayer perceptron (MLP) for forecasting tasks
using Particle swarm optimization (PSO), a swarm intelligence-based
metaheuristic optimization methodology: Proposed M-1 (PSO-LSTM), M-2 (PSO-CNN),
and M-3 (PSO-MLP). Beijing PM2.5 datasets was analyzed to measure the
performance of the proposed models. PM2.5 as a target variable was affected by
dew point, pressure, temperature, cumulated wind speed, hours of snow, and
hours of rain. The deep learning network inputs consist of three different
scenarios: daily, weekly, and monthly. The results show that the proposed M-1
with three hidden layers produces the best results of RMSE and MAPE compared to
the proposed M-2, M-3, and all the baselines. A recommendation for air
pollution management could be generated by using these optimized models
|
2306.07297 | Xishuang Dong | Shouvon Sarker, Lijun Qian, Xishuang Dong | Medical Data Augmentation via ChatGPT: A Case Study on Medication
Identification and Medication Event Classification | null | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The identification of key factors such as medications, diseases, and
relationships within electronic health records and clinical notes has a wide
range of applications in the clinical field. In the N2C2 2022 competitions,
various tasks were presented to promote the identification of key factors in
electronic health records (EHRs) using the Contextualized Medication Event
Dataset (CMED). Pretrained large language models (LLMs) demonstrated
exceptional performance in these tasks. This study aims to explore the
utilization of LLMs, specifically ChatGPT, for data augmentation to overcome
the limited availability of annotated data for identifying the key factors in
EHRs. Additionally, different pre-trained BERT models, initially trained on
extensive datasets like Wikipedia and MIMIC, were employed to develop models
for identifying these key variables in EHRs through fine-tuning on augmented
datasets. The experimental results of two EHR analysis tasks, namely medication
identification and medication event classification, indicate that data
augmentation based on ChatGPT proves beneficial in improving performance for
both medication identification and medication event classification.
| [
{
"version": "v1",
"created": "Sat, 10 Jun 2023 20:55:21 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Sarker",
"Shouvon",
""
],
[
"Qian",
"Lijun",
""
],
[
"Dong",
"Xishuang",
""
]
] | TITLE: Medical Data Augmentation via ChatGPT: A Case Study on Medication
Identification and Medication Event Classification
ABSTRACT: The identification of key factors such as medications, diseases, and
relationships within electronic health records and clinical notes has a wide
range of applications in the clinical field. In the N2C2 2022 competitions,
various tasks were presented to promote the identification of key factors in
electronic health records (EHRs) using the Contextualized Medication Event
Dataset (CMED). Pretrained large language models (LLMs) demonstrated
exceptional performance in these tasks. This study aims to explore the
utilization of LLMs, specifically ChatGPT, for data augmentation to overcome
the limited availability of annotated data for identifying the key factors in
EHRs. Additionally, different pre-trained BERT models, initially trained on
extensive datasets like Wikipedia and MIMIC, were employed to develop models
for identifying these key variables in EHRs through fine-tuning on augmented
datasets. The experimental results of two EHR analysis tasks, namely medication
identification and medication event classification, indicate that data
augmentation based on ChatGPT proves beneficial in improving performance for
both medication identification and medication event classification.
|
2306.07298 | Shruti Bhargava | Shruti Bhargava, Anand Dhoot, Ing-Marie Jonsson, Hoang Long Nguyen,
Alkesh Patel, Hong Yu, Vincent Renkens | Referring to Screen Texts with Voice Assistants | 7 pages, Accepted to ACL Industry Track 2023 | null | null | null | cs.HC cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Voice assistants help users make phone calls, send messages, create events,
navigate, and do a lot more. However, assistants have limited capacity to
understand their users' context. In this work, we aim to take a step in this
direction. Our work dives into a new experience for users to refer to phone
numbers, addresses, email addresses, URLs, and dates on their phone screens.
Our focus lies in reference understanding, which becomes particularly
interesting when multiple similar texts are present on screen, similar to
visual grounding. We collect a dataset and propose a lightweight
general-purpose model for this novel experience. Due to the high cost of
consuming pixels directly, our system is designed to rely on the extracted text
from the UI. Our model is modular, thus offering flexibility, improved
interpretability, and efficient runtime memory utilization.
| [
{
"version": "v1",
"created": "Sat, 10 Jun 2023 22:43:16 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Bhargava",
"Shruti",
""
],
[
"Dhoot",
"Anand",
""
],
[
"Jonsson",
"Ing-Marie",
""
],
[
"Nguyen",
"Hoang Long",
""
],
[
"Patel",
"Alkesh",
""
],
[
"Yu",
"Hong",
""
],
[
"Renkens",
"Vincent",
""
]
] | TITLE: Referring to Screen Texts with Voice Assistants
ABSTRACT: Voice assistants help users make phone calls, send messages, create events,
navigate, and do a lot more. However, assistants have limited capacity to
understand their users' context. In this work, we aim to take a step in this
direction. Our work dives into a new experience for users to refer to phone
numbers, addresses, email addresses, URLs, and dates on their phone screens.
Our focus lies in reference understanding, which becomes particularly
interesting when multiple similar texts are present on screen, similar to
visual grounding. We collect a dataset and propose a lightweight
general-purpose model for this novel experience. Due to the high cost of
consuming pixels directly, our system is designed to rely on the extracted text
from the UI. Our model is modular, thus offering flexibility, improved
interpretability, and efficient runtime memory utilization.
|
2306.07300 | Tariq Khan Dr | Asim Naveed, Syed S. Naqvi, Tariq M. Khan, Imran Razzak | Progressive Class-Wise Attention (PCA) Approach for Diagnosing Skin
Lesions | null | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Skin cancer holds the highest incidence rate among all cancers globally. The
importance of early detection cannot be overstated, as late-stage cases can be
lethal. Classifying skin lesions, however, presents several challenges due to
the many variations they can exhibit, such as differences in colour, shape, and
size, significant variation within the same class, and notable similarities
between different classes. This paper introduces a novel class-wise attention
technique that equally regards each class while unearthing more specific
details about skin lesions. This attention mechanism is progressively used to
amalgamate discriminative feature details from multiple scales. The introduced
technique demonstrated impressive performance, surpassing more than 15
cutting-edge methods including the winners of HAM1000 and ISIC 2019
leaderboards. It achieved an impressive accuracy rate of 97.40% on the HAM10000
dataset and 94.9% on the ISIC 2019 dataset.
| [
{
"version": "v1",
"created": "Sun, 11 Jun 2023 04:58:31 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Naveed",
"Asim",
""
],
[
"Naqvi",
"Syed S.",
""
],
[
"Khan",
"Tariq M.",
""
],
[
"Razzak",
"Imran",
""
]
] | TITLE: Progressive Class-Wise Attention (PCA) Approach for Diagnosing Skin
Lesions
ABSTRACT: Skin cancer holds the highest incidence rate among all cancers globally. The
importance of early detection cannot be overstated, as late-stage cases can be
lethal. Classifying skin lesions, however, presents several challenges due to
the many variations they can exhibit, such as differences in colour, shape, and
size, significant variation within the same class, and notable similarities
between different classes. This paper introduces a novel class-wise attention
technique that equally regards each class while unearthing more specific
details about skin lesions. This attention mechanism is progressively used to
amalgamate discriminative feature details from multiple scales. The introduced
technique demonstrated impressive performance, surpassing more than 15
cutting-edge methods including the winners of HAM1000 and ISIC 2019
leaderboards. It achieved an impressive accuracy rate of 97.40% on the HAM10000
dataset and 94.9% on the ISIC 2019 dataset.
|
2306.07301 | Dhanalakshmi M Mrs | Dhanalakshmi M and Radha V | Novel Regression and Least Square Support Vector Machine Learning
Technique for Air Pollution Forecasting | 11 pages, 7 figures, 3 tables, Article Published in April 2023,
Volume 71, Issue 04, of SSRG-International Journal of Engineering Trends and
Technology (IJETT)", ISSN: 2231-5381 | null | 10.14445/22315381/IJETT-V71I4P214 | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Air pollution is the origination of particulate matter, chemicals, or
biological substances that brings pain to either humans or other living
creatures or instigates discomfort to the natural habitat and the airspace.
Hence, air pollution remains one of the paramount environmental issues as far
as metropolitan cities are concerned. Several air pollution benchmarks are even
said to have a negative influence on human health. Also, improper detection of
air pollution benchmarks results in severe complications for humans and living
creatures. To address this aspect, a novel technique called, Discretized
Regression and Least Square Support Vector (DR-LSSV) based air pollution
forecasting is proposed. The results indicate that the proposed DR-LSSV
Technique can efficiently enhance air pollution forecasting performance and
outperforms the conventional machine learning methods in terms of air pollution
forecasting accuracy, air pollution forecasting time, and false positive rate.
| [
{
"version": "v1",
"created": "Sun, 11 Jun 2023 06:56:00 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"M",
"Dhanalakshmi",
""
],
[
"V",
"Radha",
""
]
] | TITLE: Novel Regression and Least Square Support Vector Machine Learning
Technique for Air Pollution Forecasting
ABSTRACT: Air pollution is the origination of particulate matter, chemicals, or
biological substances that brings pain to either humans or other living
creatures or instigates discomfort to the natural habitat and the airspace.
Hence, air pollution remains one of the paramount environmental issues as far
as metropolitan cities are concerned. Several air pollution benchmarks are even
said to have a negative influence on human health. Also, improper detection of
air pollution benchmarks results in severe complications for humans and living
creatures. To address this aspect, a novel technique called, Discretized
Regression and Least Square Support Vector (DR-LSSV) based air pollution
forecasting is proposed. The results indicate that the proposed DR-LSSV
Technique can efficiently enhance air pollution forecasting performance and
outperforms the conventional machine learning methods in terms of air pollution
forecasting accuracy, air pollution forecasting time, and false positive rate.
|
2306.07352 | Fransisca Susan | Fransisca Susan, Negin Golrezaei, Okke Schrijvers | Multi-Platform Budget Management in Ad Markets with Non-IC Auctions | 34 pages, 5 figures | null | null | null | cs.GT cs.LG math.OC stat.ML | http://creativecommons.org/licenses/by-sa/4.0/ | In online advertising markets, budget-constrained advertisers acquire ad
placements through repeated bidding in auctions on various platforms. We
present a strategy for bidding optimally in a set of auctions that may or may
not be incentive-compatible under the presence of budget constraints. Our
strategy maximizes the expected total utility across auctions while satisfying
the advertiser's budget constraints in expectation. Additionally, we
investigate the online setting where the advertiser must submit bids across
platforms while learning about other bidders' bids over time. Our algorithm has
$O(T^{3/4})$ regret under the full-information setting. Finally, we demonstrate
that our algorithms have superior cumulative regret on both synthetic and
real-world datasets of ad placement auctions, compared to existing adaptive
pacing algorithms.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 18:21:10 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Susan",
"Fransisca",
""
],
[
"Golrezaei",
"Negin",
""
],
[
"Schrijvers",
"Okke",
""
]
] | TITLE: Multi-Platform Budget Management in Ad Markets with Non-IC Auctions
ABSTRACT: In online advertising markets, budget-constrained advertisers acquire ad
placements through repeated bidding in auctions on various platforms. We
present a strategy for bidding optimally in a set of auctions that may or may
not be incentive-compatible under the presence of budget constraints. Our
strategy maximizes the expected total utility across auctions while satisfying
the advertiser's budget constraints in expectation. Additionally, we
investigate the online setting where the advertiser must submit bids across
platforms while learning about other bidders' bids over time. Our algorithm has
$O(T^{3/4})$ regret under the full-information setting. Finally, we demonstrate
that our algorithms have superior cumulative regret on both synthetic and
real-world datasets of ad placement auctions, compared to existing adaptive
pacing algorithms.
|
2306.07363 | Douaa Khalil | Georges Younes, Douaa Khalil, John Zelek, Daniel Asmar | H-SLAM: Hybrid Direct-Indirect Visual SLAM | null | null | null | null | cs.RO cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The recent success of hybrid methods in monocular odometry has led to many
attempts to generalize the performance gains to hybrid monocular SLAM. However,
most attempts fall short in several respects, with the most prominent issue
being the need for two different map representations (local and global maps),
with each requiring different, computationally expensive, and often redundant
processes to maintain. Moreover, these maps tend to drift with respect to each
other, resulting in contradicting pose and scene estimates, and leading to
catastrophic failure. In this paper, we propose a novel approach that makes use
of descriptor sharing to generate a single inverse depth scene representation.
This representation can be used locally, queried globally to perform loop
closure, and has the ability to re-activate previously observed map points
after redundant points are marginalized from the local map, eliminating the
need for separate and redundant map maintenance processes. The maps generated
by our method exhibit no drift between each other, and can be computed at a
fraction of the computational cost and memory footprint required by other
monocular SLAM systems. Despite the reduced resource requirements, the proposed
approach maintains its robustness and accuracy, delivering performance
comparable to state-of-the-art SLAM methods (e.g., LDSO, ORB-SLAM3) on the
majority of sequences from well-known datasets like EuRoC, KITTI, and TUM VI.
The source code is available at: https://github.com/AUBVRL/fslam_ros_docker.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 18:39:03 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Younes",
"Georges",
""
],
[
"Khalil",
"Douaa",
""
],
[
"Zelek",
"John",
""
],
[
"Asmar",
"Daniel",
""
]
] | TITLE: H-SLAM: Hybrid Direct-Indirect Visual SLAM
ABSTRACT: The recent success of hybrid methods in monocular odometry has led to many
attempts to generalize the performance gains to hybrid monocular SLAM. However,
most attempts fall short in several respects, with the most prominent issue
being the need for two different map representations (local and global maps),
with each requiring different, computationally expensive, and often redundant
processes to maintain. Moreover, these maps tend to drift with respect to each
other, resulting in contradicting pose and scene estimates, and leading to
catastrophic failure. In this paper, we propose a novel approach that makes use
of descriptor sharing to generate a single inverse depth scene representation.
This representation can be used locally, queried globally to perform loop
closure, and has the ability to re-activate previously observed map points
after redundant points are marginalized from the local map, eliminating the
need for separate and redundant map maintenance processes. The maps generated
by our method exhibit no drift between each other, and can be computed at a
fraction of the computational cost and memory footprint required by other
monocular SLAM systems. Despite the reduced resource requirements, the proposed
approach maintains its robustness and accuracy, delivering performance
comparable to state-of-the-art SLAM methods (e.g., LDSO, ORB-SLAM3) on the
majority of sequences from well-known datasets like EuRoC, KITTI, and TUM VI.
The source code is available at: https://github.com/AUBVRL/fslam_ros_docker.
|
2306.07372 | Dustin Morrill | Dustin Morrill, Thomas J. Walsh, Daniel Hernandez, Peter R. Wurman,
Peter Stone | Composing Efficient, Robust Tests for Policy Selection | 26 pages, 13 figures. To appear in Proceedings of the Thirty-Ninth
Conference on Uncertainty in Artificial Intelligence (UAI 2023) | null | null | null | cs.LG cs.AI cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern reinforcement learning systems produce many high-quality policies
throughout the learning process. However, to choose which policy to actually
deploy in the real world, they must be tested under an intractable number of
environmental conditions. We introduce RPOSST, an algorithm to select a small
set of test cases from a larger pool based on a relatively small number of
sample evaluations. RPOSST treats the test case selection problem as a
two-player game and optimizes a solution with provable $k$-of-$N$ robustness,
bounding the error relative to a test that used all the test cases in the pool.
Empirical results demonstrate that RPOSST finds a small set of test cases that
identify high quality policies in a toy one-shot game, poker datasets, and a
high-fidelity racing simulator.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 18:55:56 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Morrill",
"Dustin",
""
],
[
"Walsh",
"Thomas J.",
""
],
[
"Hernandez",
"Daniel",
""
],
[
"Wurman",
"Peter R.",
""
],
[
"Stone",
"Peter",
""
]
] | TITLE: Composing Efficient, Robust Tests for Policy Selection
ABSTRACT: Modern reinforcement learning systems produce many high-quality policies
throughout the learning process. However, to choose which policy to actually
deploy in the real world, they must be tested under an intractable number of
environmental conditions. We introduce RPOSST, an algorithm to select a small
set of test cases from a larger pool based on a relatively small number of
sample evaluations. RPOSST treats the test case selection problem as a
two-player game and optimizes a solution with provable $k$-of-$N$ robustness,
bounding the error relative to a test that used all the test cases in the pool.
Empirical results demonstrate that RPOSST finds a small set of test cases that
identify high quality policies in a toy one-shot game, poker datasets, and a
high-fidelity racing simulator.
|
2306.07381 | Yuqing Zhu | Yuqing Zhu, Xuandong Zhao, Chuan Guo, Yu-Xiang Wang | "Private Prediction Strikes Back!'' Private Kernelized Nearest Neighbors
with Individual Renyi Filter | null | null | null | null | cs.LG cs.CR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Most existing approaches of differentially private (DP) machine learning
focus on private training. Despite its many advantages, private training lacks
the flexibility in adapting to incremental changes to the training dataset such
as deletion requests from exercising GDPR's right to be forgotten. We revisit a
long-forgotten alternative, known as private prediction, and propose a new
algorithm named Individual Kernelized Nearest Neighbor (Ind-KNN). Ind-KNN is
easily updatable over dataset changes and it allows precise control of the
R\'{e}nyi DP at an individual user level -- a user's privacy loss is measured
by the exact amount of her contribution to predictions; and a user is removed
if her prescribed privacy budget runs out. Our results show that Ind-KNN
consistently improves the accuracy over existing private prediction methods for
a wide range of $\epsilon$ on four vision and language tasks. We also
illustrate several cases under which Ind-KNN is preferable over private
training with NoisySGD.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 19:14:45 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Zhu",
"Yuqing",
""
],
[
"Zhao",
"Xuandong",
""
],
[
"Guo",
"Chuan",
""
],
[
"Wang",
"Yu-Xiang",
""
]
] | TITLE: "Private Prediction Strikes Back!'' Private Kernelized Nearest Neighbors
with Individual Renyi Filter
ABSTRACT: Most existing approaches of differentially private (DP) machine learning
focus on private training. Despite its many advantages, private training lacks
the flexibility in adapting to incremental changes to the training dataset such
as deletion requests from exercising GDPR's right to be forgotten. We revisit a
long-forgotten alternative, known as private prediction, and propose a new
algorithm named Individual Kernelized Nearest Neighbor (Ind-KNN). Ind-KNN is
easily updatable over dataset changes and it allows precise control of the
R\'{e}nyi DP at an individual user level -- a user's privacy loss is measured
by the exact amount of her contribution to predictions; and a user is removed
if her prescribed privacy budget runs out. Our results show that Ind-KNN
consistently improves the accuracy over existing private prediction methods for
a wide range of $\epsilon$ on four vision and language tasks. We also
illustrate several cases under which Ind-KNN is preferable over private
training with NoisySGD.
|
2306.07383 | Shadrokh Samavi | MohammadHossein Givkashi, MohammadReza Naderi, Nader Karimi, Shahram
Shirani, Shadrokh Samavi | Supervised Deep Learning for Content-Aware Image Retargeting with
Fourier Convolutions | 18 pages, 5 figures | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Image retargeting aims to alter the size of the image with attention to the
contents. One of the main obstacles to training deep learning models for image
retargeting is the need for a vast labeled dataset. Labeled datasets are
unavailable for training deep learning models in the image retargeting tasks.
As a result, we present a new supervised approach for training deep learning
models. We use the original images as ground truth and create inputs for the
model by resizing and cropping the original images. A second challenge is
generating different image sizes in inference time. However, regular
convolutional neural networks cannot generate images of different sizes than
the input image. To address this issue, we introduced a new method for
supervised learning. In our approach, a mask is generated to show the desired
size and location of the object. Then the mask and the input image are fed to
the network. Comparing image retargeting methods and our proposed method
demonstrates the model's ability to produce high-quality retargeted images.
Afterward, we compute the image quality assessment score for each output image
based on different techniques and illustrate the effectiveness of our approach.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 19:17:44 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Givkashi",
"MohammadHossein",
""
],
[
"Naderi",
"MohammadReza",
""
],
[
"Karimi",
"Nader",
""
],
[
"Shirani",
"Shahram",
""
],
[
"Samavi",
"Shadrokh",
""
]
] | TITLE: Supervised Deep Learning for Content-Aware Image Retargeting with
Fourier Convolutions
ABSTRACT: Image retargeting aims to alter the size of the image with attention to the
contents. One of the main obstacles to training deep learning models for image
retargeting is the need for a vast labeled dataset. Labeled datasets are
unavailable for training deep learning models in the image retargeting tasks.
As a result, we present a new supervised approach for training deep learning
models. We use the original images as ground truth and create inputs for the
model by resizing and cropping the original images. A second challenge is
generating different image sizes in inference time. However, regular
convolutional neural networks cannot generate images of different sizes than
the input image. To address this issue, we introduced a new method for
supervised learning. In our approach, a mask is generated to show the desired
size and location of the object. Then the mask and the input image are fed to
the network. Comparing image retargeting methods and our proposed method
demonstrates the model's ability to produce high-quality retargeted images.
Afterward, we compute the image quality assessment score for each output image
based on different techniques and illustrate the effectiveness of our approach.
|
2306.07399 | Stefanie Wuhrer | Matthieu Armando, Laurence Boissieux, Edmond Boyer, Jean-Sebastien
Franco, Martin Humenberger, Christophe Legras, Vincent Leroy, Mathieu Marsot,
Julien Pansiot, Sergi Pujades, Rim Rekik, Gregory Rogez, Anilkumar Swamy,
Stefanie Wuhrer | 4DHumanOutfit: a multi-subject 4D dataset of human motion sequences in
varying outfits exhibiting large displacements | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work presents 4DHumanOutfit, a new dataset of densely sampled
spatio-temporal 4D human motion data of different actors, outfits and motions.
The dataset is designed to contain different actors wearing different outfits
while performing different motions in each outfit. In this way, the dataset can
be seen as a cube of data containing 4D motion sequences along 3 axes with
identity, outfit and motion. This rich dataset has numerous potential
applications for the processing and creation of digital humans, e.g. augmented
reality, avatar creation and virtual try on. 4DHumanOutfit is released for
research purposes at https://kinovis.inria.fr/4dhumanoutfit/. In addition to
image data and 4D reconstructions, the dataset includes reference solutions for
each axis. We present independent baselines along each axis that demonstrate
the value of these reference solutions for evaluation tasks.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 19:59:27 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Armando",
"Matthieu",
""
],
[
"Boissieux",
"Laurence",
""
],
[
"Boyer",
"Edmond",
""
],
[
"Franco",
"Jean-Sebastien",
""
],
[
"Humenberger",
"Martin",
""
],
[
"Legras",
"Christophe",
""
],
[
"Leroy",
"Vincent",
""
],
[
"Marsot",
"Mathieu",
""
],
[
"Pansiot",
"Julien",
""
],
[
"Pujades",
"Sergi",
""
],
[
"Rekik",
"Rim",
""
],
[
"Rogez",
"Gregory",
""
],
[
"Swamy",
"Anilkumar",
""
],
[
"Wuhrer",
"Stefanie",
""
]
] | TITLE: 4DHumanOutfit: a multi-subject 4D dataset of human motion sequences in
varying outfits exhibiting large displacements
ABSTRACT: This work presents 4DHumanOutfit, a new dataset of densely sampled
spatio-temporal 4D human motion data of different actors, outfits and motions.
The dataset is designed to contain different actors wearing different outfits
while performing different motions in each outfit. In this way, the dataset can
be seen as a cube of data containing 4D motion sequences along 3 axes with
identity, outfit and motion. This rich dataset has numerous potential
applications for the processing and creation of digital humans, e.g. augmented
reality, avatar creation and virtual try on. 4DHumanOutfit is released for
research purposes at https://kinovis.inria.fr/4dhumanoutfit/. In addition to
image data and 4D reconstructions, the dataset includes reference solutions for
each axis. We present independent baselines along each axis that demonstrate
the value of these reference solutions for evaluation tasks.
|
2306.07414 | Vukosi Marivate | Catherine Gitau and VUkosi Marivate | Textual Augmentation Techniques Applied to Low Resource Machine
Translation: Case of Swahili | Accepted for Third workshop on Resources for African Indigenous
Languages (RAIL) | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work we investigate the impact of applying textual data augmentation
tasks to low resource machine translation. There has been recent interest in
investigating approaches for training systems for languages with limited
resources and one popular approach is the use of data augmentation techniques.
Data augmentation aims to increase the quantity of data that is available to
train the system. In machine translation, majority of the language pairs around
the world are considered low resource because they have little parallel data
available and the quality of neural machine translation (NMT) systems depend a
lot on the availability of sizable parallel corpora. We study and apply three
simple data augmentation techniques popularly used in text classification
tasks; synonym replacement, random insertion and contextual data augmentation
and compare their performance with baseline neural machine translation for
English-Swahili (En-Sw) datasets. We also present results in BLEU, ChrF and
Meteor scores. Overall, the contextual data augmentation technique shows some
improvements both in the $EN \rightarrow SW$ and $SW \rightarrow EN$
directions. We see that there is potential to use these methods in neural
machine translation when more extensive experiments are done with diverse
datasets.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 20:43:24 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Gitau",
"Catherine",
""
],
[
"Marivate",
"VUkosi",
""
]
] | TITLE: Textual Augmentation Techniques Applied to Low Resource Machine
Translation: Case of Swahili
ABSTRACT: In this work we investigate the impact of applying textual data augmentation
tasks to low resource machine translation. There has been recent interest in
investigating approaches for training systems for languages with limited
resources and one popular approach is the use of data augmentation techniques.
Data augmentation aims to increase the quantity of data that is available to
train the system. In machine translation, majority of the language pairs around
the world are considered low resource because they have little parallel data
available and the quality of neural machine translation (NMT) systems depend a
lot on the availability of sizable parallel corpora. We study and apply three
simple data augmentation techniques popularly used in text classification
tasks; synonym replacement, random insertion and contextual data augmentation
and compare their performance with baseline neural machine translation for
English-Swahili (En-Sw) datasets. We also present results in BLEU, ChrF and
Meteor scores. Overall, the contextual data augmentation technique shows some
improvements both in the $EN \rightarrow SW$ and $SW \rightarrow EN$
directions. We see that there is potential to use these methods in neural
machine translation when more extensive experiments are done with diverse
datasets.
|
2306.07415 | Gunnar Lund | Gunnar Lund, Kostiantyn Omelianchuk, Igor Samokhin | Gender-Inclusive Grammatical Error Correction through Augmentation | 15 pages. Accepted for publication in BEA workshop (18th Workshop on
Innovative Use of NLP for Building Educational Applications; co-located with
ACL) | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | In this paper we show that GEC systems display gender bias related to the use
of masculine and feminine terms and the gender-neutral singular "they". We
develop parallel datasets of texts with masculine and feminine terms and
singular "they" and use them to quantify gender bias in three competitive GEC
systems. We contribute a novel data augmentation technique for singular "they"
leveraging linguistic insights about its distribution relative to plural
"they". We demonstrate that both this data augmentation technique and a
refinement of a similar augmentation technique for masculine and feminine terms
can generate training data that reduces bias in GEC systems, especially with
respect to singular "they" while maintaining the same level of quality.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 20:44:47 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Lund",
"Gunnar",
""
],
[
"Omelianchuk",
"Kostiantyn",
""
],
[
"Samokhin",
"Igor",
""
]
] | TITLE: Gender-Inclusive Grammatical Error Correction through Augmentation
ABSTRACT: In this paper we show that GEC systems display gender bias related to the use
of masculine and feminine terms and the gender-neutral singular "they". We
develop parallel datasets of texts with masculine and feminine terms and
singular "they" and use them to quantify gender bias in three competitive GEC
systems. We contribute a novel data augmentation technique for singular "they"
leveraging linguistic insights about its distribution relative to plural
"they". We demonstrate that both this data augmentation technique and a
refinement of a similar augmentation technique for masculine and feminine terms
can generate training data that reduces bias in GEC systems, especially with
respect to singular "they" while maintaining the same level of quality.
|
2306.07426 | Vukosi Marivate | Andani Madodonga, Vukosi Marivate, Matthew Adendorff | Izindaba-Tindzaba: Machine learning news categorisation for Long and
Short Text for isiZulu and Siswati | Accepted for Third workshop on Resources for African Indigenous
Languages (RAIL) | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Local/Native South African languages are classified as low-resource
languages. As such, it is essential to build the resources for these languages
so that they can benefit from advances in the field of natural language
processing. In this work, the focus was to create annotated news datasets for
the isiZulu and Siswati native languages based on news topic classification
tasks and present the findings from these baseline classification models. Due
to the shortage of data for these native South African languages, the datasets
that were created were augmented and oversampled to increase data size and
overcome class classification imbalance. In total, four different
classification models were used namely Logistic regression, Naive bayes,
XGBoost and LSTM. These models were trained on three different word embeddings
namely Bag-Of-Words, TFIDF and Word2vec. The results of this study showed that
XGBoost, Logistic Regression and LSTM, trained from Word2vec performed better
than the other combinations.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 21:02:12 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Madodonga",
"Andani",
""
],
[
"Marivate",
"Vukosi",
""
],
[
"Adendorff",
"Matthew",
""
]
] | TITLE: Izindaba-Tindzaba: Machine learning news categorisation for Long and
Short Text for isiZulu and Siswati
ABSTRACT: Local/Native South African languages are classified as low-resource
languages. As such, it is essential to build the resources for these languages
so that they can benefit from advances in the field of natural language
processing. In this work, the focus was to create annotated news datasets for
the isiZulu and Siswati native languages based on news topic classification
tasks and present the findings from these baseline classification models. Due
to the shortage of data for these native South African languages, the datasets
that were created were augmented and oversampled to increase data size and
overcome class classification imbalance. In total, four different
classification models were used namely Logistic regression, Naive bayes,
XGBoost and LSTM. These models were trained on three different word embeddings
namely Bag-Of-Words, TFIDF and Word2vec. The results of this study showed that
XGBoost, Logistic Regression and LSTM, trained from Word2vec performed better
than the other combinations.
|
2306.07427 | Bhavya Ghai | Bhavya Ghai | Towards Fair and Explainable AI using a Human-Centered AI Approach | PhD Thesis | null | null | null | cs.CY cs.HC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rise of machine learning (ML) is accompanied by several high-profile
cases that have stressed the need for fairness, accountability, explainability
and trust in ML systems. The existing literature has largely focused on fully
automated ML approaches that try to optimize for some performance metric.
However, human-centric measures like fairness, trust, explainability, etc. are
subjective in nature, context-dependent, and might not correlate with
conventional performance metrics. To deal with these challenges, we explore a
human-centered AI approach that empowers people by providing more transparency
and human control.
In this dissertation, we present 5 research projects that aim to enhance
explainability and fairness in classification systems and word embeddings. The
first project explores the utility/downsides of introducing local model
explanations as interfaces for machine teachers (crowd workers). Our study
found that adding explanations supports trust calibration for the resulting ML
model and enables rich forms of teaching feedback. The second project presents
D-BIAS, a causality-based human-in-the-loop visual tool for identifying and
mitigating social biases in tabular datasets. Apart from fairness, we found
that our tool also enhances trust and accountability. The third project
presents WordBias, a visual interactive tool that helps audit pre-trained
static word embeddings for biases against groups, such as females, or
subgroups, such as Black Muslim females. The fourth project presents DramatVis
Personae, a visual analytics tool that helps identify social biases in creative
writing. Finally, the last project presents an empirical study aimed at
understanding the cumulative impact of multiple fairness-enhancing
interventions at different stages of the ML pipeline on fairness, utility and
different population groups. We conclude by discussing some of the future
directions.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 21:08:55 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Ghai",
"Bhavya",
""
]
] | TITLE: Towards Fair and Explainable AI using a Human-Centered AI Approach
ABSTRACT: The rise of machine learning (ML) is accompanied by several high-profile
cases that have stressed the need for fairness, accountability, explainability
and trust in ML systems. The existing literature has largely focused on fully
automated ML approaches that try to optimize for some performance metric.
However, human-centric measures like fairness, trust, explainability, etc. are
subjective in nature, context-dependent, and might not correlate with
conventional performance metrics. To deal with these challenges, we explore a
human-centered AI approach that empowers people by providing more transparency
and human control.
In this dissertation, we present 5 research projects that aim to enhance
explainability and fairness in classification systems and word embeddings. The
first project explores the utility/downsides of introducing local model
explanations as interfaces for machine teachers (crowd workers). Our study
found that adding explanations supports trust calibration for the resulting ML
model and enables rich forms of teaching feedback. The second project presents
D-BIAS, a causality-based human-in-the-loop visual tool for identifying and
mitigating social biases in tabular datasets. Apart from fairness, we found
that our tool also enhances trust and accountability. The third project
presents WordBias, a visual interactive tool that helps audit pre-trained
static word embeddings for biases against groups, such as females, or
subgroups, such as Black Muslim females. The fourth project presents DramatVis
Personae, a visual analytics tool that helps identify social biases in creative
writing. Finally, the last project presents an empirical study aimed at
understanding the cumulative impact of multiple fairness-enhancing
interventions at different stages of the ML pipeline on fairness, utility and
different population groups. We conclude by discussing some of the future
directions.
|
2306.07437 | Timo Bolkart | Timo Bolkart and Tianye Li and Michael J. Black | Instant Multi-View Head Capture through Learnable Registration | Conference on Computer Vision and Pattern Recognition (CVPR) 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing methods for capturing datasets of 3D heads in dense semantic
correspondence are slow, and commonly address the problem in two separate
steps; multi-view stereo (MVS) reconstruction followed by non-rigid
registration. To simplify this process, we introduce TEMPEH (Towards Estimation
of 3D Meshes from Performances of Expressive Heads) to directly infer 3D heads
in dense correspondence from calibrated multi-view images. Registering datasets
of 3D scans typically requires manual parameter tuning to find the right
balance between accurately fitting the scans surfaces and being robust to
scanning noise and outliers. Instead, we propose to jointly register a 3D head
dataset while training TEMPEH. Specifically, during training we minimize a
geometric loss commonly used for surface registration, effectively leveraging
TEMPEH as a regularizer. Our multi-view head inference builds on a volumetric
feature representation that samples and fuses features from each view using
camera calibration information. To account for partial occlusions and a large
capture volume that enables head movements, we use view- and surface-aware
feature fusion, and a spatial transformer-based head localization module,
respectively. We use raw MVS scans as supervision during training, but, once
trained, TEMPEH directly predicts 3D heads in dense correspondence without
requiring scans. Predicting one head takes about 0.3 seconds with a median
reconstruction error of 0.26 mm, 64% lower than the current state-of-the-art.
This enables the efficient capture of large datasets containing multiple people
and diverse facial motions. Code, model, and data are publicly available at
https://tempeh.is.tue.mpg.de.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 21:45:18 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Bolkart",
"Timo",
""
],
[
"Li",
"Tianye",
""
],
[
"Black",
"Michael J.",
""
]
] | TITLE: Instant Multi-View Head Capture through Learnable Registration
ABSTRACT: Existing methods for capturing datasets of 3D heads in dense semantic
correspondence are slow, and commonly address the problem in two separate
steps; multi-view stereo (MVS) reconstruction followed by non-rigid
registration. To simplify this process, we introduce TEMPEH (Towards Estimation
of 3D Meshes from Performances of Expressive Heads) to directly infer 3D heads
in dense correspondence from calibrated multi-view images. Registering datasets
of 3D scans typically requires manual parameter tuning to find the right
balance between accurately fitting the scans surfaces and being robust to
scanning noise and outliers. Instead, we propose to jointly register a 3D head
dataset while training TEMPEH. Specifically, during training we minimize a
geometric loss commonly used for surface registration, effectively leveraging
TEMPEH as a regularizer. Our multi-view head inference builds on a volumetric
feature representation that samples and fuses features from each view using
camera calibration information. To account for partial occlusions and a large
capture volume that enables head movements, we use view- and surface-aware
feature fusion, and a spatial transformer-based head localization module,
respectively. We use raw MVS scans as supervision during training, but, once
trained, TEMPEH directly predicts 3D heads in dense correspondence without
requiring scans. Predicting one head takes about 0.3 seconds with a median
reconstruction error of 0.26 mm, 64% lower than the current state-of-the-art.
This enables the efficient capture of large datasets containing multiple people
and diverse facial motions. Code, model, and data are publicly available at
https://tempeh.is.tue.mpg.de.
|
2306.07471 | Jimmy Lin | Ehsan Kamalloo, Nandan Thakur, Carlos Lassance, Xueguang Ma,
Jheng-Hong Yang, Jimmy Lin | Resources for Brewing BEIR: Reproducible Reference Models and an
Official Leaderboard | null | null | null | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | BEIR is a benchmark dataset for zero-shot evaluation of information retrieval
models across 18 different domain/task combinations. In recent years, we have
witnessed the growing popularity of a representation learning approach to
building retrieval models, typically using pretrained transformers in a
supervised setting. This naturally begs the question: How effective are these
models when presented with queries and documents that differ from the training
data? Examples include searching in different domains (e.g., medical or legal
text) and with different types of queries (e.g., keywords vs. well-formed
questions). While BEIR was designed to answer these questions, our work
addresses two shortcomings that prevent the benchmark from achieving its full
potential: First, the sophistication of modern neural methods and the
complexity of current software infrastructure create barriers to entry for
newcomers. To this end, we provide reproducible reference implementations that
cover the two main classes of approaches: learned dense and sparse models.
Second, there does not exist a single authoritative nexus for reporting the
effectiveness of different models on BEIR, which has led to difficulty in
comparing different methods. To remedy this, we present an official
self-service BEIR leaderboard that provides fair and consistent comparisons of
retrieval models. By addressing both shortcomings, our work facilitates future
explorations in a range of interesting research questions that BEIR enables.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 00:26:18 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Kamalloo",
"Ehsan",
""
],
[
"Thakur",
"Nandan",
""
],
[
"Lassance",
"Carlos",
""
],
[
"Ma",
"Xueguang",
""
],
[
"Yang",
"Jheng-Hong",
""
],
[
"Lin",
"Jimmy",
""
]
] | TITLE: Resources for Brewing BEIR: Reproducible Reference Models and an
Official Leaderboard
ABSTRACT: BEIR is a benchmark dataset for zero-shot evaluation of information retrieval
models across 18 different domain/task combinations. In recent years, we have
witnessed the growing popularity of a representation learning approach to
building retrieval models, typically using pretrained transformers in a
supervised setting. This naturally begs the question: How effective are these
models when presented with queries and documents that differ from the training
data? Examples include searching in different domains (e.g., medical or legal
text) and with different types of queries (e.g., keywords vs. well-formed
questions). While BEIR was designed to answer these questions, our work
addresses two shortcomings that prevent the benchmark from achieving its full
potential: First, the sophistication of modern neural methods and the
complexity of current software infrastructure create barriers to entry for
newcomers. To this end, we provide reproducible reference implementations that
cover the two main classes of approaches: learned dense and sparse models.
Second, there does not exist a single authoritative nexus for reporting the
effectiveness of different models on BEIR, which has led to difficulty in
comparing different methods. To remedy this, we present an official
self-service BEIR leaderboard that provides fair and consistent comparisons of
retrieval models. By addressing both shortcomings, our work facilitates future
explorations in a range of interesting research questions that BEIR enables.
|
2306.07487 | Yangruibo Ding | Yangruibo Ding, Ben Steenhoek, Kexin Pei, Gail Kaiser, Wei Le,
Baishakhi Ray | TRACED: Execution-aware Pre-training for Source Code | Accepted by ICSE 2024 (Early Cycle). Camera-ready is in preparation | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | Most existing pre-trained language models for source code focus on learning
the static code text, typically augmented with static code structures (abstract
syntax tree, dependency graphs, etc.). However, program semantics will not be
fully exposed before the real execution. Without an understanding of the
program execution, statically pre-trained models fail to comprehensively
capture the dynamic code properties, such as the branch coverage and the
runtime variable values, and they are consequently less effective at code
understanding tasks, such as retrieving semantic clones and detecting software
vulnerabilities.
To close the gap between the static nature of language models and the dynamic
characteristics of programs, we introduce TRACED, an execution-aware
pre-training strategy for source code. Specifically, we pre-train code language
models with a combination of source code, executable inputs, and corresponding
execution traces. Our goal is to teach code models the complicated execution
logic during the pre-training, enabling the model to statically estimate the
dynamic code properties without repeatedly executing code during task-specific
fine-tuning.
To illustrate the effectiveness of our proposed approach, we fine-tune and
evaluate TRACED on three downstream tasks: static execution estimation, clone
retrieval, and vulnerability detection. The empirical results show that TRACED
relatively improves the statically pre-trained code models by 12.4% for
complete execution path prediction and by 25.2% for runtime variable value
predictions. TRACED also significantly outperforms statically pre-trained
models in clone retrieval and vulnerability detection across four public
benchmarks.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 01:30:14 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Ding",
"Yangruibo",
""
],
[
"Steenhoek",
"Ben",
""
],
[
"Pei",
"Kexin",
""
],
[
"Kaiser",
"Gail",
""
],
[
"Le",
"Wei",
""
],
[
"Ray",
"Baishakhi",
""
]
] | TITLE: TRACED: Execution-aware Pre-training for Source Code
ABSTRACT: Most existing pre-trained language models for source code focus on learning
the static code text, typically augmented with static code structures (abstract
syntax tree, dependency graphs, etc.). However, program semantics will not be
fully exposed before the real execution. Without an understanding of the
program execution, statically pre-trained models fail to comprehensively
capture the dynamic code properties, such as the branch coverage and the
runtime variable values, and they are consequently less effective at code
understanding tasks, such as retrieving semantic clones and detecting software
vulnerabilities.
To close the gap between the static nature of language models and the dynamic
characteristics of programs, we introduce TRACED, an execution-aware
pre-training strategy for source code. Specifically, we pre-train code language
models with a combination of source code, executable inputs, and corresponding
execution traces. Our goal is to teach code models the complicated execution
logic during the pre-training, enabling the model to statically estimate the
dynamic code properties without repeatedly executing code during task-specific
fine-tuning.
To illustrate the effectiveness of our proposed approach, we fine-tune and
evaluate TRACED on three downstream tasks: static execution estimation, clone
retrieval, and vulnerability detection. The empirical results show that TRACED
relatively improves the statically pre-trained code models by 12.4% for
complete execution path prediction and by 25.2% for runtime variable value
predictions. TRACED also significantly outperforms statically pre-trained
models in clone retrieval and vulnerability detection across four public
benchmarks.
|
2306.07499 | Xiao Yang | Xiao Yang, Ahmed K. Mohamed, Shashank Jain, Stanislav Peshterliev,
Debojeet Chatterjee, Hanwen Zha, Nikita Bhalla, Gagan Aneja and Pranab
Mohanty | Improving Opinion-based Question Answering Systems Through Label Error
Detection and Overwrite | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Label error is a ubiquitous problem in annotated data. Large amounts of label
error substantially degrades the quality of deep learning models. Existing
methods to tackle the label error problem largely focus on the classification
task, and either rely on task specific architecture or require non-trivial
additional computations, which is undesirable or even unattainable for industry
usage. In this paper, we propose LEDO: a model-agnostic and computationally
efficient framework for Label Error Detection and Overwrite. LEDO is based on
Monte Carlo Dropout combined with uncertainty metrics, and can be easily
generalized to multiple tasks and data sets. Applying LEDO to an industry
opinion-based question answering system demonstrates it is effective at
improving accuracy in all the core models. Specifically, LEDO brings 1.1% MRR
gain for the retrieval model, 1.5% PR AUC improvement for the machine reading
comprehension model, and 0.9% rise in the Average Precision for the ranker, on
top of the strong baselines with a large-scale social media dataset.
Importantly, LEDO is computationally efficient compared to methods that require
loss function change, and cost-effective as the resulting data can be used in
the same continuous training pipeline for production. Further analysis shows
that these gains come from an improved decision boundary after cleaning the
label errors existed in the training data.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 02:20:58 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Yang",
"Xiao",
""
],
[
"Mohamed",
"Ahmed K.",
""
],
[
"Jain",
"Shashank",
""
],
[
"Peshterliev",
"Stanislav",
""
],
[
"Chatterjee",
"Debojeet",
""
],
[
"Zha",
"Hanwen",
""
],
[
"Bhalla",
"Nikita",
""
],
[
"Aneja",
"Gagan",
""
],
[
"Mohanty",
"Pranab",
""
]
] | TITLE: Improving Opinion-based Question Answering Systems Through Label Error
Detection and Overwrite
ABSTRACT: Label error is a ubiquitous problem in annotated data. Large amounts of label
error substantially degrades the quality of deep learning models. Existing
methods to tackle the label error problem largely focus on the classification
task, and either rely on task specific architecture or require non-trivial
additional computations, which is undesirable or even unattainable for industry
usage. In this paper, we propose LEDO: a model-agnostic and computationally
efficient framework for Label Error Detection and Overwrite. LEDO is based on
Monte Carlo Dropout combined with uncertainty metrics, and can be easily
generalized to multiple tasks and data sets. Applying LEDO to an industry
opinion-based question answering system demonstrates it is effective at
improving accuracy in all the core models. Specifically, LEDO brings 1.1% MRR
gain for the retrieval model, 1.5% PR AUC improvement for the machine reading
comprehension model, and 0.9% rise in the Average Precision for the ranker, on
top of the strong baselines with a large-scale social media dataset.
Importantly, LEDO is computationally efficient compared to methods that require
loss function change, and cost-effective as the resulting data can be used in
the same continuous training pipeline for production. Further analysis shows
that these gains come from an improved decision boundary after cleaning the
label errors existed in the training data.
|
2306.07501 | Vishwanath Pratap Singh | Vishwanath Pratap Singh, Md Sahidullah, and Tomi Kinnunen | Speaker Verification Across Ages: Investigating Deep Speaker Embedding
Sensitivity to Age Mismatch in Enrollment and Test Speech | null | Interspeech 2023 | null | null | eess.AS cs.SD | http://creativecommons.org/licenses/by/4.0/ | In this paper, we study the impact of the ageing on modern deep speaker
embedding based automatic speaker verification (ASV) systems. We have selected
two different datasets to examine ageing on the state-of-the-art ECAPA-TDNN
system. The first dataset, used for addressing short-term ageing (up to 10
years time difference between enrollment and test) under uncontrolled
conditions, is VoxCeleb. The second dataset, used for addressing long-term
ageing effect (up to 40 years difference) of Finnish speakers under a more
controlled setup, is Longitudinal Corpus of Finnish Spoken in Helsinki (LCFSH).
Our study provides new insights into the impact of speaker ageing on modern ASV
systems. Specifically, we establish a quantitative measure between ageing and
ASV scores. Further, our research indicates that ageing affects female English
speakers to a greater degree than male English speakers, while in the case of
Finnish, it has a greater impact on male speakers than female speakers.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 02:23:55 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Singh",
"Vishwanath Pratap",
""
],
[
"Sahidullah",
"Md",
""
],
[
"Kinnunen",
"Tomi",
""
]
] | TITLE: Speaker Verification Across Ages: Investigating Deep Speaker Embedding
Sensitivity to Age Mismatch in Enrollment and Test Speech
ABSTRACT: In this paper, we study the impact of the ageing on modern deep speaker
embedding based automatic speaker verification (ASV) systems. We have selected
two different datasets to examine ageing on the state-of-the-art ECAPA-TDNN
system. The first dataset, used for addressing short-term ageing (up to 10
years time difference between enrollment and test) under uncontrolled
conditions, is VoxCeleb. The second dataset, used for addressing long-term
ageing effect (up to 40 years difference) of Finnish speakers under a more
controlled setup, is Longitudinal Corpus of Finnish Spoken in Helsinki (LCFSH).
Our study provides new insights into the impact of speaker ageing on modern ASV
systems. Specifically, we establish a quantitative measure between ageing and
ASV scores. Further, our research indicates that ageing affects female English
speakers to a greater degree than male English speakers, while in the case of
Finnish, it has a greater impact on male speakers than female speakers.
|
2306.07503 | Lin Ma | Lin Ma and Conan Liu and Tiefeng Ma and Shuangzhe Liu | PaVa: a novel Path-based Valley-seeking clustering algorithm | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Clustering methods are being applied to a wider range of scenarios involving
more complex datasets, where the shapes of clusters tend to be arbitrary. In
this paper, we propose a novel Path-based Valley-seeking clustering algorithm
for arbitrarily shaped clusters. This work aims to seek the valleys among
clusters and then individually extract clusters. Three vital techniques are
used in this algorithm. First, path distance (minmax distance) is employed to
transform the irregular boundaries among clusters, that is density valleys,
into perfect spherical shells. Second, a suitable density measurement,
$k$-distance, is employed to make adjustment on Minimum Spanning Tree, by which
a robust minmax distance is calculated. Third, we seek the transformed density
valleys by determining their centers and radius. First, the clusters are
wrapped in spherical shells after the distance transformation, making the
extraction process efficient even with clusters of arbitrary shape. Second,
adjusted Minimum Spanning Tree enhances the robustness of minmax distance under
different kinds of noise. Last, the number of clusters does not need to be
inputted or decided manually due to the individual extraction process. After
applying the proposed algorithm to several commonly used synthetic datasets,
the results indicate that the Path-based Valley-seeking algorithm is accurate
and efficient. The algorithm is based on the dissimilarity of objects, so it
can be applied to a wide range of fields. Its performance on real-world
datasets illustrates its versatility.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 02:29:34 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Ma",
"Lin",
""
],
[
"Liu",
"Conan",
""
],
[
"Ma",
"Tiefeng",
""
],
[
"Liu",
"Shuangzhe",
""
]
] | TITLE: PaVa: a novel Path-based Valley-seeking clustering algorithm
ABSTRACT: Clustering methods are being applied to a wider range of scenarios involving
more complex datasets, where the shapes of clusters tend to be arbitrary. In
this paper, we propose a novel Path-based Valley-seeking clustering algorithm
for arbitrarily shaped clusters. This work aims to seek the valleys among
clusters and then individually extract clusters. Three vital techniques are
used in this algorithm. First, path distance (minmax distance) is employed to
transform the irregular boundaries among clusters, that is density valleys,
into perfect spherical shells. Second, a suitable density measurement,
$k$-distance, is employed to make adjustment on Minimum Spanning Tree, by which
a robust minmax distance is calculated. Third, we seek the transformed density
valleys by determining their centers and radius. First, the clusters are
wrapped in spherical shells after the distance transformation, making the
extraction process efficient even with clusters of arbitrary shape. Second,
adjusted Minimum Spanning Tree enhances the robustness of minmax distance under
different kinds of noise. Last, the number of clusters does not need to be
inputted or decided manually due to the individual extraction process. After
applying the proposed algorithm to several commonly used synthetic datasets,
the results indicate that the Path-based Valley-seeking algorithm is accurate
and efficient. The algorithm is based on the dissimilarity of objects, so it
can be applied to a wide range of fields. Its performance on real-world
datasets illustrates its versatility.
|
2306.07512 | Wang Ruijie | Ruijie Wang, Baoyu Li, Yichen Lu, Dachun Sun, Jinning Li, Yuchen Yan,
Shengzhong Liu, Hanghang Tong, Tarek F. Abdelzaher | Noisy Positive-Unlabeled Learning with Self-Training for Speculative
Knowledge Graph Reasoning | This paper is accepted by ACL-Findings 2023 | null | null | null | cs.LG cs.AI cs.CL cs.SI | http://creativecommons.org/licenses/by/4.0/ | This paper studies speculative reasoning task on real-world knowledge graphs
(KG) that contain both \textit{false negative issue} (i.e., potential true
facts being excluded) and \textit{false positive issue} (i.e., unreliable or
outdated facts being included). State-of-the-art methods fall short in the
speculative reasoning ability, as they assume the correctness of a fact is
solely determined by its presence in KG, making them vulnerable to false
negative/positive issues. The new reasoning task is formulated as a noisy
Positive-Unlabeled learning problem. We propose a variational framework, namely
nPUGraph, that jointly estimates the correctness of both collected and
uncollected facts (which we call \textit{label posterior}) and updates model
parameters during training. The label posterior estimation facilitates
speculative reasoning from two perspectives. First, it improves the robustness
of a label posterior-aware graph encoder against false positive links. Second,
it identifies missing facts to provide high-quality grounds of reasoning. They
are unified in a simple yet effective self-training procedure. Empirically,
extensive experiments on three benchmark KG and one Twitter dataset with
various degrees of false negative/positive cases demonstrate the effectiveness
of nPUGraph.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 02:43:21 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Wang",
"Ruijie",
""
],
[
"Li",
"Baoyu",
""
],
[
"Lu",
"Yichen",
""
],
[
"Sun",
"Dachun",
""
],
[
"Li",
"Jinning",
""
],
[
"Yan",
"Yuchen",
""
],
[
"Liu",
"Shengzhong",
""
],
[
"Tong",
"Hanghang",
""
],
[
"Abdelzaher",
"Tarek F.",
""
]
] | TITLE: Noisy Positive-Unlabeled Learning with Self-Training for Speculative
Knowledge Graph Reasoning
ABSTRACT: This paper studies speculative reasoning task on real-world knowledge graphs
(KG) that contain both \textit{false negative issue} (i.e., potential true
facts being excluded) and \textit{false positive issue} (i.e., unreliable or
outdated facts being included). State-of-the-art methods fall short in the
speculative reasoning ability, as they assume the correctness of a fact is
solely determined by its presence in KG, making them vulnerable to false
negative/positive issues. The new reasoning task is formulated as a noisy
Positive-Unlabeled learning problem. We propose a variational framework, namely
nPUGraph, that jointly estimates the correctness of both collected and
uncollected facts (which we call \textit{label posterior}) and updates model
parameters during training. The label posterior estimation facilitates
speculative reasoning from two perspectives. First, it improves the robustness
of a label posterior-aware graph encoder against false positive links. Second,
it identifies missing facts to provide high-quality grounds of reasoning. They
are unified in a simple yet effective self-training procedure. Empirically,
extensive experiments on three benchmark KG and one Twitter dataset with
various degrees of false negative/positive cases demonstrate the effectiveness
of nPUGraph.
|
2306.07515 | Meng Liu | Meng Liu, Liqiang Nie, Yunxiao Wang, Meng Wang, Yong Rui | A Survey on Video Moment Localization | null | ACM Comput. Surv. 55, 9, Article 188 (September 2023) | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Video moment localization, also known as video moment retrieval, aiming to
search a target segment within a video described by a given natural language
query. Beyond the task of temporal action localization whereby the target
actions are pre-defined, video moment retrieval can query arbitrary complex
activities. In this survey paper, we aim to present a comprehensive review of
existing video moment localization techniques, including supervised, weakly
supervised, and unsupervised ones. We also review the datasets available for
video moment localization and group results of related work. In addition, we
discuss promising future directions for this field, in particular large-scale
datasets and interpretable video moment localization models.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 02:57:32 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Liu",
"Meng",
""
],
[
"Nie",
"Liqiang",
""
],
[
"Wang",
"Yunxiao",
""
],
[
"Wang",
"Meng",
""
],
[
"Rui",
"Yong",
""
]
] | TITLE: A Survey on Video Moment Localization
ABSTRACT: Video moment localization, also known as video moment retrieval, aiming to
search a target segment within a video described by a given natural language
query. Beyond the task of temporal action localization whereby the target
actions are pre-defined, video moment retrieval can query arbitrary complex
activities. In this survey paper, we aim to present a comprehensive review of
existing video moment localization techniques, including supervised, weakly
supervised, and unsupervised ones. We also review the datasets available for
video moment localization and group results of related work. In addition, we
discuss promising future directions for this field, in particular large-scale
datasets and interpretable video moment localization models.
|
2306.07559 | Xiangchun Xu | Xiangchun Xu | Marking anything: application of point cloud in extracting video target
features | 8 pages | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Extracting retrievable features from video is of great significance for
structured video database construction, video copyright protection and fake
video rumor refutation. Inspired by point cloud data processing, this paper
proposes a method for marking anything (MA) in the video, which can extract the
contour features of any target in the video and convert it into a feature
vector with a length of 256 that can be retrieved. The algorithm uses YOLO-v8
algorithm, multi-object tracking algorithm and PointNet++ to extract contour of
the video detection target to form spatial point cloud data. Then extract the
point cloud feature vector and use it as the retrievable feature of the video
detection target. In order to verify the effectiveness and robustness of
contour feature, some datasets are crawled from Dou Yin and Kinetics-700
dataset as experimental data. For Dou Yin's homogenized videos, the proposed
contour features achieve retrieval accuracy higher than 97% in Top1 return
mode. For videos from Kinetics 700, the contour feature also showed good
robustness for partial clip mode video tracing.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 06:16:49 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Xu",
"Xiangchun",
""
]
] | TITLE: Marking anything: application of point cloud in extracting video target
features
ABSTRACT: Extracting retrievable features from video is of great significance for
structured video database construction, video copyright protection and fake
video rumor refutation. Inspired by point cloud data processing, this paper
proposes a method for marking anything (MA) in the video, which can extract the
contour features of any target in the video and convert it into a feature
vector with a length of 256 that can be retrieved. The algorithm uses YOLO-v8
algorithm, multi-object tracking algorithm and PointNet++ to extract contour of
the video detection target to form spatial point cloud data. Then extract the
point cloud feature vector and use it as the retrievable feature of the video
detection target. In order to verify the effectiveness and robustness of
contour feature, some datasets are crawled from Dou Yin and Kinetics-700
dataset as experimental data. For Dou Yin's homogenized videos, the proposed
contour features achieve retrieval accuracy higher than 97% in Top1 return
mode. For videos from Kinetics 700, the contour feature also showed good
robustness for partial clip mode video tracing.
|
2306.07597 | Xiang Huang | Xiang Huang, Sitao Cheng, Yiheng Shu, Yuheng Bao, Yuzhong Qu | Question Decomposition Tree for Answering Complex Questions over
Knowledge Bases | Accepted by AAAI2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Knowledge base question answering (KBQA) has attracted a lot of interest in
recent years, especially for complex questions which require multiple facts to
answer. Question decomposition is a promising way to answer complex questions.
Existing decomposition methods split the question into sub-questions according
to a single compositionality type, which is not sufficient for questions
involving multiple compositionality types. In this paper, we propose Question
Decomposition Tree (QDT) to represent the structure of complex questions.
Inspired by recent advances in natural language generation (NLG), we present a
two-staged method called Clue-Decipher to generate QDT. It can leverage the
strong ability of NLG model and simultaneously preserve the original questions.
To verify that QDT can enhance KBQA task, we design a decomposition-based KBQA
system called QDTQA. Extensive experiments show that QDTQA outperforms previous
state-of-the-art methods on ComplexWebQuestions dataset. Besides, our
decomposition method improves an existing KBQA system by 12% and sets a new
state-of-the-art on LC-QuAD 1.0.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 07:44:29 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Huang",
"Xiang",
""
],
[
"Cheng",
"Sitao",
""
],
[
"Shu",
"Yiheng",
""
],
[
"Bao",
"Yuheng",
""
],
[
"Qu",
"Yuzhong",
""
]
] | TITLE: Question Decomposition Tree for Answering Complex Questions over
Knowledge Bases
ABSTRACT: Knowledge base question answering (KBQA) has attracted a lot of interest in
recent years, especially for complex questions which require multiple facts to
answer. Question decomposition is a promising way to answer complex questions.
Existing decomposition methods split the question into sub-questions according
to a single compositionality type, which is not sufficient for questions
involving multiple compositionality types. In this paper, we propose Question
Decomposition Tree (QDT) to represent the structure of complex questions.
Inspired by recent advances in natural language generation (NLG), we present a
two-staged method called Clue-Decipher to generate QDT. It can leverage the
strong ability of NLG model and simultaneously preserve the original questions.
To verify that QDT can enhance KBQA task, we design a decomposition-based KBQA
system called QDTQA. Extensive experiments show that QDTQA outperforms previous
state-of-the-art methods on ComplexWebQuestions dataset. Besides, our
decomposition method improves an existing KBQA system by 12% and sets a new
state-of-the-art on LC-QuAD 1.0.
|
2306.07598 | Zhiwen Fan | Panwang Pan, Zhiwen Fan, Brandon Y. Feng, Peihao Wang, Chenxin Li,
Zhangyang Wang | Learning to Estimate 6DoF Pose from Limited Data: A Few-Shot,
Generalizable Approach using RGB Images | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The accurate estimation of six degrees-of-freedom (6DoF) object poses is
essential for many applications in robotics and augmented reality. However,
existing methods for 6DoF pose estimation often depend on CAD templates or
dense support views, restricting their usefulness in realworld situations. In
this study, we present a new cascade framework named Cas6D for few-shot 6DoF
pose estimation that is generalizable and uses only RGB images. To address the
false positives of target object detection in the extreme few-shot setting, our
framework utilizes a selfsupervised pre-trained ViT to learn robust feature
representations. Then, we initialize the nearest top-K pose candidates based on
similarity score and refine the initial poses using feature pyramids to
formulate and update the cascade warped feature volume, which encodes context
at increasingly finer scales. By discretizing the pose search range using
multiple pose bins and progressively narrowing the pose search range in each
stage using predictions from the previous stage, Cas6D can overcome the large
gap between pose candidates and ground truth poses, which is a common failure
mode in sparse-view scenarios. Experimental results on the LINEMOD and GenMOP
datasets demonstrate that Cas6D outperforms state-of-the-art methods by 9.2%
and 3.8% accuracy (Proj-5) under the 32-shot setting compared to OnePose++ and
Gen6D.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 07:45:42 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Pan",
"Panwang",
""
],
[
"Fan",
"Zhiwen",
""
],
[
"Feng",
"Brandon Y.",
""
],
[
"Wang",
"Peihao",
""
],
[
"Li",
"Chenxin",
""
],
[
"Wang",
"Zhangyang",
""
]
] | TITLE: Learning to Estimate 6DoF Pose from Limited Data: A Few-Shot,
Generalizable Approach using RGB Images
ABSTRACT: The accurate estimation of six degrees-of-freedom (6DoF) object poses is
essential for many applications in robotics and augmented reality. However,
existing methods for 6DoF pose estimation often depend on CAD templates or
dense support views, restricting their usefulness in realworld situations. In
this study, we present a new cascade framework named Cas6D for few-shot 6DoF
pose estimation that is generalizable and uses only RGB images. To address the
false positives of target object detection in the extreme few-shot setting, our
framework utilizes a selfsupervised pre-trained ViT to learn robust feature
representations. Then, we initialize the nearest top-K pose candidates based on
similarity score and refine the initial poses using feature pyramids to
formulate and update the cascade warped feature volume, which encodes context
at increasingly finer scales. By discretizing the pose search range using
multiple pose bins and progressively narrowing the pose search range in each
stage using predictions from the previous stage, Cas6D can overcome the large
gap between pose candidates and ground truth poses, which is a common failure
mode in sparse-view scenarios. Experimental results on the LINEMOD and GenMOP
datasets demonstrate that Cas6D outperforms state-of-the-art methods by 9.2%
and 3.8% accuracy (Proj-5) under the 32-shot setting compared to OnePose++ and
Gen6D.
|
2306.07607 | Ping Li | Ping Li, Weijie Zhao, Chao Wang, Qi Xia, Alice Wu, Lijun Peng | Practice with Graph-based ANN Algorithms on Sparse Data: Chi-square
Two-tower model, HNSW, Sign Cauchy Projections | null | null | null | null | cs.IR stat.ML | http://creativecommons.org/licenses/by/4.0/ | Sparse data are common. The traditional ``handcrafted'' features are often
sparse. Embedding vectors from trained models can also be very sparse, for
example, embeddings trained via the ``ReLu'' activation function. In this
paper, we report our exploration of efficient search in sparse data with
graph-based ANN algorithms (e.g., HNSW, or SONG which is the GPU version of
HNSW), which are popular in industrial practice, e.g., search and ads
(advertising).
We experiment with the proprietary ads targeting application, as well as
benchmark public datasets. For ads targeting, we train embeddings with the
standard ``cosine two-tower'' model and we also develop the ``chi-square
two-tower'' model. Both models produce (highly) sparse embeddings when they are
integrated with the ``ReLu'' activation function. In EBR (embedding-based
retrieval) applications, after we the embeddings are trained, the next crucial
task is the approximate near neighbor (ANN) search for serving. While there are
many ANN algorithms we can choose from, in this study, we focus on the
graph-based ANN algorithm (e.g., HNSW-type).
Sparse embeddings should help improve the efficiency of EBR. One benefit is
the reduced memory cost for the embeddings. The other obvious benefit is the
reduced computational time for evaluating similarities, because, for
graph-based ANN algorithms such as HNSW, computing similarities is often the
dominating cost. In addition to the effort on leveraging data sparsity for
storage and computation, we also integrate ``sign cauchy random projections''
(SignCRP) to hash vectors to bits, to further reduce the memory cost and speed
up the ANN search. In NIPS'13, SignCRP was proposed to hash the chi-square
similarity, which is a well-adopted nonlinear kernel in NLP and computer
vision. Therefore, the chi-square two-tower model, SignCRP, and HNSW are now
tightly integrated.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 08:05:30 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Li",
"Ping",
""
],
[
"Zhao",
"Weijie",
""
],
[
"Wang",
"Chao",
""
],
[
"Xia",
"Qi",
""
],
[
"Wu",
"Alice",
""
],
[
"Peng",
"Lijun",
""
]
] | TITLE: Practice with Graph-based ANN Algorithms on Sparse Data: Chi-square
Two-tower model, HNSW, Sign Cauchy Projections
ABSTRACT: Sparse data are common. The traditional ``handcrafted'' features are often
sparse. Embedding vectors from trained models can also be very sparse, for
example, embeddings trained via the ``ReLu'' activation function. In this
paper, we report our exploration of efficient search in sparse data with
graph-based ANN algorithms (e.g., HNSW, or SONG which is the GPU version of
HNSW), which are popular in industrial practice, e.g., search and ads
(advertising).
We experiment with the proprietary ads targeting application, as well as
benchmark public datasets. For ads targeting, we train embeddings with the
standard ``cosine two-tower'' model and we also develop the ``chi-square
two-tower'' model. Both models produce (highly) sparse embeddings when they are
integrated with the ``ReLu'' activation function. In EBR (embedding-based
retrieval) applications, after we the embeddings are trained, the next crucial
task is the approximate near neighbor (ANN) search for serving. While there are
many ANN algorithms we can choose from, in this study, we focus on the
graph-based ANN algorithm (e.g., HNSW-type).
Sparse embeddings should help improve the efficiency of EBR. One benefit is
the reduced memory cost for the embeddings. The other obvious benefit is the
reduced computational time for evaluating similarities, because, for
graph-based ANN algorithms such as HNSW, computing similarities is often the
dominating cost. In addition to the effort on leveraging data sparsity for
storage and computation, we also integrate ``sign cauchy random projections''
(SignCRP) to hash vectors to bits, to further reduce the memory cost and speed
up the ANN search. In NIPS'13, SignCRP was proposed to hash the chi-square
similarity, which is a well-adopted nonlinear kernel in NLP and computer
vision. Therefore, the chi-square two-tower model, SignCRP, and HNSW are now
tightly integrated.
|
2306.07608 | Yizhen Zheng | Yizhen Zheng, He Zhang, Vincent CS Lee, Yu Zheng, Xiao Wang, Shirui
Pan | Finding the Missing-half: Graph Complementary Learning for
Homophily-prone and Heterophily-prone Graphs | Accepted by ICML 2023 | null | null | null | cs.SI cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Real-world graphs generally have only one kind of tendency in their
connections. These connections are either homophily-prone or heterophily-prone.
While graphs with homophily-prone edges tend to connect nodes with the same
class (i.e., intra-class nodes), heterophily-prone edges tend to build
relationships between nodes with different classes (i.e., inter-class nodes).
Existing GNNs only take the original graph during training. The problem with
this approach is that it forgets to take into consideration the ``missing-half"
structural information, that is, heterophily-prone topology for homophily-prone
graphs and homophily-prone topology for heterophily-prone graphs. In our paper,
we introduce Graph cOmplementAry Learning, namely GOAL, which consists of two
components: graph complementation and complemented graph convolution. The first
component finds the missing-half structural information for a given graph to
complement it. The complemented graph has two sets of graphs including both
homophily- and heterophily-prone topology. In the latter component, to handle
complemented graphs, we design a new graph convolution from the perspective of
optimisation. The experiment results show that GOAL consistently outperforms
all baselines in eight real-world datasets.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 08:06:10 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Zheng",
"Yizhen",
""
],
[
"Zhang",
"He",
""
],
[
"Lee",
"Vincent CS",
""
],
[
"Zheng",
"Yu",
""
],
[
"Wang",
"Xiao",
""
],
[
"Pan",
"Shirui",
""
]
] | TITLE: Finding the Missing-half: Graph Complementary Learning for
Homophily-prone and Heterophily-prone Graphs
ABSTRACT: Real-world graphs generally have only one kind of tendency in their
connections. These connections are either homophily-prone or heterophily-prone.
While graphs with homophily-prone edges tend to connect nodes with the same
class (i.e., intra-class nodes), heterophily-prone edges tend to build
relationships between nodes with different classes (i.e., inter-class nodes).
Existing GNNs only take the original graph during training. The problem with
this approach is that it forgets to take into consideration the ``missing-half"
structural information, that is, heterophily-prone topology for homophily-prone
graphs and homophily-prone topology for heterophily-prone graphs. In our paper,
we introduce Graph cOmplementAry Learning, namely GOAL, which consists of two
components: graph complementation and complemented graph convolution. The first
component finds the missing-half structural information for a given graph to
complement it. The complemented graph has two sets of graphs including both
homophily- and heterophily-prone topology. In the latter component, to handle
complemented graphs, we design a new graph convolution from the perspective of
optimisation. The experiment results show that GOAL consistently outperforms
all baselines in eight real-world datasets.
|
2306.07621 | Murtadha Ahmed | Ahmed Murtadha, Shengfeng Pan, Wen Bo, Jianlin Su, Xinxin Cao, Wenze
Zhang, Yunfeng Liu | Rank-Aware Negative Training for Semi-Supervised Text Classification | TACL 2023 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Semi-supervised text classification-based paradigms (SSTC) typically employ
the spirit of self-training. The key idea is to train a deep classifier on
limited labeled texts and then iteratively predict the unlabeled texts as their
pseudo-labels for further training. However, the performance is largely
affected by the accuracy of pseudo-labels, which may not be significant in
real-world scenarios. This paper presents a Rank-aware Negative Training (RNT)
framework to address SSTC in learning with noisy label manner. To alleviate the
noisy information, we adapt a reasoning with uncertainty-based approach to rank
the unlabeled texts based on the evidential support received from the labeled
texts. Moreover, we propose the use of negative training to train RNT based on
the concept that ``the input instance does not belong to the complementary
label''. A complementary label is randomly selected from all labels except the
label on-target. Intuitively, the probability of a true label serving as a
complementary label is low and thus provides less noisy information during the
training, resulting in better performance on the test data. Finally, we
evaluate the proposed solution on various text classification benchmark
datasets. Our extensive experiments show that it consistently overcomes the
state-of-the-art alternatives in most scenarios and achieves competitive
performance in the others. The code of RNT is publicly available
at:https://github.com/amurtadha/RNT.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 08:41:36 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Murtadha",
"Ahmed",
""
],
[
"Pan",
"Shengfeng",
""
],
[
"Bo",
"Wen",
""
],
[
"Su",
"Jianlin",
""
],
[
"Cao",
"Xinxin",
""
],
[
"Zhang",
"Wenze",
""
],
[
"Liu",
"Yunfeng",
""
]
] | TITLE: Rank-Aware Negative Training for Semi-Supervised Text Classification
ABSTRACT: Semi-supervised text classification-based paradigms (SSTC) typically employ
the spirit of self-training. The key idea is to train a deep classifier on
limited labeled texts and then iteratively predict the unlabeled texts as their
pseudo-labels for further training. However, the performance is largely
affected by the accuracy of pseudo-labels, which may not be significant in
real-world scenarios. This paper presents a Rank-aware Negative Training (RNT)
framework to address SSTC in learning with noisy label manner. To alleviate the
noisy information, we adapt a reasoning with uncertainty-based approach to rank
the unlabeled texts based on the evidential support received from the labeled
texts. Moreover, we propose the use of negative training to train RNT based on
the concept that ``the input instance does not belong to the complementary
label''. A complementary label is randomly selected from all labels except the
label on-target. Intuitively, the probability of a true label serving as a
complementary label is low and thus provides less noisy information during the
training, resulting in better performance on the test data. Finally, we
evaluate the proposed solution on various text classification benchmark
datasets. Our extensive experiments show that it consistently overcomes the
state-of-the-art alternatives in most scenarios and achieves competitive
performance in the others. The code of RNT is publicly available
at:https://github.com/amurtadha/RNT.
|
2306.07636 | P\'eter Berkecz | P\'eter Berkecz, Gy\"orgy Orosz, Zsolt Sz\'ant\'o, Gerg\H{o} Szab\'o,
Rich\'ard Farkas | Hybrid lemmatization in HuSpaCy | published at the conference XIX. Magyar Sz\'am\'it\'og\'epes
Nyelv\'eszeti Konferencia (XIX. Hungarian Computational Linguistics
Conference) | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Lemmatization is still not a trivial task for morphologically rich languages.
Previous studies showed that hybrid architectures usually work better for these
languages and can yield great results. This paper presents a hybrid lemmatizer
utilizing both a neural model, dictionaries and hand-crafted rules. We
introduce a hybrid architecture along with empirical results on a widely used
Hungarian dataset. The presented methods are published as three HuSpaCy models.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 09:15:40 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Berkecz",
"Péter",
""
],
[
"Orosz",
"György",
""
],
[
"Szántó",
"Zsolt",
""
],
[
"Szabó",
"Gergő",
""
],
[
"Farkas",
"Richárd",
""
]
] | TITLE: Hybrid lemmatization in HuSpaCy
ABSTRACT: Lemmatization is still not a trivial task for morphologically rich languages.
Previous studies showed that hybrid architectures usually work better for these
languages and can yield great results. This paper presents a hybrid lemmatizer
utilizing both a neural model, dictionaries and hand-crafted rules. We
introduce a hybrid architecture along with empirical results on a widely used
Hungarian dataset. The presented methods are published as three HuSpaCy models.
|
2306.07644 | Tanguy Marchand | Tanguy Marchand, R\'egis Loeb, Ulysse Marteau-Ferey, Jean Ogier du
Terrail, Arthur Pignet | SRATTA : Sample Re-ATTribution Attack of Secure Aggregation in Federated
Learning | Accepted to ICML2023 | null | null | null | cs.LG cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider a cross-silo federated learning (FL) setting where a machine
learning model with a fully connected first layer is trained between different
clients and a central server using FedAvg, and where the aggregation step can
be performed with secure aggregation (SA). We present SRATTA an attack relying
only on aggregated models which, under realistic assumptions, (i) recovers data
samples from the different clients, and (ii) groups data samples coming from
the same client together. While sample recovery has already been explored in an
FL setting, the ability to group samples per client, despite the use of SA, is
novel. This poses a significant unforeseen security threat to FL and
effectively breaks SA. We show that SRATTA is both theoretically grounded and
can be used in practice on realistic models and datasets. We also propose
counter-measures, and claim that clients should play an active role to
guarantee their privacy during training.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 09:33:26 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Marchand",
"Tanguy",
""
],
[
"Loeb",
"Régis",
""
],
[
"Marteau-Ferey",
"Ulysse",
""
],
[
"Terrail",
"Jean Ogier du",
""
],
[
"Pignet",
"Arthur",
""
]
] | TITLE: SRATTA : Sample Re-ATTribution Attack of Secure Aggregation in Federated
Learning
ABSTRACT: We consider a cross-silo federated learning (FL) setting where a machine
learning model with a fully connected first layer is trained between different
clients and a central server using FedAvg, and where the aggregation step can
be performed with secure aggregation (SA). We present SRATTA an attack relying
only on aggregated models which, under realistic assumptions, (i) recovers data
samples from the different clients, and (ii) groups data samples coming from
the same client together. While sample recovery has already been explored in an
FL setting, the ability to group samples per client, despite the use of SA, is
novel. This poses a significant unforeseen security threat to FL and
effectively breaks SA. We show that SRATTA is both theoretically grounded and
can be used in practice on realistic models and datasets. We also propose
counter-measures, and claim that clients should play an active role to
guarantee their privacy during training.
|
2306.07646 | Mengxi Chen | Mengxi Chen, Linyu Xing, Yu Wang, Ya Zhang | Enhanced Multimodal Representation Learning with Cross-modal KD | Accepted by CVPR2023 | null | null | null | cs.CV cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper explores the tasks of leveraging auxiliary modalities which are
only available at training to enhance multimodal representation learning
through cross-modal Knowledge Distillation (KD). The widely adopted mutual
information maximization-based objective leads to a short-cut solution of the
weak teacher, i.e., achieving the maximum mutual information by simply making
the teacher model as weak as the student model. To prevent such a weak
solution, we introduce an additional objective term, i.e., the mutual
information between the teacher and the auxiliary modality model. Besides, to
narrow down the information gap between the student and teacher, we further
propose to minimize the conditional entropy of the teacher given the student.
Novel training schemes based on contrastive learning and adversarial learning
are designed to optimize the mutual information and the conditional entropy,
respectively. Experimental results on three popular multimodal benchmark
datasets have shown that the proposed method outperforms a range of
state-of-the-art approaches for video recognition, video retrieval and emotion
classification.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 09:35:37 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Chen",
"Mengxi",
""
],
[
"Xing",
"Linyu",
""
],
[
"Wang",
"Yu",
""
],
[
"Zhang",
"Ya",
""
]
] | TITLE: Enhanced Multimodal Representation Learning with Cross-modal KD
ABSTRACT: This paper explores the tasks of leveraging auxiliary modalities which are
only available at training to enhance multimodal representation learning
through cross-modal Knowledge Distillation (KD). The widely adopted mutual
information maximization-based objective leads to a short-cut solution of the
weak teacher, i.e., achieving the maximum mutual information by simply making
the teacher model as weak as the student model. To prevent such a weak
solution, we introduce an additional objective term, i.e., the mutual
information between the teacher and the auxiliary modality model. Besides, to
narrow down the information gap between the student and teacher, we further
propose to minimize the conditional entropy of the teacher given the student.
Novel training schemes based on contrastive learning and adversarial learning
are designed to optimize the mutual information and the conditional entropy,
respectively. Experimental results on three popular multimodal benchmark
datasets have shown that the proposed method outperforms a range of
state-of-the-art approaches for video recognition, video retrieval and emotion
classification.
|
2306.07650 | Yuchen Han | Yuchen Han, Chen Xu, Tong Xiao and Jingbo Zhu | Modality Adaption or Regularization? A Case Study on End-to-End Speech
Translation | ACL 2023 Main Conference | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Pre-training and fine-tuning is a paradigm for alleviating the data scarcity
problem in end-to-end speech translation (E2E ST). The commonplace "modality
gap" between speech and text data often leads to inconsistent inputs between
pre-training and fine-tuning. However, we observe that this gap occurs in the
early stages of fine-tuning, but does not have a major impact on the final
performance. On the other hand, we find that there has another gap, which we
call the "capacity gap": high resource tasks (such as ASR and MT) always
require a large model to fit, when the model is reused for a low resource task
(E2E ST), it will get a sub-optimal performance due to the over-fitting. In a
case study, we find that the regularization plays a more important role than
the well-designed modality adaption method, which achieves 29.0 for en-de and
40.3 for en-fr on the MuST-C dataset. Code and models are available at
https://github.com/hannlp/TAB.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 09:42:48 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Han",
"Yuchen",
""
],
[
"Xu",
"Chen",
""
],
[
"Xiao",
"Tong",
""
],
[
"Zhu",
"Jingbo",
""
]
] | TITLE: Modality Adaption or Regularization? A Case Study on End-to-End Speech
Translation
ABSTRACT: Pre-training and fine-tuning is a paradigm for alleviating the data scarcity
problem in end-to-end speech translation (E2E ST). The commonplace "modality
gap" between speech and text data often leads to inconsistent inputs between
pre-training and fine-tuning. However, we observe that this gap occurs in the
early stages of fine-tuning, but does not have a major impact on the final
performance. On the other hand, we find that there has another gap, which we
call the "capacity gap": high resource tasks (such as ASR and MT) always
require a large model to fit, when the model is reused for a low resource task
(E2E ST), it will get a sub-optimal performance due to the over-fitting. In a
case study, we find that the regularization plays a more important role than
the well-designed modality adaption method, which achieves 29.0 for en-de and
40.3 for en-fr on the MuST-C dataset. Code and models are available at
https://github.com/hannlp/TAB.
|
2306.07678 | Guangan Chen | Guangan Chen, Hanhe Lin, Oliver Wiedemann, Dietmar Saupe | Localization of Just Noticeable Difference for Image Compression | null | null | null | null | cs.CV cs.MM eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The just noticeable difference (JND) is the minimal difference between
stimuli that can be detected by a person. The picture-wise just noticeable
difference (PJND) for a given reference image and a compression algorithm
represents the minimal level of compression that causes noticeable differences
in the reconstruction. These differences can only be observed in some specific
regions within the image, dubbed as JND-critical regions. Identifying these
regions can improve the development of image compression algorithms. Due to the
fact that visual perception varies among individuals, determining the PJND
values and JND-critical regions for a target population of consumers requires
subjective assessment experiments involving a sufficiently large number of
observers. In this paper, we propose a novel framework for conducting such
experiments using crowdsourcing. By applying this framework, we created a novel
PJND dataset, KonJND++, consisting of 300 source images, compressed versions
thereof under JPEG or BPG compression, and an average of 43 ratings of PJND and
129 self-reported locations of JND-critical regions for each source image. Our
experiments demonstrate the effectiveness and reliability of our proposed
framework, which is easy to be adapted for collecting a large-scale dataset.
The source code and dataset are available at
https://github.com/angchen-dev/LocJND.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 10:45:24 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Chen",
"Guangan",
""
],
[
"Lin",
"Hanhe",
""
],
[
"Wiedemann",
"Oliver",
""
],
[
"Saupe",
"Dietmar",
""
]
] | TITLE: Localization of Just Noticeable Difference for Image Compression
ABSTRACT: The just noticeable difference (JND) is the minimal difference between
stimuli that can be detected by a person. The picture-wise just noticeable
difference (PJND) for a given reference image and a compression algorithm
represents the minimal level of compression that causes noticeable differences
in the reconstruction. These differences can only be observed in some specific
regions within the image, dubbed as JND-critical regions. Identifying these
regions can improve the development of image compression algorithms. Due to the
fact that visual perception varies among individuals, determining the PJND
values and JND-critical regions for a target population of consumers requires
subjective assessment experiments involving a sufficiently large number of
observers. In this paper, we propose a novel framework for conducting such
experiments using crowdsourcing. By applying this framework, we created a novel
PJND dataset, KonJND++, consisting of 300 source images, compressed versions
thereof under JPEG or BPG compression, and an average of 43 ratings of PJND and
129 self-reported locations of JND-critical regions for each source image. Our
experiments demonstrate the effectiveness and reliability of our proposed
framework, which is easy to be adapted for collecting a large-scale dataset.
The source code and dataset are available at
https://github.com/angchen-dev/LocJND.
|
2306.07709 | Yurong Chen | Yurong Chen, Qian Wang, Zhijian Duan, Haoran Sun, Zhaohua Chen, Xiang
Yan, Xiaotie Deng | Coordinated Dynamic Bidding in Repeated Second-Price Auctions with
Budgets | 43 pages, 12 figures | null | null | null | cs.GT cs.LG econ.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In online ad markets, a rising number of advertisers are employing bidding
agencies to participate in ad auctions. These agencies are specialized in
designing online algorithms and bidding on behalf of their clients. Typically,
an agency usually has information on multiple advertisers, so she can
potentially coordinate bids to help her clients achieve higher utilities than
those under independent bidding.
In this paper, we study coordinated online bidding algorithms in repeated
second-price auctions with budgets. We propose algorithms that guarantee every
client a higher utility than the best she can get under independent bidding. We
show that these algorithms achieve maximal coalition welfare and discuss
bidders' incentives to misreport their budgets, in symmetric cases. Our proofs
combine the techniques of online learning and equilibrium analysis, overcoming
the difficulty of competing with a multi-dimensional benchmark. The performance
of our algorithms is further evaluated by experiments on both synthetic and
real data. To the best of our knowledge, we are the first to consider bidder
coordination in online repeated auctions with constraints.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 11:55:04 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Chen",
"Yurong",
""
],
[
"Wang",
"Qian",
""
],
[
"Duan",
"Zhijian",
""
],
[
"Sun",
"Haoran",
""
],
[
"Chen",
"Zhaohua",
""
],
[
"Yan",
"Xiang",
""
],
[
"Deng",
"Xiaotie",
""
]
] | TITLE: Coordinated Dynamic Bidding in Repeated Second-Price Auctions with
Budgets
ABSTRACT: In online ad markets, a rising number of advertisers are employing bidding
agencies to participate in ad auctions. These agencies are specialized in
designing online algorithms and bidding on behalf of their clients. Typically,
an agency usually has information on multiple advertisers, so she can
potentially coordinate bids to help her clients achieve higher utilities than
those under independent bidding.
In this paper, we study coordinated online bidding algorithms in repeated
second-price auctions with budgets. We propose algorithms that guarantee every
client a higher utility than the best she can get under independent bidding. We
show that these algorithms achieve maximal coalition welfare and discuss
bidders' incentives to misreport their budgets, in symmetric cases. Our proofs
combine the techniques of online learning and equilibrium analysis, overcoming
the difficulty of competing with a multi-dimensional benchmark. The performance
of our algorithms is further evaluated by experiments on both synthetic and
real data. To the best of our knowledge, we are the first to consider bidder
coordination in online repeated auctions with constraints.
|
2306.07719 | Jining Wang | Jining Wang, Delai Qiu, YouMing Liu, Yining Wang, Chuan Chen, Zibin
Zheng, Yuren Zhou | Contextual Dictionary Lookup for Knowledge Graph Completion | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Knowledge graph completion (KGC) aims to solve the incompleteness of
knowledge graphs (KGs) by predicting missing links from known triples, numbers
of knowledge graph embedding (KGE) models have been proposed to perform KGC by
learning embeddings. Nevertheless, most existing embedding models map each
relation into a unique vector, overlooking the specific fine-grained semantics
of them under different entities. Additionally, the few available fine-grained
semantic models rely on clustering algorithms, resulting in limited performance
and applicability due to the cumbersome two-stage training process. In this
paper, we present a novel method utilizing contextual dictionary lookup,
enabling conventional embedding models to learn fine-grained semantics of
relations in an end-to-end manner. More specifically, we represent each
relation using a dictionary that contains multiple latent semantics. The
composition of a given entity and the dictionary's central semantics serves as
the context for generating a lookup, thus determining the fine-grained
semantics of the relation adaptively. The proposed loss function optimizes both
the central and fine-grained semantics simultaneously to ensure their semantic
consistency. Besides, we introduce two metrics to assess the validity and
accuracy of the dictionary lookup operation. We extend several KGE models with
the method, resulting in substantial performance improvements on widely-used
benchmark datasets.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 12:13:41 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Wang",
"Jining",
""
],
[
"Qiu",
"Delai",
""
],
[
"Liu",
"YouMing",
""
],
[
"Wang",
"Yining",
""
],
[
"Chen",
"Chuan",
""
],
[
"Zheng",
"Zibin",
""
],
[
"Zhou",
"Yuren",
""
]
] | TITLE: Contextual Dictionary Lookup for Knowledge Graph Completion
ABSTRACT: Knowledge graph completion (KGC) aims to solve the incompleteness of
knowledge graphs (KGs) by predicting missing links from known triples, numbers
of knowledge graph embedding (KGE) models have been proposed to perform KGC by
learning embeddings. Nevertheless, most existing embedding models map each
relation into a unique vector, overlooking the specific fine-grained semantics
of them under different entities. Additionally, the few available fine-grained
semantic models rely on clustering algorithms, resulting in limited performance
and applicability due to the cumbersome two-stage training process. In this
paper, we present a novel method utilizing contextual dictionary lookup,
enabling conventional embedding models to learn fine-grained semantics of
relations in an end-to-end manner. More specifically, we represent each
relation using a dictionary that contains multiple latent semantics. The
composition of a given entity and the dictionary's central semantics serves as
the context for generating a lookup, thus determining the fine-grained
semantics of the relation adaptively. The proposed loss function optimizes both
the central and fine-grained semantics simultaneously to ensure their semantic
consistency. Besides, we introduce two metrics to assess the validity and
accuracy of the dictionary lookup operation. We extend several KGE models with
the method, resulting in substantial performance improvements on widely-used
benchmark datasets.
|
2306.07724 | Hakan Temiz | Hakan Temiz | Effects of Data Enrichment with Image Transformations on the Performance
of Deep Networks | null | The European Journal of Research and Development, 2(2), 23-33
(2022) | 10.56038/ejrnd.v2i2.23 | null | cs.CV cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Images cannot always be expected to come in a certain standard format and
orientation. Deep networks need to be trained to take into account unexpected
variations in orientation or format. For this purpose, training data should be
enriched to include different conditions. In this study, the effects of data
enrichment on the performance of deep networks in the super resolution problem
were investigated experimentally. A total of six basic image transformations
were used for the enrichment procedures. In the experiments, two deep network
models were trained with variants of the ILSVRC2012 dataset enriched by these
six image transformation processes. Considering a single image transformation,
it has been observed that the data enriched with 180 degree rotation provides
the best results. The most unsuccessful result was obtained when the models
were trained on the enriched data generated by the flip upside down process.
Models scored highest when trained with a mix of all transformations.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 12:22:54 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Temiz",
"Hakan",
""
]
] | TITLE: Effects of Data Enrichment with Image Transformations on the Performance
of Deep Networks
ABSTRACT: Images cannot always be expected to come in a certain standard format and
orientation. Deep networks need to be trained to take into account unexpected
variations in orientation or format. For this purpose, training data should be
enriched to include different conditions. In this study, the effects of data
enrichment on the performance of deep networks in the super resolution problem
were investigated experimentally. A total of six basic image transformations
were used for the enrichment procedures. In the experiments, two deep network
models were trained with variants of the ILSVRC2012 dataset enriched by these
six image transformation processes. Considering a single image transformation,
it has been observed that the data enriched with 180 degree rotation provides
the best results. The most unsuccessful result was obtained when the models
were trained on the enriched data generated by the flip upside down process.
Models scored highest when trained with a mix of all transformations.
|
2306.07737 | Oliver Niggemann | Alexander Windmann and Henrik Steude and Oliver Niggemann | Robustness and Generalization Performance of Deep Learning Models on
Cyber-Physical Systems: A Comparative Study | Accepted at the IJCAI 2023 Workshop of Artificial Intelligence for
Time Series Analysis (AI4TS) | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep learning (DL) models have seen increased attention for time series
forecasting, yet the application on cyber-physical systems (CPS) is hindered by
the lacking robustness of these methods. Thus, this study evaluates the
robustness and generalization performance of DL architectures on multivariate
time series data from CPS. Our investigation focuses on the models' ability to
handle a range of perturbations, such as sensor faults and noise, and assesses
their impact on overall performance. Furthermore, we test the generalization
and transfer learning capabilities of these models by exposing them to
out-of-distribution (OOD) samples. These include deviations from standard
system operations, while the core dynamics of the underlying physical system
are preserved. Additionally, we test how well the models respond to several
data augmentation techniques, including added noise and time warping. Our
experimental framework utilizes a simulated three-tank system, proposed as a
novel benchmark for evaluating the robustness and generalization performance of
DL algorithms in CPS data contexts. The findings reveal that certain DL model
architectures and training techniques exhibit superior effectiveness in
handling OOD samples and various perturbations. These insights have significant
implications for the development of DL models that deliver reliable and robust
performance in real-world CPS applications.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 12:43:59 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Windmann",
"Alexander",
""
],
[
"Steude",
"Henrik",
""
],
[
"Niggemann",
"Oliver",
""
]
] | TITLE: Robustness and Generalization Performance of Deep Learning Models on
Cyber-Physical Systems: A Comparative Study
ABSTRACT: Deep learning (DL) models have seen increased attention for time series
forecasting, yet the application on cyber-physical systems (CPS) is hindered by
the lacking robustness of these methods. Thus, this study evaluates the
robustness and generalization performance of DL architectures on multivariate
time series data from CPS. Our investigation focuses on the models' ability to
handle a range of perturbations, such as sensor faults and noise, and assesses
their impact on overall performance. Furthermore, we test the generalization
and transfer learning capabilities of these models by exposing them to
out-of-distribution (OOD) samples. These include deviations from standard
system operations, while the core dynamics of the underlying physical system
are preserved. Additionally, we test how well the models respond to several
data augmentation techniques, including added noise and time warping. Our
experimental framework utilizes a simulated three-tank system, proposed as a
novel benchmark for evaluating the robustness and generalization performance of
DL algorithms in CPS data contexts. The findings reveal that certain DL model
architectures and training techniques exhibit superior effectiveness in
handling OOD samples and various perturbations. These insights have significant
implications for the development of DL models that deliver reliable and robust
performance in real-world CPS applications.
|
2306.07744 | Daniel Stoller | Simon Durand, Daniel Stoller, Sebastian Ewert | Contrastive Learning-Based Audio to Lyrics Alignment for Multiple
Languages | 5 pages, accepted at the International Conference on Acoustics,
Speech, and Signal Processing (ICASSP) 2023 | ICASSP 2023 - 2023 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP), Rhodes Island, Greece, 2023, pp. 1-5 | 10.1109/ICASSP49357.2023.10096725 | null | cs.SD cs.LG eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Lyrics alignment gained considerable attention in recent years.
State-of-the-art systems either re-use established speech recognition toolkits,
or design end-to-end solutions involving a Connectionist Temporal
Classification (CTC) loss. However, both approaches suffer from specific
weaknesses: toolkits are known for their complexity, and CTC systems use a loss
designed for transcription which can limit alignment accuracy. In this paper,
we use instead a contrastive learning procedure that derives cross-modal
embeddings linking the audio and text domains. This way, we obtain a novel
system that is simple to train end-to-end, can make use of weakly annotated
training data, jointly learns a powerful text model, and is tailored to
alignment. The system is not only the first to yield an average absolute error
below 0.2 seconds on the standard Jamendo dataset but it is also robust to
other languages, even when trained on English data only. Finally, we release
word-level alignments for the JamendoLyrics Multi-Lang dataset.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 13:01:02 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Durand",
"Simon",
""
],
[
"Stoller",
"Daniel",
""
],
[
"Ewert",
"Sebastian",
""
]
] | TITLE: Contrastive Learning-Based Audio to Lyrics Alignment for Multiple
Languages
ABSTRACT: Lyrics alignment gained considerable attention in recent years.
State-of-the-art systems either re-use established speech recognition toolkits,
or design end-to-end solutions involving a Connectionist Temporal
Classification (CTC) loss. However, both approaches suffer from specific
weaknesses: toolkits are known for their complexity, and CTC systems use a loss
designed for transcription which can limit alignment accuracy. In this paper,
we use instead a contrastive learning procedure that derives cross-modal
embeddings linking the audio and text domains. This way, we obtain a novel
system that is simple to train end-to-end, can make use of weakly annotated
training data, jointly learns a powerful text model, and is tailored to
alignment. The system is not only the first to yield an average absolute error
below 0.2 seconds on the standard Jamendo dataset but it is also robust to
other languages, even when trained on English data only. Finally, we release
word-level alignments for the JamendoLyrics Multi-Lang dataset.
|
2306.07758 | Yihan Ma | Yihan Ma, Zhikun Zhang, Ning Yu, Xinlei He, Michael Backes, Yun Shen,
Yang Zhang | Generated Graph Detection | Accepted by ICML 2023 | null | null | null | cs.CR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Graph generative models become increasingly effective for data distribution
approximation and data augmentation. While they have aroused public concerns
about their malicious misuses or misinformation broadcasts, just as what
Deepfake visual and auditory media has been delivering to society. Hence it is
essential to regulate the prevalence of generated graphs. To tackle this
problem, we pioneer the formulation of the generated graph detection problem to
distinguish generated graphs from real ones. We propose the first framework to
systematically investigate a set of sophisticated models and their performance
in four classification scenarios. Each scenario switches between seen and
unseen datasets/generators during testing to get closer to real-world settings
and progressively challenge the classifiers. Extensive experiments evidence
that all the models are qualified for generated graph detection, with specific
models having advantages in specific scenarios. Resulting from the validated
generality and oblivion of the classifiers to unseen datasets/generators, we
draw a safe conclusion that our solution can sustain for a decent while to curb
generated graph misuses.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 13:18:04 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Ma",
"Yihan",
""
],
[
"Zhang",
"Zhikun",
""
],
[
"Yu",
"Ning",
""
],
[
"He",
"Xinlei",
""
],
[
"Backes",
"Michael",
""
],
[
"Shen",
"Yun",
""
],
[
"Zhang",
"Yang",
""
]
] | TITLE: Generated Graph Detection
ABSTRACT: Graph generative models become increasingly effective for data distribution
approximation and data augmentation. While they have aroused public concerns
about their malicious misuses or misinformation broadcasts, just as what
Deepfake visual and auditory media has been delivering to society. Hence it is
essential to regulate the prevalence of generated graphs. To tackle this
problem, we pioneer the formulation of the generated graph detection problem to
distinguish generated graphs from real ones. We propose the first framework to
systematically investigate a set of sophisticated models and their performance
in four classification scenarios. Each scenario switches between seen and
unseen datasets/generators during testing to get closer to real-world settings
and progressively challenge the classifiers. Extensive experiments evidence
that all the models are qualified for generated graph detection, with specific
models having advantages in specific scenarios. Resulting from the validated
generality and oblivion of the classifiers to unseen datasets/generators, we
draw a safe conclusion that our solution can sustain for a decent while to curb
generated graph misuses.
|
2306.07760 | Yi Guo | Yi Guo, Nan Cao, Xiaoyu Qi, Haoyang Li, Danqing Shi, Jing Zhang, Qing
Chen, and Daniel Weiskopf | Urania: Visualizing Data Analysis Pipelines for Natural Language-Based
Data Exploration | null | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Exploratory Data Analysis (EDA) is an essential yet tedious process for
examining a new dataset. To facilitate it, natural language interfaces (NLIs)
can help people intuitively explore the dataset via data-oriented questions.
However, existing NLIs primarily focus on providing accurate answers to
questions, with few offering explanations or presentations of the data analysis
pipeline used to uncover the answer. Such presentations are crucial for EDA as
they enhance the interpretability and reliability of the answer, while also
helping users understand the analysis process and derive insights. To fill this
gap, we introduce Urania, a natural language interactive system that is able to
visualize the data analysis pipelines used to resolve input questions. It
integrates a natural language interface that allows users to explore data via
questions, and a novel data-aware question decomposition algorithm that
resolves each input question into a data analysis pipeline. This pipeline is
visualized in the form of a datamation, with animated presentations of analysis
operations and their corresponding data changes. Through two quantitative
experiments and expert interviews, we demonstrated that our data-aware question
decomposition algorithm outperforms the state-of-the-art technique in terms of
execution accuracy, and that Urania can help people explore datasets better. In
the end, we discuss the observations from the studies and the potential future
works.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 13:18:47 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Guo",
"Yi",
""
],
[
"Cao",
"Nan",
""
],
[
"Qi",
"Xiaoyu",
""
],
[
"Li",
"Haoyang",
""
],
[
"Shi",
"Danqing",
""
],
[
"Zhang",
"Jing",
""
],
[
"Chen",
"Qing",
""
],
[
"Weiskopf",
"Daniel",
""
]
] | TITLE: Urania: Visualizing Data Analysis Pipelines for Natural Language-Based
Data Exploration
ABSTRACT: Exploratory Data Analysis (EDA) is an essential yet tedious process for
examining a new dataset. To facilitate it, natural language interfaces (NLIs)
can help people intuitively explore the dataset via data-oriented questions.
However, existing NLIs primarily focus on providing accurate answers to
questions, with few offering explanations or presentations of the data analysis
pipeline used to uncover the answer. Such presentations are crucial for EDA as
they enhance the interpretability and reliability of the answer, while also
helping users understand the analysis process and derive insights. To fill this
gap, we introduce Urania, a natural language interactive system that is able to
visualize the data analysis pipelines used to resolve input questions. It
integrates a natural language interface that allows users to explore data via
questions, and a novel data-aware question decomposition algorithm that
resolves each input question into a data analysis pipeline. This pipeline is
visualized in the form of a datamation, with animated presentations of analysis
operations and their corresponding data changes. Through two quantitative
experiments and expert interviews, we demonstrated that our data-aware question
decomposition algorithm outperforms the state-of-the-art technique in terms of
execution accuracy, and that Urania can help people explore datasets better. In
the end, we discuss the observations from the studies and the potential future
works.
|
2306.07785 | T. N. Vijaykumar | Conor Green, Cole Nelson, Mithuna Thottethodi, and T. N. Vijaykumar | SafeBet: Secure, Simple, and Fast Speculative Execution | null | null | null | null | cs.AR cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spectre attacks exploit microprocessor speculative execution to read and
transmit forbidden data outside the attacker's trust domain and sandbox. Recent
hardware schemes allow potentially-unsafe speculative accesses but prevent the
secret's transmission by delaying most access-dependent instructions even in
the predominantly-common, no-attack case, which incurs performance loss and
hardware complexity. Instead, we propose SafeBet which allows only, and does
not delay most, safe accesses, achieving both security and high performance.
SafeBet is based on the key observation that speculatively accessing a
destination location is safe if the location's access by the same static trust
domain has been committed previously; and potentially unsafe, otherwise. We
extend this observation to handle inter trust-domain code and data
interactions. SafeBet employs the Speculative Memory Access Control Table
(SMACT) to track non-speculative trust domain code region-destination pairs.
Disallowed accesses wait until reaching commit to trigger well-known replay,
with virtually no change to the pipeline. Software simulations using SpecCPU
benchmarks show that SafeBet uses an 8.3-KB SMACT per core to perform within 6%
on average (63% at worst) of the unsafe baseline behind which NDA-restrictive,
a previous scheme of security and hardware complexity comparable to SafeBet's,
lags by 83% on average.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 14:07:52 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Green",
"Conor",
""
],
[
"Nelson",
"Cole",
""
],
[
"Thottethodi",
"Mithuna",
""
],
[
"Vijaykumar",
"T. N.",
""
]
] | TITLE: SafeBet: Secure, Simple, and Fast Speculative Execution
ABSTRACT: Spectre attacks exploit microprocessor speculative execution to read and
transmit forbidden data outside the attacker's trust domain and sandbox. Recent
hardware schemes allow potentially-unsafe speculative accesses but prevent the
secret's transmission by delaying most access-dependent instructions even in
the predominantly-common, no-attack case, which incurs performance loss and
hardware complexity. Instead, we propose SafeBet which allows only, and does
not delay most, safe accesses, achieving both security and high performance.
SafeBet is based on the key observation that speculatively accessing a
destination location is safe if the location's access by the same static trust
domain has been committed previously; and potentially unsafe, otherwise. We
extend this observation to handle inter trust-domain code and data
interactions. SafeBet employs the Speculative Memory Access Control Table
(SMACT) to track non-speculative trust domain code region-destination pairs.
Disallowed accesses wait until reaching commit to trigger well-known replay,
with virtually no change to the pipeline. Software simulations using SpecCPU
benchmarks show that SafeBet uses an 8.3-KB SMACT per core to perform within 6%
on average (63% at worst) of the unsafe baseline behind which NDA-restrictive,
a previous scheme of security and hardware complexity comparable to SafeBet's,
lags by 83% on average.
|
2306.07790 | David Samuel | Matias Jentoft and David Samuel | NoCoLA: The Norwegian Corpus of Linguistic Acceptability | Published at NoDaLiDa 2023 | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While there has been a surge of large language models for Norwegian in recent
years, we lack any tool to evaluate their understanding of grammaticality. We
present two new Norwegian datasets for this task. NoCoLA_class is a supervised
binary classification task where the goal is to discriminate between acceptable
and non-acceptable sentences. On the other hand, NoCoLA_zero is a purely
diagnostic task for evaluating the grammatical judgement of a language model in
a completely zero-shot manner, i.e. without any further training. In this
paper, we describe both datasets in detail, show how to use them for different
flavors of language models, and conduct a comparative study of the existing
Norwegian language models.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 14:11:19 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Jentoft",
"Matias",
""
],
[
"Samuel",
"David",
""
]
] | TITLE: NoCoLA: The Norwegian Corpus of Linguistic Acceptability
ABSTRACT: While there has been a surge of large language models for Norwegian in recent
years, we lack any tool to evaluate their understanding of grammaticality. We
present two new Norwegian datasets for this task. NoCoLA_class is a supervised
binary classification task where the goal is to discriminate between acceptable
and non-acceptable sentences. On the other hand, NoCoLA_zero is a purely
diagnostic task for evaluating the grammatical judgement of a language model in
a completely zero-shot manner, i.e. without any further training. In this
paper, we describe both datasets in detail, show how to use them for different
flavors of language models, and conduct a comparative study of the existing
Norwegian language models.
|
2306.07803 | Dhananjay Bhaskar | Dhananjay Bhaskar, Sumner Magruder, Edward De Brouwer, Aarthi Venkat,
Frederik Wenkel, Guy Wolf, Smita Krishnaswamy | Inferring dynamic regulatory interaction graphs from time series data
with perturbations | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Complex systems are characterized by intricate interactions between entities
that evolve dynamically over time. Accurate inference of these dynamic
relationships is crucial for understanding and predicting system behavior. In
this paper, we propose Regulatory Temporal Interaction Network Inference
(RiTINI) for inferring time-varying interaction graphs in complex systems using
a novel combination of space-and-time graph attentions and graph neural
ordinary differential equations (ODEs). RiTINI leverages time-lapse signals on
a graph prior, as well as perturbations of signals at various nodes in order to
effectively capture the dynamics of the underlying system. This approach is
distinct from traditional causal inference networks, which are limited to
inferring acyclic and static graphs. In contrast, RiTINI can infer cyclic,
directed, and time-varying graphs, providing a more comprehensive and accurate
representation of complex systems. The graph attention mechanism in RiTINI
allows the model to adaptively focus on the most relevant interactions in time
and space, while the graph neural ODEs enable continuous-time modeling of the
system's dynamics. We evaluate RiTINI's performance on various simulated and
real-world datasets, demonstrating its state-of-the-art capability in inferring
interaction graphs compared to previous methods.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 14:25:26 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Bhaskar",
"Dhananjay",
""
],
[
"Magruder",
"Sumner",
""
],
[
"De Brouwer",
"Edward",
""
],
[
"Venkat",
"Aarthi",
""
],
[
"Wenkel",
"Frederik",
""
],
[
"Wolf",
"Guy",
""
],
[
"Krishnaswamy",
"Smita",
""
]
] | TITLE: Inferring dynamic regulatory interaction graphs from time series data
with perturbations
ABSTRACT: Complex systems are characterized by intricate interactions between entities
that evolve dynamically over time. Accurate inference of these dynamic
relationships is crucial for understanding and predicting system behavior. In
this paper, we propose Regulatory Temporal Interaction Network Inference
(RiTINI) for inferring time-varying interaction graphs in complex systems using
a novel combination of space-and-time graph attentions and graph neural
ordinary differential equations (ODEs). RiTINI leverages time-lapse signals on
a graph prior, as well as perturbations of signals at various nodes in order to
effectively capture the dynamics of the underlying system. This approach is
distinct from traditional causal inference networks, which are limited to
inferring acyclic and static graphs. In contrast, RiTINI can infer cyclic,
directed, and time-varying graphs, providing a more comprehensive and accurate
representation of complex systems. The graph attention mechanism in RiTINI
allows the model to adaptively focus on the most relevant interactions in time
and space, while the graph neural ODEs enable continuous-time modeling of the
system's dynamics. We evaluate RiTINI's performance on various simulated and
real-world datasets, demonstrating its state-of-the-art capability in inferring
interaction graphs compared to previous methods.
|
2306.07809 | Diogo Lavado | Diogo Lavado, Cl\'audia Soares, Alessandra Micheletti, Giovanni
Bocchi, Alex Coronati, Manuel Silva and Patrizio Frosini | Low-Resource White-Box Semantic Segmentation of Supporting Towers on 3D
Point Clouds via Signature Shape Identification | null | null | null | null | cs.CV cs.LG math.GT | http://creativecommons.org/licenses/by/4.0/ | Research in 3D semantic segmentation has been increasing performance metrics,
like the IoU, by scaling model complexity and computational resources, leaving
behind researchers and practitioners that (1) cannot access the necessary
resources and (2) do need transparency on the model decision mechanisms. In
this paper, we propose SCENE-Net, a low-resource white-box model for 3D point
cloud semantic segmentation. SCENE-Net identifies signature shapes on the point
cloud via group equivariant non-expansive operators (GENEOs), providing
intrinsic geometric interpretability. Our training time on a laptop is 85~min,
and our inference time is 20~ms. SCENE-Net has 11 trainable geometrical
parameters and requires fewer data than black-box models. SCENE--Net offers
robustness to noisy labeling and data imbalance and has comparable IoU to
state-of-the-art methods. With this paper, we release a 40~000 Km labeled
dataset of rural terrain point clouds and our code implementation.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 14:36:06 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Lavado",
"Diogo",
""
],
[
"Soares",
"Cláudia",
""
],
[
"Micheletti",
"Alessandra",
""
],
[
"Bocchi",
"Giovanni",
""
],
[
"Coronati",
"Alex",
""
],
[
"Silva",
"Manuel",
""
],
[
"Frosini",
"Patrizio",
""
]
] | TITLE: Low-Resource White-Box Semantic Segmentation of Supporting Towers on 3D
Point Clouds via Signature Shape Identification
ABSTRACT: Research in 3D semantic segmentation has been increasing performance metrics,
like the IoU, by scaling model complexity and computational resources, leaving
behind researchers and practitioners that (1) cannot access the necessary
resources and (2) do need transparency on the model decision mechanisms. In
this paper, we propose SCENE-Net, a low-resource white-box model for 3D point
cloud semantic segmentation. SCENE-Net identifies signature shapes on the point
cloud via group equivariant non-expansive operators (GENEOs), providing
intrinsic geometric interpretability. Our training time on a laptop is 85~min,
and our inference time is 20~ms. SCENE-Net has 11 trainable geometrical
parameters and requires fewer data than black-box models. SCENE--Net offers
robustness to noisy labeling and data imbalance and has comparable IoU to
state-of-the-art methods. With this paper, we release a 40~000 Km labeled
dataset of rural terrain point clouds and our code implementation.
|
2306.07831 | Bowen Chen | Ming Y. Lu, Bowen Chen, Andrew Zhang, Drew F.K. Williamson, Richard J.
Chen, Tong Ding, Long Phi Le, Yung-Sung Chuang, Faisal Mahmood | Visual Language Pretrained Multiple Instance Zero-Shot Transfer for
Histopathology Images | Accepted to CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Contrastive visual language pretraining has emerged as a powerful method for
either training new language-aware image encoders or augmenting existing
pretrained models with zero-shot visual recognition capabilities. However,
existing works typically train on large datasets of image-text pairs and have
been designed to perform downstream tasks involving only small to medium
sized-images, neither of which are applicable to the emerging field of
computational pathology where there are limited publicly available paired
image-text datasets and each image can span up to 100,000 x 100,000 pixels. In
this paper we present MI-Zero, a simple and intuitive framework for unleashing
the zero-shot transfer capabilities of contrastively aligned image and text
models on gigapixel histopathology whole slide images, enabling multiple
downstream diagnostic tasks to be carried out by pretrained encoders without
requiring any additional labels. MI-Zero reformulates zero-shot transfer under
the framework of multiple instance learning to overcome the computational
challenge of inference on extremely large images. We used over 550k pathology
reports and other available in-domain text corpora to pre-train our text
encoder. By effectively leveraging strong pre-trained encoders, our best model
pretrained on over 33k histopathology image-caption pairs achieves an average
median zero-shot accuracy of 70.2% across three different real-world cancer
subtyping tasks. Our code is available at:
https://github.com/mahmoodlab/MI-Zero.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 15:05:24 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Lu",
"Ming Y.",
""
],
[
"Chen",
"Bowen",
""
],
[
"Zhang",
"Andrew",
""
],
[
"Williamson",
"Drew F. K.",
""
],
[
"Chen",
"Richard J.",
""
],
[
"Ding",
"Tong",
""
],
[
"Le",
"Long Phi",
""
],
[
"Chuang",
"Yung-Sung",
""
],
[
"Mahmood",
"Faisal",
""
]
] | TITLE: Visual Language Pretrained Multiple Instance Zero-Shot Transfer for
Histopathology Images
ABSTRACT: Contrastive visual language pretraining has emerged as a powerful method for
either training new language-aware image encoders or augmenting existing
pretrained models with zero-shot visual recognition capabilities. However,
existing works typically train on large datasets of image-text pairs and have
been designed to perform downstream tasks involving only small to medium
sized-images, neither of which are applicable to the emerging field of
computational pathology where there are limited publicly available paired
image-text datasets and each image can span up to 100,000 x 100,000 pixels. In
this paper we present MI-Zero, a simple and intuitive framework for unleashing
the zero-shot transfer capabilities of contrastively aligned image and text
models on gigapixel histopathology whole slide images, enabling multiple
downstream diagnostic tasks to be carried out by pretrained encoders without
requiring any additional labels. MI-Zero reformulates zero-shot transfer under
the framework of multiple instance learning to overcome the computational
challenge of inference on extremely large images. We used over 550k pathology
reports and other available in-domain text corpora to pre-train our text
encoder. By effectively leveraging strong pre-trained encoders, our best model
pretrained on over 33k histopathology image-caption pairs achieves an average
median zero-shot accuracy of 70.2% across three different real-world cancer
subtyping tasks. Our code is available at:
https://github.com/mahmoodlab/MI-Zero.
|
2306.07842 | Guangtao Lyu | Guangtao Lyu, Anna Zhu | PSSTRNet: Progressive Segmentation-guided Scene Text Removal Network | Accepted by ICME2022 | 2022 IEEE International Conference on Multimedia and Expo (ICME) | 10.1109/ICME52920.2022.9859792 | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Scene text removal (STR) is a challenging task due to the complex text fonts,
colors, sizes, and background textures in scene images. However, most previous
methods learn both text location and background inpainting implicitly within a
single network, which weakens the text localization mechanism and makes a lossy
background. To tackle these problems, we propose a simple Progressive
Segmentation-guided Scene Text Removal Network(PSSTRNet) to remove the text in
the image iteratively. It contains two decoder branches, a text segmentation
branch, and a text removal branch, with a shared encoder. The text segmentation
branch generates text mask maps as the guidance for the regional removal
branch. In each iteration, the original image, previous text removal result,
and text mask are input to the network to extract the rest part of the text
segments and cleaner text removal result. To get a more accurate text mask map,
an update module is developed to merge the mask map in the current and previous
stages. The final text removal result is obtained by adaptive fusion of results
from all previous stages. A sufficient number of experiments and ablation
studies conducted on the real and synthetic public datasets demonstrate our
proposed method achieves state-of-the-art performance. The source code of our
work is available at:
\href{https://github.com/GuangtaoLyu/PSSTRNet}{https://github.com/GuangtaoLyu/PSSTRNet.}
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 15:20:37 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Lyu",
"Guangtao",
""
],
[
"Zhu",
"Anna",
""
]
] | TITLE: PSSTRNet: Progressive Segmentation-guided Scene Text Removal Network
ABSTRACT: Scene text removal (STR) is a challenging task due to the complex text fonts,
colors, sizes, and background textures in scene images. However, most previous
methods learn both text location and background inpainting implicitly within a
single network, which weakens the text localization mechanism and makes a lossy
background. To tackle these problems, we propose a simple Progressive
Segmentation-guided Scene Text Removal Network(PSSTRNet) to remove the text in
the image iteratively. It contains two decoder branches, a text segmentation
branch, and a text removal branch, with a shared encoder. The text segmentation
branch generates text mask maps as the guidance for the regional removal
branch. In each iteration, the original image, previous text removal result,
and text mask are input to the network to extract the rest part of the text
segments and cleaner text removal result. To get a more accurate text mask map,
an update module is developed to merge the mask map in the current and previous
stages. The final text removal result is obtained by adaptive fusion of results
from all previous stages. A sufficient number of experiments and ablation
studies conducted on the real and synthetic public datasets demonstrate our
proposed method achieves state-of-the-art performance. The source code of our
work is available at:
\href{https://github.com/GuangtaoLyu/PSSTRNet}{https://github.com/GuangtaoLyu/PSSTRNet.}
|
2306.07845 | R\u{a}zvan-Alexandru Sm\u{a}du | Sebastian-Vasile Echim, R\u{a}zvan-Alexandru Sm\u{a}du, Andrei-Marius
Avram, Dumitru-Clementin Cercel, Florin Pop | Adversarial Capsule Networks for Romanian Satire Detection and Sentiment
Analysis | 15 pages, 3 figures, Accepted by NLDB 2023 | null | 10.1007/978-3-031-35320-8_31 | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Satire detection and sentiment analysis are intensively explored natural
language processing (NLP) tasks that study the identification of the satirical
tone from texts and extracting sentiments in relationship with their targets.
In languages with fewer research resources, an alternative is to produce
artificial examples based on character-level adversarial processes to overcome
dataset size limitations. Such samples are proven to act as a regularization
method, thus improving the robustness of models. In this work, we improve the
well-known NLP models (i.e., Convolutional Neural Networks, Long Short-Term
Memory (LSTM), Bidirectional LSTM, Gated Recurrent Units (GRUs), and
Bidirectional GRUs) with adversarial training and capsule networks. The
fine-tuned models are used for satire detection and sentiment analysis tasks in
the Romanian language. The proposed framework outperforms the existing methods
for the two tasks, achieving up to 99.08% accuracy, thus confirming the
improvements added by the capsule layers and the adversarial training in NLP
approaches.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 15:23:44 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Echim",
"Sebastian-Vasile",
""
],
[
"Smădu",
"Răzvan-Alexandru",
""
],
[
"Avram",
"Andrei-Marius",
""
],
[
"Cercel",
"Dumitru-Clementin",
""
],
[
"Pop",
"Florin",
""
]
] | TITLE: Adversarial Capsule Networks for Romanian Satire Detection and Sentiment
Analysis
ABSTRACT: Satire detection and sentiment analysis are intensively explored natural
language processing (NLP) tasks that study the identification of the satirical
tone from texts and extracting sentiments in relationship with their targets.
In languages with fewer research resources, an alternative is to produce
artificial examples based on character-level adversarial processes to overcome
dataset size limitations. Such samples are proven to act as a regularization
method, thus improving the robustness of models. In this work, we improve the
well-known NLP models (i.e., Convolutional Neural Networks, Long Short-Term
Memory (LSTM), Bidirectional LSTM, Gated Recurrent Units (GRUs), and
Bidirectional GRUs) with adversarial training and capsule networks. The
fine-tuned models are used for satire detection and sentiment analysis tasks in
the Romanian language. The proposed framework outperforms the existing methods
for the two tasks, achieving up to 99.08% accuracy, thus confirming the
improvements added by the capsule layers and the adversarial training in NLP
approaches.
|
2306.07883 | Bowen Li | Bowen Li, Hanlin Gu, Ruoxin Chen, Jie Li, Chentao Wu, Na Ruan, Xueming
Si, Lixin Fan | Temporal Gradient Inversion Attacks with Robust Optimization | 24 pages | null | null | null | cs.LG cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Federated Learning (FL) has emerged as a promising approach for collaborative
model training without sharing private data. However, privacy concerns
regarding information exchanged during FL have received significant research
attention. Gradient Inversion Attacks (GIAs) have been proposed to reconstruct
the private data retained by local clients from the exchanged gradients. While
recovering private data, the data dimensions and the model complexity increase,
which thwart data reconstruction by GIAs. Existing methods adopt prior
knowledge about private data to overcome those challenges. In this paper, we
first observe that GIAs with gradients from a single iteration fail to
reconstruct private data due to insufficient dimensions of leaked gradients,
complex model architectures, and invalid gradient information. We investigate a
Temporal Gradient Inversion Attack with a Robust Optimization framework, called
TGIAs-RO, which recovers private data without any prior knowledge by leveraging
multiple temporal gradients. To eliminate the negative impacts of outliers,
e.g., invalid gradients for collaborative optimization, robust statistics are
proposed. Theoretical guarantees on the recovery performance and robustness of
TGIAs-RO against invalid gradients are also provided. Extensive empirical
results on MNIST, CIFAR10, ImageNet and Reuters 21578 datasets show that the
proposed TGIAs-RO with 10 temporal gradients improves reconstruction
performance compared to state-of-the-art methods, even for large batch sizes
(up to 128), complex models like ResNet18, and large datasets like ImageNet
(224*224 pixels). Furthermore, the proposed attack method inspires further
exploration of privacy-preserving methods in the context of FL.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 16:21:34 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Li",
"Bowen",
""
],
[
"Gu",
"Hanlin",
""
],
[
"Chen",
"Ruoxin",
""
],
[
"Li",
"Jie",
""
],
[
"Wu",
"Chentao",
""
],
[
"Ruan",
"Na",
""
],
[
"Si",
"Xueming",
""
],
[
"Fan",
"Lixin",
""
]
] | TITLE: Temporal Gradient Inversion Attacks with Robust Optimization
ABSTRACT: Federated Learning (FL) has emerged as a promising approach for collaborative
model training without sharing private data. However, privacy concerns
regarding information exchanged during FL have received significant research
attention. Gradient Inversion Attacks (GIAs) have been proposed to reconstruct
the private data retained by local clients from the exchanged gradients. While
recovering private data, the data dimensions and the model complexity increase,
which thwart data reconstruction by GIAs. Existing methods adopt prior
knowledge about private data to overcome those challenges. In this paper, we
first observe that GIAs with gradients from a single iteration fail to
reconstruct private data due to insufficient dimensions of leaked gradients,
complex model architectures, and invalid gradient information. We investigate a
Temporal Gradient Inversion Attack with a Robust Optimization framework, called
TGIAs-RO, which recovers private data without any prior knowledge by leveraging
multiple temporal gradients. To eliminate the negative impacts of outliers,
e.g., invalid gradients for collaborative optimization, robust statistics are
proposed. Theoretical guarantees on the recovery performance and robustness of
TGIAs-RO against invalid gradients are also provided. Extensive empirical
results on MNIST, CIFAR10, ImageNet and Reuters 21578 datasets show that the
proposed TGIAs-RO with 10 temporal gradients improves reconstruction
performance compared to state-of-the-art methods, even for large batch sizes
(up to 128), complex models like ResNet18, and large datasets like ImageNet
(224*224 pixels). Furthermore, the proposed attack method inspires further
exploration of privacy-preserving methods in the context of FL.
|
2306.07902 | Lukasz Augustyniak | {\L}ukasz Augustyniak, Szymon Wo\'zniak, Marcin Gruza, Piotr Gramacki,
Krzysztof Rajda, Miko{\l}aj Morzy, Tomasz Kajdanowicz | Massively Multilingual Corpus of Sentiment Datasets and Multi-faceted
Sentiment Classification Benchmark | submitted to NeurIPS 2023 Datasets and Benchmarks track. Dataset:
https://huggingface.co/datasets/Brand24/mms Code:
https://github.com/Brand24-AI/mms_benchmark | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Despite impressive advancements in multilingual corpora collection and model
training, developing large-scale deployments of multilingual models still
presents a significant challenge. This is particularly true for language tasks
that are culture-dependent. One such example is the area of multilingual
sentiment analysis, where affective markers can be subtle and deeply ensconced
in culture. This work presents the most extensive open massively multilingual
corpus of datasets for training sentiment models. The corpus consists of 79
manually selected datasets from over 350 datasets reported in the scientific
literature based on strict quality criteria. The corpus covers 27 languages
representing 6 language families. Datasets can be queried using several
linguistic and functional features. In addition, we present a multi-faceted
sentiment classification benchmark summarizing hundreds of experiments
conducted on different base models, training objectives, dataset collections,
and fine-tuning strategies.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 16:54:13 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Augustyniak",
"Łukasz",
""
],
[
"Woźniak",
"Szymon",
""
],
[
"Gruza",
"Marcin",
""
],
[
"Gramacki",
"Piotr",
""
],
[
"Rajda",
"Krzysztof",
""
],
[
"Morzy",
"Mikołaj",
""
],
[
"Kajdanowicz",
"Tomasz",
""
]
] | TITLE: Massively Multilingual Corpus of Sentiment Datasets and Multi-faceted
Sentiment Classification Benchmark
ABSTRACT: Despite impressive advancements in multilingual corpora collection and model
training, developing large-scale deployments of multilingual models still
presents a significant challenge. This is particularly true for language tasks
that are culture-dependent. One such example is the area of multilingual
sentiment analysis, where affective markers can be subtle and deeply ensconced
in culture. This work presents the most extensive open massively multilingual
corpus of datasets for training sentiment models. The corpus consists of 79
manually selected datasets from over 350 datasets reported in the scientific
literature based on strict quality criteria. The corpus covers 27 languages
representing 6 language families. Datasets can be queried using several
linguistic and functional features. In addition, we present a multi-faceted
sentiment classification benchmark summarizing hundreds of experiments
conducted on different base models, training objectives, dataset collections,
and fine-tuning strategies.
|
2306.07919 | Tianxiang Zhao | Tianxiang Zhao, Wenchao Yu, Suhang Wang, Lu Wang, Xiang Zhang, Yuncong
Chen, Yanchi Liu, Wei Cheng, Haifeng Chen | Skill Disentanglement for Imitation Learning from Suboptimal
Demonstrations | null | Proceedings of the 29th ACM SIGKDD Conference on Knowledge
Discovery and Data Mining (KDD '23), August 6--10, 2023, Long Beach, CA, USA | 10.1145/3580305.3599506 | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Imitation learning has achieved great success in many sequential
decision-making tasks, in which a neural agent is learned by imitating
collected human demonstrations. However, existing algorithms typically require
a large number of high-quality demonstrations that are difficult and expensive
to collect. Usually, a trade-off needs to be made between demonstration quality
and quantity in practice. Targeting this problem, in this work we consider the
imitation of sub-optimal demonstrations, with both a small clean demonstration
set and a large noisy set. Some pioneering works have been proposed, but they
suffer from many limitations, e.g., assuming a demonstration to be of the same
optimality throughout time steps and failing to provide any interpretation
w.r.t knowledge learned from the noisy set. Addressing these problems, we
propose {\method} by evaluating and imitating at the sub-demonstration level,
encoding action primitives of varying quality into different skills.
Concretely, {\method} consists of a high-level controller to discover skills
and a skill-conditioned module to capture action-taking policies, and is
trained following a two-phase pipeline by first discovering skills with all
demonstrations and then adapting the controller to only the clean set. A
mutual-information-based regularization and a dynamic sub-demonstration
optimality estimator are designed to promote disentanglement in the skill
space. Extensive experiments are conducted over two gym environments and a
real-world healthcare dataset to demonstrate the superiority of {\method} in
learning from sub-optimal demonstrations and its improved interpretability by
examining learned skills.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 17:24:37 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Zhao",
"Tianxiang",
""
],
[
"Yu",
"Wenchao",
""
],
[
"Wang",
"Suhang",
""
],
[
"Wang",
"Lu",
""
],
[
"Zhang",
"Xiang",
""
],
[
"Chen",
"Yuncong",
""
],
[
"Liu",
"Yanchi",
""
],
[
"Cheng",
"Wei",
""
],
[
"Chen",
"Haifeng",
""
]
] | TITLE: Skill Disentanglement for Imitation Learning from Suboptimal
Demonstrations
ABSTRACT: Imitation learning has achieved great success in many sequential
decision-making tasks, in which a neural agent is learned by imitating
collected human demonstrations. However, existing algorithms typically require
a large number of high-quality demonstrations that are difficult and expensive
to collect. Usually, a trade-off needs to be made between demonstration quality
and quantity in practice. Targeting this problem, in this work we consider the
imitation of sub-optimal demonstrations, with both a small clean demonstration
set and a large noisy set. Some pioneering works have been proposed, but they
suffer from many limitations, e.g., assuming a demonstration to be of the same
optimality throughout time steps and failing to provide any interpretation
w.r.t knowledge learned from the noisy set. Addressing these problems, we
propose {\method} by evaluating and imitating at the sub-demonstration level,
encoding action primitives of varying quality into different skills.
Concretely, {\method} consists of a high-level controller to discover skills
and a skill-conditioned module to capture action-taking policies, and is
trained following a two-phase pipeline by first discovering skills with all
demonstrations and then adapting the controller to only the clean set. A
mutual-information-based regularization and a dynamic sub-demonstration
optimality estimator are designed to promote disentanglement in the skill
space. Extensive experiments are conducted over two gym environments and a
real-world healthcare dataset to demonstrate the superiority of {\method} in
learning from sub-optimal demonstrations and its improved interpretability by
examining learned skills.
|
2306.07928 | Shuo Han | Shuo Han, Yinan Chen, Jiacheng Liu | Optimizing Investment Strategies with Lazy Factor and Probability
Weighting: A Price Portfolio Forecasting and Mean-Variance Model with
Transaction Costs Approach | Accepted by IEEE SMC 2023 | null | null | null | q-fin.PM cs.CE | http://creativecommons.org/licenses/by/4.0/ | Market traders often engage in the frequent transaction of volatile assets to
optimize their total return. In this study, we introduce a novel investment
strategy model, anchored on the 'lazy factor.' Our approach bifurcates into a
Price Portfolio Forecasting Model and a Mean-Variance Model with Transaction
Costs, utilizing probability weights as the coefficients of laziness factors.
The Price Portfolio Forecasting Model, leveraging the EXPMA Mean Method, plots
the long-term price trend line and forecasts future price movements,
incorporating the tangent slope and rate of change. For short-term investments,
we apply the ARIMA Model to predict ensuing prices. The Mean-Variance Model
with Transaction Costs employs the Monte Carlo Method to formulate the feasible
region. To strike an optimal balance between risk and return, equal probability
weights are incorporated as coefficients of the laziness factor. To assess the
efficacy of this combined strategy, we executed extensive experiments on a
specified dataset. Our findings underscore the model's adaptability and
generalizability, indicating its potential to transform investment strategies.
| [
{
"version": "v1",
"created": "Mon, 12 Jun 2023 15:08:10 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Han",
"Shuo",
""
],
[
"Chen",
"Yinan",
""
],
[
"Liu",
"Jiacheng",
""
]
] | TITLE: Optimizing Investment Strategies with Lazy Factor and Probability
Weighting: A Price Portfolio Forecasting and Mean-Variance Model with
Transaction Costs Approach
ABSTRACT: Market traders often engage in the frequent transaction of volatile assets to
optimize their total return. In this study, we introduce a novel investment
strategy model, anchored on the 'lazy factor.' Our approach bifurcates into a
Price Portfolio Forecasting Model and a Mean-Variance Model with Transaction
Costs, utilizing probability weights as the coefficients of laziness factors.
The Price Portfolio Forecasting Model, leveraging the EXPMA Mean Method, plots
the long-term price trend line and forecasts future price movements,
incorporating the tangent slope and rate of change. For short-term investments,
we apply the ARIMA Model to predict ensuing prices. The Mean-Variance Model
with Transaction Costs employs the Monte Carlo Method to formulate the feasible
region. To strike an optimal balance between risk and return, equal probability
weights are incorporated as coefficients of the laziness factor. To assess the
efficacy of this combined strategy, we executed extensive experiments on a
specified dataset. Our findings underscore the model's adaptability and
generalizability, indicating its potential to transform investment strategies.
|
2306.07934 | Mehran Kazemi | Mehran Kazemi, Quan Yuan, Deepti Bhatia, Najoung Kim, Xin Xu, Vaiva
Imbrasaite, Deepak Ramachandran | BoardgameQA: A Dataset for Natural Language Reasoning with Contradictory
Information | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Automated reasoning with unstructured natural text is a key requirement for
many potential applications of NLP and for developing robust AI systems.
Recently, Language Models (LMs) have demonstrated complex reasoning capacities
even without any finetuning. However, existing evaluation for automated
reasoning assumes access to a consistent and coherent set of information over
which models reason. When reasoning in the real-world, the available
information is frequently inconsistent or contradictory, and therefore models
need to be equipped with a strategy to resolve such conflicts when they arise.
One widely-applicable way of resolving conflicts is to impose preferences over
information sources (e.g., based on source credibility or information recency)
and adopt the source with higher preference. In this paper, we formulate the
problem of reasoning with contradictory information guided by preferences over
sources as the classical problem of defeasible reasoning, and develop a dataset
called BoardgameQA for measuring the reasoning capacity of LMs in this setting.
BoardgameQA also incorporates reasoning with implicit background knowledge, to
better reflect reasoning problems in downstream applications. We benchmark
various LMs on BoardgameQA and the results reveal a significant gap in the
reasoning capacity of state-of-the-art LMs on this problem, showing that
reasoning with conflicting information does not surface out-of-the-box in LMs.
While performance can be improved with finetuning, it nevertheless remains
poor.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 17:39:20 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Kazemi",
"Mehran",
""
],
[
"Yuan",
"Quan",
""
],
[
"Bhatia",
"Deepti",
""
],
[
"Kim",
"Najoung",
""
],
[
"Xu",
"Xin",
""
],
[
"Imbrasaite",
"Vaiva",
""
],
[
"Ramachandran",
"Deepak",
""
]
] | TITLE: BoardgameQA: A Dataset for Natural Language Reasoning with Contradictory
Information
ABSTRACT: Automated reasoning with unstructured natural text is a key requirement for
many potential applications of NLP and for developing robust AI systems.
Recently, Language Models (LMs) have demonstrated complex reasoning capacities
even without any finetuning. However, existing evaluation for automated
reasoning assumes access to a consistent and coherent set of information over
which models reason. When reasoning in the real-world, the available
information is frequently inconsistent or contradictory, and therefore models
need to be equipped with a strategy to resolve such conflicts when they arise.
One widely-applicable way of resolving conflicts is to impose preferences over
information sources (e.g., based on source credibility or information recency)
and adopt the source with higher preference. In this paper, we formulate the
problem of reasoning with contradictory information guided by preferences over
sources as the classical problem of defeasible reasoning, and develop a dataset
called BoardgameQA for measuring the reasoning capacity of LMs in this setting.
BoardgameQA also incorporates reasoning with implicit background knowledge, to
better reflect reasoning problems in downstream applications. We benchmark
various LMs on BoardgameQA and the results reveal a significant gap in the
reasoning capacity of state-of-the-art LMs on this problem, showing that
reasoning with conflicting information does not surface out-of-the-box in LMs.
While performance can be improved with finetuning, it nevertheless remains
poor.
|
2306.07935 | Jiayi Luo | Ruiting Dai, Jiayi Luo, Xucheng Luo, Lisi Mo, Wanlun Ma, Fan Zhou | Multi-modal Representation Learning for Social Post Location Inference | 6 pages, 2023 International Conference on Communications | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Inferring geographic locations via social posts is essential for many
practical location-based applications such as product marketing,
point-of-interest recommendation, and infector tracking for COVID-19. Unlike
image-based location retrieval or social-post text embedding-based location
inference, the combined effect of multi-modal information (i.e., post images,
text, and hashtags) for social post positioning receives less attention. In
this work, we collect real datasets of social posts with images, texts, and
hashtags from Instagram and propose a novel Multi-modal Representation Learning
Framework (MRLF) capable of fusing different modalities of social posts for
location inference. MRLF integrates a multi-head attention mechanism to enhance
location-salient information extraction while significantly improving location
inference compared with single domain-based methods. To overcome the noisy
user-generated textual content, we introduce a novel attention-based
character-aware module that considers the relative dependencies between
characters of social post texts and hashtags for flexible multi-model
information fusion. The experimental results show that MRLF can make accurate
location predictions and open a new door to understanding the multi-modal data
of social posts for online inference tasks.
| [
{
"version": "v1",
"created": "Sun, 11 Jun 2023 02:35:48 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Dai",
"Ruiting",
""
],
[
"Luo",
"Jiayi",
""
],
[
"Luo",
"Xucheng",
""
],
[
"Mo",
"Lisi",
""
],
[
"Ma",
"Wanlun",
""
],
[
"Zhou",
"Fan",
""
]
] | TITLE: Multi-modal Representation Learning for Social Post Location Inference
ABSTRACT: Inferring geographic locations via social posts is essential for many
practical location-based applications such as product marketing,
point-of-interest recommendation, and infector tracking for COVID-19. Unlike
image-based location retrieval or social-post text embedding-based location
inference, the combined effect of multi-modal information (i.e., post images,
text, and hashtags) for social post positioning receives less attention. In
this work, we collect real datasets of social posts with images, texts, and
hashtags from Instagram and propose a novel Multi-modal Representation Learning
Framework (MRLF) capable of fusing different modalities of social posts for
location inference. MRLF integrates a multi-head attention mechanism to enhance
location-salient information extraction while significantly improving location
inference compared with single domain-based methods. To overcome the noisy
user-generated textual content, we introduce a novel attention-based
character-aware module that considers the relative dependencies between
characters of social post texts and hashtags for flexible multi-model
information fusion. The experimental results show that MRLF can make accurate
location predictions and open a new door to understanding the multi-modal data
of social posts for online inference tasks.
|
2306.07944 | Izhak Shafran | Mingqiu Wang, Izhak Shafran, Hagen Soltau, Wei Han, Yuan Cao, Dian Yu,
Laurent El Shafey | Speech-to-Text Adapter and Speech-to-Entity Retriever Augmented LLMs for
Speech Understanding | null | null | null | null | eess.AS cs.AI cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Large Language Models (LLMs) have been applied in the speech domain, often
incurring a performance drop due to misaligned between speech and language
representations. To bridge this gap, we propose a joint speech and language
model (SLM) using a Speech2Text adapter, which maps speech into text token
embedding space without speech information loss. Additionally, using a
CTC-based blank-filtering, we can reduce the speech sequence length to that of
text. In speech MultiWoz dataset (DSTC11 challenge), SLM largely improves the
dialog state tracking (DST) performance (24.7% to 28.4% accuracy). Further to
address errors on rare entities, we augment SLM with a Speech2Entity retriever,
which uses speech to retrieve relevant entities, and then adds them to the
original SLM input as a prefix. With this retrieval-augmented SLM (ReSLM), the
DST performance jumps to 34.6% accuracy. Moreover, augmenting the ASR task with
the dialog understanding task improves the ASR performance from 9.4% to 8.5%
WER.
| [
{
"version": "v1",
"created": "Thu, 8 Jun 2023 22:33:22 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Wang",
"Mingqiu",
""
],
[
"Shafran",
"Izhak",
""
],
[
"Soltau",
"Hagen",
""
],
[
"Han",
"Wei",
""
],
[
"Cao",
"Yuan",
""
],
[
"Yu",
"Dian",
""
],
[
"Shafey",
"Laurent El",
""
]
] | TITLE: Speech-to-Text Adapter and Speech-to-Entity Retriever Augmented LLMs for
Speech Understanding
ABSTRACT: Large Language Models (LLMs) have been applied in the speech domain, often
incurring a performance drop due to misaligned between speech and language
representations. To bridge this gap, we propose a joint speech and language
model (SLM) using a Speech2Text adapter, which maps speech into text token
embedding space without speech information loss. Additionally, using a
CTC-based blank-filtering, we can reduce the speech sequence length to that of
text. In speech MultiWoz dataset (DSTC11 challenge), SLM largely improves the
dialog state tracking (DST) performance (24.7% to 28.4% accuracy). Further to
address errors on rare entities, we augment SLM with a Speech2Entity retriever,
which uses speech to retrieve relevant entities, and then adds them to the
original SLM input as a prefix. With this retrieval-augmented SLM (ReSLM), the
DST performance jumps to 34.6% accuracy. Moreover, augmenting the ASR task with
the dialog understanding task improves the ASR performance from 9.4% to 8.5%
WER.
|
2306.07959 | Disha Makhija | Disha Makhija and Joydeep Ghosh and Nhat Ho | Privacy Preserving Bayesian Federated Learning in Heterogeneous Settings | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In several practical applications of federated learning (FL), the clients are
highly heterogeneous in terms of both their data and compute resources, and
therefore enforcing the same model architecture for each client is very
limiting. Moreover, the need for uncertainty quantification and data privacy
constraints are often particularly amplified for clients that have limited
local data. This paper presents a unified FL framework to simultaneously
address all these constraints and concerns, based on training customized local
Bayesian models that learn well even in the absence of large local datasets. A
Bayesian framework provides a natural way of incorporating supervision in the
form of prior distributions. We use priors in the functional (output) space of
the networks to facilitate collaboration across heterogeneous clients.
Moreover, formal differential privacy guarantees are provided for this
framework. Experiments on standard FL datasets demonstrate that our approach
outperforms strong baselines in both homogeneous and heterogeneous settings and
under strict privacy constraints, while also providing characterizations of
model uncertainties.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 17:55:30 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Makhija",
"Disha",
""
],
[
"Ghosh",
"Joydeep",
""
],
[
"Ho",
"Nhat",
""
]
] | TITLE: Privacy Preserving Bayesian Federated Learning in Heterogeneous Settings
ABSTRACT: In several practical applications of federated learning (FL), the clients are
highly heterogeneous in terms of both their data and compute resources, and
therefore enforcing the same model architecture for each client is very
limiting. Moreover, the need for uncertainty quantification and data privacy
constraints are often particularly amplified for clients that have limited
local data. This paper presents a unified FL framework to simultaneously
address all these constraints and concerns, based on training customized local
Bayesian models that learn well even in the absence of large local datasets. A
Bayesian framework provides a natural way of incorporating supervision in the
form of prior distributions. We use priors in the functional (output) space of
the networks to facilitate collaboration across heterogeneous clients.
Moreover, formal differential privacy guarantees are provided for this
framework. Experiments on standard FL datasets demonstrate that our approach
outperforms strong baselines in both homogeneous and heterogeneous settings and
under strict privacy constraints, while also providing characterizations of
model uncertainties.
|
2306.07968 | Gyungin Shin | Gyungin Shin, Weidi Xie, Samuel Albanie | arXiVeri: Automatic table verification with GPT | Tech report | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Without accurate transcription of numerical data in scientific documents, a
scientist cannot draw accurate conclusions. Unfortunately, the process of
copying numerical data from one paper to another is prone to human error. In
this paper, we propose to meet this challenge through the novel task of
automatic table verification (AutoTV), in which the objective is to verify the
accuracy of numerical data in tables by cross-referencing cited sources. To
support this task, we propose a new benchmark, arXiVeri, which comprises
tabular data drawn from open-access academic papers on arXiv. We introduce
metrics to evaluate the performance of a table verifier in two key areas: (i)
table matching, which aims to identify the source table in a cited document
that corresponds to a target table, and (ii) cell matching, which aims to
locate shared cells between a target and source table and identify their row
and column indices accurately. By leveraging the flexible capabilities of
modern large language models (LLMs), we propose simple baselines for table
verification. Our findings highlight the complexity of this task, even for
state-of-the-art LLMs like OpenAI's GPT-4. The code and benchmark will be made
publicly available.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 17:59:57 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Shin",
"Gyungin",
""
],
[
"Xie",
"Weidi",
""
],
[
"Albanie",
"Samuel",
""
]
] | TITLE: arXiVeri: Automatic table verification with GPT
ABSTRACT: Without accurate transcription of numerical data in scientific documents, a
scientist cannot draw accurate conclusions. Unfortunately, the process of
copying numerical data from one paper to another is prone to human error. In
this paper, we propose to meet this challenge through the novel task of
automatic table verification (AutoTV), in which the objective is to verify the
accuracy of numerical data in tables by cross-referencing cited sources. To
support this task, we propose a new benchmark, arXiVeri, which comprises
tabular data drawn from open-access academic papers on arXiv. We introduce
metrics to evaluate the performance of a table verifier in two key areas: (i)
table matching, which aims to identify the source table in a cited document
that corresponds to a target table, and (ii) cell matching, which aims to
locate shared cells between a target and source table and identify their row
and column indices accurately. By leveraging the flexible capabilities of
modern large language models (LLMs), we propose simple baselines for table
verification. Our findings highlight the complexity of this task, even for
state-of-the-art LLMs like OpenAI's GPT-4. The code and benchmark will be made
publicly available.
|
2306.07969 | Sagar Vaze | Sagar Vaze, Nicolas Carion, Ishan Misra | GeneCIS: A Benchmark for General Conditional Image Similarity | CVPR 2023 (Highlighted Paper). Project page at
https://sgvaze.github.io/genecis/ | null | null | null | cs.CV cs.AI cs.LG cs.MM | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We argue that there are many notions of 'similarity' and that models, like
humans, should be able to adapt to these dynamically. This contrasts with most
representation learning methods, supervised or self-supervised, which learn a
fixed embedding function and hence implicitly assume a single notion of
similarity. For instance, models trained on ImageNet are biased towards object
categories, while a user might prefer the model to focus on colors, textures or
specific elements in the scene. In this paper, we propose the GeneCIS
('genesis') benchmark, which measures models' ability to adapt to a range of
similarity conditions. Extending prior work, our benchmark is designed for
zero-shot evaluation only, and hence considers an open-set of similarity
conditions. We find that baselines from powerful CLIP models struggle on
GeneCIS and that performance on the benchmark is only weakly correlated with
ImageNet accuracy, suggesting that simply scaling existing methods is not
fruitful. We further propose a simple, scalable solution based on automatically
mining information from existing image-caption datasets. We find our method
offers a substantial boost over the baselines on GeneCIS, and further improves
zero-shot performance on related image retrieval benchmarks. In fact, though
evaluated zero-shot, our model surpasses state-of-the-art supervised models on
MIT-States. Project page at https://sgvaze.github.io/genecis/.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 17:59:58 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Vaze",
"Sagar",
""
],
[
"Carion",
"Nicolas",
""
],
[
"Misra",
"Ishan",
""
]
] | TITLE: GeneCIS: A Benchmark for General Conditional Image Similarity
ABSTRACT: We argue that there are many notions of 'similarity' and that models, like
humans, should be able to adapt to these dynamically. This contrasts with most
representation learning methods, supervised or self-supervised, which learn a
fixed embedding function and hence implicitly assume a single notion of
similarity. For instance, models trained on ImageNet are biased towards object
categories, while a user might prefer the model to focus on colors, textures or
specific elements in the scene. In this paper, we propose the GeneCIS
('genesis') benchmark, which measures models' ability to adapt to a range of
similarity conditions. Extending prior work, our benchmark is designed for
zero-shot evaluation only, and hence considers an open-set of similarity
conditions. We find that baselines from powerful CLIP models struggle on
GeneCIS and that performance on the benchmark is only weakly correlated with
ImageNet accuracy, suggesting that simply scaling existing methods is not
fruitful. We further propose a simple, scalable solution based on automatically
mining information from existing image-caption datasets. We find our method
offers a substantial boost over the baselines on GeneCIS, and further improves
zero-shot performance on related image retrieval benchmarks. In fact, though
evaluated zero-shot, our model surpasses state-of-the-art supervised models on
MIT-States. Project page at https://sgvaze.github.io/genecis/.
|
2306.07970 | Haotong Lin | Haotong Lin, Qianqian Wang, Ruojin Cai, Sida Peng, Hadar
Averbuch-Elor, Xiaowei Zhou, Noah Snavely | Neural Scene Chronology | CVPR 2023; Project page: https://zju3dv.github.io/neusc/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we aim to reconstruct a time-varying 3D model, capable of
rendering photo-realistic renderings with independent control of viewpoint,
illumination, and time, from Internet photos of large-scale landmarks. The core
challenges are twofold. First, different types of temporal changes, such as
illumination and changes to the underlying scene itself (such as replacing one
graffiti artwork with another) are entangled together in the imagery. Second,
scene-level temporal changes are often discrete and sporadic over time, rather
than continuous. To tackle these problems, we propose a new scene
representation equipped with a novel temporal step function encoding method
that can model discrete scene-level content changes as piece-wise constant
functions over time. Specifically, we represent the scene as a space-time
radiance field with a per-image illumination embedding, where
temporally-varying scene changes are encoded using a set of learned step
functions. To facilitate our task of chronology reconstruction from Internet
imagery, we also collect a new dataset of four scenes that exhibit various
changes over time. We demonstrate that our method exhibits state-of-the-art
view synthesis results on this dataset, while achieving independent control of
viewpoint, time, and illumination.
| [
{
"version": "v1",
"created": "Tue, 13 Jun 2023 17:59:58 GMT"
}
] | 2023-06-14T00:00:00 | [
[
"Lin",
"Haotong",
""
],
[
"Wang",
"Qianqian",
""
],
[
"Cai",
"Ruojin",
""
],
[
"Peng",
"Sida",
""
],
[
"Averbuch-Elor",
"Hadar",
""
],
[
"Zhou",
"Xiaowei",
""
],
[
"Snavely",
"Noah",
""
]
] | TITLE: Neural Scene Chronology
ABSTRACT: In this work, we aim to reconstruct a time-varying 3D model, capable of
rendering photo-realistic renderings with independent control of viewpoint,
illumination, and time, from Internet photos of large-scale landmarks. The core
challenges are twofold. First, different types of temporal changes, such as
illumination and changes to the underlying scene itself (such as replacing one
graffiti artwork with another) are entangled together in the imagery. Second,
scene-level temporal changes are often discrete and sporadic over time, rather
than continuous. To tackle these problems, we propose a new scene
representation equipped with a novel temporal step function encoding method
that can model discrete scene-level content changes as piece-wise constant
functions over time. Specifically, we represent the scene as a space-time
radiance field with a per-image illumination embedding, where
temporally-varying scene changes are encoded using a set of learned step
functions. To facilitate our task of chronology reconstruction from Internet
imagery, we also collect a new dataset of four scenes that exhibit various
changes over time. We demonstrate that our method exhibits state-of-the-art
view synthesis results on this dataset, while achieving independent control of
viewpoint, time, and illumination.
|
2007.02622 | Albert Bou | Albert Bou, Sebastian Dittert and Gianni De Fabritiis | Integrating Distributed Architectures in Highly Modular RL Libraries | 9 pages, 9 figures | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Advancing reinforcement learning (RL) requires tools that are flexible enough
to easily prototype new methods while avoiding impractically slow experimental
turnaround times. To match the first requirement, the most popular RL libraries
advocate for highly modular agent composability, which facilitates
experimentation and development. To solve challenging environments within
reasonable time frames, scaling RL to large sampling and computing resources
has proved a successful strategy. However, this capability has been so far
difficult to combine with modularity. In this work, we explore design choices
to allow agent composability both at a local and distributed level of
execution. We propose a versatile approach that allows the definition of RL
agents at different scales through independent reusable components. We
demonstrate experimentally that our design choices allow us to reproduce
classical benchmarks, explore multiple distributed architectures, and solve
novel and complex environments while giving full control to the user in the
agent definition and training scheme definition. We believe this work can
provide useful insights to the next generation of RL libraries.
| [
{
"version": "v1",
"created": "Mon, 6 Jul 2020 10:22:07 GMT"
},
{
"version": "v2",
"created": "Wed, 10 Feb 2021 16:39:37 GMT"
},
{
"version": "v3",
"created": "Mon, 12 Jun 2023 08:40:02 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Bou",
"Albert",
""
],
[
"Dittert",
"Sebastian",
""
],
[
"De Fabritiis",
"Gianni",
""
]
] | TITLE: Integrating Distributed Architectures in Highly Modular RL Libraries
ABSTRACT: Advancing reinforcement learning (RL) requires tools that are flexible enough
to easily prototype new methods while avoiding impractically slow experimental
turnaround times. To match the first requirement, the most popular RL libraries
advocate for highly modular agent composability, which facilitates
experimentation and development. To solve challenging environments within
reasonable time frames, scaling RL to large sampling and computing resources
has proved a successful strategy. However, this capability has been so far
difficult to combine with modularity. In this work, we explore design choices
to allow agent composability both at a local and distributed level of
execution. We propose a versatile approach that allows the definition of RL
agents at different scales through independent reusable components. We
demonstrate experimentally that our design choices allow us to reproduce
classical benchmarks, explore multiple distributed architectures, and solve
novel and complex environments while giving full control to the user in the
agent definition and training scheme definition. We believe this work can
provide useful insights to the next generation of RL libraries.
|
2010.12427 | Shiqi Yang | Shiqi Yang, Yaxing Wang, Joost van de Weijer, Luis Herranz and
Shangling Jui | Casting a BAIT for Offline and Online Source-free Domain Adaptation | Accepted by Computer Vision and Image Understanding | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We address the source-free domain adaptation (SFDA) problem, where only the
source model is available during adaptation to the target domain. We consider
two settings: the offline setting where all target data can be visited multiple
times (epochs) to arrive at a prediction for each target sample, and the online
setting where the target data needs to be directly classified upon arrival.
Inspired by diverse classifier based domain adaptation methods, in this paper
we introduce a second classifier, but with another classifier head fixed. When
adapting to the target domain, the additional classifier initialized from
source classifier is expected to find misclassified features. Next, when
updating the feature extractor, those features will be pushed towards the right
side of the source decision boundary, thus achieving source-free domain
adaptation. Experimental results show that the proposed method achieves
competitive results for offline SFDA on several benchmark datasets compared
with existing DA and SFDA methods, and our method surpasses by a large margin
other SFDA methods under online source-free domain adaptation setting.
| [
{
"version": "v1",
"created": "Fri, 23 Oct 2020 14:18:42 GMT"
},
{
"version": "v2",
"created": "Wed, 28 Oct 2020 14:07:09 GMT"
},
{
"version": "v3",
"created": "Thu, 19 Nov 2020 16:48:38 GMT"
},
{
"version": "v4",
"created": "Fri, 29 Oct 2021 18:47:21 GMT"
},
{
"version": "v5",
"created": "Sat, 10 Jun 2023 05:24:15 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Yang",
"Shiqi",
""
],
[
"Wang",
"Yaxing",
""
],
[
"van de Weijer",
"Joost",
""
],
[
"Herranz",
"Luis",
""
],
[
"Jui",
"Shangling",
""
]
] | TITLE: Casting a BAIT for Offline and Online Source-free Domain Adaptation
ABSTRACT: We address the source-free domain adaptation (SFDA) problem, where only the
source model is available during adaptation to the target domain. We consider
two settings: the offline setting where all target data can be visited multiple
times (epochs) to arrive at a prediction for each target sample, and the online
setting where the target data needs to be directly classified upon arrival.
Inspired by diverse classifier based domain adaptation methods, in this paper
we introduce a second classifier, but with another classifier head fixed. When
adapting to the target domain, the additional classifier initialized from
source classifier is expected to find misclassified features. Next, when
updating the feature extractor, those features will be pushed towards the right
side of the source decision boundary, thus achieving source-free domain
adaptation. Experimental results show that the proposed method achieves
competitive results for offline SFDA on several benchmark datasets compared
with existing DA and SFDA methods, and our method surpasses by a large margin
other SFDA methods under online source-free domain adaptation setting.
|
2012.13188 | Yalda Foroutan | Yalda Foroutan, Ahmad Kalhor, Saeid Mohammadi Nejati, Samad Sheikhaei | Control of Computer Pointer Using Hand Gesture Recognition in Motion
Pictures | 9 pages, 6 figures, 2 tables | null | null | null | cs.CV cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a user interface designed to enable computer cursor
control through hand detection and gesture classification. A comprehensive hand
dataset comprising 6720 image samples was collected, encompassing four distinct
classes: fist, palm, pointing to the left, and pointing to the right. The
images were captured from 15 individuals in various settings, including simple
backgrounds with different perspectives and lighting conditions. A
convolutional neural network (CNN) was trained on this dataset to accurately
predict labels for each captured image and measure their similarity. The system
incorporates defined commands for cursor movement, left-click, and right-click
actions. Experimental results indicate that the proposed algorithm achieves a
remarkable accuracy of 91.88% and demonstrates its potential applicability
across diverse backgrounds.
| [
{
"version": "v1",
"created": "Thu, 24 Dec 2020 10:24:51 GMT"
},
{
"version": "v2",
"created": "Fri, 9 Jun 2023 21:49:33 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Foroutan",
"Yalda",
""
],
[
"Kalhor",
"Ahmad",
""
],
[
"Nejati",
"Saeid Mohammadi",
""
],
[
"Sheikhaei",
"Samad",
""
]
] | TITLE: Control of Computer Pointer Using Hand Gesture Recognition in Motion
Pictures
ABSTRACT: This paper presents a user interface designed to enable computer cursor
control through hand detection and gesture classification. A comprehensive hand
dataset comprising 6720 image samples was collected, encompassing four distinct
classes: fist, palm, pointing to the left, and pointing to the right. The
images were captured from 15 individuals in various settings, including simple
backgrounds with different perspectives and lighting conditions. A
convolutional neural network (CNN) was trained on this dataset to accurately
predict labels for each captured image and measure their similarity. The system
incorporates defined commands for cursor movement, left-click, and right-click
actions. Experimental results indicate that the proposed algorithm achieves a
remarkable accuracy of 91.88% and demonstrates its potential applicability
across diverse backgrounds.
|
2103.13725 | Haipeng Li | Haipeng Li and Kunming Luo and Shuaicheng Liu | GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning | null | 2021 IEEE/CVF International Conference on Computer Vision (ICCV) | 10.1109/ICCV48922.2021.01263 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing optical flow methods are erroneous in challenging scenes, such as
fog, rain, and night because the basic optical flow assumptions such as
brightness and gradient constancy are broken. To address this problem, we
present an unsupervised learning approach that fuses gyroscope into optical
flow learning. Specifically, we first convert gyroscope readings into motion
fields named gyro field. Second, we design a self-guided fusion module to fuse
the background motion extracted from the gyro field with the optical flow and
guide the network to focus on motion details. To the best of our knowledge,
this is the first deep learning-based framework that fuses gyroscope data and
image content for optical flow learning. To validate our method, we propose a
new dataset that covers regular and challenging scenes. Experiments show that
our method outperforms the state-of-art methods in both regular and challenging
scenes. Code and dataset are available at
https://github.com/megvii-research/GyroFlow.
| [
{
"version": "v1",
"created": "Thu, 25 Mar 2021 10:14:57 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Aug 2021 07:46:31 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Li",
"Haipeng",
""
],
[
"Luo",
"Kunming",
""
],
[
"Liu",
"Shuaicheng",
""
]
] | TITLE: GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning
ABSTRACT: Existing optical flow methods are erroneous in challenging scenes, such as
fog, rain, and night because the basic optical flow assumptions such as
brightness and gradient constancy are broken. To address this problem, we
present an unsupervised learning approach that fuses gyroscope into optical
flow learning. Specifically, we first convert gyroscope readings into motion
fields named gyro field. Second, we design a self-guided fusion module to fuse
the background motion extracted from the gyro field with the optical flow and
guide the network to focus on motion details. To the best of our knowledge,
this is the first deep learning-based framework that fuses gyroscope data and
image content for optical flow learning. To validate our method, we propose a
new dataset that covers regular and challenging scenes. Experiments show that
our method outperforms the state-of-art methods in both regular and challenging
scenes. Code and dataset are available at
https://github.com/megvii-research/GyroFlow.
|
2104.01393 | Shigehiko Schamoni | Tsz Kin Lam, Mayumi Ohta, Shigehiko Schamoni, Stefan Riezler | On-the-Fly Aligned Data Augmentation for Sequence-to-Sequence ASR | Accepted at INTERSPEECH 2021 | null | 10.21437/Interspeech.2021-1679 | null | cs.CL eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose an on-the-fly data augmentation method for automatic speech
recognition (ASR) that uses alignment information to generate effective
training samples. Our method, called Aligned Data Augmentation (ADA) for ASR,
replaces transcribed tokens and the speech representations in an aligned manner
to generate previously unseen training pairs. The speech representations are
sampled from an audio dictionary that has been extracted from the training
corpus and inject speaker variations into the training examples. The
transcribed tokens are either predicted by a language model such that the
augmented data pairs are semantically close to the original data, or randomly
sampled. Both strategies result in training pairs that improve robustness in
ASR training. Our experiments on a Seq-to-Seq architecture show that ADA can be
applied on top of SpecAugment, and achieves about 9-23% and 4-15% relative
improvements in WER over SpecAugment alone on LibriSpeech 100h and LibriSpeech
960h test datasets, respectively.
| [
{
"version": "v1",
"created": "Sat, 3 Apr 2021 13:00:00 GMT"
},
{
"version": "v2",
"created": "Wed, 9 Jun 2021 19:48:46 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Lam",
"Tsz Kin",
""
],
[
"Ohta",
"Mayumi",
""
],
[
"Schamoni",
"Shigehiko",
""
],
[
"Riezler",
"Stefan",
""
]
] | TITLE: On-the-Fly Aligned Data Augmentation for Sequence-to-Sequence ASR
ABSTRACT: We propose an on-the-fly data augmentation method for automatic speech
recognition (ASR) that uses alignment information to generate effective
training samples. Our method, called Aligned Data Augmentation (ADA) for ASR,
replaces transcribed tokens and the speech representations in an aligned manner
to generate previously unseen training pairs. The speech representations are
sampled from an audio dictionary that has been extracted from the training
corpus and inject speaker variations into the training examples. The
transcribed tokens are either predicted by a language model such that the
augmented data pairs are semantically close to the original data, or randomly
sampled. Both strategies result in training pairs that improve robustness in
ASR training. Our experiments on a Seq-to-Seq architecture show that ADA can be
applied on top of SpecAugment, and achieves about 9-23% and 4-15% relative
improvements in WER over SpecAugment alone on LibriSpeech 100h and LibriSpeech
960h test datasets, respectively.
|
2108.07401 | Shengcheng Yu | Shengcheng Yu, Chunrong Fang, Quanjun Zhang, Zhihao Cao, Yexiao Yun,
Zhenfei Cao, Kai Mei, Zhenyu Chen | Mobile App Crowdsourced Test Report Consistency Detection via Deep
Image-and-Text Fusion Understanding | null | null | 10.1109/TSE.2023.3285787 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Crowdsourced testing, as a distinct testing paradigm, has attracted much
attention in software testing, especially in mobile application (app) testing
field. Compared with in-house testing, crowdsourced testing shows superiority
with the diverse testing environments when faced with the mobile testing
fragmentation problem. However, crowdsourced testing also encounters the
low-quality test report problem caused by unprofessional crowdworkers involved
with different expertise. In order to handle the submitted reports of uneven
quality, app developers have to distinguish high-quality reports from
low-quality ones to help the bug inspection. One kind of typical low-quality
test report is inconsistent test reports, which means the textual descriptions
are not focusing on the attached bug-occurring screenshots. According to our
empirical survey, only 18.07% crowdsourced test reports are consistent.
Inconsistent reports cause waste on mobile app testing.
To solve the inconsistency problem, we propose ReCoDe to detect the
consistency of crowdsourced test reports via deep image-and-text fusion
understanding. ReCoDe is a two-stage approach that first classifies the reports
based on textual descriptions into different categories according to the bug
feature. In the second stage, ReCoDe has a deep understanding of the GUI image
features of the app screenshots and then applies different strategies to handle
different types of bugs to detect the consistency of the crowdsourced test
reports. We conduct an experiment on a dataset with over 22k test reports to
evaluate ReCoDe, and the results show the effectiveness of ReCoDe in detecting
the consistency of crowdsourced test reports. Besides, a user study is
conducted to prove the practical value of ReCoDe in effectively helping app
developers improve the efficiency of reviewing the crowdsourced test reports.
| [
{
"version": "v1",
"created": "Tue, 17 Aug 2021 02:02:56 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Aug 2021 02:22:06 GMT"
},
{
"version": "v3",
"created": "Mon, 12 Jun 2023 08:26:44 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Yu",
"Shengcheng",
""
],
[
"Fang",
"Chunrong",
""
],
[
"Zhang",
"Quanjun",
""
],
[
"Cao",
"Zhihao",
""
],
[
"Yun",
"Yexiao",
""
],
[
"Cao",
"Zhenfei",
""
],
[
"Mei",
"Kai",
""
],
[
"Chen",
"Zhenyu",
""
]
] | TITLE: Mobile App Crowdsourced Test Report Consistency Detection via Deep
Image-and-Text Fusion Understanding
ABSTRACT: Crowdsourced testing, as a distinct testing paradigm, has attracted much
attention in software testing, especially in mobile application (app) testing
field. Compared with in-house testing, crowdsourced testing shows superiority
with the diverse testing environments when faced with the mobile testing
fragmentation problem. However, crowdsourced testing also encounters the
low-quality test report problem caused by unprofessional crowdworkers involved
with different expertise. In order to handle the submitted reports of uneven
quality, app developers have to distinguish high-quality reports from
low-quality ones to help the bug inspection. One kind of typical low-quality
test report is inconsistent test reports, which means the textual descriptions
are not focusing on the attached bug-occurring screenshots. According to our
empirical survey, only 18.07% crowdsourced test reports are consistent.
Inconsistent reports cause waste on mobile app testing.
To solve the inconsistency problem, we propose ReCoDe to detect the
consistency of crowdsourced test reports via deep image-and-text fusion
understanding. ReCoDe is a two-stage approach that first classifies the reports
based on textual descriptions into different categories according to the bug
feature. In the second stage, ReCoDe has a deep understanding of the GUI image
features of the app screenshots and then applies different strategies to handle
different types of bugs to detect the consistency of the crowdsourced test
reports. We conduct an experiment on a dataset with over 22k test reports to
evaluate ReCoDe, and the results show the effectiveness of ReCoDe in detecting
the consistency of crowdsourced test reports. Besides, a user study is
conducted to prove the practical value of ReCoDe in effectively helping app
developers improve the efficiency of reviewing the crowdsourced test reports.
|
2108.10714 | Labib Chowdhury | Labib Chowdhury, Mustafa Kamal, Najia Hasan and Nabeel Mohammed | Curricular SincNet: Towards Robust Deep Speaker Recognition by
Emphasizing Hard Samples in Latent Space | Accepted at 20th International Conference of the Biometrics Special
Interest Group (BIOSIG 2021) | null | 10.1109/BIOSIG52210.2021.9548296 | null | eess.AS cs.LG cs.SD | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Deep learning models have become an increasingly preferred option for
biometric recognition systems, such as speaker recognition. SincNet, a deep
neural network architecture, gained popularity in speaker recognition tasks due
to its parameterized sinc functions that allow it to work directly on the
speech signal. The original SincNet architecture uses the softmax loss, which
may not be the most suitable choice for recognition-based tasks. Such loss
functions do not impose inter-class margins nor differentiate between easy and
hard training samples. Curriculum learning, particularly those leveraging
angular margin-based losses, has proven very successful in other biometric
applications such as face recognition. The advantage of such a curriculum
learning-based techniques is that it will impose inter-class margins as well as
taking to account easy and hard samples. In this paper, we propose Curricular
SincNet (CL-SincNet), an improved SincNet model where we use a curricular loss
function to train the SincNet architecture. The proposed model is evaluated on
multiple datasets using intra-dataset and inter-dataset evaluation protocols.
In both settings, the model performs competitively with other previously
published work. In the case of inter-dataset testing, it achieves the best
overall results with a reduction of 4\% error rate compare to SincNet and other
published work.
| [
{
"version": "v1",
"created": "Sat, 21 Aug 2021 09:13:45 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Chowdhury",
"Labib",
""
],
[
"Kamal",
"Mustafa",
""
],
[
"Hasan",
"Najia",
""
],
[
"Mohammed",
"Nabeel",
""
]
] | TITLE: Curricular SincNet: Towards Robust Deep Speaker Recognition by
Emphasizing Hard Samples in Latent Space
ABSTRACT: Deep learning models have become an increasingly preferred option for
biometric recognition systems, such as speaker recognition. SincNet, a deep
neural network architecture, gained popularity in speaker recognition tasks due
to its parameterized sinc functions that allow it to work directly on the
speech signal. The original SincNet architecture uses the softmax loss, which
may not be the most suitable choice for recognition-based tasks. Such loss
functions do not impose inter-class margins nor differentiate between easy and
hard training samples. Curriculum learning, particularly those leveraging
angular margin-based losses, has proven very successful in other biometric
applications such as face recognition. The advantage of such a curriculum
learning-based techniques is that it will impose inter-class margins as well as
taking to account easy and hard samples. In this paper, we propose Curricular
SincNet (CL-SincNet), an improved SincNet model where we use a curricular loss
function to train the SincNet architecture. The proposed model is evaluated on
multiple datasets using intra-dataset and inter-dataset evaluation protocols.
In both settings, the model performs competitively with other previously
published work. In the case of inter-dataset testing, it achieves the best
overall results with a reduction of 4\% error rate compare to SincNet and other
published work.
|
2110.03997 | Shozo Saeki | Shozo Saeki, Minoru Kawahara, and Hirohisa Aman | Multi Proxy Anchor Family Loss for Several Types of Gradients | null | Computer Vision and Image Understanding. 229 (2023) 103654 | 10.1016/j.cviu.2023.103654 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The deep metric learning (DML) objective is to learn a neural network that
maps into an embedding space where similar data are near and dissimilar data
are far. However, conventional proxy-based losses for DML have two problems:
gradient problem and application of the real-world dataset with multiple local
centers. Additionally, the performance metrics of DML also have some issues
with stability and flexibility. This paper proposes three multi-proxies anchor
(MPA) family losses and a normalized discounted cumulative gain (nDCG@k)
metric. This paper makes three contributions. (1) MPA-family losses can learn
using a real-world dataset with multi-local centers. (2) MPA-family losses
improve the training capacity of a neural network owing to solving the gradient
problem. (3) MPA-family losses have data-wise or class-wise characteristics
with respect to gradient generation. Finally, we demonstrate the effectiveness
of MPA-family losses, and MPA-family losses achieves higher accuracy on two
datasets for fine-grained images.
| [
{
"version": "v1",
"created": "Fri, 8 Oct 2021 09:34:38 GMT"
},
{
"version": "v2",
"created": "Mon, 11 Oct 2021 01:42:25 GMT"
},
{
"version": "v3",
"created": "Mon, 8 Nov 2021 08:58:14 GMT"
},
{
"version": "v4",
"created": "Sun, 19 Dec 2021 15:00:06 GMT"
},
{
"version": "v5",
"created": "Fri, 22 Apr 2022 12:09:08 GMT"
},
{
"version": "v6",
"created": "Sat, 18 Feb 2023 02:51:48 GMT"
},
{
"version": "v7",
"created": "Wed, 22 Feb 2023 10:05:18 GMT"
},
{
"version": "v8",
"created": "Sun, 11 Jun 2023 10:10:56 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Saeki",
"Shozo",
""
],
[
"Kawahara",
"Minoru",
""
],
[
"Aman",
"Hirohisa",
""
]
] | TITLE: Multi Proxy Anchor Family Loss for Several Types of Gradients
ABSTRACT: The deep metric learning (DML) objective is to learn a neural network that
maps into an embedding space where similar data are near and dissimilar data
are far. However, conventional proxy-based losses for DML have two problems:
gradient problem and application of the real-world dataset with multiple local
centers. Additionally, the performance metrics of DML also have some issues
with stability and flexibility. This paper proposes three multi-proxies anchor
(MPA) family losses and a normalized discounted cumulative gain (nDCG@k)
metric. This paper makes three contributions. (1) MPA-family losses can learn
using a real-world dataset with multi-local centers. (2) MPA-family losses
improve the training capacity of a neural network owing to solving the gradient
problem. (3) MPA-family losses have data-wise or class-wise characteristics
with respect to gradient generation. Finally, we demonstrate the effectiveness
of MPA-family losses, and MPA-family losses achieves higher accuracy on two
datasets for fine-grained images.
|
2111.03362 | Moran Baruch | Moran Baruch, Nir Drucker, Lev Greenberg and Guy Moshkowich | A methodology for training homomorphicencryption friendly neural
networks | null | null | 10.1007/978-3-031-16815-4_29 | null | cs.CR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Privacy-preserving deep neural network (DNN) inference is a necessity in
different regulated industries such as healthcare, finance and retail.
Recently, homomorphic encryption (HE) has been used as a method to enable
analytics while addressing privacy concerns. HE enables secure predictions over
encrypted data. However, there are several challenges related to the use of HE,
including DNN size limitations and the lack of support for some operation
types. Most notably, the commonly used ReLU activation is not supported under
some HE schemes. We propose a structured methodology to replace ReLU with a
quadratic polynomial activation. To address the accuracy degradation issue, we
use a pre-trained model that trains another HE-friendly model, using techniques
such as trainable activation functions and knowledge distillation. We
demonstrate our methodology on the AlexNet architecture, using the chest X-Ray
and CT datasets for COVID-19 detection. Experiments using our approach reduced
the gap between the F1 score and accuracy of the models trained with ReLU and
the HE-friendly model to within a mere 0.32-5.3 percent degradation. We also
demonstrate our methodology using the SqueezeNet architecture, for which we
observed 7 percent accuracy and F1 improvements over training similar networks
with other HE-friendly training methods.
| [
{
"version": "v1",
"created": "Fri, 5 Nov 2021 10:04:15 GMT"
},
{
"version": "v2",
"created": "Wed, 17 Nov 2021 08:22:48 GMT"
},
{
"version": "v3",
"created": "Thu, 7 Jul 2022 19:22:02 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Baruch",
"Moran",
""
],
[
"Drucker",
"Nir",
""
],
[
"Greenberg",
"Lev",
""
],
[
"Moshkowich",
"Guy",
""
]
] | TITLE: A methodology for training homomorphicencryption friendly neural
networks
ABSTRACT: Privacy-preserving deep neural network (DNN) inference is a necessity in
different regulated industries such as healthcare, finance and retail.
Recently, homomorphic encryption (HE) has been used as a method to enable
analytics while addressing privacy concerns. HE enables secure predictions over
encrypted data. However, there are several challenges related to the use of HE,
including DNN size limitations and the lack of support for some operation
types. Most notably, the commonly used ReLU activation is not supported under
some HE schemes. We propose a structured methodology to replace ReLU with a
quadratic polynomial activation. To address the accuracy degradation issue, we
use a pre-trained model that trains another HE-friendly model, using techniques
such as trainable activation functions and knowledge distillation. We
demonstrate our methodology on the AlexNet architecture, using the chest X-Ray
and CT datasets for COVID-19 detection. Experiments using our approach reduced
the gap between the F1 score and accuracy of the models trained with ReLU and
the HE-friendly model to within a mere 0.32-5.3 percent degradation. We also
demonstrate our methodology using the SqueezeNet architecture, for which we
observed 7 percent accuracy and F1 improvements over training similar networks
with other HE-friendly training methods.
|
2112.06028 | Hankz Hankui Zhuo | Siqi Hong, Hankz Hankui Zhuo, Kebing Jin, Guang Shao, Zhanwen Zhou | Retrosynthetic Planning with Experience-Guided Monte Carlo Tree Search | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | In retrosynthetic planning, the huge number of possible routes to synthesize
a complex molecule using simple building blocks leads to a combinatorial
explosion of possibilities. Even experienced chemists often have difficulty to
select the most promising transformations. The current approaches rely on
human-defined or machine-trained score functions which have limited chemical
knowledge or use expensive estimation methods for guiding. Here we an propose
experience-guided Monte Carlo tree search (EG-MCTS) to deal with this problem.
Instead of rollout, we build an experience guidance network to learn knowledge
from synthetic experiences during the search. Experiments on benchmark USPTO
datasets show that, EG-MCTS gains significant improvement over state-of-the-art
approaches both in efficiency and effectiveness. In a comparative experiment
with the literature, our computer-generated routes mostly matched the reported
routes. Routes designed for real drug compounds exhibit the effectiveness of
EG-MCTS on assisting chemists performing retrosynthetic analysis.
| [
{
"version": "v1",
"created": "Sat, 11 Dec 2021 17:14:15 GMT"
},
{
"version": "v2",
"created": "Sat, 10 Jun 2023 03:13:46 GMT"
}
] | 2023-06-13T00:00:00 | [
[
"Hong",
"Siqi",
""
],
[
"Zhuo",
"Hankz Hankui",
""
],
[
"Jin",
"Kebing",
""
],
[
"Shao",
"Guang",
""
],
[
"Zhou",
"Zhanwen",
""
]
] | TITLE: Retrosynthetic Planning with Experience-Guided Monte Carlo Tree Search
ABSTRACT: In retrosynthetic planning, the huge number of possible routes to synthesize
a complex molecule using simple building blocks leads to a combinatorial
explosion of possibilities. Even experienced chemists often have difficulty to
select the most promising transformations. The current approaches rely on
human-defined or machine-trained score functions which have limited chemical
knowledge or use expensive estimation methods for guiding. Here we an propose
experience-guided Monte Carlo tree search (EG-MCTS) to deal with this problem.
Instead of rollout, we build an experience guidance network to learn knowledge
from synthetic experiences during the search. Experiments on benchmark USPTO
datasets show that, EG-MCTS gains significant improvement over state-of-the-art
approaches both in efficiency and effectiveness. In a comparative experiment
with the literature, our computer-generated routes mostly matched the reported
routes. Routes designed for real drug compounds exhibit the effectiveness of
EG-MCTS on assisting chemists performing retrosynthetic analysis.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.