id
large_stringlengths 9
16
| submitter
large_stringlengths 3
64
⌀ | authors
large_stringlengths 5
20.2k
| title
large_stringlengths 5
245
| comments
large_stringlengths 1
482
⌀ | journal-ref
large_stringlengths 4
382
⌀ | doi
large_stringlengths 9
151
⌀ | report-no
large_stringlengths 2
509
⌀ | categories
large_stringlengths 5
78
| license
large_stringclasses 9
values | abstract
large_stringlengths 83
3.55k
| versions
large listlengths 1
21
| update_date
timestamp[ms]date 2007-05-23 00:00:00
2025-05-30 00:00:00
| authors_parsed
large listlengths 1
1.35k
| prompt
large_stringlengths 166
3.64k
| label
stringclasses 2
values | prob
float64 0.5
0.73
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2208.00050 | Mohammed Daoudi | Naima Otberdout, Claudio Ferrari, Mohamed Daoudi, Stefano Berretti,
Alberto Del Bimbo | Generating Multiple 4D Expression Transitions by Learning Face Landmark
Trajectories | This preprint is an extension of CVPR 2022 paper arXiv:2105.07463 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | In this work, we address the problem of 4D facial expressions generation.
This is usually addressed by animating a neutral 3D face to reach an expression
peak, and then get back to the neutral state. In the real world though, people
show more complex expressions, and switch from one expression to another. We
thus propose a new model that generates transitions between different
expressions, and synthesizes long and composed 4D expressions. This involves
three sub-problems: (i) modeling the temporal dynamics of expressions, (ii)
learning transitions between them, and (iii) deforming a generic mesh. We
propose to encode the temporal evolution of expressions using the motion of a
set of 3D landmarks, that we learn to generate by training a manifold-valued
GAN (Motion3DGAN). To allow the generation of composed expressions, this model
accepts two labels encoding the starting and the ending expressions. The final
sequence of meshes is generated by a Sparse2Dense mesh Decoder (S2D-Dec) that
maps the landmark displacements to a dense, per-vertex displacement of a known
mesh topology. By explicitly working with motion trajectories, the model is
totally independent from the identity. Extensive experiments on five public
datasets show that our proposed approach brings significant improvements with
respect to previous solutions, while retaining good generalization to unseen
data.
| [
{
"version": "v1",
"created": "Fri, 29 Jul 2022 19:33:56 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 07:25:05 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Otberdout",
"Naima",
""
],
[
"Ferrari",
"Claudio",
""
],
[
"Daoudi",
"Mohamed",
""
],
[
"Berretti",
"Stefano",
""
],
[
"Del Bimbo",
"Alberto",
""
]
] | TITLE: Generating Multiple 4D Expression Transitions by Learning Face Landmark
Trajectories
ABSTRACT: In this work, we address the problem of 4D facial expressions generation.
This is usually addressed by animating a neutral 3D face to reach an expression
peak, and then get back to the neutral state. In the real world though, people
show more complex expressions, and switch from one expression to another. We
thus propose a new model that generates transitions between different
expressions, and synthesizes long and composed 4D expressions. This involves
three sub-problems: (i) modeling the temporal dynamics of expressions, (ii)
learning transitions between them, and (iii) deforming a generic mesh. We
propose to encode the temporal evolution of expressions using the motion of a
set of 3D landmarks, that we learn to generate by training a manifold-valued
GAN (Motion3DGAN). To allow the generation of composed expressions, this model
accepts two labels encoding the starting and the ending expressions. The final
sequence of meshes is generated by a Sparse2Dense mesh Decoder (S2D-Dec) that
maps the landmark displacements to a dense, per-vertex displacement of a known
mesh topology. By explicitly working with motion trajectories, the model is
totally independent from the identity. Extensive experiments on five public
datasets show that our proposed approach brings significant improvements with
respect to previous solutions, while retaining good generalization to unseen
data.
| no_new_dataset | 0.709416 |
2208.07360 | Kevin Musgrave | Kevin Musgrave, Serge Belongie, Ser-Nam Lim | Three New Validators and a Large-Scale Benchmark Ranking for
Unsupervised Domain Adaptation | This paper was previously titled Benchmarking Validation Methods for
Unsupervised Domain Adaptation. This version contains new experiments,
analysis, and figures | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Changes to hyperparameters can have a dramatic effect on model accuracy.
Thus, the tuning of hyperparameters plays an important role in optimizing
machine-learning models. An integral part of the hyperparameter-tuning process
is the evaluation of model checkpoints, which is done through the use of
"validators". In a supervised setting, these validators evaluate checkpoints by
computing accuracy on a validation set that has labels. In contrast, in an
unsupervised setting, the validation set has no such labels. Without any
labels, it is impossible to compute accuracy, so validators must estimate
accuracy instead. But what is the best approach to estimating accuracy? In this
paper, we consider this question in the context of unsupervised domain
adaptation (UDA). Specifically, we propose three new validators, and we compare
and rank them against five other existing validators, on a large dataset of
1,000,000 checkpoints. Extensive experimental results show that two of our
proposed validators achieve state-of-the-art performance in various settings.
Finally, we find that in many cases, the state-of-the-art is obtained by a
simple baseline method. To the best of our knowledge, this is the largest
empirical study of UDA validators to date. Code is available at
https://www.github.com/KevinMusgrave/powerful-benchmarker.
| [
{
"version": "v1",
"created": "Mon, 15 Aug 2022 17:55:26 GMT"
},
{
"version": "v2",
"created": "Fri, 20 Jan 2023 14:13:08 GMT"
},
{
"version": "v3",
"created": "Thu, 9 Mar 2023 16:42:28 GMT"
},
{
"version": "v4",
"created": "Wed, 17 May 2023 23:24:06 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Musgrave",
"Kevin",
""
],
[
"Belongie",
"Serge",
""
],
[
"Lim",
"Ser-Nam",
""
]
] | TITLE: Three New Validators and a Large-Scale Benchmark Ranking for
Unsupervised Domain Adaptation
ABSTRACT: Changes to hyperparameters can have a dramatic effect on model accuracy.
Thus, the tuning of hyperparameters plays an important role in optimizing
machine-learning models. An integral part of the hyperparameter-tuning process
is the evaluation of model checkpoints, which is done through the use of
"validators". In a supervised setting, these validators evaluate checkpoints by
computing accuracy on a validation set that has labels. In contrast, in an
unsupervised setting, the validation set has no such labels. Without any
labels, it is impossible to compute accuracy, so validators must estimate
accuracy instead. But what is the best approach to estimating accuracy? In this
paper, we consider this question in the context of unsupervised domain
adaptation (UDA). Specifically, we propose three new validators, and we compare
and rank them against five other existing validators, on a large dataset of
1,000,000 checkpoints. Extensive experimental results show that two of our
proposed validators achieve state-of-the-art performance in various settings.
Finally, we find that in many cases, the state-of-the-art is obtained by a
simple baseline method. To the best of our knowledge, this is the largest
empirical study of UDA validators to date. Code is available at
https://www.github.com/KevinMusgrave/powerful-benchmarker.
| no_new_dataset | 0.708175 |
2208.10364 | Jintang Li | Jintang Li, Zhouxin Yu, Zulun Zhu, Liang Chen, Qi Yu, Zibin Zheng,
Sheng Tian, Ruofan Wu, Changhua Meng | Scaling Up Dynamic Graph Representation Learning via Spiking Neural
Networks | Accepted by AAAI 2023.Contains appendix with additional details about
algorithms and experiments. Code available at
https://github.com/EdisonLeeeee/SpikeNet | null | null | null | cs.NE cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent years have seen a surge in research on dynamic graph representation
learning, which aims to model temporal graphs that are dynamic and evolving
constantly over time. However, current work typically models graph dynamics
with recurrent neural networks (RNNs), making them suffer seriously from
computation and memory overheads on large temporal graphs. So far, scalability
of dynamic graph representation learning on large temporal graphs remains one
of the major challenges. In this paper, we present a scalable framework, namely
SpikeNet, to efficiently capture the temporal and structural patterns of
temporal graphs. We explore a new direction in that we can capture the evolving
dynamics of temporal graphs with spiking neural networks (SNNs) instead of
RNNs. As a low-power alternative to RNNs, SNNs explicitly model graph dynamics
as spike trains of neuron populations and enable spike-based propagation in an
efficient way. Experiments on three large real-world temporal graph datasets
demonstrate that SpikeNet outperforms strong baselines on the temporal node
classification task with lower computational costs. Particularly, SpikeNet
generalizes to a large temporal graph (2.7M nodes and 13.9M edges) with
significantly fewer parameters and computation overheads.Our code is publicly
available at \url{https://github.com/EdisonLeeeee/SpikeNet}.
| [
{
"version": "v1",
"created": "Mon, 15 Aug 2022 09:22:15 GMT"
},
{
"version": "v2",
"created": "Sat, 26 Nov 2022 02:59:08 GMT"
},
{
"version": "v3",
"created": "Thu, 18 May 2023 11:57:10 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Jintang",
""
],
[
"Yu",
"Zhouxin",
""
],
[
"Zhu",
"Zulun",
""
],
[
"Chen",
"Liang",
""
],
[
"Yu",
"Qi",
""
],
[
"Zheng",
"Zibin",
""
],
[
"Tian",
"Sheng",
""
],
[
"Wu",
"Ruofan",
""
],
[
"Meng",
"Changhua",
""
]
] | TITLE: Scaling Up Dynamic Graph Representation Learning via Spiking Neural
Networks
ABSTRACT: Recent years have seen a surge in research on dynamic graph representation
learning, which aims to model temporal graphs that are dynamic and evolving
constantly over time. However, current work typically models graph dynamics
with recurrent neural networks (RNNs), making them suffer seriously from
computation and memory overheads on large temporal graphs. So far, scalability
of dynamic graph representation learning on large temporal graphs remains one
of the major challenges. In this paper, we present a scalable framework, namely
SpikeNet, to efficiently capture the temporal and structural patterns of
temporal graphs. We explore a new direction in that we can capture the evolving
dynamics of temporal graphs with spiking neural networks (SNNs) instead of
RNNs. As a low-power alternative to RNNs, SNNs explicitly model graph dynamics
as spike trains of neuron populations and enable spike-based propagation in an
efficient way. Experiments on three large real-world temporal graph datasets
demonstrate that SpikeNet outperforms strong baselines on the temporal node
classification task with lower computational costs. Particularly, SpikeNet
generalizes to a large temporal graph (2.7M nodes and 13.9M edges) with
significantly fewer parameters and computation overheads.Our code is publicly
available at \url{https://github.com/EdisonLeeeee/SpikeNet}.
| no_new_dataset | 0.708408 |
2209.04333 | Yeon Seonwoo | Yeon Seonwoo, Guoyin Wang, Changmin Seo, Sajal Choudhary, Jiwei Li,
Xiang Li, Puyang Xu, Sunghyun Park, Alice Oh | Ranking-Enhanced Unsupervised Sentence Representation Learning | ACL 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Unsupervised sentence representation learning has progressed through
contrastive learning and data augmentation methods such as dropout masking.
Despite this progress, sentence encoders are still limited to using only an
input sentence when predicting its semantic vector. In this work, we show that
the semantic meaning of a sentence is also determined by nearest-neighbor
sentences that are similar to the input sentence. Based on this finding, we
propose a novel unsupervised sentence encoder, RankEncoder. RankEncoder
predicts the semantic vector of an input sentence by leveraging its
relationship with other sentences in an external corpus, as well as the input
sentence itself. We evaluate RankEncoder on semantic textual benchmark
datasets. From the experimental results, we verify that 1) RankEncoder achieves
80.07% Spearman's correlation, a 1.1% absolute improvement compared to the
previous state-of-the-art performance, 2) RankEncoder is universally applicable
to existing unsupervised sentence embedding methods, and 3) RankEncoder is
specifically effective for predicting the similarity scores of similar sentence
pairs.
| [
{
"version": "v1",
"created": "Fri, 9 Sep 2022 14:45:16 GMT"
},
{
"version": "v2",
"created": "Wed, 21 Sep 2022 08:43:01 GMT"
},
{
"version": "v3",
"created": "Thu, 18 May 2023 08:50:22 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Seonwoo",
"Yeon",
""
],
[
"Wang",
"Guoyin",
""
],
[
"Seo",
"Changmin",
""
],
[
"Choudhary",
"Sajal",
""
],
[
"Li",
"Jiwei",
""
],
[
"Li",
"Xiang",
""
],
[
"Xu",
"Puyang",
""
],
[
"Park",
"Sunghyun",
""
],
[
"Oh",
"Alice",
""
]
] | TITLE: Ranking-Enhanced Unsupervised Sentence Representation Learning
ABSTRACT: Unsupervised sentence representation learning has progressed through
contrastive learning and data augmentation methods such as dropout masking.
Despite this progress, sentence encoders are still limited to using only an
input sentence when predicting its semantic vector. In this work, we show that
the semantic meaning of a sentence is also determined by nearest-neighbor
sentences that are similar to the input sentence. Based on this finding, we
propose a novel unsupervised sentence encoder, RankEncoder. RankEncoder
predicts the semantic vector of an input sentence by leveraging its
relationship with other sentences in an external corpus, as well as the input
sentence itself. We evaluate RankEncoder on semantic textual benchmark
datasets. From the experimental results, we verify that 1) RankEncoder achieves
80.07% Spearman's correlation, a 1.1% absolute improvement compared to the
previous state-of-the-art performance, 2) RankEncoder is universally applicable
to existing unsupervised sentence embedding methods, and 3) RankEncoder is
specifically effective for predicting the similarity scores of similar sentence
pairs.
| no_new_dataset | 0.710057 |
2210.08855 | Weiwen Xu | Weiwen Xu, Xin Li, Yang Deng, Wai Lam, Lidong Bing | PeerDA: Data Augmentation via Modeling Peer Relation for Span
Identification Tasks | To appear at ACL 2023 main conference | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Span identification aims at identifying specific text spans from text input
and classifying them into pre-defined categories. Different from previous works
that merely leverage the Subordinate (SUB) relation (i.e. if a span is an
instance of a certain category) to train models, this paper for the first time
explores the Peer (PR) relation, which indicates that two spans are instances
of the same category and share similar features. Specifically, a novel Peer
Data Augmentation (PeerDA) approach is proposed which employs span pairs with
the PR relation as the augmentation data for training. PeerDA has two unique
advantages: (1) There are a large number of PR span pairs for augmenting the
training data. (2) The augmented data can prevent the trained model from
over-fitting the superficial span-category mapping by pushing the model to
leverage the span semantics. Experimental results on ten datasets over four
diverse tasks across seven domains demonstrate the effectiveness of PeerDA.
Notably, PeerDA achieves state-of-the-art results on six of them.
| [
{
"version": "v1",
"created": "Mon, 17 Oct 2022 08:51:30 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 12:11:08 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Xu",
"Weiwen",
""
],
[
"Li",
"Xin",
""
],
[
"Deng",
"Yang",
""
],
[
"Lam",
"Wai",
""
],
[
"Bing",
"Lidong",
""
]
] | TITLE: PeerDA: Data Augmentation via Modeling Peer Relation for Span
Identification Tasks
ABSTRACT: Span identification aims at identifying specific text spans from text input
and classifying them into pre-defined categories. Different from previous works
that merely leverage the Subordinate (SUB) relation (i.e. if a span is an
instance of a certain category) to train models, this paper for the first time
explores the Peer (PR) relation, which indicates that two spans are instances
of the same category and share similar features. Specifically, a novel Peer
Data Augmentation (PeerDA) approach is proposed which employs span pairs with
the PR relation as the augmentation data for training. PeerDA has two unique
advantages: (1) There are a large number of PR span pairs for augmenting the
training data. (2) The augmented data can prevent the trained model from
over-fitting the superficial span-category mapping by pushing the model to
leverage the span semantics. Experimental results on ten datasets over four
diverse tasks across seven domains demonstrate the effectiveness of PeerDA.
Notably, PeerDA achieves state-of-the-art results on six of them.
| no_new_dataset | 0.711663 |
2210.15195 | Ahmed Attia | Ahmed Adel Attia, Carol Espy-Wilson | Masked Autoencoders Are Articulatory Learners | null | null | 10.1109/ICASSP49357.2023.10096209 | null | eess.AS cs.LG cs.SD | http://creativecommons.org/licenses/by/4.0/ | Articulatory recordings track the positions and motion of different
articulators along the vocal tract and are widely used to study speech
production and to develop speech technologies such as articulatory based speech
synthesizers and speech inversion systems. The University of Wisconsin X-Ray
microbeam (XRMB) dataset is one of various datasets that provide articulatory
recordings synced with audio recordings. The XRMB articulatory recordings
employ pellets placed on a number of articulators which can be tracked by the
microbeam. However, a significant portion of the articulatory recordings are
mistracked, and have been so far unsuable. In this work, we present a deep
learning based approach using Masked Autoencoders to accurately reconstruct the
mistracked articulatory recordings for 41 out of 47 speakers of the XRMB
dataset. Our model is able to reconstruct articulatory trajectories that
closely match ground truth, even when three out of eight articulators are
mistracked, and retrieve 3.28 out of 3.4 hours of previously unusable
recordings.
| [
{
"version": "v1",
"created": "Thu, 27 Oct 2022 06:03:47 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Apr 2023 02:21:22 GMT"
},
{
"version": "v3",
"created": "Thu, 18 May 2023 07:38:34 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Attia",
"Ahmed Adel",
""
],
[
"Espy-Wilson",
"Carol",
""
]
] | TITLE: Masked Autoencoders Are Articulatory Learners
ABSTRACT: Articulatory recordings track the positions and motion of different
articulators along the vocal tract and are widely used to study speech
production and to develop speech technologies such as articulatory based speech
synthesizers and speech inversion systems. The University of Wisconsin X-Ray
microbeam (XRMB) dataset is one of various datasets that provide articulatory
recordings synced with audio recordings. The XRMB articulatory recordings
employ pellets placed on a number of articulators which can be tracked by the
microbeam. However, a significant portion of the articulatory recordings are
mistracked, and have been so far unsuable. In this work, we present a deep
learning based approach using Masked Autoencoders to accurately reconstruct the
mistracked articulatory recordings for 41 out of 47 speakers of the XRMB
dataset. Our model is able to reconstruct articulatory trajectories that
closely match ground truth, even when three out of eight articulators are
mistracked, and retrieve 3.28 out of 3.4 hours of previously unusable
recordings.
| no_new_dataset | 0.707967 |
2211.15661 | Ekin Aky\"urek | Ekin Aky\"urek, Dale Schuurmans, Jacob Andreas, Tengyu Ma, Denny Zhou | What learning algorithm is in-context learning? Investigations with
linear models | ICLR2023 Camera Ready | null | null | null | cs.LG cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural sequence models, especially transformers, exhibit a remarkable
capacity for in-context learning. They can construct new predictors from
sequences of labeled examples $(x, f(x))$ presented in the input without
further parameter updates. We investigate the hypothesis that transformer-based
in-context learners implement standard learning algorithms implicitly, by
encoding smaller models in their activations, and updating these implicit
models as new examples appear in the context. Using linear regression as a
prototypical problem, we offer three sources of evidence for this hypothesis.
First, we prove by construction that transformers can implement learning
algorithms for linear models based on gradient descent and closed-form ridge
regression. Second, we show that trained in-context learners closely match the
predictors computed by gradient descent, ridge regression, and exact
least-squares regression, transitioning between different predictors as
transformer depth and dataset noise vary, and converging to Bayesian estimators
for large widths and depths. Third, we present preliminary evidence that
in-context learners share algorithmic features with these predictors: learners'
late layers non-linearly encode weight vectors and moment matrices. These
results suggest that in-context learning is understandable in algorithmic
terms, and that (at least in the linear case) learners may rediscover standard
estimation algorithms. Code and reference implementations are released at
https://github.com/ekinakyurek/google-research/blob/master/incontext.
| [
{
"version": "v1",
"created": "Mon, 28 Nov 2022 18:59:51 GMT"
},
{
"version": "v2",
"created": "Tue, 29 Nov 2022 02:21:00 GMT"
},
{
"version": "v3",
"created": "Wed, 17 May 2023 21:08:32 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Akyürek",
"Ekin",
""
],
[
"Schuurmans",
"Dale",
""
],
[
"Andreas",
"Jacob",
""
],
[
"Ma",
"Tengyu",
""
],
[
"Zhou",
"Denny",
""
]
] | TITLE: What learning algorithm is in-context learning? Investigations with
linear models
ABSTRACT: Neural sequence models, especially transformers, exhibit a remarkable
capacity for in-context learning. They can construct new predictors from
sequences of labeled examples $(x, f(x))$ presented in the input without
further parameter updates. We investigate the hypothesis that transformer-based
in-context learners implement standard learning algorithms implicitly, by
encoding smaller models in their activations, and updating these implicit
models as new examples appear in the context. Using linear regression as a
prototypical problem, we offer three sources of evidence for this hypothesis.
First, we prove by construction that transformers can implement learning
algorithms for linear models based on gradient descent and closed-form ridge
regression. Second, we show that trained in-context learners closely match the
predictors computed by gradient descent, ridge regression, and exact
least-squares regression, transitioning between different predictors as
transformer depth and dataset noise vary, and converging to Bayesian estimators
for large widths and depths. Third, we present preliminary evidence that
in-context learners share algorithmic features with these predictors: learners'
late layers non-linearly encode weight vectors and moment matrices. These
results suggest that in-context learning is understandable in algorithmic
terms, and that (at least in the linear case) learners may rediscover standard
estimation algorithms. Code and reference implementations are released at
https://github.com/ekinakyurek/google-research/blob/master/incontext.
| no_new_dataset | 0.708635 |
2211.16553 | Amanda M. Buch Ph.D. | Amanda M. Buch, Conor Liston, and Logan Grosenick | Simple and Scalable Algorithms for Cluster-Aware Precision Medicine | 15 pages, 3 figures, 3 tables | null | null | null | cs.LG q-bio.QM stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | AI-enabled precision medicine promises a transformational improvement in
healthcare outcomes by enabling data-driven personalized diagnosis, prognosis,
and treatment. However, the well-known "curse of dimensionality" and the
clustered structure of biomedical data together interact to present a joint
challenge in the high dimensional, limited observation precision medicine
regime. To overcome both issues simultaneously we propose a simple and scalable
approach to joint clustering and embedding that combines standard embedding
methods with a convex clustering penalty in a modular way. This novel,
cluster-aware embedding approach overcomes the complexity and limitations of
current joint embedding and clustering methods, which we show with
straightforward implementations of hierarchically clustered principal component
analysis (PCA), locally linear embedding (LLE), and canonical correlation
analysis (CCA). Through both numerical experiments and real-world examples, we
demonstrate that our approach outperforms traditional and contemporary
clustering methods on highly underdetermined problems (e.g., with just tens of
observations) as well as on large sample datasets. Importantly, our approach
does not require the user to choose the desired number of clusters, but instead
yields interpretable dendrograms of hierarchically clustered embeddings. Thus
our approach improves significantly on existing methods for identifying patient
subgroups in multiomics and neuroimaging data, enabling scalable and
interpretable biomarkers for precision medicine.
| [
{
"version": "v1",
"created": "Tue, 29 Nov 2022 19:27:26 GMT"
},
{
"version": "v2",
"created": "Tue, 31 Jan 2023 02:21:28 GMT"
},
{
"version": "v3",
"created": "Wed, 17 May 2023 22:49:42 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Buch",
"Amanda M.",
""
],
[
"Liston",
"Conor",
""
],
[
"Grosenick",
"Logan",
""
]
] | TITLE: Simple and Scalable Algorithms for Cluster-Aware Precision Medicine
ABSTRACT: AI-enabled precision medicine promises a transformational improvement in
healthcare outcomes by enabling data-driven personalized diagnosis, prognosis,
and treatment. However, the well-known "curse of dimensionality" and the
clustered structure of biomedical data together interact to present a joint
challenge in the high dimensional, limited observation precision medicine
regime. To overcome both issues simultaneously we propose a simple and scalable
approach to joint clustering and embedding that combines standard embedding
methods with a convex clustering penalty in a modular way. This novel,
cluster-aware embedding approach overcomes the complexity and limitations of
current joint embedding and clustering methods, which we show with
straightforward implementations of hierarchically clustered principal component
analysis (PCA), locally linear embedding (LLE), and canonical correlation
analysis (CCA). Through both numerical experiments and real-world examples, we
demonstrate that our approach outperforms traditional and contemporary
clustering methods on highly underdetermined problems (e.g., with just tens of
observations) as well as on large sample datasets. Importantly, our approach
does not require the user to choose the desired number of clusters, but instead
yields interpretable dendrograms of hierarchically clustered embeddings. Thus
our approach improves significantly on existing methods for identifying patient
subgroups in multiomics and neuroimaging data, enabling scalable and
interpretable biomarkers for precision medicine.
| no_new_dataset | 0.710641 |
2212.00193 | Kumar Shridhar | Kumar Shridhar, Alessandro Stolfo, Mrinmaya Sachan | Distilling Reasoning Capabilities into Smaller Language Models | Accepted at ACL 2023 (Findings) | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | Step-by-step reasoning approaches like chain of thought (CoT) have proved to
be very effective in inducing reasoning capabilities in large language models.
However, the success of the CoT approach is fundamentally tied to the model
size, and billion parameter-scale models are often needed to get CoT to work.
In this paper, we propose a knowledge distillation approach that leverages the
step-by-step CoT reasoning capabilities of larger models and distills these
abilities into smaller models.
In this work, we propose an alternative reasoning scheme, Socratic CoT, that
learns a decomposition of the original problem into a sequence of subproblems
and uses it to guide the intermediate reasoning steps. We use Socratic CoT to
train a combination of two small distilled models: a problem decomposer and a
subproblem solver. In practice, given a new problem, the two distilled models
work in sync to decompose and solve complex problems. On multiple reasoning
datasets (GSM8K, StrategyQA, and SVAMP), our proposed distillation strategies
boosts the performance of smaller models over 70% compared to the baselines.
Finally, we investigate when Socratic CoT is an effective alternative to CoT,
demonstrating cases where a much smaller model (GPT-2 large) can outperform a
10X larger model (GPT-3 6B). Our code is available here:
https://github.com/kumar-shridhar/Distiiling-LM
| [
{
"version": "v1",
"created": "Thu, 1 Dec 2022 00:39:56 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 04:44:51 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Shridhar",
"Kumar",
""
],
[
"Stolfo",
"Alessandro",
""
],
[
"Sachan",
"Mrinmaya",
""
]
] | TITLE: Distilling Reasoning Capabilities into Smaller Language Models
ABSTRACT: Step-by-step reasoning approaches like chain of thought (CoT) have proved to
be very effective in inducing reasoning capabilities in large language models.
However, the success of the CoT approach is fundamentally tied to the model
size, and billion parameter-scale models are often needed to get CoT to work.
In this paper, we propose a knowledge distillation approach that leverages the
step-by-step CoT reasoning capabilities of larger models and distills these
abilities into smaller models.
In this work, we propose an alternative reasoning scheme, Socratic CoT, that
learns a decomposition of the original problem into a sequence of subproblems
and uses it to guide the intermediate reasoning steps. We use Socratic CoT to
train a combination of two small distilled models: a problem decomposer and a
subproblem solver. In practice, given a new problem, the two distilled models
work in sync to decompose and solve complex problems. On multiple reasoning
datasets (GSM8K, StrategyQA, and SVAMP), our proposed distillation strategies
boosts the performance of smaller models over 70% compared to the baselines.
Finally, we investigate when Socratic CoT is an effective alternative to CoT,
demonstrating cases where a much smaller model (GPT-2 large) can outperform a
10X larger model (GPT-3 6B). Our code is available here:
https://github.com/kumar-shridhar/Distiiling-LM
| no_new_dataset | 0.708767 |
2212.04205 | Jianhao Yan | Jianhao Yan, Jin Xu, Fandong Meng, Jie Zhou, Yue Zhang | DC-MBR: Distributional Cooling for Minimum Bayesian Risk Decoding | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Minimum Bayesian Risk Decoding (MBR) emerges as a promising decoding
algorithm in Neural Machine Translation. However, MBR performs poorly with
label smoothing, which is surprising as label smoothing provides decent
improvement with beam search and improves generality in various tasks. In this
work, we show that the issue arises from the un-consistency of label smoothing
on the token-level and sequence-level distributions. We demonstrate that even
though label smoothing only causes a slight change in the token-level, the
sequence-level distribution is highly skewed. We coin the issue
\emph{autoregressive over-smoothness}. To address this issue, we propose a
simple and effective method, Distributional Cooling MBR (DC-MBR), which
manipulates the entropy of output distributions by tuning down the Softmax
temperature. We theoretically prove the equivalence between pre-tuning label
smoothing factor and distributional cooling. Extensive experiments on NMT
benchmarks validate that distributional cooling improves MBR in various
settings.
| [
{
"version": "v1",
"created": "Thu, 8 Dec 2022 11:40:31 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 04:06:50 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yan",
"Jianhao",
""
],
[
"Xu",
"Jin",
""
],
[
"Meng",
"Fandong",
""
],
[
"Zhou",
"Jie",
""
],
[
"Zhang",
"Yue",
""
]
] | TITLE: DC-MBR: Distributional Cooling for Minimum Bayesian Risk Decoding
ABSTRACT: Minimum Bayesian Risk Decoding (MBR) emerges as a promising decoding
algorithm in Neural Machine Translation. However, MBR performs poorly with
label smoothing, which is surprising as label smoothing provides decent
improvement with beam search and improves generality in various tasks. In this
work, we show that the issue arises from the un-consistency of label smoothing
on the token-level and sequence-level distributions. We demonstrate that even
though label smoothing only causes a slight change in the token-level, the
sequence-level distribution is highly skewed. We coin the issue
\emph{autoregressive over-smoothness}. To address this issue, we propose a
simple and effective method, Distributional Cooling MBR (DC-MBR), which
manipulates the entropy of output distributions by tuning down the Softmax
temperature. We theoretically prove the equivalence between pre-tuning label
smoothing factor and distributional cooling. Extensive experiments on NMT
benchmarks validate that distributional cooling improves MBR in various
settings.
| no_new_dataset | 0.709831 |
2301.09833 | Zhiwei Zhang | Moshe Y. Vardi and Zhiwei Zhang | Solving Quantum-Inspired Perfect Matching Problems via Tutte's
Theorem-Based Hybrid Boolean Constraints | Accepted by IJCAI'23 | null | null | null | cs.AI cs.LO math.CO quant-ph | http://creativecommons.org/licenses/by/4.0/ | Determining the satisfiability of Boolean constraint-satisfaction problems
with different types of constraints, that is hybrid constraints, is a
well-studied problem with important applications. We study here a new
application of hybrid Boolean constraints, which arises in quantum computing.
The problem relates to constrained perfect matching in edge-colored graphs.
While general-purpose hybrid constraint solvers can be powerful, we show that
direct encodings of the constrained-matching problem as hybrid constraints
scale poorly and special techniques are still needed. We propose a novel
encoding based on Tutte's Theorem in graph theory as well as optimization
techniques. Empirical results demonstrate that our encoding, in suitable
languages with advanced SAT solvers, scales significantly better than a number
of competing approaches on constrained-matching benchmarks. Our study
identifies the necessity of designing problem-specific encodings when applying
powerful general-purpose constraint solvers.
| [
{
"version": "v1",
"created": "Tue, 24 Jan 2023 06:14:24 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 02:46:09 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Vardi",
"Moshe Y.",
""
],
[
"Zhang",
"Zhiwei",
""
]
] | TITLE: Solving Quantum-Inspired Perfect Matching Problems via Tutte's
Theorem-Based Hybrid Boolean Constraints
ABSTRACT: Determining the satisfiability of Boolean constraint-satisfaction problems
with different types of constraints, that is hybrid constraints, is a
well-studied problem with important applications. We study here a new
application of hybrid Boolean constraints, which arises in quantum computing.
The problem relates to constrained perfect matching in edge-colored graphs.
While general-purpose hybrid constraint solvers can be powerful, we show that
direct encodings of the constrained-matching problem as hybrid constraints
scale poorly and special techniques are still needed. We propose a novel
encoding based on Tutte's Theorem in graph theory as well as optimization
techniques. Empirical results demonstrate that our encoding, in suitable
languages with advanced SAT solvers, scales significantly better than a number
of competing approaches on constrained-matching benchmarks. Our study
identifies the necessity of designing problem-specific encodings when applying
powerful general-purpose constraint solvers.
| no_new_dataset | 0.705189 |
2301.12077 | Zheng Lian | Mingyu Xu, Zheng Lian, Lei Feng, Bin Liu, Jianhua Tao | ALIM: Adjusting Label Importance Mechanism for Noisy Partial Label
Learning | null | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Noisy partial label learning (noisy PLL) is an important branch of weakly
supervised learning. Unlike PLL where the ground-truth label must conceal in
the candidate label set, noisy PLL relaxes this constraint and allows the
ground-truth label may not be in the candidate label set. To address this
challenging problem, most of the existing works attempt to detect noisy samples
and estimate the ground-truth label for each noisy sample. However, detection
errors are unavoidable. These errors can accumulate during training and
continuously affect model optimization. To this end, we propose a novel
framework for noisy PLL with theoretical guarantees, called ``Adjusting Label
Importance Mechanism (ALIM)''. It aims to reduce the negative impact of
detection errors by trading off the initial candidate set and model outputs.
ALIM is a plug-in strategy that can be integrated with existing PLL approaches.
Experimental results on benchmark datasets demonstrate that our method can
achieve state-of-the-art performance on noisy PLL.
\textcolor[rgb]{0.93,0.0,0.47}{Our code can be found in Supplementary
Material}.
| [
{
"version": "v1",
"created": "Sat, 28 Jan 2023 03:42:53 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 02:07:54 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Xu",
"Mingyu",
""
],
[
"Lian",
"Zheng",
""
],
[
"Feng",
"Lei",
""
],
[
"Liu",
"Bin",
""
],
[
"Tao",
"Jianhua",
""
]
] | TITLE: ALIM: Adjusting Label Importance Mechanism for Noisy Partial Label
Learning
ABSTRACT: Noisy partial label learning (noisy PLL) is an important branch of weakly
supervised learning. Unlike PLL where the ground-truth label must conceal in
the candidate label set, noisy PLL relaxes this constraint and allows the
ground-truth label may not be in the candidate label set. To address this
challenging problem, most of the existing works attempt to detect noisy samples
and estimate the ground-truth label for each noisy sample. However, detection
errors are unavoidable. These errors can accumulate during training and
continuously affect model optimization. To this end, we propose a novel
framework for noisy PLL with theoretical guarantees, called ``Adjusting Label
Importance Mechanism (ALIM)''. It aims to reduce the negative impact of
detection errors by trading off the initial candidate set and model outputs.
ALIM is a plug-in strategy that can be integrated with existing PLL approaches.
Experimental results on benchmark datasets demonstrate that our method can
achieve state-of-the-art performance on noisy PLL.
\textcolor[rgb]{0.93,0.0,0.47}{Our code can be found in Supplementary
Material}.
| no_new_dataset | 0.709007 |
2301.12586 | Giorgio Giannone | Dimitrios Christofidellis, Giorgio Giannone, Jannis Born, Ole Winther,
Teodoro Laino, Matteo Manica | Unifying Molecular and Textual Representations via Multi-task Language
Modelling | ICML 2023 | null | null | null | cs.LG cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The recent advances in neural language models have also been successfully
applied to the field of chemistry, offering generative solutions for classical
problems in molecular design and synthesis planning. These new methods have the
potential to fuel a new era of data-driven automation in scientific discovery.
However, specialized models are still typically required for each task, leading
to the need for problem-specific fine-tuning and neglecting task
interrelations. The main obstacle in this field is the lack of a unified
representation between natural language and chemical representations,
complicating and limiting human-machine interaction. Here, we propose the first
multi-domain, multi-task language model that can solve a wide range of tasks in
both the chemical and natural language domains. Our model can handle chemical
and natural language concurrently, without requiring expensive pre-training on
single domains or task-specific models. Interestingly, sharing weights across
domains remarkably improves our model when benchmarked against state-of-the-art
baselines on single-domain and cross-domain tasks. In particular, sharing
information across domains and tasks gives rise to large improvements in
cross-domain tasks, the magnitude of which increase with scale, as measured by
more than a dozen of relevant metrics. Our work suggests that such models can
robustly and efficiently accelerate discovery in physical sciences by
superseding problem-specific fine-tuning and enhancing human-model
interactions.
| [
{
"version": "v1",
"created": "Sun, 29 Jan 2023 23:56:45 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 00:37:00 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Christofidellis",
"Dimitrios",
""
],
[
"Giannone",
"Giorgio",
""
],
[
"Born",
"Jannis",
""
],
[
"Winther",
"Ole",
""
],
[
"Laino",
"Teodoro",
""
],
[
"Manica",
"Matteo",
""
]
] | TITLE: Unifying Molecular and Textual Representations via Multi-task Language
Modelling
ABSTRACT: The recent advances in neural language models have also been successfully
applied to the field of chemistry, offering generative solutions for classical
problems in molecular design and synthesis planning. These new methods have the
potential to fuel a new era of data-driven automation in scientific discovery.
However, specialized models are still typically required for each task, leading
to the need for problem-specific fine-tuning and neglecting task
interrelations. The main obstacle in this field is the lack of a unified
representation between natural language and chemical representations,
complicating and limiting human-machine interaction. Here, we propose the first
multi-domain, multi-task language model that can solve a wide range of tasks in
both the chemical and natural language domains. Our model can handle chemical
and natural language concurrently, without requiring expensive pre-training on
single domains or task-specific models. Interestingly, sharing weights across
domains remarkably improves our model when benchmarked against state-of-the-art
baselines on single-domain and cross-domain tasks. In particular, sharing
information across domains and tasks gives rise to large improvements in
cross-domain tasks, the magnitude of which increase with scale, as measured by
more than a dozen of relevant metrics. Our work suggests that such models can
robustly and efficiently accelerate discovery in physical sciences by
superseding problem-specific fine-tuning and enhancing human-model
interactions.
| no_new_dataset | 0.709453 |
2302.05328 | Jingnan Zheng | An Zhang, Jingnan Zheng, Xiang Wang, Yancheng Yuan, Tat-Seng Chua | Invariant Collaborative Filtering to Popularity Distribution Shift | null | 2023 TheWebConf | 10.1145/3543507.3583461 | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Collaborative Filtering (CF) models, despite their great success, suffer from
severe performance drops due to popularity distribution shifts, where these
changes are ubiquitous and inevitable in real-world scenarios. Unfortunately,
most leading popularity debiasing strategies, rather than tackling the
vulnerability of CF models to varying popularity distributions, require prior
knowledge of the test distribution to identify the degree of bias and further
learn the popularity-entangled representations to mitigate the bias.
Consequently, these models result in significant performance benefits in the
target test set, while dramatically deviating the recommendation from users'
true interests without knowing the popularity distribution in advance. In this
work, we propose a novel learning framework, Invariant Collaborative Filtering
(InvCF), to discover disentangled representations that faithfully reveal the
latent preference and popularity semantics without making any assumption about
the popularity distribution. At its core is the distillation of unbiased
preference representations (i.e., user preference on item property), which are
invariant to the change of popularity semantics, while filtering out the
popularity feature that is unstable or outdated. Extensive experiments on five
benchmark datasets and four evaluation settings (i.e., synthetic long-tail,
unbiased, temporal split, and out-of-distribution evaluations) demonstrate that
InvCF outperforms the state-of-the-art baselines in terms of popularity
generalization ability on real recommendations. Visualization studies shed
light on the advantages of InvCF for disentangled representation learning. Our
codes are available at https://github.com/anzhang314/InvCF.
| [
{
"version": "v1",
"created": "Fri, 10 Feb 2023 15:45:59 GMT"
},
{
"version": "v2",
"created": "Mon, 13 Feb 2023 11:00:47 GMT"
},
{
"version": "v3",
"created": "Thu, 18 May 2023 04:01:30 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Zhang",
"An",
""
],
[
"Zheng",
"Jingnan",
""
],
[
"Wang",
"Xiang",
""
],
[
"Yuan",
"Yancheng",
""
],
[
"Chua",
"Tat-Seng",
""
]
] | TITLE: Invariant Collaborative Filtering to Popularity Distribution Shift
ABSTRACT: Collaborative Filtering (CF) models, despite their great success, suffer from
severe performance drops due to popularity distribution shifts, where these
changes are ubiquitous and inevitable in real-world scenarios. Unfortunately,
most leading popularity debiasing strategies, rather than tackling the
vulnerability of CF models to varying popularity distributions, require prior
knowledge of the test distribution to identify the degree of bias and further
learn the popularity-entangled representations to mitigate the bias.
Consequently, these models result in significant performance benefits in the
target test set, while dramatically deviating the recommendation from users'
true interests without knowing the popularity distribution in advance. In this
work, we propose a novel learning framework, Invariant Collaborative Filtering
(InvCF), to discover disentangled representations that faithfully reveal the
latent preference and popularity semantics without making any assumption about
the popularity distribution. At its core is the distillation of unbiased
preference representations (i.e., user preference on item property), which are
invariant to the change of popularity semantics, while filtering out the
popularity feature that is unstable or outdated. Extensive experiments on five
benchmark datasets and four evaluation settings (i.e., synthetic long-tail,
unbiased, temporal split, and out-of-distribution evaluations) demonstrate that
InvCF outperforms the state-of-the-art baselines in terms of popularity
generalization ability on real recommendations. Visualization studies shed
light on the advantages of InvCF for disentangled representation learning. Our
codes are available at https://github.com/anzhang314/InvCF.
| no_new_dataset | 0.709372 |
2302.05918 | Biao Xu | Biao Xu, Yao Wang, Xiuwu Liao, Kaidong Wang | Efficient Fraud Detection Using Deep Boosting Decision Trees | 34 pages, 8 figures | null | null | null | stat.ML cs.LG stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fraud detection is to identify, monitor, and prevent potentially fraudulent
activities from complex data. The recent development and success in AI,
especially machine learning, provides a new data-driven way to deal with fraud.
From a methodological point of view, machine learning based fraud detection can
be divided into two categories, i.e., conventional methods (decision tree,
boosting...) and deep learning, both of which have significant limitations in
terms of the lack of representation learning ability for the former and
interpretability for the latter. Furthermore, due to the rarity of detected
fraud cases, the associated data is usually imbalanced, which seriously
degrades the performance of classification algorithms. In this paper, we
propose deep boosting decision trees (DBDT), a novel approach for fraud
detection based on gradient boosting and neural networks. In order to combine
the advantages of both conventional methods and deep learning, we first
construct soft decision tree (SDT), a decision tree structured model with
neural networks as its nodes, and then ensemble SDTs using the idea of gradient
boosting. In this way we embed neural networks into gradient boosting to
improve its representation learning capability and meanwhile maintain the
interpretability. Furthermore, aiming at the rarity of detected fraud cases, in
the model training phase we propose a compositional AUC maximization approach
to deal with data imbalances at algorithm level. Extensive experiments on
several real-life fraud detection datasets show that DBDT can significantly
improve the performance and meanwhile maintain good interpretability. Our code
is available at https://github.com/freshmanXB/DBDT.
| [
{
"version": "v1",
"created": "Sun, 12 Feb 2023 14:02:58 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 11:19:24 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Xu",
"Biao",
""
],
[
"Wang",
"Yao",
""
],
[
"Liao",
"Xiuwu",
""
],
[
"Wang",
"Kaidong",
""
]
] | TITLE: Efficient Fraud Detection Using Deep Boosting Decision Trees
ABSTRACT: Fraud detection is to identify, monitor, and prevent potentially fraudulent
activities from complex data. The recent development and success in AI,
especially machine learning, provides a new data-driven way to deal with fraud.
From a methodological point of view, machine learning based fraud detection can
be divided into two categories, i.e., conventional methods (decision tree,
boosting...) and deep learning, both of which have significant limitations in
terms of the lack of representation learning ability for the former and
interpretability for the latter. Furthermore, due to the rarity of detected
fraud cases, the associated data is usually imbalanced, which seriously
degrades the performance of classification algorithms. In this paper, we
propose deep boosting decision trees (DBDT), a novel approach for fraud
detection based on gradient boosting and neural networks. In order to combine
the advantages of both conventional methods and deep learning, we first
construct soft decision tree (SDT), a decision tree structured model with
neural networks as its nodes, and then ensemble SDTs using the idea of gradient
boosting. In this way we embed neural networks into gradient boosting to
improve its representation learning capability and meanwhile maintain the
interpretability. Furthermore, aiming at the rarity of detected fraud cases, in
the model training phase we propose a compositional AUC maximization approach
to deal with data imbalances at algorithm level. Extensive experiments on
several real-life fraud detection datasets show that DBDT can significantly
improve the performance and meanwhile maintain good interpretability. Our code
is available at https://github.com/freshmanXB/DBDT.
| no_new_dataset | 0.708635 |
2302.10756 | Feng Qian | Feng Qian, Yuehua Yue, Yu He, Hongtao Yu, Yingjie Zhou, Jinliang Tang,
and Guangmin Hu | Unsupervised Seismic Footprint Removal With Physical Prior Augmented
Deep Autoencoder | null | IEEE Transactions on Geoscience and Remote Sensing,2023 | 10.1109/TGRS.2023.3277973 | 2302.10756 | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Seismic acquisition footprints appear as stably faint and dim structures and
emerge fully spatially coherent, causing inevitable damage to useful signals
during the suppression process. Various footprint removal methods, including
filtering and sparse representation (SR), have been reported to attain
promising results for surmounting this challenge. However, these methods, e.g.,
SR, rely solely on the handcrafted image priors of useful signals, which is
sometimes an unreasonable demand if complex geological structures are contained
in the given seismic data. As an alternative, this article proposes a footprint
removal network (dubbed FR-Net) for the unsupervised suppression of acquired
footprints without any assumptions regarding valuable signals. The key to the
FR-Net is to design a unidirectional total variation (UTV) model for footprint
acquisition according to the intrinsically directional property of noise. By
strongly regularizing a deep convolutional autoencoder (DCAE) using the UTV
model, our FR-Net transforms the DCAE from an entirely data-driven model to a
\textcolor{black}{prior-augmented} approach, inheriting the superiority of the
DCAE and our footprint model. Subsequently, the complete separation of the
footprint noise and useful signals is projected in an unsupervised manner,
specifically by optimizing the FR-Net via the backpropagation (BP) algorithm.
We provide qualitative and quantitative evaluations conducted on three
synthetic and field datasets, demonstrating that our FR-Net surpasses the
previous state-of-the-art (SOTA) methods.
| [
{
"version": "v1",
"created": "Wed, 8 Feb 2023 07:46:28 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Qian",
"Feng",
""
],
[
"Yue",
"Yuehua",
""
],
[
"He",
"Yu",
""
],
[
"Yu",
"Hongtao",
""
],
[
"Zhou",
"Yingjie",
""
],
[
"Tang",
"Jinliang",
""
],
[
"Hu",
"Guangmin",
""
]
] | TITLE: Unsupervised Seismic Footprint Removal With Physical Prior Augmented
Deep Autoencoder
ABSTRACT: Seismic acquisition footprints appear as stably faint and dim structures and
emerge fully spatially coherent, causing inevitable damage to useful signals
during the suppression process. Various footprint removal methods, including
filtering and sparse representation (SR), have been reported to attain
promising results for surmounting this challenge. However, these methods, e.g.,
SR, rely solely on the handcrafted image priors of useful signals, which is
sometimes an unreasonable demand if complex geological structures are contained
in the given seismic data. As an alternative, this article proposes a footprint
removal network (dubbed FR-Net) for the unsupervised suppression of acquired
footprints without any assumptions regarding valuable signals. The key to the
FR-Net is to design a unidirectional total variation (UTV) model for footprint
acquisition according to the intrinsically directional property of noise. By
strongly regularizing a deep convolutional autoencoder (DCAE) using the UTV
model, our FR-Net transforms the DCAE from an entirely data-driven model to a
\textcolor{black}{prior-augmented} approach, inheriting the superiority of the
DCAE and our footprint model. Subsequently, the complete separation of the
footprint noise and useful signals is projected in an unsupervised manner,
specifically by optimizing the FR-Net via the backpropagation (BP) algorithm.
We provide qualitative and quantitative evaluations conducted on three
synthetic and field datasets, demonstrating that our FR-Net surpasses the
previous state-of-the-art (SOTA) methods.
| no_new_dataset | 0.711186 |
2302.11716 | Zheng Lian | Mingyu Xu, Zheng Lian, Bin Liu, Jianhua Tao | VRA: Variational Rectified Activation for Out-of-distribution Detection | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Out-of-distribution (OOD) detection is critical to building reliable machine
learning systems in the open world. Researchers have proposed various
strategies to reduce model overconfidence on OOD data. Among them, ReAct is a
typical and effective technique to deal with model overconfidence, which
truncates high activations to increase the gap between in-distribution and OOD.
Despite its promising results, is this technique the best choice for widening
the gap? To answer this question, we leverage the variational method to find
the optimal operation and verify the necessity of suppressing abnormally low
and high activations and amplifying intermediate activations in OOD detection,
rather than focusing only on high activations like ReAct. This motivates us to
propose a novel technique called ``Variational Rectified Activation (VRA)'',
which simulates these suppression and amplification operations using piecewise
functions. Experimental results on multiple benchmark datasets demonstrate that
our method outperforms existing post-hoc strategies. Meanwhile, VRA is
compatible with different scoring functions and network architectures.
\textcolor[rgb]{0.93,0.0,0.47}{Our code can be found in Supplementary
Material}.
| [
{
"version": "v1",
"created": "Thu, 23 Feb 2023 00:45:14 GMT"
},
{
"version": "v2",
"created": "Tue, 28 Feb 2023 12:21:42 GMT"
},
{
"version": "v3",
"created": "Fri, 3 Mar 2023 11:57:27 GMT"
},
{
"version": "v4",
"created": "Thu, 18 May 2023 02:05:22 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Xu",
"Mingyu",
""
],
[
"Lian",
"Zheng",
""
],
[
"Liu",
"Bin",
""
],
[
"Tao",
"Jianhua",
""
]
] | TITLE: VRA: Variational Rectified Activation for Out-of-distribution Detection
ABSTRACT: Out-of-distribution (OOD) detection is critical to building reliable machine
learning systems in the open world. Researchers have proposed various
strategies to reduce model overconfidence on OOD data. Among them, ReAct is a
typical and effective technique to deal with model overconfidence, which
truncates high activations to increase the gap between in-distribution and OOD.
Despite its promising results, is this technique the best choice for widening
the gap? To answer this question, we leverage the variational method to find
the optimal operation and verify the necessity of suppressing abnormally low
and high activations and amplifying intermediate activations in OOD detection,
rather than focusing only on high activations like ReAct. This motivates us to
propose a novel technique called ``Variational Rectified Activation (VRA)'',
which simulates these suppression and amplification operations using piecewise
functions. Experimental results on multiple benchmark datasets demonstrate that
our method outperforms existing post-hoc strategies. Meanwhile, VRA is
compatible with different scoring functions and network architectures.
\textcolor[rgb]{0.93,0.0,0.47}{Our code can be found in Supplementary
Material}.
| no_new_dataset | 0.709069 |
2303.05668 | Sreyan Ghosh | Ashish Seth and Sreyan Ghosh and S. Umesh and Dinesh Manocha | UNFUSED: UNsupervised Finetuning Using SElf supervised Distillation | ICASSP 2023 SASB Workshop | null | null | null | eess.AS cs.AI | http://creativecommons.org/licenses/by/4.0/ | In this paper, we introduce UnFuSeD, a novel approach to leverage
self-supervised learning and reduce the need for large amounts of labeled data
for audio classification. Unlike prior works, which directly fine-tune a
self-supervised pre-trained encoder on a target dataset, we use the encoder to
generate pseudo-labels for unsupervised fine-tuning before the actual
fine-tuning step. We first train an encoder using a novel self-supervised
learning algorithm (SSL) on an unlabeled audio dataset. Then, we use that
encoder to generate pseudo-labels on our target task dataset via clustering the
extracted representations. These pseudo-labels are then used to guide
self-distillation on a randomly initialized model, which we call unsupervised
fine-tuning. Finally, the resultant encoder is then fine-tuned on our target
task dataset. Through UnFuSeD, we propose the first system that moves away from
generic SSL paradigms in literature, which pre-train and fine-tune the same
encoder, and present a novel self-distillation-based system to leverage SSL
pre-training for low-resource audio classification. In practice, UnFuSeD
achieves state-of-the-art results on the LAPE Benchmark, significantly
outperforming all our baselines. Additionally, UnFuSeD allows us to achieve
this at a 40% reduction in the number of parameters over the previous
state-of-the-art system. We make all our codes publicly available.
| [
{
"version": "v1",
"created": "Fri, 10 Mar 2023 02:43:36 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 01:28:01 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Seth",
"Ashish",
""
],
[
"Ghosh",
"Sreyan",
""
],
[
"Umesh",
"S.",
""
],
[
"Manocha",
"Dinesh",
""
]
] | TITLE: UNFUSED: UNsupervised Finetuning Using SElf supervised Distillation
ABSTRACT: In this paper, we introduce UnFuSeD, a novel approach to leverage
self-supervised learning and reduce the need for large amounts of labeled data
for audio classification. Unlike prior works, which directly fine-tune a
self-supervised pre-trained encoder on a target dataset, we use the encoder to
generate pseudo-labels for unsupervised fine-tuning before the actual
fine-tuning step. We first train an encoder using a novel self-supervised
learning algorithm (SSL) on an unlabeled audio dataset. Then, we use that
encoder to generate pseudo-labels on our target task dataset via clustering the
extracted representations. These pseudo-labels are then used to guide
self-distillation on a randomly initialized model, which we call unsupervised
fine-tuning. Finally, the resultant encoder is then fine-tuned on our target
task dataset. Through UnFuSeD, we propose the first system that moves away from
generic SSL paradigms in literature, which pre-train and fine-tune the same
encoder, and present a novel self-distillation-based system to leverage SSL
pre-training for low-resource audio classification. In practice, UnFuSeD
achieves state-of-the-art results on the LAPE Benchmark, significantly
outperforming all our baselines. Additionally, UnFuSeD allows us to achieve
this at a 40% reduction in the number of parameters over the previous
state-of-the-art system. We make all our codes publicly available.
| no_new_dataset | 0.710245 |
2303.17995 | Andrei Velichko | Andrei Velichko, Maksim Belyaev, Yuriy Izotov, Murugappan Murugappan
and Hanif Heidari | Neural Network Entropy (NNetEn): Entropy-Based EEG Signal and Chaotic
Time Series Classification, Python Package for NNetEn Calculation | 26 pages, 18 figures, 2 tables | Algorithms 2023, 16, 255 | 10.3390/a16050255 | null | cs.LG cs.AI eess.SP nlin.CD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Entropy measures are effective features for time series classification
problems. Traditional entropy measures, such as Shannon entropy, use
probability distribution function. However, for the effective separation of
time series, new entropy estimation methods are required to characterize the
chaotic dynamic of the system. Our concept of Neural Network Entropy (NNetEn)
is based on the classification of special datasets in relation to the entropy
of the time series recorded in the reservoir of the neural network. NNetEn
estimates the chaotic dynamics of time series in an original way and does not
take into account probability distribution functions. We propose two new
classification metrics: R2 Efficiency and Pearson Efficiency. The efficiency of
NNetEn is verified on separation of two chaotic time series of sine mapping
using dispersion analysis. For two close dynamic time series (r = 1.1918 and r
= 1.2243), the F-ratio has reached the value of 124 and reflects high
efficiency of the introduced method in classification problems. The
electroenceph-alography signal classification for healthy persons and patients
with Alzheimer disease illustrates the practical application of the NNetEn
features. Our computations demonstrate the synergistic effect of increasing
classification accuracy when applying traditional entropy measures and the
NNetEn concept conjointly. An implementation of the algorithms in Python is
presented.
| [
{
"version": "v1",
"created": "Fri, 31 Mar 2023 12:11:21 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 15:57:24 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Velichko",
"Andrei",
""
],
[
"Belyaev",
"Maksim",
""
],
[
"Izotov",
"Yuriy",
""
],
[
"Murugappan",
"Murugappan",
""
],
[
"Heidari",
"Hanif",
""
]
] | TITLE: Neural Network Entropy (NNetEn): Entropy-Based EEG Signal and Chaotic
Time Series Classification, Python Package for NNetEn Calculation
ABSTRACT: Entropy measures are effective features for time series classification
problems. Traditional entropy measures, such as Shannon entropy, use
probability distribution function. However, for the effective separation of
time series, new entropy estimation methods are required to characterize the
chaotic dynamic of the system. Our concept of Neural Network Entropy (NNetEn)
is based on the classification of special datasets in relation to the entropy
of the time series recorded in the reservoir of the neural network. NNetEn
estimates the chaotic dynamics of time series in an original way and does not
take into account probability distribution functions. We propose two new
classification metrics: R2 Efficiency and Pearson Efficiency. The efficiency of
NNetEn is verified on separation of two chaotic time series of sine mapping
using dispersion analysis. For two close dynamic time series (r = 1.1918 and r
= 1.2243), the F-ratio has reached the value of 124 and reflects high
efficiency of the introduced method in classification problems. The
electroenceph-alography signal classification for healthy persons and patients
with Alzheimer disease illustrates the practical application of the NNetEn
features. Our computations demonstrate the synergistic effect of increasing
classification accuracy when applying traditional entropy measures and the
NNetEn concept conjointly. An implementation of the algorithms in Python is
presented.
| no_new_dataset | 0.711268 |
2304.01890 | Antonios Maronikolakis | Antonis Maronikolakis, Abdullatif K\"oksal, Hinrich Sch\"utze | Sociocultural knowledge is needed for selection of shots in hate speech
detection tasks | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | We introduce HATELEXICON, a lexicon of slurs and targets of hate speech for
the countries of Brazil, Germany, India and Kenya, to aid training and
interpretability of models. We demonstrate how our lexicon can be used to
interpret model predictions, showing that models developed to classify extreme
speech rely heavily on target words when making predictions. Further, we
propose a method to aid shot selection for training in low-resource settings
via HATELEXICON. In few-shot learning, the selection of shots is of paramount
importance to model performance. In our work, we simulate a few-shot setting
for German and Hindi, using HASOC data for training and the Multilingual
HateCheck (MHC) as a benchmark. We show that selecting shots based on our
lexicon leads to models performing better on MHC than models trained on shots
sampled randomly. Thus, when given only a few training examples, using our
lexicon to select shots containing more sociocultural information leads to
better few-shot performance.
| [
{
"version": "v1",
"created": "Tue, 4 Apr 2023 15:42:08 GMT"
},
{
"version": "v2",
"created": "Wed, 5 Apr 2023 06:37:05 GMT"
},
{
"version": "v3",
"created": "Tue, 11 Apr 2023 19:42:48 GMT"
},
{
"version": "v4",
"created": "Wed, 17 May 2023 23:01:20 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Maronikolakis",
"Antonis",
""
],
[
"Köksal",
"Abdullatif",
""
],
[
"Schütze",
"Hinrich",
""
]
] | TITLE: Sociocultural knowledge is needed for selection of shots in hate speech
detection tasks
ABSTRACT: We introduce HATELEXICON, a lexicon of slurs and targets of hate speech for
the countries of Brazil, Germany, India and Kenya, to aid training and
interpretability of models. We demonstrate how our lexicon can be used to
interpret model predictions, showing that models developed to classify extreme
speech rely heavily on target words when making predictions. Further, we
propose a method to aid shot selection for training in low-resource settings
via HATELEXICON. In few-shot learning, the selection of shots is of paramount
importance to model performance. In our work, we simulate a few-shot setting
for German and Hindi, using HASOC data for training and the Multilingual
HateCheck (MHC) as a benchmark. We show that selecting shots based on our
lexicon leads to models performing better on MHC than models trained on shots
sampled randomly. Thus, when given only a few training examples, using our
lexicon to select shots containing more sociocultural information leads to
better few-shot performance.
| no_new_dataset | 0.702556 |
2304.12550 | Xiaoling Zhou | Xiaoling Zhou, Nan Yang, Ou Wu | Combining Adversaries with Anti-adversaries in Training | 8 pages, 5 figures | AAAI2023 | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Adversarial training is an effective learning technique to improve the
robustness of deep neural networks. In this study, the influence of adversarial
training on deep learning models in terms of fairness, robustness, and
generalization is theoretically investigated under more general perturbation
scope that different samples can have different perturbation directions (the
adversarial and anti-adversarial directions) and varied perturbation bounds.
Our theoretical explorations suggest that the combination of adversaries and
anti-adversaries (samples with anti-adversarial perturbations) in training can
be more effective in achieving better fairness between classes and a better
tradeoff between robustness and generalization in some typical learning
scenarios (e.g., noisy label learning and imbalance learning) compared with
standard adversarial training. On the basis of our theoretical findings, a more
general learning objective that combines adversaries and anti-adversaries with
varied bounds on each training sample is presented. Meta learning is utilized
to optimize the combination weights. Experiments on benchmark datasets under
different learning scenarios verify our theoretical findings and the
effectiveness of the proposed methodology.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 03:34:35 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 04:39:45 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Zhou",
"Xiaoling",
""
],
[
"Yang",
"Nan",
""
],
[
"Wu",
"Ou",
""
]
] | TITLE: Combining Adversaries with Anti-adversaries in Training
ABSTRACT: Adversarial training is an effective learning technique to improve the
robustness of deep neural networks. In this study, the influence of adversarial
training on deep learning models in terms of fairness, robustness, and
generalization is theoretically investigated under more general perturbation
scope that different samples can have different perturbation directions (the
adversarial and anti-adversarial directions) and varied perturbation bounds.
Our theoretical explorations suggest that the combination of adversaries and
anti-adversaries (samples with anti-adversarial perturbations) in training can
be more effective in achieving better fairness between classes and a better
tradeoff between robustness and generalization in some typical learning
scenarios (e.g., noisy label learning and imbalance learning) compared with
standard adversarial training. On the basis of our theoretical findings, a more
general learning objective that combines adversaries and anti-adversaries with
varied bounds on each training sample is presented. Meta learning is utilized
to optimize the combination weights. Experiments on benchmark datasets under
different learning scenarios verify our theoretical findings and the
effectiveness of the proposed methodology.
| no_new_dataset | 0.71086 |
2304.12955 | Brian DuSell | Brian DuSell | Nondeterministic Stacks in Neural Networks | 158 pages, 24 figures. PhD thesis | null | 10.7274/jh343r10k4d | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human language is full of compositional syntactic structures, and although
neural networks have contributed to groundbreaking improvements in computer
systems that process language, widely-used neural network architectures still
exhibit limitations in their ability to process syntax. To address this issue,
prior work has proposed adding stack data structures to neural networks,
drawing inspiration from theoretical connections between syntax and stacks.
However, these methods employ deterministic stacks that are designed to track
one parse at a time, whereas syntactic ambiguity, which requires a
nondeterministic stack to parse, is extremely common in language. In this
dissertation, we remedy this discrepancy by proposing a method of incorporating
nondeterministic stacks into neural networks. We develop a differentiable data
structure that efficiently simulates a nondeterministic pushdown automaton,
representing an exponential number of computations with a dynamic programming
algorithm. We incorporate this module into two predominant architectures:
recurrent neural networks (RNNs) and transformers. We show that this raises
their formal recognition power to arbitrary context-free languages, and also
aids training, even on deterministic context-free languages. Empirically,
neural networks with nondeterministic stacks learn context-free languages much
more effectively than prior stack-augmented models, including a language with
theoretically maximal parsing difficulty. We also show that an RNN augmented
with a nondeterministic stack is capable of surprisingly powerful behavior,
such as learning cross-serial dependencies, a well-known non-context-free
pattern. We demonstrate improvements on natural language modeling and provide
analysis on a syntactic generalization benchmark. This work represents an
important step toward building systems that learn to use syntax in more
human-like fashion.
| [
{
"version": "v1",
"created": "Tue, 25 Apr 2023 16:00:40 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 01:56:21 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"DuSell",
"Brian",
""
]
] | TITLE: Nondeterministic Stacks in Neural Networks
ABSTRACT: Human language is full of compositional syntactic structures, and although
neural networks have contributed to groundbreaking improvements in computer
systems that process language, widely-used neural network architectures still
exhibit limitations in their ability to process syntax. To address this issue,
prior work has proposed adding stack data structures to neural networks,
drawing inspiration from theoretical connections between syntax and stacks.
However, these methods employ deterministic stacks that are designed to track
one parse at a time, whereas syntactic ambiguity, which requires a
nondeterministic stack to parse, is extremely common in language. In this
dissertation, we remedy this discrepancy by proposing a method of incorporating
nondeterministic stacks into neural networks. We develop a differentiable data
structure that efficiently simulates a nondeterministic pushdown automaton,
representing an exponential number of computations with a dynamic programming
algorithm. We incorporate this module into two predominant architectures:
recurrent neural networks (RNNs) and transformers. We show that this raises
their formal recognition power to arbitrary context-free languages, and also
aids training, even on deterministic context-free languages. Empirically,
neural networks with nondeterministic stacks learn context-free languages much
more effectively than prior stack-augmented models, including a language with
theoretically maximal parsing difficulty. We also show that an RNN augmented
with a nondeterministic stack is capable of surprisingly powerful behavior,
such as learning cross-serial dependencies, a well-known non-context-free
pattern. We demonstrate improvements on natural language modeling and provide
analysis on a syntactic generalization benchmark. This work represents an
important step toward building systems that learn to use syntax in more
human-like fashion.
| no_new_dataset | 0.708421 |
2305.00477 | Remo Sasso | Remo Sasso, Michelangelo Conserva, Paulo Rauber | Posterior Sampling for Deep Reinforcement Learning | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite remarkable successes, deep reinforcement learning algorithms remain
sample inefficient: they require an enormous amount of trial and error to find
good policies. Model-based algorithms promise sample efficiency by building an
environment model that can be used for planning. Posterior Sampling for
Reinforcement Learning is such a model-based algorithm that has attracted
significant interest due to its performance in the tabular setting. This paper
introduces Posterior Sampling for Deep Reinforcement Learning (PSDRL), the
first truly scalable approximation of Posterior Sampling for Reinforcement
Learning that retains its model-based essence. PSDRL combines efficient
uncertainty quantification over latent state space models with a specially
tailored continual planning algorithm based on value-function approximation.
Extensive experiments on the Atari benchmark show that PSDRL significantly
outperforms previous state-of-the-art attempts at scaling up posterior sampling
while being competitive with a state-of-the-art (model-based) reinforcement
learning method, both in sample efficiency and computational efficiency.
| [
{
"version": "v1",
"created": "Sun, 30 Apr 2023 13:23:50 GMT"
},
{
"version": "v2",
"created": "Wed, 17 May 2023 18:51:19 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Sasso",
"Remo",
""
],
[
"Conserva",
"Michelangelo",
""
],
[
"Rauber",
"Paulo",
""
]
] | TITLE: Posterior Sampling for Deep Reinforcement Learning
ABSTRACT: Despite remarkable successes, deep reinforcement learning algorithms remain
sample inefficient: they require an enormous amount of trial and error to find
good policies. Model-based algorithms promise sample efficiency by building an
environment model that can be used for planning. Posterior Sampling for
Reinforcement Learning is such a model-based algorithm that has attracted
significant interest due to its performance in the tabular setting. This paper
introduces Posterior Sampling for Deep Reinforcement Learning (PSDRL), the
first truly scalable approximation of Posterior Sampling for Reinforcement
Learning that retains its model-based essence. PSDRL combines efficient
uncertainty quantification over latent state space models with a specially
tailored continual planning algorithm based on value-function approximation.
Extensive experiments on the Atari benchmark show that PSDRL significantly
outperforms previous state-of-the-art attempts at scaling up posterior sampling
while being competitive with a state-of-the-art (model-based) reinforcement
learning method, both in sample efficiency and computational efficiency.
| no_new_dataset | 0.709831 |
2305.02777 | Yunlong Liang | Yunlong Liang, Fandong Meng, Jinan Xu, Jiaan Wang, Yufeng Chen and Jie
Zhou | Unified Model Learning for Various Neural Machine Translation | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Existing neural machine translation (NMT) studies mainly focus on developing
dataset-specific models based on data from different tasks (e.g., document
translation and chat translation). Although the dataset-specific models have
achieved impressive performance, it is cumbersome as each dataset demands a
model to be designed, trained, and stored. In this work, we aim to unify these
translation tasks into a more general setting. Specifically, we propose a
``versatile'' model, i.e., the Unified Model Learning for NMT (UMLNMT) that
works with data from different tasks, and can translate well in multiple
settings simultaneously, and theoretically it can be as many as possible.
Through unified learning, UMLNMT is able to jointly train across multiple
tasks, implementing intelligent on-demand translation. On seven widely-used
translation tasks, including sentence translation, document translation, and
chat translation, our UMLNMT results in substantial improvements over
dataset-specific models with significantly reduced model deployment costs.
Furthermore, UMLNMT can achieve competitive or better performance than
state-of-the-art dataset-specific methods. Human evaluation and in-depth
analysis also demonstrate the superiority of our approach on generating diverse
and high-quality translations. Additionally, we provide a new genre translation
dataset about famous aphorisms with 186k Chinese->English sentence pairs.
| [
{
"version": "v1",
"created": "Thu, 4 May 2023 12:21:52 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 11:53:14 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Liang",
"Yunlong",
""
],
[
"Meng",
"Fandong",
""
],
[
"Xu",
"Jinan",
""
],
[
"Wang",
"Jiaan",
""
],
[
"Chen",
"Yufeng",
""
],
[
"Zhou",
"Jie",
""
]
] | TITLE: Unified Model Learning for Various Neural Machine Translation
ABSTRACT: Existing neural machine translation (NMT) studies mainly focus on developing
dataset-specific models based on data from different tasks (e.g., document
translation and chat translation). Although the dataset-specific models have
achieved impressive performance, it is cumbersome as each dataset demands a
model to be designed, trained, and stored. In this work, we aim to unify these
translation tasks into a more general setting. Specifically, we propose a
``versatile'' model, i.e., the Unified Model Learning for NMT (UMLNMT) that
works with data from different tasks, and can translate well in multiple
settings simultaneously, and theoretically it can be as many as possible.
Through unified learning, UMLNMT is able to jointly train across multiple
tasks, implementing intelligent on-demand translation. On seven widely-used
translation tasks, including sentence translation, document translation, and
chat translation, our UMLNMT results in substantial improvements over
dataset-specific models with significantly reduced model deployment costs.
Furthermore, UMLNMT can achieve competitive or better performance than
state-of-the-art dataset-specific methods. Human evaluation and in-depth
analysis also demonstrate the superiority of our approach on generating diverse
and high-quality translations. Additionally, we provide a new genre translation
dataset about famous aphorisms with 186k Chinese->English sentence pairs.
| no_new_dataset | 0.708213 |
2305.06638 | arXiv Admin | Yafeng Wu, Lan Liu, Yongjie Yu, Guiming Chen, Junhan Hu | An Adaptive Ensemble Framework for Addressing Concept Drift in IoT Data
Streams | This submission has been withdrawn by arXiv administrators due to
inappropriate text overlap with external sources | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the modern era of digital transformation, the evolution of the
fifth-generation (5G) wireless network has played a pivotal role in
revolutionizing communication technology and accelerating the growth of smart
technology applications. Enabled by the high-speed, low-latency characteristics
of 5G, these applications have shown significant potential in various sectors,
from healthcare and transportation to energy management and beyond. As a
crucial component of smart technology, IoT systems for service delivery often
face concept drift issues in network data stream analytics due to dynamic IoT
environments, resulting in performance degradation. In this article, we propose
a drift-adaptive framework called Adaptive Exponentially Weighted Average
Ensemble (AEWAE) consisting of three stages: IoT data preprocessing, base model
learning, and online ensembling. It is a data stream analytics framework that
integrates dynamic adjustments of ensemble methods to tackle various scenarios.
Experimental results on two public IoT datasets demonstrate that our proposed
framework outperforms state-of-the-art methods, achieving high accuracy and
efficiency in IoT data stream analytics.
| [
{
"version": "v1",
"created": "Thu, 11 May 2023 08:16:09 GMT"
},
{
"version": "v2",
"created": "Sat, 13 May 2023 01:30:30 GMT"
},
{
"version": "v3",
"created": "Thu, 18 May 2023 12:42:06 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Wu",
"Yafeng",
""
],
[
"Liu",
"Lan",
""
],
[
"Yu",
"Yongjie",
""
],
[
"Chen",
"Guiming",
""
],
[
"Hu",
"Junhan",
""
]
] | TITLE: An Adaptive Ensemble Framework for Addressing Concept Drift in IoT Data
Streams
ABSTRACT: In the modern era of digital transformation, the evolution of the
fifth-generation (5G) wireless network has played a pivotal role in
revolutionizing communication technology and accelerating the growth of smart
technology applications. Enabled by the high-speed, low-latency characteristics
of 5G, these applications have shown significant potential in various sectors,
from healthcare and transportation to energy management and beyond. As a
crucial component of smart technology, IoT systems for service delivery often
face concept drift issues in network data stream analytics due to dynamic IoT
environments, resulting in performance degradation. In this article, we propose
a drift-adaptive framework called Adaptive Exponentially Weighted Average
Ensemble (AEWAE) consisting of three stages: IoT data preprocessing, base model
learning, and online ensembling. It is a data stream analytics framework that
integrates dynamic adjustments of ensemble methods to tackle various scenarios.
Experimental results on two public IoT datasets demonstrate that our proposed
framework outperforms state-of-the-art methods, achieving high accuracy and
efficiency in IoT data stream analytics.
| no_new_dataset | 0.70724 |
2305.07031 | Seung Ji Kook | Minju Jo, Seungji Kook, Noseong Park | Hawkes Process Based on Controlled Differential Equations | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Hawkes processes are a popular framework to model the occurrence of
sequential events, i.e., occurrence dynamics, in several fields such as social
diffusion. In real-world scenarios, the inter-arrival time among events is
irregular. However, existing neural network-based Hawkes process models not
only i) fail to capture such complicated irregular dynamics, but also ii)
resort to heuristics to calculate the log-likelihood of events since they are
mostly based on neural networks designed for regular discrete inputs. To this
end, we present the concept of Hawkes process based on controlled differential
equations (HP-CDE), by adopting the neural controlled differential equation
(neural CDE) technology which is an analogue to continuous RNNs. Since HP-CDE
continuously reads data, i) irregular time-series datasets can be properly
treated preserving their uneven temporal spaces, and ii) the log-likelihood can
be exactly computed. Moreover, as both Hawkes processes and neural CDEs are
first developed to model complicated human behavioral dynamics, neural
CDE-based Hawkes processes are successful in modeling such occurrence dynamics.
In our experiments with 4 real-world datasets, our method outperforms existing
methods by non-trivial margins.
| [
{
"version": "v1",
"created": "Tue, 9 May 2023 07:52:56 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 05:34:39 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Jo",
"Minju",
""
],
[
"Kook",
"Seungji",
""
],
[
"Park",
"Noseong",
""
]
] | TITLE: Hawkes Process Based on Controlled Differential Equations
ABSTRACT: Hawkes processes are a popular framework to model the occurrence of
sequential events, i.e., occurrence dynamics, in several fields such as social
diffusion. In real-world scenarios, the inter-arrival time among events is
irregular. However, existing neural network-based Hawkes process models not
only i) fail to capture such complicated irregular dynamics, but also ii)
resort to heuristics to calculate the log-likelihood of events since they are
mostly based on neural networks designed for regular discrete inputs. To this
end, we present the concept of Hawkes process based on controlled differential
equations (HP-CDE), by adopting the neural controlled differential equation
(neural CDE) technology which is an analogue to continuous RNNs. Since HP-CDE
continuously reads data, i) irregular time-series datasets can be properly
treated preserving their uneven temporal spaces, and ii) the log-likelihood can
be exactly computed. Moreover, as both Hawkes processes and neural CDEs are
first developed to model complicated human behavioral dynamics, neural
CDE-based Hawkes processes are successful in modeling such occurrence dynamics.
In our experiments with 4 real-world datasets, our method outperforms existing
methods by non-trivial margins.
| no_new_dataset | 0.710415 |
2305.07854 | Yan Qin | Anushiya Arunan, Yan Qin, Xiaoli Li, and Chau Yuen | A Federated Learning-based Industrial Health Prognostics for
Heterogeneous Edge Devices using Matched Feature Extraction | 17 pages, 11 figures, and 6 tables | Aeecpted by IEEE TASE 2023 | 10.1109/TASE.2023.3274648 | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data-driven industrial health prognostics require rich training data to
develop accurate and reliable predictive models. However, stringent data
privacy laws and the abundance of edge industrial data necessitate
decentralized data utilization. Thus, the industrial health prognostics field
is well suited to significantly benefit from federated learning (FL), a
decentralized and privacy-preserving learning technique. However, FL-based
health prognostics tasks have hardly been investigated due to the complexities
of meaningfully aggregating model parameters trained from heterogeneous data to
form a high performing federated model. Specifically, data heterogeneity among
edge devices, stemming from dissimilar degradation mechanisms and unequal
dataset sizes, poses a critical statistical challenge for developing accurate
federated models. We propose a pioneering FL-based health prognostic model with
a feature similarity-matched parameter aggregation algorithm to
discriminatingly learn from heterogeneous edge data. The algorithm searches
across the heterogeneous locally trained models and matches neurons with
probabilistically similar feature extraction functions first, before
selectively averaging them to form the federated model parameters. As the
algorithm only averages similar neurons, as opposed to conventional naive
averaging of coordinate-wise neurons, the distinct feature extractors of local
models are carried over with less dilution to the resultant federated model.
Using both cyclic degradation data of Li-ion batteries and non-cyclic data of
turbofan engines, we demonstrate that the proposed method yields accuracy
improvements as high as 44.5\% and 39.3\% for state-of-health estimation and
remaining useful life estimation, respectively.
| [
{
"version": "v1",
"created": "Sat, 13 May 2023 07:20:31 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 06:03:58 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Arunan",
"Anushiya",
""
],
[
"Qin",
"Yan",
""
],
[
"Li",
"Xiaoli",
""
],
[
"Yuen",
"Chau",
""
]
] | TITLE: A Federated Learning-based Industrial Health Prognostics for
Heterogeneous Edge Devices using Matched Feature Extraction
ABSTRACT: Data-driven industrial health prognostics require rich training data to
develop accurate and reliable predictive models. However, stringent data
privacy laws and the abundance of edge industrial data necessitate
decentralized data utilization. Thus, the industrial health prognostics field
is well suited to significantly benefit from federated learning (FL), a
decentralized and privacy-preserving learning technique. However, FL-based
health prognostics tasks have hardly been investigated due to the complexities
of meaningfully aggregating model parameters trained from heterogeneous data to
form a high performing federated model. Specifically, data heterogeneity among
edge devices, stemming from dissimilar degradation mechanisms and unequal
dataset sizes, poses a critical statistical challenge for developing accurate
federated models. We propose a pioneering FL-based health prognostic model with
a feature similarity-matched parameter aggregation algorithm to
discriminatingly learn from heterogeneous edge data. The algorithm searches
across the heterogeneous locally trained models and matches neurons with
probabilistically similar feature extraction functions first, before
selectively averaging them to form the federated model parameters. As the
algorithm only averages similar neurons, as opposed to conventional naive
averaging of coordinate-wise neurons, the distinct feature extractors of local
models are carried over with less dilution to the resultant federated model.
Using both cyclic degradation data of Li-ion batteries and non-cyclic data of
turbofan engines, we demonstrate that the proposed method yields accuracy
improvements as high as 44.5\% and 39.3\% for state-of-health estimation and
remaining useful life estimation, respectively.
| no_new_dataset | 0.71247 |
2305.07969 | Yutian Chen | Yutian Chen, Hao Kang, Vivian Zhai, Liangze Li, Rita Singh, Bhiksha
Raj | GPT-Sentinel: Distinguishing Human and ChatGPT Generated Content | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a novel approach for detecting ChatGPT-generated vs.
human-written text using language models. To this end, we first collected and
released a pre-processed dataset named OpenGPTText, which consists of rephrased
content generated using ChatGPT. We then designed, implemented, and trained two
different models for text classification, using Robustly Optimized BERT
Pretraining Approach (RoBERTa) and Text-to-Text Transfer Transformer (T5),
respectively. Our models achieved remarkable results, with an accuracy of over
97% on the test dataset, as evaluated through various metrics. Furthermore, we
conducted an interpretability study to showcase our model's ability to extract
and differentiate key features between human-written and ChatGPT-generated
text. Our findings provide important insights into the effective use of
language models to detect generated text.
| [
{
"version": "v1",
"created": "Sat, 13 May 2023 17:12:11 GMT"
},
{
"version": "v2",
"created": "Wed, 17 May 2023 18:21:03 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Chen",
"Yutian",
""
],
[
"Kang",
"Hao",
""
],
[
"Zhai",
"Vivian",
""
],
[
"Li",
"Liangze",
""
],
[
"Singh",
"Rita",
""
],
[
"Raj",
"Bhiksha",
""
]
] | TITLE: GPT-Sentinel: Distinguishing Human and ChatGPT Generated Content
ABSTRACT: This paper presents a novel approach for detecting ChatGPT-generated vs.
human-written text using language models. To this end, we first collected and
released a pre-processed dataset named OpenGPTText, which consists of rephrased
content generated using ChatGPT. We then designed, implemented, and trained two
different models for text classification, using Robustly Optimized BERT
Pretraining Approach (RoBERTa) and Text-to-Text Transfer Transformer (T5),
respectively. Our models achieved remarkable results, with an accuracy of over
97% on the test dataset, as evaluated through various metrics. Furthermore, we
conducted an interpretability study to showcase our model's ability to extract
and differentiate key features between human-written and ChatGPT-generated
text. Our findings provide important insights into the effective use of
language models to detect generated text.
| new_dataset | 0.713619 |
2305.09041 | Antoine Theberge | Antoine Th\'eberge, Christian Desrosiers, Maxime Descoteaux,
Pierre-Marc Jodoin | What Matters in Reinforcement Learning for Tractography | null | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recently, deep reinforcement learning (RL) has been proposed to learn the
tractography procedure and train agents to reconstruct the structure of the
white matter without manually curated reference streamlines. While the
performances reported were competitive, the proposed framework is complex, and
little is still known about the role and impact of its multiple parts. In this
work, we thoroughly explore the different components of the proposed framework,
such as the choice of the RL algorithm, seeding strategy, the input signal and
reward function, and shed light on their impact. Approximately 7,400 models
were trained for this work, totalling nearly 41,000 hours of GPU time. Our goal
is to guide researchers eager to explore the possibilities of deep RL for
tractography by exposing what works and what does not work with the category of
approach. As such, we ultimately propose a series of recommendations concerning
the choice of RL algorithm, the input to the agents, the reward function and
more to help future work using reinforcement learning for tractography. We also
release the open source codebase, trained models, and datasets for users and
researchers wanting to explore reinforcement learning for tractography.
| [
{
"version": "v1",
"created": "Mon, 15 May 2023 22:01:48 GMT"
},
{
"version": "v2",
"created": "Wed, 17 May 2023 19:30:27 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Théberge",
"Antoine",
""
],
[
"Desrosiers",
"Christian",
""
],
[
"Descoteaux",
"Maxime",
""
],
[
"Jodoin",
"Pierre-Marc",
""
]
] | TITLE: What Matters in Reinforcement Learning for Tractography
ABSTRACT: Recently, deep reinforcement learning (RL) has been proposed to learn the
tractography procedure and train agents to reconstruct the structure of the
white matter without manually curated reference streamlines. While the
performances reported were competitive, the proposed framework is complex, and
little is still known about the role and impact of its multiple parts. In this
work, we thoroughly explore the different components of the proposed framework,
such as the choice of the RL algorithm, seeding strategy, the input signal and
reward function, and shed light on their impact. Approximately 7,400 models
were trained for this work, totalling nearly 41,000 hours of GPU time. Our goal
is to guide researchers eager to explore the possibilities of deep RL for
tractography by exposing what works and what does not work with the category of
approach. As such, we ultimately propose a series of recommendations concerning
the choice of RL algorithm, the input to the agents, the reward function and
more to help future work using reinforcement learning for tractography. We also
release the open source codebase, trained models, and datasets for users and
researchers wanting to explore reinforcement learning for tractography.
| no_new_dataset | 0.703549 |
2305.09134 | Aditya Kalapaaking | Aditya Pribadi Kalapaaking, Ibrahim Khalil, and Mohammed Atiquzzaman | Smart Policy Control for Securing Federated Learning Management System | null | IEEE Transactions on Network and Service Management, 2023 | 10.1109/TNSM.2023.3276594 | null | cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The widespread adoption of Internet of Things (IoT) devices in smart cities,
intelligent healthcare systems, and various real-world applications have
resulted in the generation of vast amounts of data, often analyzed using
different Machine Learning (ML) models. Federated learning (FL) has been
acknowledged as a privacy-preserving machine learning technology, where
multiple parties cooperatively train ML models without exchanging raw data.
However, the current FL architecture does not allow for an audit of the
training process due to the various data-protection policies implemented by
each FL participant. Furthermore, there is no global model verifiability
available in the current architecture. This paper proposes a smart
contract-based policy control for securing the Federated Learning (FL)
management system. First, we develop and deploy a smart contract-based local
training policy control on the FL participants' side. This policy control is
used to verify the training process, ensuring that the evaluation process
follows the same rules for all FL participants. We then enforce a smart
contract-based aggregation policy to manage the global model aggregation
process. Upon completion, the aggregated model and policy are stored on
blockchain-based storage. Subsequently, we distribute the aggregated global
model and the smart contract to all FL participants. Our proposed method uses
smart policy control to manage access and verify the integrity of machine
learning models. We conducted multiple experiments with various machine
learning architectures and datasets to evaluate our proposed framework, such as
MNIST and CIFAR-10.
| [
{
"version": "v1",
"created": "Tue, 16 May 2023 03:34:49 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 11:54:47 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Kalapaaking",
"Aditya Pribadi",
""
],
[
"Khalil",
"Ibrahim",
""
],
[
"Atiquzzaman",
"Mohammed",
""
]
] | TITLE: Smart Policy Control for Securing Federated Learning Management System
ABSTRACT: The widespread adoption of Internet of Things (IoT) devices in smart cities,
intelligent healthcare systems, and various real-world applications have
resulted in the generation of vast amounts of data, often analyzed using
different Machine Learning (ML) models. Federated learning (FL) has been
acknowledged as a privacy-preserving machine learning technology, where
multiple parties cooperatively train ML models without exchanging raw data.
However, the current FL architecture does not allow for an audit of the
training process due to the various data-protection policies implemented by
each FL participant. Furthermore, there is no global model verifiability
available in the current architecture. This paper proposes a smart
contract-based policy control for securing the Federated Learning (FL)
management system. First, we develop and deploy a smart contract-based local
training policy control on the FL participants' side. This policy control is
used to verify the training process, ensuring that the evaluation process
follows the same rules for all FL participants. We then enforce a smart
contract-based aggregation policy to manage the global model aggregation
process. Upon completion, the aggregated model and policy are stored on
blockchain-based storage. Subsequently, we distribute the aggregated global
model and the smart contract to all FL participants. Our proposed method uses
smart policy control to manage access and verify the integrity of machine
learning models. We conducted multiple experiments with various machine
learning architectures and datasets to evaluate our proposed framework, such as
MNIST and CIFAR-10.
| no_new_dataset | 0.710666 |
2305.09251 | Arsenia (Ersi) Chorti | Amitha Mayya, Miroslav Mitev, Arsenia Chorti, Gerhard Fettweis | A SKG Security Challenge: Indoor SKG Under an On-The-Shoulder
Eavesdropping Attack | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | Physical layer security (PLS) is seen as the means to enhance physical layer
trustworthiness in 6G. This work provides a proof-of-concept for one of the
most mature PLS technologies, i.e., secret key generation (SKG) from wireless
fading coefficients during the channel's coherence time. As opposed to other
works, where only specific parts of the protocol are typically investigated,
here, we implement the full SKG chain in four indoor experimental campaigns. In
detail, we consider two legitimate nodes, who use the wireless channel to
extract secret keys and a malicious node placed in the immediate vicinity of
one of them, who acts as a passive eavesdropper. To estimate the final SKG rate
we evaluate the conditional min-entropy by taking into account all information
available at the eavesdropper. Finally, we use this paper to announce the first
ever physical layer security challenge, mirroring practices in cryptography. We
call the community to scrutinize the presented results and try to ``break'' our
SKG implementation. To this end, we provide, i) the full dataset observed by
the eavesdroppers, ii) $20$ blocks of $16-$byte long ciphertexts, encrypted
using one-time pad with $20$ distilled secret keys, and, iii) all codes and
software used in our SKG implementation. An attack will be considered
successful if any part(s) of the plaintext are successfully retrieved.
| [
{
"version": "v1",
"created": "Tue, 16 May 2023 07:59:21 GMT"
},
{
"version": "v2",
"created": "Wed, 17 May 2023 13:11:14 GMT"
},
{
"version": "v3",
"created": "Thu, 18 May 2023 11:14:51 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Mayya",
"Amitha",
""
],
[
"Mitev",
"Miroslav",
""
],
[
"Chorti",
"Arsenia",
""
],
[
"Fettweis",
"Gerhard",
""
]
] | TITLE: A SKG Security Challenge: Indoor SKG Under an On-The-Shoulder
Eavesdropping Attack
ABSTRACT: Physical layer security (PLS) is seen as the means to enhance physical layer
trustworthiness in 6G. This work provides a proof-of-concept for one of the
most mature PLS technologies, i.e., secret key generation (SKG) from wireless
fading coefficients during the channel's coherence time. As opposed to other
works, where only specific parts of the protocol are typically investigated,
here, we implement the full SKG chain in four indoor experimental campaigns. In
detail, we consider two legitimate nodes, who use the wireless channel to
extract secret keys and a malicious node placed in the immediate vicinity of
one of them, who acts as a passive eavesdropper. To estimate the final SKG rate
we evaluate the conditional min-entropy by taking into account all information
available at the eavesdropper. Finally, we use this paper to announce the first
ever physical layer security challenge, mirroring practices in cryptography. We
call the community to scrutinize the presented results and try to ``break'' our
SKG implementation. To this end, we provide, i) the full dataset observed by
the eavesdroppers, ii) $20$ blocks of $16-$byte long ciphertexts, encrypted
using one-time pad with $20$ distilled secret keys, and, iii) all codes and
software used in our SKG implementation. An attack will be considered
successful if any part(s) of the plaintext are successfully retrieved.
| no_new_dataset | 0.698818 |
2305.10049 | Hao Li | Hao Li, Peng Jin, Zesen Cheng, Songyang Zhang, Kai Chen, Zhennan Wang,
Chang Liu, Jie Chen | TG-VQA: Ternary Game of Video Question Answering | IJCAI 2023 | null | null | null | cs.CV cs.MM | http://creativecommons.org/licenses/by/4.0/ | Video question answering aims at answering a question about the video content
by reasoning the alignment semantics within them. However, since relying
heavily on human instructions, i.e., annotations or priors, current contrastive
learning-based VideoQA methods remains challenging to perform fine-grained
visual-linguistic alignments. In this work, we innovatively resort to game
theory, which can simulate complicated relationships among multiple players
with specific interaction strategies, e.g., video, question, and answer as
ternary players, to achieve fine-grained alignment for VideoQA task.
Specifically, we carefully design a VideoQA-specific interaction strategy to
tailor the characteristics of VideoQA, which can mathematically generate the
fine-grained visual-linguistic alignment label without label-intensive efforts.
Our TG-VQA outperforms existing state-of-the-art by a large margin (more than
5%) on long-term and short-term VideoQA datasets, verifying its effectiveness
and generalization ability. Thanks to the guidance of game-theoretic
interaction, our model impressively convergences well on limited data (${10}^4
~videos$), surpassing most of those pre-trained on large-scale data
($10^7~videos$).
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 08:42:53 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 06:27:06 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Hao",
""
],
[
"Jin",
"Peng",
""
],
[
"Cheng",
"Zesen",
""
],
[
"Zhang",
"Songyang",
""
],
[
"Chen",
"Kai",
""
],
[
"Wang",
"Zhennan",
""
],
[
"Liu",
"Chang",
""
],
[
"Chen",
"Jie",
""
]
] | TITLE: TG-VQA: Ternary Game of Video Question Answering
ABSTRACT: Video question answering aims at answering a question about the video content
by reasoning the alignment semantics within them. However, since relying
heavily on human instructions, i.e., annotations or priors, current contrastive
learning-based VideoQA methods remains challenging to perform fine-grained
visual-linguistic alignments. In this work, we innovatively resort to game
theory, which can simulate complicated relationships among multiple players
with specific interaction strategies, e.g., video, question, and answer as
ternary players, to achieve fine-grained alignment for VideoQA task.
Specifically, we carefully design a VideoQA-specific interaction strategy to
tailor the characteristics of VideoQA, which can mathematically generate the
fine-grained visual-linguistic alignment label without label-intensive efforts.
Our TG-VQA outperforms existing state-of-the-art by a large margin (more than
5%) on long-term and short-term VideoQA datasets, verifying its effectiveness
and generalization ability. Thanks to the guidance of game-theoretic
interaction, our model impressively convergences well on limited data (${10}^4
~videos$), surpassing most of those pre-trained on large-scale data
($10^7~videos$).
| no_new_dataset | 0.709573 |
2305.10169 | Xiaocui Yang | Xiaocui Yang, Shi Feng, Daling Wang, Sun Qi, Wenfang Wu, Yifei Zhang,
Pengfei Hong, Soujanya Poria | Few-shot Joint Multimodal Aspect-Sentiment Analysis Based on Generative
Multimodal Prompt | 13 pages, 7 figures, 6 tables, ACL 2023 Long Paper (Findings) | null | null | null | cs.MM | http://creativecommons.org/licenses/by/4.0/ | We have witnessed the rapid proliferation of multimodal data on numerous
social media platforms. Conventional studies typically require massive labeled
data to train models for Multimodal Aspect-Based Sentiment Analysis (MABSA).
However, collecting and annotating fine-grained multimodal data for MABSA is
tough. To alleviate the above issue, we perform three MABSA-related tasks with
quite a small number of labeled multimodal samples. We first build diverse and
comprehensive multimodal few-shot datasets according to the data distribution.
To capture the specific prompt for each aspect term in a few-shot scenario, we
propose a novel Generative Multimodal Prompt (GMP) model for MABSA, which
includes the Multimodal Encoder module and the N-Stream Decoders module. We
further introduce a subtask to predict the number of aspect terms in each
instance to construct the multimodal prompt. Extensive experiments on two
datasets demonstrate that our approach outperforms strong baselines on two
MABSA-related tasks in the few-shot setting.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 12:46:37 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 13:28:22 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yang",
"Xiaocui",
""
],
[
"Feng",
"Shi",
""
],
[
"Wang",
"Daling",
""
],
[
"Qi",
"Sun",
""
],
[
"Wu",
"Wenfang",
""
],
[
"Zhang",
"Yifei",
""
],
[
"Hong",
"Pengfei",
""
],
[
"Poria",
"Soujanya",
""
]
] | TITLE: Few-shot Joint Multimodal Aspect-Sentiment Analysis Based on Generative
Multimodal Prompt
ABSTRACT: We have witnessed the rapid proliferation of multimodal data on numerous
social media platforms. Conventional studies typically require massive labeled
data to train models for Multimodal Aspect-Based Sentiment Analysis (MABSA).
However, collecting and annotating fine-grained multimodal data for MABSA is
tough. To alleviate the above issue, we perform three MABSA-related tasks with
quite a small number of labeled multimodal samples. We first build diverse and
comprehensive multimodal few-shot datasets according to the data distribution.
To capture the specific prompt for each aspect term in a few-shot scenario, we
propose a novel Generative Multimodal Prompt (GMP) model for MABSA, which
includes the Multimodal Encoder module and the N-Stream Decoders module. We
further introduce a subtask to predict the number of aspect terms in each
instance to construct the multimodal prompt. Extensive experiments on two
datasets demonstrate that our approach outperforms strong baselines on two
MABSA-related tasks in the few-shot setting.
| no_new_dataset | 0.702352 |
2305.10198 | Hanxiao Liu | Chenyang Shi, Hanxiao Liu, Jing Jin, Wenzhuo Li, Yuzhen Li, Boyi Wei,
Yibo Zhang | IDO-VFI: Identifying Dynamics via Optical Flow Guidance for Video Frame
Interpolation with Events | null | null | null | null | cs.CV eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Video frame interpolation aims to generate high-quality intermediate frames
from boundary frames and increase frame rate. While existing linear, symmetric
and nonlinear models are used to bridge the gap from the lack of inter-frame
motion, they cannot reconstruct real motions. Event cameras, however, are ideal
for capturing inter-frame dynamics with their extremely high temporal
resolution. In this paper, we propose an event-and-frame-based video frame
interpolation method named IDO-VFI that assigns varying amounts of computation
for different sub-regions via optical flow guidance. The proposed method first
estimates the optical flow based on frames and events, and then decides whether
to further calculate the residual optical flow in those sub-regions via a
Gumbel gating module according to the optical flow amplitude. Intermediate
frames are eventually generated through a concise Transformer-based fusion
network. Our proposed method maintains high-quality performance while reducing
computation time and computational effort by 10% and 17% respectively on
Vimeo90K datasets, compared with a unified process on the whole region.
Moreover, our method outperforms state-of-the-art frame-only and
frames-plus-events methods on multiple video frame interpolation benchmarks.
Codes and models are available at https://github.com/shicy17/IDO-VFI.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 13:22:21 GMT"
},
{
"version": "v2",
"created": "Thu, 18 May 2023 07:06:39 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Shi",
"Chenyang",
""
],
[
"Liu",
"Hanxiao",
""
],
[
"Jin",
"Jing",
""
],
[
"Li",
"Wenzhuo",
""
],
[
"Li",
"Yuzhen",
""
],
[
"Wei",
"Boyi",
""
],
[
"Zhang",
"Yibo",
""
]
] | TITLE: IDO-VFI: Identifying Dynamics via Optical Flow Guidance for Video Frame
Interpolation with Events
ABSTRACT: Video frame interpolation aims to generate high-quality intermediate frames
from boundary frames and increase frame rate. While existing linear, symmetric
and nonlinear models are used to bridge the gap from the lack of inter-frame
motion, they cannot reconstruct real motions. Event cameras, however, are ideal
for capturing inter-frame dynamics with their extremely high temporal
resolution. In this paper, we propose an event-and-frame-based video frame
interpolation method named IDO-VFI that assigns varying amounts of computation
for different sub-regions via optical flow guidance. The proposed method first
estimates the optical flow based on frames and events, and then decides whether
to further calculate the residual optical flow in those sub-regions via a
Gumbel gating module according to the optical flow amplitude. Intermediate
frames are eventually generated through a concise Transformer-based fusion
network. Our proposed method maintains high-quality performance while reducing
computation time and computational effort by 10% and 17% respectively on
Vimeo90K datasets, compared with a unified process on the whole region.
Moreover, our method outperforms state-of-the-art frame-only and
frames-plus-events methods on multiple video frame interpolation benchmarks.
Codes and models are available at https://github.com/shicy17/IDO-VFI.
| no_new_dataset | 0.710019 |
2305.10432 | Haotian Chen | Chang'an Yi, Haotian Chen, Yonghui Xu, Yifan Zhang | Model-Contrastive Federated Domain Adaptation | 13 pages | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated domain adaptation (FDA) aims to collaboratively transfer knowledge
from source clients (domains) to the related but different target client,
without communicating the local data of any client. Moreover, the source
clients have different data distributions, leading to extremely challenging in
knowledge transfer. Despite the recent progress in FDA, we empirically find
that existing methods can not leverage models of heterogeneous domains and thus
they fail to achieve excellent performance. In this paper, we propose a
model-based method named FDAC, aiming to address {\bf F}ederated {\bf D}omain
{\bf A}daptation based on {\bf C}ontrastive learning and Vision Transformer
(ViT). In particular, contrastive learning can leverage the unlabeled data to
train excellent models and the ViT architecture performs better than
convolutional neural networks (CNNs) in extracting adaptable features. To the
best of our knowledge, FDAC is the first attempt to learn transferable
representations by manipulating the latent architecture of ViT under the
federated setting. Furthermore, FDAC can increase the target data diversity by
compensating from each source model with insufficient knowledge of samples and
features, based on domain augmentation and semantic matching. Extensive
experiments on several real datasets demonstrate that FDAC outperforms all the
comparative methods in most conditions. Moreover, FDCA can also improve
communication efficiency which is another key factor in the federated setting.
| [
{
"version": "v1",
"created": "Sun, 7 May 2023 23:48:03 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yi",
"Chang'an",
""
],
[
"Chen",
"Haotian",
""
],
[
"Xu",
"Yonghui",
""
],
[
"Zhang",
"Yifan",
""
]
] | TITLE: Model-Contrastive Federated Domain Adaptation
ABSTRACT: Federated domain adaptation (FDA) aims to collaboratively transfer knowledge
from source clients (domains) to the related but different target client,
without communicating the local data of any client. Moreover, the source
clients have different data distributions, leading to extremely challenging in
knowledge transfer. Despite the recent progress in FDA, we empirically find
that existing methods can not leverage models of heterogeneous domains and thus
they fail to achieve excellent performance. In this paper, we propose a
model-based method named FDAC, aiming to address {\bf F}ederated {\bf D}omain
{\bf A}daptation based on {\bf C}ontrastive learning and Vision Transformer
(ViT). In particular, contrastive learning can leverage the unlabeled data to
train excellent models and the ViT architecture performs better than
convolutional neural networks (CNNs) in extracting adaptable features. To the
best of our knowledge, FDAC is the first attempt to learn transferable
representations by manipulating the latent architecture of ViT under the
federated setting. Furthermore, FDAC can increase the target data diversity by
compensating from each source model with insufficient knowledge of samples and
features, based on domain augmentation and semantic matching. Extensive
experiments on several real datasets demonstrate that FDAC outperforms all the
comparative methods in most conditions. Moreover, FDCA can also improve
communication efficiency which is another key factor in the federated setting.
| no_new_dataset | 0.708168 |
2305.10433 | Abeer AlDayel | Huriyyah Althunayan, Rahaf Bahlas, Manar Alharbi, Lena Alsuwailem,
Abeer Aldayel, Rehab ALahmadi | Toxicity Inspector: A Framework to Evaluate Ground Truth in Toxicity
Detection Through Feedback | To appear in Workshop on 2nd Workshop on Novel Evaluation Approaches
for Text Classification Systems (NEATCLasS-2023).ICWSM, AAAI, 2023 | null | null | null | cs.CL cs.SI | http://creativecommons.org/licenses/by/4.0/ | Toxic language is difficult to define, as it is not monolithic and has many
variations in perceptions of toxicity. This challenge of detecting toxic
language is increased by the highly contextual and subjectivity of its
interpretation, which can degrade the reliability of datasets and negatively
affect detection model performance. To fill this void, this paper introduces a
toxicity inspector framework that incorporates a human-in-the-loop pipeline
with the aim of enhancing the reliability of toxicity benchmark datasets by
centering the evaluator's values through an iterative feedback cycle. The
centerpiece of this framework is the iterative feedback process, which is
guided by two metric types (hard and soft) that provide evaluators and dataset
creators with insightful examination to balance the tradeoff between
performance gains and toxicity avoidance.
| [
{
"version": "v1",
"created": "Thu, 11 May 2023 11:56:42 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Althunayan",
"Huriyyah",
""
],
[
"Bahlas",
"Rahaf",
""
],
[
"Alharbi",
"Manar",
""
],
[
"Alsuwailem",
"Lena",
""
],
[
"Aldayel",
"Abeer",
""
],
[
"ALahmadi",
"Rehab",
""
]
] | TITLE: Toxicity Inspector: A Framework to Evaluate Ground Truth in Toxicity
Detection Through Feedback
ABSTRACT: Toxic language is difficult to define, as it is not monolithic and has many
variations in perceptions of toxicity. This challenge of detecting toxic
language is increased by the highly contextual and subjectivity of its
interpretation, which can degrade the reliability of datasets and negatively
affect detection model performance. To fill this void, this paper introduces a
toxicity inspector framework that incorporates a human-in-the-loop pipeline
with the aim of enhancing the reliability of toxicity benchmark datasets by
centering the evaluator's values through an iterative feedback cycle. The
centerpiece of this framework is the iterative feedback process, which is
guided by two metric types (hard and soft) that provide evaluators and dataset
creators with insightful examination to balance the tradeoff between
performance gains and toxicity avoidance.
| no_new_dataset | 0.711481 |
2305.10442 | Abhinav Sagar | Abhinav Sagar, Sai Teja Gilukara | CBAGAN-RRT: Convolutional Block Attention Generative Adversarial Network
for Sampling-Based Path Planning | 11 pages, 6 figures, 8 tables | null | null | null | cs.RO cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Sampling-based path planning algorithms play an important role in autonomous
robotics. However, a common problem among the RRT-based algorithms is that the
initial path generated is not optimal and the convergence is too slow to be
used in real-world applications. In this paper, we propose a novel image-based
learning algorithm (CBAGAN-RRT) using a Convolutional Block Attention
Generative Adversarial Network with a combination of spatial and channel
attention and a novel loss function to design the heuristics, find a better
optimal path, and improve the convergence of the algorithm both concerning time
and speed. The probability distribution of the paths generated from our GAN
model is used to guide the sampling process for the RRT algorithm. We train and
test our network on the dataset generated by \cite{zhang2021generative} and
demonstrate that our algorithm outperforms the previous state-of-the-art
algorithms using both the image quality generation metrics like IOU Score, Dice
Score, FID score, and path planning metrics like time cost and the number of
nodes. We conduct detailed experiments and ablation studies to illustrate the
feasibility of our study and show that our model performs well not only on the
training dataset but also on the unseen test dataset. The advantage of our
approach is that we can avoid the complicated preprocessing in the state space,
our model can be generalized to complicated environments like those containing
turns and narrow passages without loss of accuracy, and our model can be easily
integrated with other sampling-based path planning algorithms.
| [
{
"version": "v1",
"created": "Sat, 13 May 2023 20:06:53 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Sagar",
"Abhinav",
""
],
[
"Gilukara",
"Sai Teja",
""
]
] | TITLE: CBAGAN-RRT: Convolutional Block Attention Generative Adversarial Network
for Sampling-Based Path Planning
ABSTRACT: Sampling-based path planning algorithms play an important role in autonomous
robotics. However, a common problem among the RRT-based algorithms is that the
initial path generated is not optimal and the convergence is too slow to be
used in real-world applications. In this paper, we propose a novel image-based
learning algorithm (CBAGAN-RRT) using a Convolutional Block Attention
Generative Adversarial Network with a combination of spatial and channel
attention and a novel loss function to design the heuristics, find a better
optimal path, and improve the convergence of the algorithm both concerning time
and speed. The probability distribution of the paths generated from our GAN
model is used to guide the sampling process for the RRT algorithm. We train and
test our network on the dataset generated by \cite{zhang2021generative} and
demonstrate that our algorithm outperforms the previous state-of-the-art
algorithms using both the image quality generation metrics like IOU Score, Dice
Score, FID score, and path planning metrics like time cost and the number of
nodes. We conduct detailed experiments and ablation studies to illustrate the
feasibility of our study and show that our model performs well not only on the
training dataset but also on the unseen test dataset. The advantage of our
approach is that we can avoid the complicated preprocessing in the state space,
our model can be generalized to complicated environments like those containing
turns and narrow passages without loss of accuracy, and our model can be easily
integrated with other sampling-based path planning algorithms.
| no_new_dataset | 0.710867 |
2305.10447 | Oscar Morris | Oscar Morris | The Effectiveness of a Dynamic Loss Function in Neural Network Based
Automated Essay Scoring | 7 pages, 5 figures, 2 tables | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | Neural networks and in particular the attention mechanism have brought
significant advances to the field of Automated Essay Scoring. Many of these
systems use a regression-based model which may be prone to underfitting when
the model only predicts the mean of the training data. In this paper, we
present a dynamic loss function that creates an incentive for the model to
predict with the correct distribution, as well as predicting the correct
values. Our loss function achieves this goal without sacrificing any
performance achieving a Quadratic Weighted Kappa score of 0.752 on the
Automated Student Assessment Prize Automated Essay Scoring dataset.
| [
{
"version": "v1",
"created": "Mon, 15 May 2023 16:39:35 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Morris",
"Oscar",
""
]
] | TITLE: The Effectiveness of a Dynamic Loss Function in Neural Network Based
Automated Essay Scoring
ABSTRACT: Neural networks and in particular the attention mechanism have brought
significant advances to the field of Automated Essay Scoring. Many of these
systems use a regression-based model which may be prone to underfitting when
the model only predicts the mean of the training data. In this paper, we
present a dynamic loss function that creates an incentive for the model to
predict with the correct distribution, as well as predicting the correct
values. Our loss function achieves this goal without sacrificing any
performance achieving a Quadratic Weighted Kappa score of 0.752 on the
Automated Student Assessment Prize Automated Essay Scoring dataset.
| no_new_dataset | 0.711281 |
2305.10452 | Mario Graff | Sergio Nava-Mu\~noz and Mario Graff Guerrero and Hugo Jair Escalante | Comparison of classifiers in challenge scheme | null | null | null | null | cs.LG cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent decades, challenges have become very popular in scientific research
as these are crowdsourcing schemes. In particular, challenges are essential for
developing machine learning algorithms. For the challenges settings, it is
vital to establish the scientific question, the dataset (with adequate quality,
quantity, diversity, and complexity), performance metrics, as well as a way to
authenticate the participants' results (Gold Standard). This paper addresses
the problem of evaluating the performance of different competitors (algorithms)
under the restrictions imposed by the challenge scheme, such as the comparison
of multiple competitors with a unique dataset (with fixed size), a minimal
number of submissions and, a set of metrics chosen to assess performance. The
algorithms are sorted according to the performance metric. Still, it is common
to observe performance differences among competitors as small as hundredths or
even thousandths, so the question is whether the differences are significant.
This paper analyzes the results of the MeOffendEs@IberLEF 2021 competition and
proposes to make inference through resampling techniques (bootstrap) to support
Challenge organizers' decision-making.
| [
{
"version": "v1",
"created": "Tue, 16 May 2023 23:38:34 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Nava-Muñoz",
"Sergio",
""
],
[
"Guerrero",
"Mario Graff",
""
],
[
"Escalante",
"Hugo Jair",
""
]
] | TITLE: Comparison of classifiers in challenge scheme
ABSTRACT: In recent decades, challenges have become very popular in scientific research
as these are crowdsourcing schemes. In particular, challenges are essential for
developing machine learning algorithms. For the challenges settings, it is
vital to establish the scientific question, the dataset (with adequate quality,
quantity, diversity, and complexity), performance metrics, as well as a way to
authenticate the participants' results (Gold Standard). This paper addresses
the problem of evaluating the performance of different competitors (algorithms)
under the restrictions imposed by the challenge scheme, such as the comparison
of multiple competitors with a unique dataset (with fixed size), a minimal
number of submissions and, a set of metrics chosen to assess performance. The
algorithms are sorted according to the performance metric. Still, it is common
to observe performance differences among competitors as small as hundredths or
even thousandths, so the question is whether the differences are significant.
This paper analyzes the results of the MeOffendEs@IberLEF 2021 competition and
proposes to make inference through resampling techniques (bootstrap) to support
Challenge organizers' decision-making.
| no_new_dataset | 0.696152 |
2305.10507 | Hao Shao | Hao Shao, Letian Wang, Ruobing Chen, Steven L. Waslander, Hongsheng
Li, Yu Liu | ReasonNet: End-to-End Driving with Temporal and Global Reasoning | CVPR 2023 | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | The large-scale deployment of autonomous vehicles is yet to come, and one of
the major remaining challenges lies in urban dense traffic scenarios. In such
cases, it remains challenging to predict the future evolution of the scene and
future behaviors of objects, and to deal with rare adverse events such as the
sudden appearance of occluded objects. In this paper, we present ReasonNet, a
novel end-to-end driving framework that extensively exploits both temporal and
global information of the driving scene. By reasoning on the temporal behavior
of objects, our method can effectively process the interactions and
relationships among features in different frames. Reasoning about the global
information of the scene can also improve overall perception performance and
benefit the detection of adverse events, especially the anticipation of
potential danger from occluded objects. For comprehensive evaluation on
occlusion events, we also release publicly a driving simulation benchmark
DriveOcclusionSim consisting of diverse occlusion events. We conduct extensive
experiments on multiple CARLA benchmarks, where our model outperforms all prior
methods, ranking first on the sensor track of the public CARLA Leaderboard.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 18:24:43 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Shao",
"Hao",
""
],
[
"Wang",
"Letian",
""
],
[
"Chen",
"Ruobing",
""
],
[
"Waslander",
"Steven L.",
""
],
[
"Li",
"Hongsheng",
""
],
[
"Liu",
"Yu",
""
]
] | TITLE: ReasonNet: End-to-End Driving with Temporal and Global Reasoning
ABSTRACT: The large-scale deployment of autonomous vehicles is yet to come, and one of
the major remaining challenges lies in urban dense traffic scenarios. In such
cases, it remains challenging to predict the future evolution of the scene and
future behaviors of objects, and to deal with rare adverse events such as the
sudden appearance of occluded objects. In this paper, we present ReasonNet, a
novel end-to-end driving framework that extensively exploits both temporal and
global information of the driving scene. By reasoning on the temporal behavior
of objects, our method can effectively process the interactions and
relationships among features in different frames. Reasoning about the global
information of the scene can also improve overall perception performance and
benefit the detection of adverse events, especially the anticipation of
potential danger from occluded objects. For comprehensive evaluation on
occlusion events, we also release publicly a driving simulation benchmark
DriveOcclusionSim consisting of diverse occlusion events. We conduct extensive
experiments on multiple CARLA benchmarks, where our model outperforms all prior
methods, ranking first on the sensor track of the public CARLA Leaderboard.
| new_dataset | 0.714647 |
2305.10527 | Yiguang Zhang | Yiguang Zhang, Kristen Altenburger, Poppy Zhang, Tsutomu Okano, and
Shawndra Hill | Node Attribute Prediction on Multilayer Networks with Weighted and
Directed Edges | null | null | null | null | cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid development of digital platforms, users can now interact in
endless ways from writing business reviews and comments to sharing information
with their friends and followers. As a result, organizations have numerous
digital social networks available for graph learning problems with little
guidance on how to select the right graph or how to combine multiple edge
types. In this paper, we first describe the types of user-to-user networks
available across the Facebook (FB) and Instagram (IG) platforms. We observe
minimal edge overlap between these networks, indicating users are exhibiting
different behaviors and interaction patterns between platforms. We then compare
predictive performance metrics across various node attribute prediction tasks
for an ads click prediction task on Facebook and for a publicly available
dataset from the Open Graph Benchmark. We adapt an existing node attribute
prediction method for binary prediction, LINK-Naive Bayes, to account for both
edge direction and weights on single-layer networks. We observe meaningful
predictive performance gains when incorporating edge direction and weight. We
then introduce an approach called MultiLayerLINK-NaiveBayes that can combine
multiple network layers during training and observe superior performance over
the single-layer results. Ultimately, whether edge direction, edge weights, and
multi-layers are practically useful will depend on the particular setting. Our
approach enables practitioners to quickly combine multiple layers and
additional edge information such as direction or weight.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 19:22:23 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Zhang",
"Yiguang",
""
],
[
"Altenburger",
"Kristen",
""
],
[
"Zhang",
"Poppy",
""
],
[
"Okano",
"Tsutomu",
""
],
[
"Hill",
"Shawndra",
""
]
] | TITLE: Node Attribute Prediction on Multilayer Networks with Weighted and
Directed Edges
ABSTRACT: With the rapid development of digital platforms, users can now interact in
endless ways from writing business reviews and comments to sharing information
with their friends and followers. As a result, organizations have numerous
digital social networks available for graph learning problems with little
guidance on how to select the right graph or how to combine multiple edge
types. In this paper, we first describe the types of user-to-user networks
available across the Facebook (FB) and Instagram (IG) platforms. We observe
minimal edge overlap between these networks, indicating users are exhibiting
different behaviors and interaction patterns between platforms. We then compare
predictive performance metrics across various node attribute prediction tasks
for an ads click prediction task on Facebook and for a publicly available
dataset from the Open Graph Benchmark. We adapt an existing node attribute
prediction method for binary prediction, LINK-Naive Bayes, to account for both
edge direction and weights on single-layer networks. We observe meaningful
predictive performance gains when incorporating edge direction and weight. We
then introduce an approach called MultiLayerLINK-NaiveBayes that can combine
multiple network layers during training and observe superior performance over
the single-layer results. Ultimately, whether edge direction, edge weights, and
multi-layers are practically useful will depend on the particular setting. Our
approach enables practitioners to quickly combine multiple layers and
additional edge information such as direction or weight.
| no_new_dataset | 0.710867 |
2305.10550 | Chanwoo Chun | Chanwoo Chun, Daniel D. Lee | Sparsity-depth Tradeoff in Infinitely Wide Deep Neural Networks | null | null | null | null | cs.LG cond-mat.dis-nn q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate how sparse neural activity affects the generalization
performance of a deep Bayesian neural network at the large width limit. To this
end, we derive a neural network Gaussian Process (NNGP) kernel with rectified
linear unit (ReLU) activation and a predetermined fraction of active neurons.
Using the NNGP kernel, we observe that the sparser networks outperform the
non-sparse networks at shallow depths on a variety of datasets. We validate
this observation by extending the existing theory on the generalization error
of kernel-ridge regression.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 20:09:35 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Chun",
"Chanwoo",
""
],
[
"Lee",
"Daniel D.",
""
]
] | TITLE: Sparsity-depth Tradeoff in Infinitely Wide Deep Neural Networks
ABSTRACT: We investigate how sparse neural activity affects the generalization
performance of a deep Bayesian neural network at the large width limit. To this
end, we derive a neural network Gaussian Process (NNGP) kernel with rectified
linear unit (ReLU) activation and a predetermined fraction of active neurons.
Using the NNGP kernel, we observe that the sparser networks outperform the
non-sparse networks at shallow depths on a variety of datasets. We validate
this observation by extending the existing theory on the generalization error
of kernel-ridge regression.
| no_new_dataset | 0.710653 |
2305.10559 | Konstantin Hopf | Elena Giacomazzi, Felix Haag, Konstantin Hopf | Short-Term Electricity Load Forecasting Using the Temporal Fusion
Transformer: Effect of Grid Hierarchies and Data Sources | null | The 14th ACM International Conference on Future Energy Systems
(e-Energy '23), June 20--23, 2023, Orlando, FL, USA | 10.1145/10.1145/3575813.3597345 | null | cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recent developments related to the energy transition pose particular
challenges for distribution grids. Hence, precise load forecasts become more
and more important for effective grid management. Novel modeling approaches
such as the Transformer architecture, in particular the Temporal Fusion
Transformer (TFT), have emerged as promising methods for time series
forecasting. To date, just a handful of studies apply TFTs to electricity load
forecasting problems, mostly considering only single datasets and a few
covariates. Therefore, we examine the potential of the TFT architecture for
hourly short-term load forecasting across different time horizons (day-ahead
and week-ahead) and network levels (grid and substation level). We find that
the TFT architecture does not offer higher predictive performance than a
state-of-the-art LSTM model for day-ahead forecasting on the entire grid.
However, the results display significant improvements for the TFT when applied
at the substation level with a subsequent aggregation to the upper grid-level,
resulting in a prediction error of 2.43% (MAPE) for the best-performing
scenario. In addition, the TFT appears to offer remarkable improvements over
the LSTM approach for week-ahead forecasting (yielding a predictive error of
2.52% (MAPE) at the lowest). We outline avenues for future research using the
TFT approach for load forecasting, including the exploration of various grid
levels (e.g., grid, substation, and household level).
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 20:33:51 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Giacomazzi",
"Elena",
""
],
[
"Haag",
"Felix",
""
],
[
"Hopf",
"Konstantin",
""
]
] | TITLE: Short-Term Electricity Load Forecasting Using the Temporal Fusion
Transformer: Effect of Grid Hierarchies and Data Sources
ABSTRACT: Recent developments related to the energy transition pose particular
challenges for distribution grids. Hence, precise load forecasts become more
and more important for effective grid management. Novel modeling approaches
such as the Transformer architecture, in particular the Temporal Fusion
Transformer (TFT), have emerged as promising methods for time series
forecasting. To date, just a handful of studies apply TFTs to electricity load
forecasting problems, mostly considering only single datasets and a few
covariates. Therefore, we examine the potential of the TFT architecture for
hourly short-term load forecasting across different time horizons (day-ahead
and week-ahead) and network levels (grid and substation level). We find that
the TFT architecture does not offer higher predictive performance than a
state-of-the-art LSTM model for day-ahead forecasting on the entire grid.
However, the results display significant improvements for the TFT when applied
at the substation level with a subsequent aggregation to the upper grid-level,
resulting in a prediction error of 2.43% (MAPE) for the best-performing
scenario. In addition, the TFT appears to offer remarkable improvements over
the LSTM approach for week-ahead forecasting (yielding a predictive error of
2.52% (MAPE) at the lowest). We outline avenues for future research using the
TFT approach for load forecasting, including the exploration of various grid
levels (e.g., grid, substation, and household level).
| no_new_dataset | 0.711588 |
2305.10580 | Juncheng Li | Juncheng Li, David J. Cappelleri | Sim-MEES: Modular End-Effector System Grasping Dataset for Mobile
Manipulators in Cluttered Environments | null | null | null | null | cs.RO cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present Sim-MEES: a large-scale synthetic dataset that
contains 1,550 objects with varying difficulty levels and physics properties,
as well as 11 million grasp labels for mobile manipulators to plan grasps using
different gripper modalities in cluttered environments. Our dataset generation
process combines analytic models and dynamic simulations of the entire
cluttered environment to provide accurate grasp labels. We provide a detailed
study of our proposed labeling process for both parallel jaw grippers and
suction cup grippers, comparing them with state-of-the-art methods to
demonstrate how Sim-MEES can provide precise grasp labels in cluttered
environments.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 21:40:26 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Juncheng",
""
],
[
"Cappelleri",
"David J.",
""
]
] | TITLE: Sim-MEES: Modular End-Effector System Grasping Dataset for Mobile
Manipulators in Cluttered Environments
ABSTRACT: In this paper, we present Sim-MEES: a large-scale synthetic dataset that
contains 1,550 objects with varying difficulty levels and physics properties,
as well as 11 million grasp labels for mobile manipulators to plan grasps using
different gripper modalities in cluttered environments. Our dataset generation
process combines analytic models and dynamic simulations of the entire
cluttered environment to provide accurate grasp labels. We provide a detailed
study of our proposed labeling process for both parallel jaw grippers and
suction cup grippers, comparing them with state-of-the-art methods to
demonstrate how Sim-MEES can provide precise grasp labels in cluttered
environments.
| new_dataset | 0.712038 |
2305.10596 | Xinrui Liu | Xinrui Liu, Yajie Wang, Yu-an Tan, Kefan Qiu, Yuanzhang Li | Towards Invisible Backdoor Attacks in the Frequency Domain against Deep
Neural Networks | arXiv admin note: text overlap with arXiv:2305.09677 | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | Deep neural networks (DNNs) have made tremendous progress in the past ten
years and have been applied in various critical applications. However, recent
studies have shown that deep neural networks are vulnerable to backdoor
attacks. By injecting malicious data into the training set, an adversary can
plant the backdoor into the original model. The backdoor can remain hidden
indefinitely until activated by a sample with a specific trigger, which is
hugely concealed, bringing serious security risks to critical applications.
However, one main limitation of current backdoor attacks is that the trigger is
often visible to human perception. Therefore, it is crucial to study the
stealthiness of backdoor triggers. In this paper, we propose a novel
frequency-domain backdooring technique. In particular, our method aims to add a
backdoor trigger in the frequency domain of original images via Discrete
Fourier Transform, thus hidding the trigger. We evaluate our method on three
benchmark datasets: MNIST, CIFAR-10 and Imagenette. Our experiments show that
we can simultaneously fool human inspection and DNN models. We further apply
two image similarity evaluation metrics to illustrate that our method adds the
most subtle perturbation without compromising attack success rate and clean
sample accuracy.
| [
{
"version": "v1",
"created": "Wed, 10 May 2023 09:09:56 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Liu",
"Xinrui",
""
],
[
"Wang",
"Yajie",
""
],
[
"Tan",
"Yu-an",
""
],
[
"Qiu",
"Kefan",
""
],
[
"Li",
"Yuanzhang",
""
]
] | TITLE: Towards Invisible Backdoor Attacks in the Frequency Domain against Deep
Neural Networks
ABSTRACT: Deep neural networks (DNNs) have made tremendous progress in the past ten
years and have been applied in various critical applications. However, recent
studies have shown that deep neural networks are vulnerable to backdoor
attacks. By injecting malicious data into the training set, an adversary can
plant the backdoor into the original model. The backdoor can remain hidden
indefinitely until activated by a sample with a specific trigger, which is
hugely concealed, bringing serious security risks to critical applications.
However, one main limitation of current backdoor attacks is that the trigger is
often visible to human perception. Therefore, it is crucial to study the
stealthiness of backdoor triggers. In this paper, we propose a novel
frequency-domain backdooring technique. In particular, our method aims to add a
backdoor trigger in the frequency domain of original images via Discrete
Fourier Transform, thus hidding the trigger. We evaluate our method on three
benchmark datasets: MNIST, CIFAR-10 and Imagenette. Our experiments show that
we can simultaneously fool human inspection and DNN models. We further apply
two image similarity evaluation metrics to illustrate that our method adds the
most subtle perturbation without compromising attack success rate and clean
sample accuracy.
| no_new_dataset | 0.706792 |
2305.10610 | Saeth Wannasuphoprasit | Saeth Wannasuphoprasit, Yi Zhou, Danushka Bollegala | Solving Cosine Similarity Underestimation between High Frequency Words
by L2 Norm Discounting | 7 pages, 5 figures. To be published in the Proceedings of the 61st
Annual Meeting of the Association for Computational Linguistics, 9-14 July
2023, Toronto, Canada | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Cosine similarity between two words, computed using their contextualised
token embeddings obtained from masked language models (MLMs) such as BERT has
shown to underestimate the actual similarity between those words (Zhou et al.,
2022). This similarity underestimation problem is particularly severe for
highly frequent words. Although this problem has been noted in prior work, no
solution has been proposed thus far. We observe that the L2 norm of
contextualised embeddings of a word correlates with its log-frequency in the
pretraining corpus. Consequently, the larger L2 norms associated with the
highly frequent words reduce the cosine similarity values measured between
them, thus underestimating the similarity scores. To solve this issue, we
propose a method to discount the L2 norm of a contextualised word embedding by
the frequency of that word in a corpus when measuring the cosine similarities
between words. We show that the so called stop words behave differently from
the rest of the words, which require special consideration during their
discounting process. Experimental results on a contextualised word similarity
dataset show that our proposed discounting method accurately solves the
similarity underestimation problem.
| [
{
"version": "v1",
"created": "Wed, 17 May 2023 23:41:30 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Wannasuphoprasit",
"Saeth",
""
],
[
"Zhou",
"Yi",
""
],
[
"Bollegala",
"Danushka",
""
]
] | TITLE: Solving Cosine Similarity Underestimation between High Frequency Words
by L2 Norm Discounting
ABSTRACT: Cosine similarity between two words, computed using their contextualised
token embeddings obtained from masked language models (MLMs) such as BERT has
shown to underestimate the actual similarity between those words (Zhou et al.,
2022). This similarity underestimation problem is particularly severe for
highly frequent words. Although this problem has been noted in prior work, no
solution has been proposed thus far. We observe that the L2 norm of
contextualised embeddings of a word correlates with its log-frequency in the
pretraining corpus. Consequently, the larger L2 norms associated with the
highly frequent words reduce the cosine similarity values measured between
them, thus underestimating the similarity scores. To solve this issue, we
propose a method to discount the L2 norm of a contextualised word embedding by
the frequency of that word in a corpus when measuring the cosine similarities
between words. We show that the so called stop words behave differently from
the rest of the words, which require special consideration during their
discounting process. Experimental results on a contextualised word similarity
dataset show that our proposed discounting method accurately solves the
similarity underestimation problem.
| no_new_dataset | 0.709806 |
2305.10636 | Yue Qiu | Jiankui Zhou and Yue Qiu | Augmented Message Passing Stein Variational Gradient Descent | null | null | null | null | cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Stein Variational Gradient Descent (SVGD) is a popular particle-based method
for Bayesian inference. However, its convergence suffers from the variance
collapse, which reduces the accuracy and diversity of the estimation. In this
paper, we study the isotropy property of finite particles during the
convergence process and show that SVGD of finite particles cannot spread across
the entire sample space. Instead, all particles tend to cluster around the
particle center within a certain range and we provide an analytical bound for
this cluster. To further improve the effectiveness of SVGD for high-dimensional
problems, we propose the Augmented Message Passing SVGD (AUMP-SVGD) method,
which is a two-stage optimization procedure that does not require sparsity of
the target distribution, unlike the MP-SVGD method. Our algorithm achieves
satisfactory accuracy and overcomes the variance collapse problem in various
benchmark problems.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 01:13:04 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Zhou",
"Jiankui",
""
],
[
"Qiu",
"Yue",
""
]
] | TITLE: Augmented Message Passing Stein Variational Gradient Descent
ABSTRACT: Stein Variational Gradient Descent (SVGD) is a popular particle-based method
for Bayesian inference. However, its convergence suffers from the variance
collapse, which reduces the accuracy and diversity of the estimation. In this
paper, we study the isotropy property of finite particles during the
convergence process and show that SVGD of finite particles cannot spread across
the entire sample space. Instead, all particles tend to cluster around the
particle center within a certain range and we provide an analytical bound for
this cluster. To further improve the effectiveness of SVGD for high-dimensional
problems, we propose the Augmented Message Passing SVGD (AUMP-SVGD) method,
which is a two-stage optimization procedure that does not require sparsity of
the target distribution, unlike the MP-SVGD method. Our algorithm achieves
satisfactory accuracy and overcomes the variance collapse problem in various
benchmark problems.
| no_new_dataset | 0.710603 |
2305.10640 | Xiaoguang Li | Xiaoguang Li, Qing Guo, Pingping Cai, Wei Feng, Ivor Tsang, Song Wang | Learning Restoration is Not Enough: Transfering Identical Mapping for
Single-Image Shadow Removal | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Shadow removal is to restore shadow regions to their shadow-free counterparts
while leaving non-shadow regions unchanged. State-of-the-art shadow removal
methods train deep neural networks on collected shadow & shadow-free image
pairs, which are desired to complete two distinct tasks via shared weights,
i.e., data restoration for shadow regions and identical mapping for non-shadow
regions. We find that these two tasks exhibit poor compatibility, and using
shared weights for these two tasks could lead to the model being optimized
towards only one task instead of both during the training process. Note that
such a key issue is not identified by existing deep learning-based shadow
removal methods. To address this problem, we propose to handle these two tasks
separately and leverage the identical mapping results to guide the shadow
restoration in an iterative manner. Specifically, our method consists of three
components: an identical mapping branch (IMB) for non-shadow regions
processing, an iterative de-shadow branch (IDB) for shadow regions restoration
based on identical results, and a smart aggregation block (SAB). The IMB aims
to reconstruct an image that is identical to the input one, which can benefit
the restoration of the non-shadow regions without explicitly distinguishing
between shadow and non-shadow regions. Utilizing the multi-scale features
extracted by the IMB, the IDB can effectively transfer information from
non-shadow regions to shadow regions progressively, facilitating the process of
shadow removal. The SAB is designed to adaptive integrate features from both
IMB and IDB. Moreover, it generates a finely tuned soft shadow mask that guides
the process of removing shadows. Extensive experiments demonstrate our method
outperforms all the state-of-the-art shadow removal approaches on the widely
used shadow removal datasets.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 01:36:23 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Xiaoguang",
""
],
[
"Guo",
"Qing",
""
],
[
"Cai",
"Pingping",
""
],
[
"Feng",
"Wei",
""
],
[
"Tsang",
"Ivor",
""
],
[
"Wang",
"Song",
""
]
] | TITLE: Learning Restoration is Not Enough: Transfering Identical Mapping for
Single-Image Shadow Removal
ABSTRACT: Shadow removal is to restore shadow regions to their shadow-free counterparts
while leaving non-shadow regions unchanged. State-of-the-art shadow removal
methods train deep neural networks on collected shadow & shadow-free image
pairs, which are desired to complete two distinct tasks via shared weights,
i.e., data restoration for shadow regions and identical mapping for non-shadow
regions. We find that these two tasks exhibit poor compatibility, and using
shared weights for these two tasks could lead to the model being optimized
towards only one task instead of both during the training process. Note that
such a key issue is not identified by existing deep learning-based shadow
removal methods. To address this problem, we propose to handle these two tasks
separately and leverage the identical mapping results to guide the shadow
restoration in an iterative manner. Specifically, our method consists of three
components: an identical mapping branch (IMB) for non-shadow regions
processing, an iterative de-shadow branch (IDB) for shadow regions restoration
based on identical results, and a smart aggregation block (SAB). The IMB aims
to reconstruct an image that is identical to the input one, which can benefit
the restoration of the non-shadow regions without explicitly distinguishing
between shadow and non-shadow regions. Utilizing the multi-scale features
extracted by the IMB, the IDB can effectively transfer information from
non-shadow regions to shadow regions progressively, facilitating the process of
shadow removal. The SAB is designed to adaptive integrate features from both
IMB and IDB. Moreover, it generates a finely tuned soft shadow mask that guides
the process of removing shadows. Extensive experiments demonstrate our method
outperforms all the state-of-the-art shadow removal approaches on the widely
used shadow removal datasets.
| no_new_dataset | 0.709849 |
2305.10643 | Nathan Beck | Nathan Beck, Suraj Kothawade, Pradeep Shenoy, Rishabh Iyer | STREAMLINE: Streaming Active Learning for Realistic Multi-Distributional
Settings | 20 pages, 14 figures, 2 tables | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks have consistently shown great performance in several
real-world use cases like autonomous vehicles, satellite imaging, etc.,
effectively leveraging large corpora of labeled training data. However,
learning unbiased models depends on building a dataset that is representative
of a diverse range of realistic scenarios for a given task. This is challenging
in many settings where data comes from high-volume streams, with each scenario
occurring in random interleaved episodes at varying frequencies. We study
realistic streaming settings where data instances arrive in and are sampled
from an episodic multi-distributional data stream. Using submodular information
measures, we propose STREAMLINE, a novel streaming active learning framework
that mitigates scenario-driven slice imbalance in the working labeled data via
a three-step procedure of slice identification, slice-aware budgeting, and data
selection. We extensively evaluate STREAMLINE on real-world streaming scenarios
for image classification and object detection tasks. We observe that STREAMLINE
improves the performance on infrequent yet critical slices of the data over
current baselines by up to $5\%$ in terms of accuracy on our image
classification tasks and by up to $8\%$ in terms of mAP on our object detection
tasks.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 02:01:45 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Beck",
"Nathan",
""
],
[
"Kothawade",
"Suraj",
""
],
[
"Shenoy",
"Pradeep",
""
],
[
"Iyer",
"Rishabh",
""
]
] | TITLE: STREAMLINE: Streaming Active Learning for Realistic Multi-Distributional
Settings
ABSTRACT: Deep neural networks have consistently shown great performance in several
real-world use cases like autonomous vehicles, satellite imaging, etc.,
effectively leveraging large corpora of labeled training data. However,
learning unbiased models depends on building a dataset that is representative
of a diverse range of realistic scenarios for a given task. This is challenging
in many settings where data comes from high-volume streams, with each scenario
occurring in random interleaved episodes at varying frequencies. We study
realistic streaming settings where data instances arrive in and are sampled
from an episodic multi-distributional data stream. Using submodular information
measures, we propose STREAMLINE, a novel streaming active learning framework
that mitigates scenario-driven slice imbalance in the working labeled data via
a three-step procedure of slice identification, slice-aware budgeting, and data
selection. We extensively evaluate STREAMLINE on real-world streaming scenarios
for image classification and object detection tasks. We observe that STREAMLINE
improves the performance on infrequent yet critical slices of the data over
current baselines by up to $5\%$ in terms of accuracy on our image
classification tasks and by up to $8\%$ in terms of mAP on our object detection
tasks.
| no_new_dataset | 0.70562 |
2305.10647 | Sreyan Ghosh | Sreyan Ghosh and Utkarsh Tyagi and Sonal Kumar and Dinesh Manocha | BioAug: Conditional Generation based Data Augmentation for Low-Resource
Biomedical NER | SIGIR 2023 | null | null | null | cs.CL cs.AI cs.IR | http://creativecommons.org/licenses/by/4.0/ | Biomedical Named Entity Recognition (BioNER) is the fundamental task of
identifying named entities from biomedical text. However, BioNER suffers from
severe data scarcity and lacks high-quality labeled data due to the highly
specialized and expert knowledge required for annotation. Though data
augmentation has shown to be highly effective for low-resource NER in general,
existing data augmentation techniques fail to produce factual and diverse
augmentations for BioNER. In this paper, we present BioAug, a novel data
augmentation framework for low-resource BioNER. BioAug, built on BART, is
trained to solve a novel text reconstruction task based on selective masking
and knowledge augmentation. Post training, we perform conditional generation
and generate diverse augmentations conditioning BioAug on selectively corrupted
text similar to the training stage. We demonstrate the effectiveness of BioAug
on 5 benchmark BioNER datasets and show that BioAug outperforms all our
baselines by a significant margin (1.5%-21.5% absolute improvement) and is able
to generate augmentations that are both more factual and diverse. Code:
https://github.com/Sreyan88/BioAug.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 02:04:38 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Ghosh",
"Sreyan",
""
],
[
"Tyagi",
"Utkarsh",
""
],
[
"Kumar",
"Sonal",
""
],
[
"Manocha",
"Dinesh",
""
]
] | TITLE: BioAug: Conditional Generation based Data Augmentation for Low-Resource
Biomedical NER
ABSTRACT: Biomedical Named Entity Recognition (BioNER) is the fundamental task of
identifying named entities from biomedical text. However, BioNER suffers from
severe data scarcity and lacks high-quality labeled data due to the highly
specialized and expert knowledge required for annotation. Though data
augmentation has shown to be highly effective for low-resource NER in general,
existing data augmentation techniques fail to produce factual and diverse
augmentations for BioNER. In this paper, we present BioAug, a novel data
augmentation framework for low-resource BioNER. BioAug, built on BART, is
trained to solve a novel text reconstruction task based on selective masking
and knowledge augmentation. Post training, we perform conditional generation
and generate diverse augmentations conditioning BioAug on selectively corrupted
text similar to the training stage. We demonstrate the effectiveness of BioAug
on 5 benchmark BioNER datasets and show that BioAug outperforms all our
baselines by a significant margin (1.5%-21.5% absolute improvement) and is able
to generate augmentations that are both more factual and diverse. Code:
https://github.com/Sreyan88/BioAug.
| no_new_dataset | 0.709818 |
2305.10673 | Jintang Li | Jintang Li, Sheng Tian, Ruofan Wu, Liang Zhu, Welong Zhao, Changhua
Meng, Liang Chen, Zibin Zheng, Hongzhi Yin | Less Can Be More: Unsupervised Graph Pruning for Large-scale Dynamic
Graphs | Preprint | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The prevalence of large-scale graphs poses great challenges in time and
storage for training and deploying graph neural networks (GNNs). Several recent
works have explored solutions for pruning the large original graph into a small
and highly-informative one, such that training and inference on the pruned and
large graphs have comparable performance. Although empirically effective,
current researches focus on static or non-temporal graphs, which are not
directly applicable to dynamic scenarios. In addition, they require labels as
ground truth to learn the informative structure, limiting their applicability
to new problem domains where labels are hard to obtain. To solve the dilemma,
we propose and study the problem of unsupervised graph pruning on dynamic
graphs. We approach the problem by our proposed STEP, a self-supervised
temporal pruning framework that learns to remove potentially redundant edges
from input dynamic graphs. From a technical and industrial viewpoint, our
method overcomes the trade-offs between the performance and the time & memory
overheads. Our results on three real-world datasets demonstrate the advantages
on improving the efficacy, robustness, and efficiency of GNNs on dynamic node
classification tasks. Most notably, STEP is able to prune more than 50% of
edges on a million-scale industrial graph Alipay (7M nodes, 21M edges) while
approximating up to 98% of the original performance. Code is available at
https://github.com/EdisonLeeeee/STEP.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 03:23:53 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Jintang",
""
],
[
"Tian",
"Sheng",
""
],
[
"Wu",
"Ruofan",
""
],
[
"Zhu",
"Liang",
""
],
[
"Zhao",
"Welong",
""
],
[
"Meng",
"Changhua",
""
],
[
"Chen",
"Liang",
""
],
[
"Zheng",
"Zibin",
""
],
[
"Yin",
"Hongzhi",
""
]
] | TITLE: Less Can Be More: Unsupervised Graph Pruning for Large-scale Dynamic
Graphs
ABSTRACT: The prevalence of large-scale graphs poses great challenges in time and
storage for training and deploying graph neural networks (GNNs). Several recent
works have explored solutions for pruning the large original graph into a small
and highly-informative one, such that training and inference on the pruned and
large graphs have comparable performance. Although empirically effective,
current researches focus on static or non-temporal graphs, which are not
directly applicable to dynamic scenarios. In addition, they require labels as
ground truth to learn the informative structure, limiting their applicability
to new problem domains where labels are hard to obtain. To solve the dilemma,
we propose and study the problem of unsupervised graph pruning on dynamic
graphs. We approach the problem by our proposed STEP, a self-supervised
temporal pruning framework that learns to remove potentially redundant edges
from input dynamic graphs. From a technical and industrial viewpoint, our
method overcomes the trade-offs between the performance and the time & memory
overheads. Our results on three real-world datasets demonstrate the advantages
on improving the efficacy, robustness, and efficiency of GNNs on dynamic node
classification tasks. Most notably, STEP is able to prune more than 50% of
edges on a million-scale industrial graph Alipay (7M nodes, 21M edges) while
approximating up to 98% of the original performance. Code is available at
https://github.com/EdisonLeeeee/STEP.
| no_new_dataset | 0.708168 |
2305.10679 | Xin-Ye Li | Xin-Ye Li, Jiang-Tian Xue, Zheng Xie and Ming Li | Think Outside the Code: Brainstorming Boosts Large Language Models in
Code Generation | 13 pages, 5 figures | null | null | null | cs.AI cs.CL cs.SE | http://creativecommons.org/licenses/by/4.0/ | Code generation aims to automatically generate source code from high-level
task specifications, which can significantly increase productivity of software
engineering. Recently, approaches based on large language models (LLMs) have
shown remarkable code generation abilities on simple tasks. However, generate
code for more complex tasks, such as competition-level problems, remains
challenging. In this paper, we introduce Brainstorm framework for code
generation. It leverages a brainstorming step that generates and selects
diverse thoughts on the problem to facilitate algorithmic reasoning, where the
thoughts are possible blueprint of solving the problem. We demonstrate that
Brainstorm significantly enhances the ability of LLMs to solve
competition-level programming problems, resulting in a more than 50% increase
in the pass@$k$ metrics for ChatGPT on the CodeContests benchmark, achieving
state-of-the-art performance. Furthermore, our experiments conducted on
LeetCode contests show that our framework boosts the ability of ChatGPT to a
level comparable to that of human programmers.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 03:32:54 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Xin-Ye",
""
],
[
"Xue",
"Jiang-Tian",
""
],
[
"Xie",
"Zheng",
""
],
[
"Li",
"Ming",
""
]
] | TITLE: Think Outside the Code: Brainstorming Boosts Large Language Models in
Code Generation
ABSTRACT: Code generation aims to automatically generate source code from high-level
task specifications, which can significantly increase productivity of software
engineering. Recently, approaches based on large language models (LLMs) have
shown remarkable code generation abilities on simple tasks. However, generate
code for more complex tasks, such as competition-level problems, remains
challenging. In this paper, we introduce Brainstorm framework for code
generation. It leverages a brainstorming step that generates and selects
diverse thoughts on the problem to facilitate algorithmic reasoning, where the
thoughts are possible blueprint of solving the problem. We demonstrate that
Brainstorm significantly enhances the ability of LLMs to solve
competition-level programming problems, resulting in a more than 50% increase
in the pass@$k$ metrics for ChatGPT on the CodeContests benchmark, achieving
state-of-the-art performance. Furthermore, our experiments conducted on
LeetCode contests show that our framework boosts the ability of ChatGPT to a
level comparable to that of human programmers.
| no_new_dataset | 0.70866 |
2305.10686 | Jinzheng He | Jinzheng He, Jinglin Liu, Zhenhui Ye, Rongjie Huang, Chenye Cui,
Huadai Liu, Zhou Zhao | RMSSinger: Realistic-Music-Score based Singing Voice Synthesis | Accepted by Finding of ACL2023 | null | null | null | cs.SD cs.CL eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We are interested in a challenging task, Realistic-Music-Score based Singing
Voice Synthesis (RMS-SVS). RMS-SVS aims to generate high-quality singing voices
given realistic music scores with different note types (grace, slur, rest,
etc.). Though significant progress has been achieved, recent singing voice
synthesis (SVS) methods are limited to fine-grained music scores, which require
a complicated data collection pipeline with time-consuming manual annotation to
align music notes with phonemes. Furthermore, these manual annotation destroys
the regularity of note durations in music scores, making fine-grained music
scores inconvenient for composing. To tackle these challenges, we propose
RMSSinger, the first RMS-SVS method, which takes realistic music scores as
input, eliminating most of the tedious manual annotation and avoiding the
aforementioned inconvenience. Note that music scores are based on words rather
than phonemes, in RMSSinger, we introduce word-level modeling to avoid the
time-consuming phoneme duration annotation and the complicated phoneme-level
mel-note alignment. Furthermore, we propose the first diffusion-based pitch
modeling method, which ameliorates the naturalness of existing pitch-modeling
methods. To achieve these, we collect a new dataset containing realistic music
scores and singing voices according to these realistic music scores from
professional singers. Extensive experiments on the dataset demonstrate the
effectiveness of our methods. Audio samples are available at
https://rmssinger.github.io/.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 03:57:51 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"He",
"Jinzheng",
""
],
[
"Liu",
"Jinglin",
""
],
[
"Ye",
"Zhenhui",
""
],
[
"Huang",
"Rongjie",
""
],
[
"Cui",
"Chenye",
""
],
[
"Liu",
"Huadai",
""
],
[
"Zhao",
"Zhou",
""
]
] | TITLE: RMSSinger: Realistic-Music-Score based Singing Voice Synthesis
ABSTRACT: We are interested in a challenging task, Realistic-Music-Score based Singing
Voice Synthesis (RMS-SVS). RMS-SVS aims to generate high-quality singing voices
given realistic music scores with different note types (grace, slur, rest,
etc.). Though significant progress has been achieved, recent singing voice
synthesis (SVS) methods are limited to fine-grained music scores, which require
a complicated data collection pipeline with time-consuming manual annotation to
align music notes with phonemes. Furthermore, these manual annotation destroys
the regularity of note durations in music scores, making fine-grained music
scores inconvenient for composing. To tackle these challenges, we propose
RMSSinger, the first RMS-SVS method, which takes realistic music scores as
input, eliminating most of the tedious manual annotation and avoiding the
aforementioned inconvenience. Note that music scores are based on words rather
than phonemes, in RMSSinger, we introduce word-level modeling to avoid the
time-consuming phoneme duration annotation and the complicated phoneme-level
mel-note alignment. Furthermore, we propose the first diffusion-based pitch
modeling method, which ameliorates the naturalness of existing pitch-modeling
methods. To achieve these, we collect a new dataset containing realistic music
scores and singing voices according to these realistic music scores from
professional singers. Extensive experiments on the dataset demonstrate the
effectiveness of our methods. Audio samples are available at
https://rmssinger.github.io/.
| new_dataset | 0.713226 |
2305.10696 | Zheyu Zhang | Zheyu Zhang, Tianping Zhang, Jian Li | Unbiased Gradient Boosting Decision Tree with Unbiased Feature
Importance | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gradient Boosting Decision Tree (GBDT) has achieved remarkable success in a
wide variety of applications. The split finding algorithm, which determines the
tree construction process, is one of the most crucial components of GBDT.
However, the split finding algorithm has long been criticized for its bias
towards features with a large number of potential splits. This bias introduces
severe interpretability and overfitting issues in GBDT. To this end, we provide
a fine-grained analysis of bias in GBDT and demonstrate that the bias
originates from 1) the systematic bias in the gain estimation of each split and
2) the bias in the split finding algorithm resulting from the use of the same
data to evaluate the split improvement and determine the best split. Based on
the analysis, we propose unbiased gain, a new unbiased measurement of gain
importance using out-of-bag samples. Moreover, we incorporate the unbiased
property into the split finding algorithm and develop UnbiasedGBM to solve the
overfitting issue of GBDT. We assess the performance of UnbiasedGBM and
unbiased gain in a large-scale empirical study comprising 60 datasets and show
that: 1) UnbiasedGBM exhibits better performance than popular GBDT
implementations such as LightGBM, XGBoost, and Catboost on average on the 60
datasets and 2) unbiased gain achieves better average performance in feature
selection than popular feature importance methods. The codes are available at
https://github.com/ZheyuAqaZhang/UnbiasedGBM.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 04:17:46 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Zhang",
"Zheyu",
""
],
[
"Zhang",
"Tianping",
""
],
[
"Li",
"Jian",
""
]
] | TITLE: Unbiased Gradient Boosting Decision Tree with Unbiased Feature
Importance
ABSTRACT: Gradient Boosting Decision Tree (GBDT) has achieved remarkable success in a
wide variety of applications. The split finding algorithm, which determines the
tree construction process, is one of the most crucial components of GBDT.
However, the split finding algorithm has long been criticized for its bias
towards features with a large number of potential splits. This bias introduces
severe interpretability and overfitting issues in GBDT. To this end, we provide
a fine-grained analysis of bias in GBDT and demonstrate that the bias
originates from 1) the systematic bias in the gain estimation of each split and
2) the bias in the split finding algorithm resulting from the use of the same
data to evaluate the split improvement and determine the best split. Based on
the analysis, we propose unbiased gain, a new unbiased measurement of gain
importance using out-of-bag samples. Moreover, we incorporate the unbiased
property into the split finding algorithm and develop UnbiasedGBM to solve the
overfitting issue of GBDT. We assess the performance of UnbiasedGBM and
unbiased gain in a large-scale empirical study comprising 60 datasets and show
that: 1) UnbiasedGBM exhibits better performance than popular GBDT
implementations such as LightGBM, XGBoost, and Catboost on average on the 60
datasets and 2) unbiased gain achieves better average performance in feature
selection than popular feature importance methods. The codes are available at
https://github.com/ZheyuAqaZhang/UnbiasedGBM.
| no_new_dataset | 0.710069 |
2305.10703 | Yue Yu | Yue Yu, Yuchen Zhuang, Rongzhi Zhang, Yu Meng, Jiaming Shen, Chao
Zhang | ReGen: Zero-Shot Text Classification via Training Data Generation with
Progressive Dense Retrieval | ACL 2023 Findings (Code: https://github.com/yueyu1030/ReGen) | null | null | null | cs.CL cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the development of large language models (LLMs), zero-shot learning has
attracted much attention for various NLP tasks. Different from prior works that
generate training data with billion-scale natural language generation (NLG)
models, we propose a retrieval-enhanced framework to create training data from
a general-domain unlabeled corpus. To realize this, we first conduct
contrastive pretraining to learn an unsupervised dense retriever for extracting
the most relevant documents using class-descriptive verbalizers. We then
further propose two simple strategies, namely Verbalizer Augmentation with
Demonstrations and Self-consistency Guided Filtering to improve the topic
coverage of the dataset while removing noisy examples. Experiments on nine
datasets demonstrate that REGEN achieves 4.3% gain over the strongest baselines
and saves around 70% of the time compared to baselines using large NLG models.
Besides, REGEN can be naturally integrated with recently proposed large
language models to boost performance.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 04:30:09 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yu",
"Yue",
""
],
[
"Zhuang",
"Yuchen",
""
],
[
"Zhang",
"Rongzhi",
""
],
[
"Meng",
"Yu",
""
],
[
"Shen",
"Jiaming",
""
],
[
"Zhang",
"Chao",
""
]
] | TITLE: ReGen: Zero-Shot Text Classification via Training Data Generation with
Progressive Dense Retrieval
ABSTRACT: With the development of large language models (LLMs), zero-shot learning has
attracted much attention for various NLP tasks. Different from prior works that
generate training data with billion-scale natural language generation (NLG)
models, we propose a retrieval-enhanced framework to create training data from
a general-domain unlabeled corpus. To realize this, we first conduct
contrastive pretraining to learn an unsupervised dense retriever for extracting
the most relevant documents using class-descriptive verbalizers. We then
further propose two simple strategies, namely Verbalizer Augmentation with
Demonstrations and Self-consistency Guided Filtering to improve the topic
coverage of the dataset while removing noisy examples. Experiments on nine
datasets demonstrate that REGEN achieves 4.3% gain over the strongest baselines
and saves around 70% of the time compared to baselines using large NLG models.
Besides, REGEN can be naturally integrated with recently proposed large
language models to boost performance.
| no_new_dataset | 0.710597 |
2305.10709 | Tingting Wu | Tingting Wu, Xiao Ding, Minji Tang, Hao Zhang, Bing Qin, Ting Liu | NoisywikiHow: A Benchmark for Learning with Real-world Noisy Labels in
Natural Language Processing | The paper has been accepted by ACL 2023 Findings. Dataset and code
are available in this https URL | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large-scale datasets in the real world inevitably involve label noise. Deep
models can gradually overfit noisy labels and thus degrade model
generalization. To mitigate the effects of label noise, learning with noisy
labels (LNL) methods are designed to achieve better generalization performance.
Due to the lack of suitable datasets, previous studies have frequently employed
synthetic label noise to mimic real-world label noise. However, synthetic noise
is not instance-dependent, making this approximation not always effective in
practice. Recent research has proposed benchmarks for learning with real-world
noisy labels. However, the noise sources within may be single or fuzzy, making
benchmarks different from data with heterogeneous label noises in the real
world. To tackle these issues, we contribute NoisywikiHow, the largest NLP
benchmark built with minimal supervision. Specifically, inspired by human
cognition, we explicitly construct multiple sources of label noise to imitate
human errors throughout the annotation, replicating real-world noise, whose
corruption is affected by both ground-truth labels and instances. Moreover, we
provide a variety of noise levels to support controlled experiments on noisy
data, enabling us to evaluate LNL methods systematically and comprehensively.
After that, we conduct extensive multi-dimensional experiments on a broad range
of LNL methods, obtaining new and intriguing findings.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 05:01:04 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Wu",
"Tingting",
""
],
[
"Ding",
"Xiao",
""
],
[
"Tang",
"Minji",
""
],
[
"Zhang",
"Hao",
""
],
[
"Qin",
"Bing",
""
],
[
"Liu",
"Ting",
""
]
] | TITLE: NoisywikiHow: A Benchmark for Learning with Real-world Noisy Labels in
Natural Language Processing
ABSTRACT: Large-scale datasets in the real world inevitably involve label noise. Deep
models can gradually overfit noisy labels and thus degrade model
generalization. To mitigate the effects of label noise, learning with noisy
labels (LNL) methods are designed to achieve better generalization performance.
Due to the lack of suitable datasets, previous studies have frequently employed
synthetic label noise to mimic real-world label noise. However, synthetic noise
is not instance-dependent, making this approximation not always effective in
practice. Recent research has proposed benchmarks for learning with real-world
noisy labels. However, the noise sources within may be single or fuzzy, making
benchmarks different from data with heterogeneous label noises in the real
world. To tackle these issues, we contribute NoisywikiHow, the largest NLP
benchmark built with minimal supervision. Specifically, inspired by human
cognition, we explicitly construct multiple sources of label noise to imitate
human errors throughout the annotation, replicating real-world noise, whose
corruption is affected by both ground-truth labels and instances. Moreover, we
provide a variety of noise levels to support controlled experiments on noisy
data, enabling us to evaluate LNL methods systematically and comprehensively.
After that, we conduct extensive multi-dimensional experiments on a broad range
of LNL methods, obtaining new and intriguing findings.
| no_new_dataset | 0.710239 |
2305.10727 | Chong Yu | Chong Yu, Tao Chen, Zhongxue Gan, Jiayuan Fan | Boost Vision Transformer with GPU-Friendly Sparsity and Quantization | Accepted to CVPR 2023, 11 pages, 6 figures | null | null | null | cs.CV cs.LG cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The transformer extends its success from the language to the vision domain.
Because of the stacked self-attention and cross-attention blocks, the
acceleration deployment of vision transformer on GPU hardware is challenging
and also rarely studied. This paper thoroughly designs a compression scheme to
maximally utilize the GPU-friendly 2:4 fine-grained structured sparsity and
quantization. Specially, an original large model with dense weight parameters
is first pruned into a sparse one by 2:4 structured pruning, which considers
the GPU's acceleration of 2:4 structured sparse pattern with FP16 data type,
then the floating-point sparse model is further quantized into a fixed-point
one by sparse-distillation-aware quantization aware training, which considers
GPU can provide an extra speedup of 2:4 sparse calculation with integer
tensors. A mixed-strategy knowledge distillation is used during the pruning and
quantization process. The proposed compression scheme is flexible to support
supervised and unsupervised learning styles. Experiment results show GPUSQ-ViT
scheme achieves state-of-the-art compression by reducing vision transformer
models 6.4-12.7 times on model size and 30.3-62 times on FLOPs with negligible
accuracy degradation on ImageNet classification, COCO detection and ADE20K
segmentation benchmarking tasks. Moreover, GPUSQ-ViT can boost actual
deployment performance by 1.39-1.79 times and 3.22-3.43 times of latency and
throughput on A100 GPU, and 1.57-1.69 times and 2.11-2.51 times improvement of
latency and throughput on AGX Orin.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 05:55:48 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yu",
"Chong",
""
],
[
"Chen",
"Tao",
""
],
[
"Gan",
"Zhongxue",
""
],
[
"Fan",
"Jiayuan",
""
]
] | TITLE: Boost Vision Transformer with GPU-Friendly Sparsity and Quantization
ABSTRACT: The transformer extends its success from the language to the vision domain.
Because of the stacked self-attention and cross-attention blocks, the
acceleration deployment of vision transformer on GPU hardware is challenging
and also rarely studied. This paper thoroughly designs a compression scheme to
maximally utilize the GPU-friendly 2:4 fine-grained structured sparsity and
quantization. Specially, an original large model with dense weight parameters
is first pruned into a sparse one by 2:4 structured pruning, which considers
the GPU's acceleration of 2:4 structured sparse pattern with FP16 data type,
then the floating-point sparse model is further quantized into a fixed-point
one by sparse-distillation-aware quantization aware training, which considers
GPU can provide an extra speedup of 2:4 sparse calculation with integer
tensors. A mixed-strategy knowledge distillation is used during the pruning and
quantization process. The proposed compression scheme is flexible to support
supervised and unsupervised learning styles. Experiment results show GPUSQ-ViT
scheme achieves state-of-the-art compression by reducing vision transformer
models 6.4-12.7 times on model size and 30.3-62 times on FLOPs with negligible
accuracy degradation on ImageNet classification, COCO detection and ADE20K
segmentation benchmarking tasks. Moreover, GPUSQ-ViT can boost actual
deployment performance by 1.39-1.79 times and 3.22-3.43 times of latency and
throughput on A100 GPU, and 1.57-1.69 times and 2.11-2.51 times improvement of
latency and throughput on AGX Orin.
| no_new_dataset | 0.708648 |
2305.10736 | Chenhe Dong | Chenhe Dong, Yuexiang Xie, Yaliang Li, Ying Shen | Counterfactual Debiasing for Generating Factually Consistent Text
Summaries | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite substantial progress in abstractive text summarization to generate
fluent and informative texts, the factual inconsistency in the generated
summaries remains an important yet challenging problem to be solved. In this
paper, we construct causal graphs for abstractive text summarization and
identify the intrinsic causes of the factual inconsistency, i.e., the language
bias and irrelevancy bias, and further propose a debiasing framework, named
CoFactSum, to alleviate the causal effects of these biases by counterfactual
estimation. Specifically, the proposed CoFactSum provides two counterfactual
estimation strategies, i.e., Explicit Counterfactual Masking with an explicit
dynamic masking strategy, and Implicit Counterfactual Training with an implicit
discriminative cross-attention mechanism. Meanwhile, we design a Debiasing
Degree Adjustment mechanism to dynamically adapt the debiasing degree at each
decoding step. Extensive experiments on two widely-used summarization datasets
demonstrate the effectiveness of CoFactSum in enhancing the factual consistency
of generated summaries compared with several baselines.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 06:15:45 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Dong",
"Chenhe",
""
],
[
"Xie",
"Yuexiang",
""
],
[
"Li",
"Yaliang",
""
],
[
"Shen",
"Ying",
""
]
] | TITLE: Counterfactual Debiasing for Generating Factually Consistent Text
Summaries
ABSTRACT: Despite substantial progress in abstractive text summarization to generate
fluent and informative texts, the factual inconsistency in the generated
summaries remains an important yet challenging problem to be solved. In this
paper, we construct causal graphs for abstractive text summarization and
identify the intrinsic causes of the factual inconsistency, i.e., the language
bias and irrelevancy bias, and further propose a debiasing framework, named
CoFactSum, to alleviate the causal effects of these biases by counterfactual
estimation. Specifically, the proposed CoFactSum provides two counterfactual
estimation strategies, i.e., Explicit Counterfactual Masking with an explicit
dynamic masking strategy, and Implicit Counterfactual Training with an implicit
discriminative cross-attention mechanism. Meanwhile, we design a Debiasing
Degree Adjustment mechanism to dynamically adapt the debiasing degree at each
decoding step. Extensive experiments on two widely-used summarization datasets
demonstrate the effectiveness of CoFactSum in enhancing the factual consistency
of generated summaries compared with several baselines.
| no_new_dataset | 0.709384 |
2305.10763 | Zhenhui Ye | Zhenhui Ye, Rongjie Huang, Yi Ren, Ziyue Jiang, Jinglin Liu, Jinzheng
He, Xiang Yin, Zhou Zhao | CLAPSpeech: Learning Prosody from Text Context with Contrastive
Language-Audio Pre-training | Accepted by ACL 2023 (Main Conference) | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Improving text representation has attracted much attention to achieve
expressive text-to-speech (TTS). However, existing works only implicitly learn
the prosody with masked token reconstruction tasks, which leads to low training
efficiency and difficulty in prosody modeling. We propose CLAPSpeech, a
cross-modal contrastive pre-training framework that explicitly learns the
prosody variance of the same text token under different contexts. Specifically,
1) We encourage the model to connect the text context with its corresponding
prosody pattern in the joint multi-modal space with the elaborate design of the
encoder inputs and contrastive loss; 2) We introduce a multi-scale pre-training
pipeline to capture prosody patterns in multiple levels. We show how to
incorporate CLAPSpeech into existing TTS models for better prosody. Experiments
on three datasets not only show that CLAPSpeech could improve the prosody
prediction for existing TTS methods, but also demonstrate its generalization
ability to adapt to multiple languages and multi-speaker TTS. We also deeply
analyze the principle behind the performance of CLAPSpeech. Ablation studies
demonstrate the necessity of each component in our method. Source code and
audio samples are available at https://clapspeech.github.io.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 07:07:04 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Ye",
"Zhenhui",
""
],
[
"Huang",
"Rongjie",
""
],
[
"Ren",
"Yi",
""
],
[
"Jiang",
"Ziyue",
""
],
[
"Liu",
"Jinglin",
""
],
[
"He",
"Jinzheng",
""
],
[
"Yin",
"Xiang",
""
],
[
"Zhao",
"Zhou",
""
]
] | TITLE: CLAPSpeech: Learning Prosody from Text Context with Contrastive
Language-Audio Pre-training
ABSTRACT: Improving text representation has attracted much attention to achieve
expressive text-to-speech (TTS). However, existing works only implicitly learn
the prosody with masked token reconstruction tasks, which leads to low training
efficiency and difficulty in prosody modeling. We propose CLAPSpeech, a
cross-modal contrastive pre-training framework that explicitly learns the
prosody variance of the same text token under different contexts. Specifically,
1) We encourage the model to connect the text context with its corresponding
prosody pattern in the joint multi-modal space with the elaborate design of the
encoder inputs and contrastive loss; 2) We introduce a multi-scale pre-training
pipeline to capture prosody patterns in multiple levels. We show how to
incorporate CLAPSpeech into existing TTS models for better prosody. Experiments
on three datasets not only show that CLAPSpeech could improve the prosody
prediction for existing TTS methods, but also demonstrate its generalization
ability to adapt to multiple languages and multi-speaker TTS. We also deeply
analyze the principle behind the performance of CLAPSpeech. Ablation studies
demonstrate the necessity of each component in our method. Source code and
audio samples are available at https://clapspeech.github.io.
| no_new_dataset | 0.707657 |
2305.10772 | Mengke Li | Mengke Li, Yiu-ming Cheung, Juyong Jiang | Feature-Balanced Loss for Long-Tailed Visual Recognition | Published as a conference paper at ICME 2022 | null | 10.1109/ICME52920.2022.9860003 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Deep neural networks frequently suffer from performance degradation when the
training data is long-tailed because several majority classes dominate the
training, resulting in a biased model. Recent studies have made a great effort
in solving this issue by obtaining good representations from data space, but
few of them pay attention to the influence of feature norm on the predicted
results. In this paper, we therefore address the long-tailed problem from
feature space and thereby propose the feature-balanced loss. Specifically, we
encourage larger feature norms of tail classes by giving them relatively
stronger stimuli. Moreover, the stimuli intensity is gradually increased in the
way of curriculum learning, which improves the generalization of the tail
classes, meanwhile maintaining the performance of the head classes. Extensive
experiments on multiple popular long-tailed recognition benchmarks demonstrate
that the feature-balanced loss achieves superior performance gains compared
with the state-of-the-art methods.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 07:30:22 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Mengke",
""
],
[
"Cheung",
"Yiu-ming",
""
],
[
"Jiang",
"Juyong",
""
]
] | TITLE: Feature-Balanced Loss for Long-Tailed Visual Recognition
ABSTRACT: Deep neural networks frequently suffer from performance degradation when the
training data is long-tailed because several majority classes dominate the
training, resulting in a biased model. Recent studies have made a great effort
in solving this issue by obtaining good representations from data space, but
few of them pay attention to the influence of feature norm on the predicted
results. In this paper, we therefore address the long-tailed problem from
feature space and thereby propose the feature-balanced loss. Specifically, we
encourage larger feature norms of tail classes by giving them relatively
stronger stimuli. Moreover, the stimuli intensity is gradually increased in the
way of curriculum learning, which improves the generalization of the tail
classes, meanwhile maintaining the performance of the head classes. Extensive
experiments on multiple popular long-tailed recognition benchmarks demonstrate
that the feature-balanced loss achieves superior performance gains compared
with the state-of-the-art methods.
| no_new_dataset | 0.711205 |
2305.10783 | Julia Kiseleva | Shrestha Mohanty and Negar Arabzadeh and Julia Kiseleva and Artem
Zholus and Milagro Teruel and Ahmed Awadallah and Yuxuan Sun and Kavya Srinet
and Arthur Szlam | Transforming Human-Centered AI Collaboration: Redefining Embodied Agents
Capabilities through Interactive Grounded Language Instructions | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Human intelligence's adaptability is remarkable, allowing us to adjust to new
tasks and multi-modal environments swiftly. This skill is evident from a young
age as we acquire new abilities and solve problems by imitating others or
following natural language instructions. The research community is actively
pursuing the development of interactive "embodied agents" that can engage in
natural conversations with humans and assist them with real-world tasks. These
agents must possess the ability to promptly request feedback in case
communication breaks down or instructions are unclear. Additionally, they must
demonstrate proficiency in learning new vocabulary specific to a given domain.
In this paper, we made the following contributions: (1) a crowd-sourcing tool
for collecting grounded language instructions; (2) the largest dataset of
grounded language instructions; and (3) several state-of-the-art baselines.
These contributions are suitable as a foundation for further research.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 07:51:33 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Mohanty",
"Shrestha",
""
],
[
"Arabzadeh",
"Negar",
""
],
[
"Kiseleva",
"Julia",
""
],
[
"Zholus",
"Artem",
""
],
[
"Teruel",
"Milagro",
""
],
[
"Awadallah",
"Ahmed",
""
],
[
"Sun",
"Yuxuan",
""
],
[
"Srinet",
"Kavya",
""
],
[
"Szlam",
"Arthur",
""
]
] | TITLE: Transforming Human-Centered AI Collaboration: Redefining Embodied Agents
Capabilities through Interactive Grounded Language Instructions
ABSTRACT: Human intelligence's adaptability is remarkable, allowing us to adjust to new
tasks and multi-modal environments swiftly. This skill is evident from a young
age as we acquire new abilities and solve problems by imitating others or
following natural language instructions. The research community is actively
pursuing the development of interactive "embodied agents" that can engage in
natural conversations with humans and assist them with real-world tasks. These
agents must possess the ability to promptly request feedback in case
communication breaks down or instructions are unclear. Additionally, they must
demonstrate proficiency in learning new vocabulary specific to a given domain.
In this paper, we made the following contributions: (1) a crowd-sourcing tool
for collecting grounded language instructions; (2) the largest dataset of
grounded language instructions; and (3) several state-of-the-art baselines.
These contributions are suitable as a foundation for further research.
| new_dataset | 0.714236 |
2305.10785 | Bo Lin | Bo Lin, Shangwen Wang, Zhongxin Liu, Yepang Liu, Xin Xia and Xiaoguang
Mao | CCT5: A Code-Change-Oriented Pre-Trained Model | null | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Software is constantly changing, requiring developers to perform several
derived tasks in a timely manner, such as writing a description for the
intention of the code change, or identifying the defect-prone code changes.
Considering that the cost of dealing with these tasks can account for a large
proportion (typically around 70 percent) of the total development expenditure,
automating such processes will significantly lighten the burdens of developers.
To achieve such a target, existing approaches mainly rely on training deep
learning models from scratch or fine-tuning existing pretrained models on such
tasks, both of which have weaknesses. Specifically, the former uses
comparatively small-scale labelled data for training, making it difficult to
learn and exploit the domain knowledge of programming language hidden in the
large-amount unlabelled code in the wild; the latter is hard to fully leverage
the learned knowledge of the pre-trained model, as existing pre-trained models
are designed to encode a single code snippet rather than a code change (i.e.,
the difference between two code snippets). We propose to pre-train a model
specially designed for code changes to better support developers in software
maintenance. To this end, we first collect a large-scale dataset containing
1.5M+ pairwise data of code changes and commit messages. Based on these data,
we curate five different tasks for pre-training, which equip the model with
diverse domain knowledge about code changes. We fine-tune the pre-trained
model, CCT5, on three widely-studied tasks incurred by code changes and two
tasks specific to the code review process. Results show that CCT5 outperforms
both conventional deep learning approaches and existing pre-trained models on
these tasks.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 07:55:37 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Lin",
"Bo",
""
],
[
"Wang",
"Shangwen",
""
],
[
"Liu",
"Zhongxin",
""
],
[
"Liu",
"Yepang",
""
],
[
"Xia",
"Xin",
""
],
[
"Mao",
"Xiaoguang",
""
]
] | TITLE: CCT5: A Code-Change-Oriented Pre-Trained Model
ABSTRACT: Software is constantly changing, requiring developers to perform several
derived tasks in a timely manner, such as writing a description for the
intention of the code change, or identifying the defect-prone code changes.
Considering that the cost of dealing with these tasks can account for a large
proportion (typically around 70 percent) of the total development expenditure,
automating such processes will significantly lighten the burdens of developers.
To achieve such a target, existing approaches mainly rely on training deep
learning models from scratch or fine-tuning existing pretrained models on such
tasks, both of which have weaknesses. Specifically, the former uses
comparatively small-scale labelled data for training, making it difficult to
learn and exploit the domain knowledge of programming language hidden in the
large-amount unlabelled code in the wild; the latter is hard to fully leverage
the learned knowledge of the pre-trained model, as existing pre-trained models
are designed to encode a single code snippet rather than a code change (i.e.,
the difference between two code snippets). We propose to pre-train a model
specially designed for code changes to better support developers in software
maintenance. To this end, we first collect a large-scale dataset containing
1.5M+ pairwise data of code changes and commit messages. Based on these data,
we curate five different tasks for pre-training, which equip the model with
diverse domain knowledge about code changes. We fine-tune the pre-trained
model, CCT5, on three widely-studied tasks incurred by code changes and two
tasks specific to the code review process. Results show that CCT5 outperforms
both conventional deep learning approaches and existing pre-trained models on
these tasks.
| new_dataset | 0.668069 |
2305.10799 | Qiuhui Chen | Qiuhui Chen, Xinyue Hu, Zirui Wang, Yi Hong | MedBLIP: Bootstrapping Language-Image Pre-training from 3D Medical
Images and Texts | 11 pages, 3 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-language pre-training (VLP) models have been demonstrated to be
effective in many computer vision applications. In this paper, we consider
developing a VLP model in the medical domain for making computer-aided
diagnoses (CAD) based on image scans and text descriptions in electronic health
records, as done in practice. To achieve our goal, we present a lightweight CAD
system MedBLIP, a new paradigm for bootstrapping VLP from off-the-shelf frozen
pre-trained image encoders and frozen large language models. We design a
MedQFormer module to bridge the gap between 3D medical images and 2D
pre-trained image encoders and language models as well. To evaluate the
effectiveness of our MedBLIP, we collect more than 30,000 image volumes from
five public Alzheimer's disease (AD) datasets, i.e., ADNI, NACC, OASIS, AIBL,
and MIRIAD. On this largest AD dataset we know, our model achieves the SOTA
performance on the zero-shot classification of healthy, mild cognitive
impairment (MCI), and AD subjects, and shows its capability of making medical
visual question answering (VQA). The code and pre-trained models is available
online: https://github.com/Qybc/MedBLIP.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 08:19:33 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Chen",
"Qiuhui",
""
],
[
"Hu",
"Xinyue",
""
],
[
"Wang",
"Zirui",
""
],
[
"Hong",
"Yi",
""
]
] | TITLE: MedBLIP: Bootstrapping Language-Image Pre-training from 3D Medical
Images and Texts
ABSTRACT: Vision-language pre-training (VLP) models have been demonstrated to be
effective in many computer vision applications. In this paper, we consider
developing a VLP model in the medical domain for making computer-aided
diagnoses (CAD) based on image scans and text descriptions in electronic health
records, as done in practice. To achieve our goal, we present a lightweight CAD
system MedBLIP, a new paradigm for bootstrapping VLP from off-the-shelf frozen
pre-trained image encoders and frozen large language models. We design a
MedQFormer module to bridge the gap between 3D medical images and 2D
pre-trained image encoders and language models as well. To evaluate the
effectiveness of our MedBLIP, we collect more than 30,000 image volumes from
five public Alzheimer's disease (AD) datasets, i.e., ADNI, NACC, OASIS, AIBL,
and MIRIAD. On this largest AD dataset we know, our model achieves the SOTA
performance on the zero-shot classification of healthy, mild cognitive
impairment (MCI), and AD subjects, and shows its capability of making medical
visual question answering (VQA). The code and pre-trained models is available
online: https://github.com/Qybc/MedBLIP.
| no_new_dataset | 0.709057 |
2305.10822 | Zihua Si | Zihua Si, Zhongxiang Sun, Xiao Zhang, Jun Xu, Xiaoxue Zang, Yang Song,
Kun Gai, Ji-Rong Wen | When Search Meets Recommendation: Learning Disentangled Search
Representation for Recommendation | Accecpted by SIGIR 2023 | null | 10.1145/3539618.3591786 | null | cs.IR cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Modern online service providers such as online shopping platforms often
provide both search and recommendation (S&R) services to meet different user
needs. Rarely has there been any effective means of incorporating user behavior
data from both S&R services. Most existing approaches either simply treat S&R
behaviors separately, or jointly optimize them by aggregating data from both
services, ignoring the fact that user intents in S&R can be distinctively
different. In our paper, we propose a Search-Enhanced framework for the
Sequential Recommendation (SESRec) that leverages users' search interests for
recommendation, by disentangling similar and dissimilar representations within
S&R behaviors. Specifically, SESRec first aligns query and item embeddings
based on users' query-item interactions for the computations of their
similarities. Two transformer encoders are used to learn the contextual
representations of S&R behaviors independently. Then a contrastive learning
task is designed to supervise the disentanglement of similar and dissimilar
representations from behavior sequences of S&R. Finally, we extract user
interests by the attention mechanism from three perspectives, i.e., the
contextual representations, the two separated behaviors containing similar and
dissimilar interests. Extensive experiments on both industrial and public
datasets demonstrate that SESRec consistently outperforms state-of-the-art
models. Empirical studies further validate that SESRec successfully disentangle
similar and dissimilar user interests from their S&R behaviors.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 09:04:50 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Si",
"Zihua",
""
],
[
"Sun",
"Zhongxiang",
""
],
[
"Zhang",
"Xiao",
""
],
[
"Xu",
"Jun",
""
],
[
"Zang",
"Xiaoxue",
""
],
[
"Song",
"Yang",
""
],
[
"Gai",
"Kun",
""
],
[
"Wen",
"Ji-Rong",
""
]
] | TITLE: When Search Meets Recommendation: Learning Disentangled Search
Representation for Recommendation
ABSTRACT: Modern online service providers such as online shopping platforms often
provide both search and recommendation (S&R) services to meet different user
needs. Rarely has there been any effective means of incorporating user behavior
data from both S&R services. Most existing approaches either simply treat S&R
behaviors separately, or jointly optimize them by aggregating data from both
services, ignoring the fact that user intents in S&R can be distinctively
different. In our paper, we propose a Search-Enhanced framework for the
Sequential Recommendation (SESRec) that leverages users' search interests for
recommendation, by disentangling similar and dissimilar representations within
S&R behaviors. Specifically, SESRec first aligns query and item embeddings
based on users' query-item interactions for the computations of their
similarities. Two transformer encoders are used to learn the contextual
representations of S&R behaviors independently. Then a contrastive learning
task is designed to supervise the disentanglement of similar and dissimilar
representations from behavior sequences of S&R. Finally, we extract user
interests by the attention mechanism from three perspectives, i.e., the
contextual representations, the two separated behaviors containing similar and
dissimilar interests. Extensive experiments on both industrial and public
datasets demonstrate that SESRec consistently outperforms state-of-the-art
models. Empirical studies further validate that SESRec successfully disentangle
similar and dissimilar user interests from their S&R behaviors.
| no_new_dataset | 0.709044 |
2305.10834 | Jill Walker Rettberg | Scott Rettberg, Talan Memmott, Jill Walker Rettberg, Jason Nelson and
Patrick Lichty | AIwriting: Relations Between Image Generation and Digital Writing | Extended abstract for panel presented at ISEA 2023, Paris 16-22 May
2023 | null | null | null | cs.AI cs.CL cs.HC cs.MM | http://creativecommons.org/licenses/by/4.0/ | During 2022, both transformer-based AI text generation sys-tems such as GPT-3
and AI text-to-image generation systems such as DALL-E 2 and Stable Diffusion
made exponential leaps forward and are unquestionably altering the fields of
digital art and electronic literature. In this panel a group of electronic
literature authors and theorists consider new oppor-tunities for human
creativity presented by these systems and present new works have produced
during the past year that specifically address these systems as environments
for literary expressions that are translated through iterative interlocutive
processes into visual representations. The premise that binds these
presentations is that these systems and the works gener-ated must be considered
from a literary perspective, as they originate in human writing. In works
ranging from a visual memoir of the personal experience of a health crisis, to
interac-tive web comics, to architectures based on abstract poetic language, to
political satire, four artists explore the capabili-ties of these writing
environments for new genres of literary artist practice, while a digital
culture theorist considers the origins and effects of the particular training
datasets of human language and images on which these new hybrid forms are
based.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 09:23:05 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Rettberg",
"Scott",
""
],
[
"Memmott",
"Talan",
""
],
[
"Rettberg",
"Jill Walker",
""
],
[
"Nelson",
"Jason",
""
],
[
"Lichty",
"Patrick",
""
]
] | TITLE: AIwriting: Relations Between Image Generation and Digital Writing
ABSTRACT: During 2022, both transformer-based AI text generation sys-tems such as GPT-3
and AI text-to-image generation systems such as DALL-E 2 and Stable Diffusion
made exponential leaps forward and are unquestionably altering the fields of
digital art and electronic literature. In this panel a group of electronic
literature authors and theorists consider new oppor-tunities for human
creativity presented by these systems and present new works have produced
during the past year that specifically address these systems as environments
for literary expressions that are translated through iterative interlocutive
processes into visual representations. The premise that binds these
presentations is that these systems and the works gener-ated must be considered
from a literary perspective, as they originate in human writing. In works
ranging from a visual memoir of the personal experience of a health crisis, to
interac-tive web comics, to architectures based on abstract poetic language, to
political satire, four artists explore the capabili-ties of these writing
environments for new genres of literary artist practice, while a digital
culture theorist considers the origins and effects of the particular training
datasets of human language and images on which these new hybrid forms are
based.
| no_new_dataset | 0.710019 |
2305.10835 | Daniil Gavrilov | Daniil Gavrilov, Nikita Balagansky | Ahead-of-Time P-Tuning | null | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | In this paper, we propose Ahead-of-Time (AoT) P-Tuning, a novel
parameter-efficient fine-tuning method for pre-trained Language Models (LMs)
that adds input-dependent bias before each Transformer layer. We evaluate AoT
P-Tuning on GLUE and SuperGLUE benchmarking datasets using RoBERTa and DeBERTa
models, showing that it outperforms BitFit and is comparable or better than
other baseline methods for efficient fine-tuning. Additionally, we assess the
inference overhead of AoT P-Tuning and demonstrate that it introduces
negligible overhead compared to established baseline methods. Our method
enables multi-task inference with a single backbone LM, making it a practical
solution for real-world applications.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 09:24:53 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Gavrilov",
"Daniil",
""
],
[
"Balagansky",
"Nikita",
""
]
] | TITLE: Ahead-of-Time P-Tuning
ABSTRACT: In this paper, we propose Ahead-of-Time (AoT) P-Tuning, a novel
parameter-efficient fine-tuning method for pre-trained Language Models (LMs)
that adds input-dependent bias before each Transformer layer. We evaluate AoT
P-Tuning on GLUE and SuperGLUE benchmarking datasets using RoBERTa and DeBERTa
models, showing that it outperforms BitFit and is comparable or better than
other baseline methods for efficient fine-tuning. Additionally, we assess the
inference overhead of AoT P-Tuning and demonstrate that it introduces
negligible overhead compared to established baseline methods. Our method
enables multi-task inference with a single backbone LM, making it a practical
solution for real-world applications.
| no_new_dataset | 0.710616 |
2305.10839 | Chong-En Lin | Chong-En Lin, Kuan-Yu Chen | A Lexical-aware Non-autoregressive Transformer-based ASR Model | Accepted by Interspeech 2023 | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Non-autoregressive automatic speech recognition (ASR) has become a mainstream
of ASR modeling because of its fast decoding speed and satisfactory result. To
further boost the performance, relaxing the conditional independence assumption
and cascading large-scaled pre-trained models are two active research
directions. In addition to these strategies, we propose a lexical-aware
non-autoregressive Transformer-based (LA-NAT) ASR framework, which consists of
an acoustic encoder, a speech-text shared encoder, and a speech-text shared
decoder. The acoustic encoder is used to process the input speech features as
usual, and the speech-text shared encoder and decoder are designed to train
speech and text data simultaneously. By doing so, LA-NAT aims to make the ASR
model aware of lexical information, so the resulting model is expected to
achieve better results by leveraging the learned linguistic knowledge. A series
of experiments are conducted on the AISHELL-1, CSJ, and TEDLIUM 2 datasets.
According to the experiments, the proposed LA-NAT can provide superior results
than other recently proposed non-autoregressive ASR models. In addition, LA-NAT
is a relatively compact model than most non-autoregressive ASR models, and it
is about 58 times faster than the classic autoregressive model.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 09:50:47 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Lin",
"Chong-En",
""
],
[
"Chen",
"Kuan-Yu",
""
]
] | TITLE: A Lexical-aware Non-autoregressive Transformer-based ASR Model
ABSTRACT: Non-autoregressive automatic speech recognition (ASR) has become a mainstream
of ASR modeling because of its fast decoding speed and satisfactory result. To
further boost the performance, relaxing the conditional independence assumption
and cascading large-scaled pre-trained models are two active research
directions. In addition to these strategies, we propose a lexical-aware
non-autoregressive Transformer-based (LA-NAT) ASR framework, which consists of
an acoustic encoder, a speech-text shared encoder, and a speech-text shared
decoder. The acoustic encoder is used to process the input speech features as
usual, and the speech-text shared encoder and decoder are designed to train
speech and text data simultaneously. By doing so, LA-NAT aims to make the ASR
model aware of lexical information, so the resulting model is expected to
achieve better results by leveraging the learned linguistic knowledge. A series
of experiments are conducted on the AISHELL-1, CSJ, and TEDLIUM 2 datasets.
According to the experiments, the proposed LA-NAT can provide superior results
than other recently proposed non-autoregressive ASR models. In addition, LA-NAT
is a relatively compact model than most non-autoregressive ASR models, and it
is about 58 times faster than the classic autoregressive model.
| no_new_dataset | 0.709655 |
2305.10840 | Luigi Sbail\`o | Luigi Sbail\`o and Luca M. Ghiringhelli | Uncertainty Quantification in Deep Neural Networks through Statistical
Inference on Latent Space | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Uncertainty-quantification methods are applied to estimate the confidence of
deep-neural-networks classifiers over their predictions. However, most widely
used methods are known to be overconfident. We address this problem by
developing an algorithm that exploits the latent-space representation of data
points fed into the network, to assess the accuracy of their prediction. Using
the latent-space representation generated by the fraction of training set that
the network classifies correctly, we build a statistical model that is able to
capture the likelihood of a given prediction. We show on a synthetic dataset
that commonly used methods are mostly overconfident. Overconfidence occurs also
for predictions made on data points that are outside the distribution that
generated the training data. In contrast, our method can detect such
out-of-distribution data points as inaccurately predicted, thus aiding in the
automatic detection of outliers.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 09:52:06 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Sbailò",
"Luigi",
""
],
[
"Ghiringhelli",
"Luca M.",
""
]
] | TITLE: Uncertainty Quantification in Deep Neural Networks through Statistical
Inference on Latent Space
ABSTRACT: Uncertainty-quantification methods are applied to estimate the confidence of
deep-neural-networks classifiers over their predictions. However, most widely
used methods are known to be overconfident. We address this problem by
developing an algorithm that exploits the latent-space representation of data
points fed into the network, to assess the accuracy of their prediction. Using
the latent-space representation generated by the fraction of training set that
the network classifies correctly, we build a statistical model that is able to
capture the likelihood of a given prediction. We show on a synthetic dataset
that commonly used methods are mostly overconfident. Overconfidence occurs also
for predictions made on data points that are outside the distribution that
generated the training data. In contrast, our method can detect such
out-of-distribution data points as inaccurately predicted, thus aiding in the
automatic detection of outliers.
| no_new_dataset | 0.680016 |
2305.10848 | Dmitriy Kalugin-Balashov | Dmitriy Kalugin-Balashov | Advancing Full-Text Search Lemmatization Techniques with Paradigm
Retrieval from OpenCorpora | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | In this paper, we unveil a groundbreaking method to amplify full-text search
lemmatization, utilizing the OpenCorpora dataset and a bespoke paradigm
retrieval algorithm. Our primary aim is to streamline the extraction of a
word's primary form or lemma - a crucial factor in full-text search.
Additionally, we propose a compact dictionary storage strategy, significantly
boosting the speed and precision of lemma retrieval.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 10:07:50 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Kalugin-Balashov",
"Dmitriy",
""
]
] | TITLE: Advancing Full-Text Search Lemmatization Techniques with Paradigm
Retrieval from OpenCorpora
ABSTRACT: In this paper, we unveil a groundbreaking method to amplify full-text search
lemmatization, utilizing the OpenCorpora dataset and a bespoke paradigm
retrieval algorithm. Our primary aim is to streamline the extraction of a
word's primary form or lemma - a crucial factor in full-text search.
Additionally, we propose a compact dictionary storage strategy, significantly
boosting the speed and precision of lemma retrieval.
| no_new_dataset | 0.711049 |
2305.10868 | Yuan Zhou | Yuan Zhou, Xin Chen, Yanrong Guo, Shijie Hao, Richang Hong, Qi Tian | Advancing Incremental Few-shot Semantic Segmentation via Semantic-guided
Relation Alignment and Adaptation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Incremental few-shot semantic segmentation (IFSS) aims to incrementally
extend a semantic segmentation model to novel classes according to only a few
pixel-level annotated data, while preserving its segmentation capability on
previously learned base categories. This task faces a severe semantic-aliasing
issue between base and novel classes due to data imbalance, which makes
segmentation results unsatisfactory. To alleviate this issue, we propose the
Semantic-guided Relation Alignment and Adaptation (SRAA) method that fully
considers the guidance of prior semantic information. Specifically, we first
conduct Semantic Relation Alignment (SRA) in the base step, so as to
semantically align base class representations to their semantics. As a result,
the embeddings of base classes are constrained to have relatively low semantic
correlations to categories that are different from them. Afterwards, based on
the semantically aligned base categories, Semantic-Guided Adaptation (SGA) is
employed during the incremental learning stage. It aims to ensure affinities
between visual and semantic embeddings of encountered novel categories, thereby
making the feature representations be consistent with their semantic
information. In this way, the semantic-aliasing issue can be suppressed. We
evaluate our model on the PASCAL VOC 2012 and the COCO dataset. The
experimental results on both these two datasets exhibit its competitive
performance, which demonstrates the superiority of our method.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 10:40:52 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Zhou",
"Yuan",
""
],
[
"Chen",
"Xin",
""
],
[
"Guo",
"Yanrong",
""
],
[
"Hao",
"Shijie",
""
],
[
"Hong",
"Richang",
""
],
[
"Tian",
"Qi",
""
]
] | TITLE: Advancing Incremental Few-shot Semantic Segmentation via Semantic-guided
Relation Alignment and Adaptation
ABSTRACT: Incremental few-shot semantic segmentation (IFSS) aims to incrementally
extend a semantic segmentation model to novel classes according to only a few
pixel-level annotated data, while preserving its segmentation capability on
previously learned base categories. This task faces a severe semantic-aliasing
issue between base and novel classes due to data imbalance, which makes
segmentation results unsatisfactory. To alleviate this issue, we propose the
Semantic-guided Relation Alignment and Adaptation (SRAA) method that fully
considers the guidance of prior semantic information. Specifically, we first
conduct Semantic Relation Alignment (SRA) in the base step, so as to
semantically align base class representations to their semantics. As a result,
the embeddings of base classes are constrained to have relatively low semantic
correlations to categories that are different from them. Afterwards, based on
the semantically aligned base categories, Semantic-Guided Adaptation (SGA) is
employed during the incremental learning stage. It aims to ensure affinities
between visual and semantic embeddings of encountered novel categories, thereby
making the feature representations be consistent with their semantic
information. In this way, the semantic-aliasing issue can be suppressed. We
evaluate our model on the PASCAL VOC 2012 and the COCO dataset. The
experimental results on both these two datasets exhibit its competitive
performance, which demonstrates the superiority of our method.
| no_new_dataset | 0.711675 |
2305.10872 | Vitaly Aksenov | Vitaly Aksenov, Dmitry Ivanov, Ravil Galiev | Benchmark Framework with Skewed Workloads | null | null | null | null | cs.DC | http://creativecommons.org/licenses/by/4.0/ | In this work, we present a new benchmarking suite with new real-life inspired
skewed workloads to test the performance of concurrent index data structures.
We started this project to prepare workloads specifically for self-adjusting
data structures, i.e., they handle more frequent requests faster, and, thus,
should perform better than their standard counterparts. We looked over the
commonly used suites to test performance of concurrent indices trying to find
an inspiration: Synchrobench, Setbench, YCSB, and TPC - and we found several
issues with them.
The major problem is that they are not flexible: it is difficult to introduce
new workloads, it is difficult to set the duration of the experiments, and it
is difficult to change the parameters. We decided to solve this issue by
presenting a new suite based on Synchrobench.
Finally, we highlight the problem of measuring performance of data
structures. We show that the relative performance of data structures highly
depends on the workload: it is not clear which data structure is best. For
that, we take three state-of-the-art concurrent binary search trees and run
them on the workloads from our benchmarking suite. As a result, we get six
experiments with all possible relative performance of the chosen data
structures.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 10:54:24 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Aksenov",
"Vitaly",
""
],
[
"Ivanov",
"Dmitry",
""
],
[
"Galiev",
"Ravil",
""
]
] | TITLE: Benchmark Framework with Skewed Workloads
ABSTRACT: In this work, we present a new benchmarking suite with new real-life inspired
skewed workloads to test the performance of concurrent index data structures.
We started this project to prepare workloads specifically for self-adjusting
data structures, i.e., they handle more frequent requests faster, and, thus,
should perform better than their standard counterparts. We looked over the
commonly used suites to test performance of concurrent indices trying to find
an inspiration: Synchrobench, Setbench, YCSB, and TPC - and we found several
issues with them.
The major problem is that they are not flexible: it is difficult to introduce
new workloads, it is difficult to set the duration of the experiments, and it
is difficult to change the parameters. We decided to solve this issue by
presenting a new suite based on Synchrobench.
Finally, we highlight the problem of measuring performance of data
structures. We show that the relative performance of data structures highly
depends on the workload: it is not clear which data structure is best. For
that, we take three state-of-the-art concurrent binary search trees and run
them on the workloads from our benchmarking suite. As a result, we get six
experiments with all possible relative performance of the chosen data
structures.
| new_dataset | 0.609466 |
2305.10882 | Luigi Sigillo | Luigi Sigillo, Eleonora Grassucci, Danilo Comminiello | StawGAN: Structural-Aware Generative Adversarial Networks for Infrared
Image Translation | null | 2023 IEEE International Symposium on Circuits and Systems (ISCAS) | null | null | cs.CV cs.LG eess.IV | http://creativecommons.org/licenses/by/4.0/ | This paper addresses the problem of translating night-time thermal infrared
images, which are the most adopted image modalities to analyze night-time
scenes, to daytime color images (NTIT2DC), which provide better perceptions of
objects. We introduce a novel model that focuses on enhancing the quality of
the target generation without merely colorizing it. The proposed structural
aware (StawGAN) enables the translation of better-shaped and high-definition
objects in the target domain. We test our model on aerial images of the
DroneVeichle dataset containing RGB-IR paired images. The proposed approach
produces a more accurate translation with respect to other state-of-the-art
image translation models. The source code is available at
https://github.com/LuigiSigillo/StawGAN
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 11:22:33 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Sigillo",
"Luigi",
""
],
[
"Grassucci",
"Eleonora",
""
],
[
"Comminiello",
"Danilo",
""
]
] | TITLE: StawGAN: Structural-Aware Generative Adversarial Networks for Infrared
Image Translation
ABSTRACT: This paper addresses the problem of translating night-time thermal infrared
images, which are the most adopted image modalities to analyze night-time
scenes, to daytime color images (NTIT2DC), which provide better perceptions of
objects. We introduce a novel model that focuses on enhancing the quality of
the target generation without merely colorizing it. The proposed structural
aware (StawGAN) enables the translation of better-shaped and high-definition
objects in the target domain. We test our model on aerial images of the
DroneVeichle dataset containing RGB-IR paired images. The proposed approach
produces a more accurate translation with respect to other state-of-the-art
image translation models. The source code is available at
https://github.com/LuigiSigillo/StawGAN
| no_new_dataset | 0.711994 |
2305.10887 | Priyanka Maity | Priyanka Maity, Kunwar Pritiraj Rajput, Suraj Srivastava, Naveen K. D.
Venkategowda, Aditya K. Jagannatham and Lajos Hanzo | Robust Hybrid Transceiver Designs for Linear Decentralized Estimation in
mmWave MIMO IoT Networks in the Face of Imperfect CSI | 16 pages, 8 figures | null | null | null | cs.IT eess.SP math.IT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Hybrid transceivers are designed for linear decentralized estimation (LDE) in
a mmWave multiple-input multiple-output (MIMO) IoT network (IoTNe). For a
noiseless fusion center (FC), it is demonstrated that the MSE performance is
determined by the number of RF chains used at each IoT node (IoTNo). Next, the
minimum-MSE RF transmit precoders (TPCs) and receive combiner (RC) matrices are
designed for this setup using the dominant array response vectors, and
subsequently, a closed-form expression is obtained for the baseband (BB) TPC at
each IoTNo using Cauchy's interlacing theorem. For a realistic noisy FC, it is
shown that the resultant mean squared error (MSE) minimization problem is
non-convex. To address this challenge, a block-coordinate descent-based
iterative scheme is proposed to obtain the fully digital TPC and RC matrices
followed by the simultaneous orthogonal matching pursuit (SOMP) technique for
decomposing the fully-digital transceiver into its corresponding RF and BB
components. A theoretical proof of the convergence is also presented for the
proposed iterative design procedure. Furthermore, robust hybrid transceiver
designs are also derived for a practical scenario in the face of channel state
information (CSI) uncertainty. The centralized MMSE lower bound has also been
derived that benchmarks the performance of the proposed LDE schemes. Finally,
our numerical results characterize the performance of the proposed transceivers
as well as corroborate our various analytical propositions.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 11:32:02 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Maity",
"Priyanka",
""
],
[
"Rajput",
"Kunwar Pritiraj",
""
],
[
"Srivastava",
"Suraj",
""
],
[
"Venkategowda",
"Naveen K. D.",
""
],
[
"Jagannatham",
"Aditya K.",
""
],
[
"Hanzo",
"Lajos",
""
]
] | TITLE: Robust Hybrid Transceiver Designs for Linear Decentralized Estimation in
mmWave MIMO IoT Networks in the Face of Imperfect CSI
ABSTRACT: Hybrid transceivers are designed for linear decentralized estimation (LDE) in
a mmWave multiple-input multiple-output (MIMO) IoT network (IoTNe). For a
noiseless fusion center (FC), it is demonstrated that the MSE performance is
determined by the number of RF chains used at each IoT node (IoTNo). Next, the
minimum-MSE RF transmit precoders (TPCs) and receive combiner (RC) matrices are
designed for this setup using the dominant array response vectors, and
subsequently, a closed-form expression is obtained for the baseband (BB) TPC at
each IoTNo using Cauchy's interlacing theorem. For a realistic noisy FC, it is
shown that the resultant mean squared error (MSE) minimization problem is
non-convex. To address this challenge, a block-coordinate descent-based
iterative scheme is proposed to obtain the fully digital TPC and RC matrices
followed by the simultaneous orthogonal matching pursuit (SOMP) technique for
decomposing the fully-digital transceiver into its corresponding RF and BB
components. A theoretical proof of the convergence is also presented for the
proposed iterative design procedure. Furthermore, robust hybrid transceiver
designs are also derived for a practical scenario in the face of channel state
information (CSI) uncertainty. The centralized MMSE lower bound has also been
derived that benchmarks the performance of the proposed LDE schemes. Finally,
our numerical results characterize the performance of the proposed transceivers
as well as corroborate our various analytical propositions.
| no_new_dataset | 0.710226 |
2305.10893 | Mengyang Yuan | Mengyang Yuan, Bo Lang, Fengnan Quan | Student-friendly Knowledge Distillation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | In knowledge distillation, the knowledge from the teacher model is often too
complex for the student model to thoroughly process. However, good teachers in
real life always simplify complex material before teaching it to students.
Inspired by this fact, we propose student-friendly knowledge distillation (SKD)
to simplify teacher output into new knowledge representations, which makes the
learning of the student model easier and more effective. SKD contains a
softening processing and a learning simplifier. First, the softening processing
uses the temperature hyperparameter to soften the output logits of the teacher
model, which simplifies the output to some extent and makes it easier for the
learning simplifier to process. The learning simplifier utilizes the attention
mechanism to further simplify the knowledge of the teacher model and is jointly
trained with the student model using the distillation loss, which means that
the process of simplification is correlated with the training objective of the
student model and ensures that the simplified new teacher knowledge
representation is more suitable for the specific student model. Furthermore,
since SKD does not change the form of the distillation loss, it can be easily
combined with other distillation methods that are based on the logits or
features of intermediate layers to enhance its effectiveness. Therefore, SKD
has wide applicability. The experimental results on the CIFAR-100 and ImageNet
datasets show that our method achieves state-of-the-art performance while
maintaining high training efficiency.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 11:44:30 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yuan",
"Mengyang",
""
],
[
"Lang",
"Bo",
""
],
[
"Quan",
"Fengnan",
""
]
] | TITLE: Student-friendly Knowledge Distillation
ABSTRACT: In knowledge distillation, the knowledge from the teacher model is often too
complex for the student model to thoroughly process. However, good teachers in
real life always simplify complex material before teaching it to students.
Inspired by this fact, we propose student-friendly knowledge distillation (SKD)
to simplify teacher output into new knowledge representations, which makes the
learning of the student model easier and more effective. SKD contains a
softening processing and a learning simplifier. First, the softening processing
uses the temperature hyperparameter to soften the output logits of the teacher
model, which simplifies the output to some extent and makes it easier for the
learning simplifier to process. The learning simplifier utilizes the attention
mechanism to further simplify the knowledge of the teacher model and is jointly
trained with the student model using the distillation loss, which means that
the process of simplification is correlated with the training objective of the
student model and ensures that the simplified new teacher knowledge
representation is more suitable for the specific student model. Furthermore,
since SKD does not change the form of the distillation loss, it can be easily
combined with other distillation methods that are based on the logits or
features of intermediate layers to enhance its effectiveness. Therefore, SKD
has wide applicability. The experimental results on the CIFAR-100 and ImageNet
datasets show that our method achieves state-of-the-art performance while
maintaining high training efficiency.
| no_new_dataset | 0.709007 |
2305.10899 | Deyi Ji | Deyi Ji, Feng Zhao, Hongtao Lu, Mingyuan Tao, Jieping Ye | Ultra-High Resolution Segmentation with Ultra-Rich Context: A Novel
Benchmark | Accepted to CVPR 2023 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the increasing interest and rapid development of methods for Ultra-High
Resolution (UHR) segmentation, a large-scale benchmark covering a wide range of
scenes with full fine-grained dense annotations is urgently needed to
facilitate the field. To this end, the URUR dataset is introduced, in the
meaning of Ultra-High Resolution dataset with Ultra-Rich Context. As the name
suggests, URUR contains amounts of images with high enough resolution (3,008
images of size 5,120x5,120), a wide range of complex scenes (from 63 cities),
rich-enough context (1 million instances with 8 categories) and fine-grained
annotations (about 80 billion manually annotated pixels), which is far superior
to all the existing UHR datasets including DeepGlobe, Inria Aerial, UDD, etc..
Moreover, we also propose WSDNet, a more efficient and effective framework for
UHR segmentation especially with ultra-rich context. Specifically, multi-level
Discrete Wavelet Transform (DWT) is naturally integrated to release computation
burden while preserve more spatial details, along with a Wavelet Smooth Loss
(WSL) to reconstruct original structured context and texture with a smooth
constrain. Experiments on several UHR datasets demonstrate its state-of-the-art
performance. The dataset is available at https://github.com/jankyee/URUR.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 11:54:13 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Ji",
"Deyi",
""
],
[
"Zhao",
"Feng",
""
],
[
"Lu",
"Hongtao",
""
],
[
"Tao",
"Mingyuan",
""
],
[
"Ye",
"Jieping",
""
]
] | TITLE: Ultra-High Resolution Segmentation with Ultra-Rich Context: A Novel
Benchmark
ABSTRACT: With the increasing interest and rapid development of methods for Ultra-High
Resolution (UHR) segmentation, a large-scale benchmark covering a wide range of
scenes with full fine-grained dense annotations is urgently needed to
facilitate the field. To this end, the URUR dataset is introduced, in the
meaning of Ultra-High Resolution dataset with Ultra-Rich Context. As the name
suggests, URUR contains amounts of images with high enough resolution (3,008
images of size 5,120x5,120), a wide range of complex scenes (from 63 cities),
rich-enough context (1 million instances with 8 categories) and fine-grained
annotations (about 80 billion manually annotated pixels), which is far superior
to all the existing UHR datasets including DeepGlobe, Inria Aerial, UDD, etc..
Moreover, we also propose WSDNet, a more efficient and effective framework for
UHR segmentation especially with ultra-rich context. Specifically, multi-level
Discrete Wavelet Transform (DWT) is naturally integrated to release computation
burden while preserve more spatial details, along with a Wavelet Smooth Loss
(WSL) to reconstruct original structured context and texture with a smooth
constrain. Experiments on several UHR datasets demonstrate its state-of-the-art
performance. The dataset is available at https://github.com/jankyee/URUR.
| new_dataset | 0.716857 |
2305.10907 | Xinze Li | Xinze Li, Yixin Cao, Muhao Chen, Aixin Sun | Take a Break in the Middle: Investigating Subgoals towards Hierarchical
Script Generation | Accepted by ACL 2023 Findings | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Goal-oriented Script Generation is a new task of generating a list of steps
that can fulfill the given goal. In this paper, we propose to extend the task
from the perspective of cognitive theory. Instead of a simple flat structure,
the steps are typically organized hierarchically - Human often decompose a
complex task into subgoals, where each subgoal can be further decomposed into
steps. To establish the benchmark, we contribute a new dataset, propose several
baseline methods, and set up evaluation metrics. Both automatic and human
evaluation verify the high-quality of dataset, as well as the effectiveness of
incorporating subgoals into hierarchical script generation. Furthermore, We
also design and evaluate the model to discover subgoal, and find that it is a
bit more difficult to decompose the goals than summarizing from segmented
steps.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 12:10:06 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Li",
"Xinze",
""
],
[
"Cao",
"Yixin",
""
],
[
"Chen",
"Muhao",
""
],
[
"Sun",
"Aixin",
""
]
] | TITLE: Take a Break in the Middle: Investigating Subgoals towards Hierarchical
Script Generation
ABSTRACT: Goal-oriented Script Generation is a new task of generating a list of steps
that can fulfill the given goal. In this paper, we propose to extend the task
from the perspective of cognitive theory. Instead of a simple flat structure,
the steps are typically organized hierarchically - Human often decompose a
complex task into subgoals, where each subgoal can be further decomposed into
steps. To establish the benchmark, we contribute a new dataset, propose several
baseline methods, and set up evaluation metrics. Both automatic and human
evaluation verify the high-quality of dataset, as well as the effectiveness of
incorporating subgoals into hierarchical script generation. Furthermore, We
also design and evaluate the model to discover subgoal, and find that it is a
bit more difficult to decompose the goals than summarizing from segmented
steps.
| new_dataset | 0.71579 |
2305.10928 | Nadav Borenstein | Nadav Borenstein and Natalia da Silva Perez and Isabelle Augenstein | Multilingual Event Extraction from Historical Newspaper Adverts | Accepted to the main track of ACL2023 | null | null | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | NLP methods can aid historians in analyzing textual materials in greater
volumes than manually feasible. Developing such methods poses substantial
challenges though. First, acquiring large, annotated historical datasets is
difficult, as only domain experts can reliably label them. Second, most
available off-the-shelf NLP models are trained on modern language texts,
rendering them significantly less effective when applied to historical corpora.
This is particularly problematic for less well studied tasks, and for languages
other than English. This paper addresses these challenges while focusing on the
under-explored task of event extraction from a novel domain of historical
texts. We introduce a new multilingual dataset in English, French, and Dutch
composed of newspaper ads from the early modern colonial period reporting on
enslaved people who liberated themselves from enslavement. We find that: 1)
even with scarce annotated data, it is possible to achieve surprisingly good
results by formulating the problem as an extractive QA task and leveraging
existing datasets and models for modern languages; and 2) cross-lingual
low-resource learning for historical languages is highly challenging, and
machine translation of the historical datasets to the considered target
languages is, in practice, often the best-performing solution.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 12:40:41 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Borenstein",
"Nadav",
""
],
[
"Perez",
"Natalia da Silva",
""
],
[
"Augenstein",
"Isabelle",
""
]
] | TITLE: Multilingual Event Extraction from Historical Newspaper Adverts
ABSTRACT: NLP methods can aid historians in analyzing textual materials in greater
volumes than manually feasible. Developing such methods poses substantial
challenges though. First, acquiring large, annotated historical datasets is
difficult, as only domain experts can reliably label them. Second, most
available off-the-shelf NLP models are trained on modern language texts,
rendering them significantly less effective when applied to historical corpora.
This is particularly problematic for less well studied tasks, and for languages
other than English. This paper addresses these challenges while focusing on the
under-explored task of event extraction from a novel domain of historical
texts. We introduce a new multilingual dataset in English, French, and Dutch
composed of newspaper ads from the early modern colonial period reporting on
enslaved people who liberated themselves from enslavement. We find that: 1)
even with scarce annotated data, it is possible to achieve surprisingly good
results by formulating the problem as an extractive QA task and leveraging
existing datasets and models for modern languages; and 2) cross-lingual
low-resource learning for historical languages is highly challenging, and
machine translation of the historical datasets to the considered target
languages is, in practice, often the best-performing solution.
| new_dataset | 0.715859 |
2305.10929 | Yihao Huang | Di Yang, Yihao Huang, Qing Guo, Felix Juefei-Xu, Ming Hu, Yang Liu,
Geguang Pu | Architecture-agnostic Iterative Black-box Certified Defense against
Adversarial Patches | 9 pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The adversarial patch attack aims to fool image classifiers within a bounded,
contiguous region of arbitrary changes, posing a real threat to computer vision
systems (e.g., autonomous driving, content moderation, biometric
authentication, medical imaging) in the physical world. To address this problem
in a trustworthy way, proposals have been made for certified patch defenses
that ensure the robustness of classification models and prevent future patch
attacks from breaching the defense. State-of-the-art certified defenses can be
compatible with any model architecture, as well as achieve high clean and
certified accuracy. Although the methods are adaptive to arbitrary patch
positions, they inevitably need to access the size of the adversarial patch,
which is unreasonable and impractical in real-world attack scenarios. To
improve the feasibility of the architecture-agnostic certified defense in a
black-box setting (i.e., position and size of the patch are both unknown), we
propose a novel two-stage Iterative Black-box Certified Defense method, termed
IBCD.In the first stage, it estimates the patch size in a search-based manner
by evaluating the size relationship between the patch and mask with pixel
masking. In the second stage, the accuracy results are calculated by the
existing white-box certified defense methods with the estimated patch size. The
experiments conducted on two popular model architectures and two datasets
verify the effectiveness and efficiency of IBCD.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 12:43:04 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yang",
"Di",
""
],
[
"Huang",
"Yihao",
""
],
[
"Guo",
"Qing",
""
],
[
"Juefei-Xu",
"Felix",
""
],
[
"Hu",
"Ming",
""
],
[
"Liu",
"Yang",
""
],
[
"Pu",
"Geguang",
""
]
] | TITLE: Architecture-agnostic Iterative Black-box Certified Defense against
Adversarial Patches
ABSTRACT: The adversarial patch attack aims to fool image classifiers within a bounded,
contiguous region of arbitrary changes, posing a real threat to computer vision
systems (e.g., autonomous driving, content moderation, biometric
authentication, medical imaging) in the physical world. To address this problem
in a trustworthy way, proposals have been made for certified patch defenses
that ensure the robustness of classification models and prevent future patch
attacks from breaching the defense. State-of-the-art certified defenses can be
compatible with any model architecture, as well as achieve high clean and
certified accuracy. Although the methods are adaptive to arbitrary patch
positions, they inevitably need to access the size of the adversarial patch,
which is unreasonable and impractical in real-world attack scenarios. To
improve the feasibility of the architecture-agnostic certified defense in a
black-box setting (i.e., position and size of the patch are both unknown), we
propose a novel two-stage Iterative Black-box Certified Defense method, termed
IBCD.In the first stage, it estimates the patch size in a search-based manner
by evaluating the size relationship between the patch and mask with pixel
masking. In the second stage, the accuracy results are calculated by the
existing white-box certified defense methods with the estimated patch size. The
experiments conducted on two popular model architectures and two datasets
verify the effectiveness and efficiency of IBCD.
| no_new_dataset | 0.710396 |
2305.10975 | Samiul Based Shuvo | Syed Samiul Alam, Samiul Based Shuvo, Shams Nafisa Ali, Fardeen Ahmed,
Arbil Chakma, Yeong Min Jang | Benchmarking Deep Learning Frameworks for Automated Diagnosis of Ocular
Toxoplasmosis: A Comprehensive Approach to Classification and Segmentation | null | null | null | null | eess.IV cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Ocular Toxoplasmosis (OT), is a common eye infection caused by T. gondii that
can cause vision problems. Diagnosis is typically done through a clinical
examination and imaging, but these methods can be complicated and costly,
requiring trained personnel. To address this issue, we have created a benchmark
study that evaluates the effectiveness of existing pre-trained networks using
transfer learning techniques to detect OT from fundus images. Furthermore, we
have also analysed the performance of transfer-learning based segmentation
networks to segment lesions in the images. This research seeks to provide a
guide for future researchers looking to utilise DL techniques and develop a
cheap, automated, easy-to-use, and accurate diagnostic method. We have
performed in-depth analysis of different feature extraction techniques in order
to find the most optimal one for OT classification and segmentation of lesions.
For classification tasks, we have evaluated pre-trained models such as VGG16,
MobileNetV2, InceptionV3, ResNet50, and DenseNet121 models. Among them,
MobileNetV2 outperformed all other models in terms of Accuracy (Acc), Recall,
and F1 Score outperforming the second-best model, InceptionV3 by 0.7% higher
Acc. However, DenseNet121 achieved the best result in terms of Precision, which
was 0.1% higher than MobileNetv2. For the segmentation task, this work has
exploited U-Net architecture. In order to utilize transfer learning the encoder
block of the traditional U-Net was replaced by MobileNetV2, InceptionV3,
ResNet34, and VGG16 to evaluate different architectures moreover two different
two different loss functions (Dice loss and Jaccard loss) were exploited in
order to find the most optimal one. The MobileNetV2/U-Net outperformed ResNet34
by 0.5% and 2.1% in terms of Acc and Dice Score, respectively when Jaccard loss
function is employed during the training.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 13:42:15 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Alam",
"Syed Samiul",
""
],
[
"Shuvo",
"Samiul Based",
""
],
[
"Ali",
"Shams Nafisa",
""
],
[
"Ahmed",
"Fardeen",
""
],
[
"Chakma",
"Arbil",
""
],
[
"Jang",
"Yeong Min",
""
]
] | TITLE: Benchmarking Deep Learning Frameworks for Automated Diagnosis of Ocular
Toxoplasmosis: A Comprehensive Approach to Classification and Segmentation
ABSTRACT: Ocular Toxoplasmosis (OT), is a common eye infection caused by T. gondii that
can cause vision problems. Diagnosis is typically done through a clinical
examination and imaging, but these methods can be complicated and costly,
requiring trained personnel. To address this issue, we have created a benchmark
study that evaluates the effectiveness of existing pre-trained networks using
transfer learning techniques to detect OT from fundus images. Furthermore, we
have also analysed the performance of transfer-learning based segmentation
networks to segment lesions in the images. This research seeks to provide a
guide for future researchers looking to utilise DL techniques and develop a
cheap, automated, easy-to-use, and accurate diagnostic method. We have
performed in-depth analysis of different feature extraction techniques in order
to find the most optimal one for OT classification and segmentation of lesions.
For classification tasks, we have evaluated pre-trained models such as VGG16,
MobileNetV2, InceptionV3, ResNet50, and DenseNet121 models. Among them,
MobileNetV2 outperformed all other models in terms of Accuracy (Acc), Recall,
and F1 Score outperforming the second-best model, InceptionV3 by 0.7% higher
Acc. However, DenseNet121 achieved the best result in terms of Precision, which
was 0.1% higher than MobileNetv2. For the segmentation task, this work has
exploited U-Net architecture. In order to utilize transfer learning the encoder
block of the traditional U-Net was replaced by MobileNetV2, InceptionV3,
ResNet34, and VGG16 to evaluate different architectures moreover two different
two different loss functions (Dice loss and Jaccard loss) were exploited in
order to find the most optimal one. The MobileNetV2/U-Net outperformed ResNet34
by 0.5% and 2.1% in terms of Acc and Dice Score, respectively when Jaccard loss
function is employed during the training.
| no_new_dataset | 0.712282 |
2305.10985 | Elisa Bassignana | Elisa Bassignana, Filip Ginter, Sampo Pyysalo, Rob van der Goot, and
Barbara Plank | Multi-CrossRE A Multi-Lingual Multi-Domain Dataset for Relation
Extraction | Accepted at NoDaLiDa 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Most research in Relation Extraction (RE) involves the English language,
mainly due to the lack of multi-lingual resources. We propose Multi-CrossRE,
the broadest multi-lingual dataset for RE, including 26 languages in addition
to English, and covering six text domains. Multi-CrossRE is a machine
translated version of CrossRE (Bassignana and Plank, 2022), with a sub-portion
including more than 200 sentences in seven diverse languages checked by native
speakers. We run a baseline model over the 26 new datasets and--as sanity
check--over the 26 back-translations to English. Results on the back-translated
data are consistent with the ones on the original English CrossRE, indicating
high quality of the translation and the resulting dataset.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:01:33 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Bassignana",
"Elisa",
""
],
[
"Ginter",
"Filip",
""
],
[
"Pyysalo",
"Sampo",
""
],
[
"van der Goot",
"Rob",
""
],
[
"Plank",
"Barbara",
""
]
] | TITLE: Multi-CrossRE A Multi-Lingual Multi-Domain Dataset for Relation
Extraction
ABSTRACT: Most research in Relation Extraction (RE) involves the English language,
mainly due to the lack of multi-lingual resources. We propose Multi-CrossRE,
the broadest multi-lingual dataset for RE, including 26 languages in addition
to English, and covering six text domains. Multi-CrossRE is a machine
translated version of CrossRE (Bassignana and Plank, 2022), with a sub-portion
including more than 200 sentences in seven diverse languages checked by native
speakers. We run a baseline model over the 26 new datasets and--as sanity
check--over the 26 back-translations to English. Results on the back-translated
data are consistent with the ones on the original English CrossRE, indicating
high quality of the translation and the resulting dataset.
| new_dataset | 0.715008 |
2305.10987 | Henrique Branquinho | Henrique Branquinho, Nuno Louren\c{c}o, Ernesto Costa | SPENSER: Towards a NeuroEvolutionary Approach for Convolutional Spiking
Neural Networks | null | null | 10.1145/3583133.3596399 | null | cs.NE cs.LG | http://creativecommons.org/licenses/by/4.0/ | Spiking Neural Networks (SNNs) have attracted recent interest due to their
energy efficiency and biological plausibility. However, the performance of SNNs
still lags behind traditional Artificial Neural Networks (ANNs), as there is no
consensus on the best learning algorithm for SNNs. Best-performing SNNs are
based on ANN to SNN conversion or learning with spike-based backpropagation
through surrogate gradients. The focus of recent research has been on
developing and testing different learning strategies, with hand-tailored
architectures and parameter tuning. Neuroevolution (NE), has proven successful
as a way to automatically design ANNs and tune parameters, but its applications
to SNNs are still at an early stage. DENSER is a NE framework for the automatic
design and parametrization of ANNs, based on the principles of Genetic
Algorithms (GA) and Structured Grammatical Evolution (SGE). In this paper, we
propose SPENSER, a NE framework for SNN generation based on DENSER, for image
classification on the MNIST and Fashion-MNIST datasets. SPENSER generates
competitive performing networks with a test accuracy of 99.42% and 91.65%
respectively.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:06:37 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Branquinho",
"Henrique",
""
],
[
"Lourenço",
"Nuno",
""
],
[
"Costa",
"Ernesto",
""
]
] | TITLE: SPENSER: Towards a NeuroEvolutionary Approach for Convolutional Spiking
Neural Networks
ABSTRACT: Spiking Neural Networks (SNNs) have attracted recent interest due to their
energy efficiency and biological plausibility. However, the performance of SNNs
still lags behind traditional Artificial Neural Networks (ANNs), as there is no
consensus on the best learning algorithm for SNNs. Best-performing SNNs are
based on ANN to SNN conversion or learning with spike-based backpropagation
through surrogate gradients. The focus of recent research has been on
developing and testing different learning strategies, with hand-tailored
architectures and parameter tuning. Neuroevolution (NE), has proven successful
as a way to automatically design ANNs and tune parameters, but its applications
to SNNs are still at an early stage. DENSER is a NE framework for the automatic
design and parametrization of ANNs, based on the principles of Genetic
Algorithms (GA) and Structured Grammatical Evolution (SGE). In this paper, we
propose SPENSER, a NE framework for SNN generation based on DENSER, for image
classification on the MNIST and Fashion-MNIST datasets. SPENSER generates
competitive performing networks with a test accuracy of 99.42% and 91.65%
respectively.
| no_new_dataset | 0.709636 |
2305.10991 | Luca Herranz-Celotti | Luca Herranz-Celotti, Ermal Rrapaj | Less is More! A slim architecture for optimal language translation | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The softmax attention mechanism has emerged as a noteworthy development in
the field of Artificial Intelligence research, building on the successes of
Transformer-based architectures. However, their ever increasing sizes
necessitate ever increasing computational memory, that limits their usage. We
propose KgV, a sigmoid gating mechanism that, in conjunction with softmax
attention, significantly boosts performance without increasing architecture
size. To amend the size requirements, we leverage Tensor Chains to identify and
prune the excess parameters. We find that such excess resides primarily within
the embedding layer, and not in the output linear layer. To further improve
embedding and significantly reduce parameters, we introduce H-SoftPOS, a
hierarchical embedding layer which simultaneously enhances performance.
Remarkably, on the WMT14 English-German validation set, our approach yields a
threefold reduction in perplexity, surpassing the current state-of-the-art,
while reducing parameter counts also by a factor of 3. When we further reduce
the number of parameters up to sevenfold, we can still achieve a 21\% decrease
in perplexity with respect to the baseline Transformer. To understand
generalization capabilities, we conduct experiments on the 7 language pairs of
the WMT17 dataset. Our method outperforms existing techniques in terms of test
loss while simultaneously halving the number of parameters. Moreover, we
observe a 70 times reduction in variance with respect to the prior
state-of-the-art. In conclusion, our proposed method yields significant
improvements in performance and much lower memory cost. We call the resulting
architecture Anthe.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:09:52 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Herranz-Celotti",
"Luca",
""
],
[
"Rrapaj",
"Ermal",
""
]
] | TITLE: Less is More! A slim architecture for optimal language translation
ABSTRACT: The softmax attention mechanism has emerged as a noteworthy development in
the field of Artificial Intelligence research, building on the successes of
Transformer-based architectures. However, their ever increasing sizes
necessitate ever increasing computational memory, that limits their usage. We
propose KgV, a sigmoid gating mechanism that, in conjunction with softmax
attention, significantly boosts performance without increasing architecture
size. To amend the size requirements, we leverage Tensor Chains to identify and
prune the excess parameters. We find that such excess resides primarily within
the embedding layer, and not in the output linear layer. To further improve
embedding and significantly reduce parameters, we introduce H-SoftPOS, a
hierarchical embedding layer which simultaneously enhances performance.
Remarkably, on the WMT14 English-German validation set, our approach yields a
threefold reduction in perplexity, surpassing the current state-of-the-art,
while reducing parameter counts also by a factor of 3. When we further reduce
the number of parameters up to sevenfold, we can still achieve a 21\% decrease
in perplexity with respect to the baseline Transformer. To understand
generalization capabilities, we conduct experiments on the 7 language pairs of
the WMT17 dataset. Our method outperforms existing techniques in terms of test
loss while simultaneously halving the number of parameters. Moreover, we
observe a 70 times reduction in variance with respect to the prior
state-of-the-art. In conclusion, our proposed method yields significant
improvements in performance and much lower memory cost. We call the resulting
architecture Anthe.
| no_new_dataset | 0.708799 |
2305.10992 | Atsuki Yamaguchi | Atsuki Yamaguchi, Hiroaki Ozaki, Terufumi Morishita, Gaku Morio,
Yasuhiro Sogawa | How does the task complexity of masked pretraining objectives affect
downstream performance? | Accepted at ACL 2023 Findings | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Masked language modeling (MLM) is a widely used self-supervised pretraining
objective, where a model needs to predict an original token that is replaced
with a mask given contexts. Although simpler and computationally efficient
pretraining objectives, e.g., predicting the first character of a masked token,
have recently shown comparable results to MLM, no objectives with a masking
scheme actually outperform it in downstream tasks. Motivated by the assumption
that their lack of complexity plays a vital role in the degradation, we
validate whether more complex masked objectives can achieve better results and
investigate how much complexity they should have to perform comparably to MLM.
Our results using GLUE, SQuAD, and Universal Dependencies benchmarks
demonstrate that more complicated objectives tend to show better downstream
results with at least half of the MLM complexity needed to perform comparably
to MLM. Finally, we discuss how we should pretrain a model using a masked
objective from the task complexity perspective.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:11:57 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Yamaguchi",
"Atsuki",
""
],
[
"Ozaki",
"Hiroaki",
""
],
[
"Morishita",
"Terufumi",
""
],
[
"Morio",
"Gaku",
""
],
[
"Sogawa",
"Yasuhiro",
""
]
] | TITLE: How does the task complexity of masked pretraining objectives affect
downstream performance?
ABSTRACT: Masked language modeling (MLM) is a widely used self-supervised pretraining
objective, where a model needs to predict an original token that is replaced
with a mask given contexts. Although simpler and computationally efficient
pretraining objectives, e.g., predicting the first character of a masked token,
have recently shown comparable results to MLM, no objectives with a masking
scheme actually outperform it in downstream tasks. Motivated by the assumption
that their lack of complexity plays a vital role in the degradation, we
validate whether more complex masked objectives can achieve better results and
investigate how much complexity they should have to perform comparably to MLM.
Our results using GLUE, SQuAD, and Universal Dependencies benchmarks
demonstrate that more complicated objectives tend to show better downstream
results with at least half of the MLM complexity needed to perform comparably
to MLM. Finally, we discuss how we should pretrain a model using a masked
objective from the task complexity perspective.
| no_new_dataset | 0.707367 |
2305.11012 | Hyungseob Shin | Hyungseob Shin, Hyeongyu Kim, Sewon Kim, Yohan Jun, Taejoon Eo, Dosik
Hwang | SDC-UDA: Volumetric Unsupervised Domain Adaptation Framework for
Slice-Direction Continuous Cross-Modality Medical Image Segmentation | 10 pages, 7 figures, CVPR 2023 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advances in deep learning-based medical image segmentation studies
achieve nearly human-level performance in fully supervised manner. However,
acquiring pixel-level expert annotations is extremely expensive and laborious
in medical imaging fields. Unsupervised domain adaptation (UDA) can alleviate
this problem, which makes it possible to use annotated data in one imaging
modality to train a network that can successfully perform segmentation on
target imaging modality with no labels. In this work, we propose SDC-UDA, a
simple yet effective volumetric UDA framework for slice-direction continuous
cross-modality medical image segmentation which combines intra- and inter-slice
self-attentive image translation, uncertainty-constrained pseudo-label
refinement, and volumetric self-training. Our method is distinguished from
previous methods on UDA for medical image segmentation in that it can obtain
continuous segmentation in the slice direction, thereby ensuring higher
accuracy and potential in clinical practice. We validate SDC-UDA with multiple
publicly available cross-modality medical image segmentation datasets and
achieve state-of-the-art segmentation performance, not to mention the superior
slice-direction continuity of prediction compared to previous studies.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:44:27 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Shin",
"Hyungseob",
""
],
[
"Kim",
"Hyeongyu",
""
],
[
"Kim",
"Sewon",
""
],
[
"Jun",
"Yohan",
""
],
[
"Eo",
"Taejoon",
""
],
[
"Hwang",
"Dosik",
""
]
] | TITLE: SDC-UDA: Volumetric Unsupervised Domain Adaptation Framework for
Slice-Direction Continuous Cross-Modality Medical Image Segmentation
ABSTRACT: Recent advances in deep learning-based medical image segmentation studies
achieve nearly human-level performance in fully supervised manner. However,
acquiring pixel-level expert annotations is extremely expensive and laborious
in medical imaging fields. Unsupervised domain adaptation (UDA) can alleviate
this problem, which makes it possible to use annotated data in one imaging
modality to train a network that can successfully perform segmentation on
target imaging modality with no labels. In this work, we propose SDC-UDA, a
simple yet effective volumetric UDA framework for slice-direction continuous
cross-modality medical image segmentation which combines intra- and inter-slice
self-attentive image translation, uncertainty-constrained pseudo-label
refinement, and volumetric self-training. Our method is distinguished from
previous methods on UDA for medical image segmentation in that it can obtain
continuous segmentation in the slice direction, thereby ensuring higher
accuracy and potential in clinical practice. We validate SDC-UDA with multiple
publicly available cross-modality medical image segmentation datasets and
achieve state-of-the-art segmentation performance, not to mention the superior
slice-direction continuity of prediction compared to previous studies.
| no_new_dataset | 0.710239 |
2305.11013 | Zhifu Gao | Zhifu Gao, Zerui Li, Jiaming Wang, Haoneng Luo, Xian Shi, Mengzhe
Chen, Yabin Li, Lingyun Zuo, Zhihao Du, Zhangyu Xiao, Shiliang Zhang | FunASR: A Fundamental End-to-End Speech Recognition Toolkit | 5 pages, 3 figures, accepted by INTERSPEECH 2023 | null | null | null | cs.SD cs.CL eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces FunASR, an open-source speech recognition toolkit
designed to bridge the gap between academic research and industrial
applications. FunASR offers models trained on large-scale industrial corpora
and the ability to deploy them in applications. The toolkit's flagship model,
Paraformer, is a non-autoregressive end-to-end speech recognition model that
has been trained on a manually annotated Mandarin speech recognition dataset
that contains 60,000 hours of speech. To improve the performance of Paraformer,
we have added timestamp prediction and hotword customization capabilities to
the standard Paraformer backbone. In addition, to facilitate model deployment,
we have open-sourced a voice activity detection model based on the Feedforward
Sequential Memory Network (FSMN-VAD) and a text post-processing punctuation
model based on the controllable time-delay Transformer (CT-Transformer), both
of which were trained on industrial corpora. These functional modules provide a
solid foundation for building high-precision long audio speech recognition
services. Compared to other models trained on open datasets, Paraformer
demonstrates superior performance.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:45:09 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Gao",
"Zhifu",
""
],
[
"Li",
"Zerui",
""
],
[
"Wang",
"Jiaming",
""
],
[
"Luo",
"Haoneng",
""
],
[
"Shi",
"Xian",
""
],
[
"Chen",
"Mengzhe",
""
],
[
"Li",
"Yabin",
""
],
[
"Zuo",
"Lingyun",
""
],
[
"Du",
"Zhihao",
""
],
[
"Xiao",
"Zhangyu",
""
],
[
"Zhang",
"Shiliang",
""
]
] | TITLE: FunASR: A Fundamental End-to-End Speech Recognition Toolkit
ABSTRACT: This paper introduces FunASR, an open-source speech recognition toolkit
designed to bridge the gap between academic research and industrial
applications. FunASR offers models trained on large-scale industrial corpora
and the ability to deploy them in applications. The toolkit's flagship model,
Paraformer, is a non-autoregressive end-to-end speech recognition model that
has been trained on a manually annotated Mandarin speech recognition dataset
that contains 60,000 hours of speech. To improve the performance of Paraformer,
we have added timestamp prediction and hotword customization capabilities to
the standard Paraformer backbone. In addition, to facilitate model deployment,
we have open-sourced a voice activity detection model based on the Feedforward
Sequential Memory Network (FSMN-VAD) and a text post-processing punctuation
model based on the controllable time-delay Transformer (CT-Transformer), both
of which were trained on industrial corpora. These functional modules provide a
solid foundation for building high-precision long audio speech recognition
services. Compared to other models trained on open datasets, Paraformer
demonstrates superior performance.
| no_new_dataset | 0.708799 |
2305.11016 | Elisa Bassignana | Elisa Bassignana, Filip Ginter, Sampo Pyysalo, Rob van der Goot, and
Barbara Plank | Silver Syntax Pre-training for Cross-Domain Relation Extraction | Accepted in Findings of the Association for Computational
Linguistics: ACL 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Relation Extraction (RE) remains a challenging task, especially when
considering realistic out-of-domain evaluations. One of the main reasons for
this is the limited training size of current RE datasets: obtaining
high-quality (manually annotated) data is extremely expensive and cannot
realistically be repeated for each new domain. An intermediate training step on
data from related tasks has shown to be beneficial across many NLP
tasks.However, this setup still requires supplementary annotated data, which is
often not available. In this paper, we investigate intermediate pre-training
specifically for RE. We exploit the affinity between syntactic structure and
semantic RE, and identify the syntactic relations which are closely related to
RE by being on the shortest dependency path between two entities. We then take
advantage of the high accuracy of current syntactic parsers in order to
automatically obtain large amounts of low-cost pre-training data. By
pre-training our RE model on the relevant syntactic relations, we are able to
outperform the baseline in five out of six cross-domain setups, without any
additional annotated data.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:49:19 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Bassignana",
"Elisa",
""
],
[
"Ginter",
"Filip",
""
],
[
"Pyysalo",
"Sampo",
""
],
[
"van der Goot",
"Rob",
""
],
[
"Plank",
"Barbara",
""
]
] | TITLE: Silver Syntax Pre-training for Cross-Domain Relation Extraction
ABSTRACT: Relation Extraction (RE) remains a challenging task, especially when
considering realistic out-of-domain evaluations. One of the main reasons for
this is the limited training size of current RE datasets: obtaining
high-quality (manually annotated) data is extremely expensive and cannot
realistically be repeated for each new domain. An intermediate training step on
data from related tasks has shown to be beneficial across many NLP
tasks.However, this setup still requires supplementary annotated data, which is
often not available. In this paper, we investigate intermediate pre-training
specifically for RE. We exploit the affinity between syntactic structure and
semantic RE, and identify the syntactic relations which are closely related to
RE by being on the shortest dependency path between two entities. We then take
advantage of the high accuracy of current syntactic parsers in order to
automatically obtain large amounts of low-cost pre-training data. By
pre-training our RE model on the relevant syntactic relations, we are able to
outperform the baseline in five out of six cross-domain setups, without any
additional annotated data.
| no_new_dataset | 0.708824 |
2305.11017 | Gang Chen | Gang Chen and Victoria Huang | Deep Metric Tensor Regularized Policy Gradient | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Policy gradient algorithms are an important family of deep reinforcement
learning techniques. Many past research endeavors focused on using the
first-order policy gradient information to train policy networks. Different
from these works, we conduct research in this paper driven by the believe that
properly utilizing and controlling Hessian information associated with the
policy gradient can noticeably improve the performance of policy gradient
algorithms. One key Hessian information that attracted our attention is the
Hessian trace, which gives the divergence of the policy gradient vector field
in the Euclidean policy parametric space. We set the goal to generalize this
Euclidean policy parametric space into a general Riemmanian manifold by
introducing a metric tensor field $g_ab$ in the parametric space. This is
achieved through newly developed mathematical tools, deep learning algorithms,
and metric tensor deep neural networks (DNNs). Armed with these technical
developments, we propose a new policy gradient algorithm that learns to
minimize the absolute divergence in the Riemannian manifold as an important
regularization mechanism, allowing the Riemannian manifold to smoothen its
policy gradient vector field. The newly developed algorithm is experimentally
studied on several benchmark reinforcement learning problems. Our experiments
clearly show that the new metric tensor regularized algorithm can significantly
outperform its counterpart that does not use our regularization technique.
Additional experimental analysis further suggests that the trained metric
tensor DNN and the corresponding metric tensor $g_{ab}$ can effectively reduce
the absolute divergence towards zero in the Riemannian manifold.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 14:50:00 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Chen",
"Gang",
""
],
[
"Huang",
"Victoria",
""
]
] | TITLE: Deep Metric Tensor Regularized Policy Gradient
ABSTRACT: Policy gradient algorithms are an important family of deep reinforcement
learning techniques. Many past research endeavors focused on using the
first-order policy gradient information to train policy networks. Different
from these works, we conduct research in this paper driven by the believe that
properly utilizing and controlling Hessian information associated with the
policy gradient can noticeably improve the performance of policy gradient
algorithms. One key Hessian information that attracted our attention is the
Hessian trace, which gives the divergence of the policy gradient vector field
in the Euclidean policy parametric space. We set the goal to generalize this
Euclidean policy parametric space into a general Riemmanian manifold by
introducing a metric tensor field $g_ab$ in the parametric space. This is
achieved through newly developed mathematical tools, deep learning algorithms,
and metric tensor deep neural networks (DNNs). Armed with these technical
developments, we propose a new policy gradient algorithm that learns to
minimize the absolute divergence in the Riemannian manifold as an important
regularization mechanism, allowing the Riemannian manifold to smoothen its
policy gradient vector field. The newly developed algorithm is experimentally
studied on several benchmark reinforcement learning problems. Our experiments
clearly show that the new metric tensor regularized algorithm can significantly
outperform its counterpart that does not use our regularization technique.
Additional experimental analysis further suggests that the trained metric
tensor DNN and the corresponding metric tensor $g_{ab}$ can effectively reduce
the absolute divergence towards zero in the Riemannian manifold.
| no_new_dataset | 0.710057 |
2305.11023 | Harshil Shah | Harshil Shah, Arthur Wilcke, Marius Cobzarenco, Cristi Cobzarenco,
Edward Challis, David Barber | Generalized Multiple Intent Conditioned Slot Filling | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Natural language understanding includes the tasks of intent detection
(identifying a user's objectives) and slot filling (extracting the entities
relevant to those objectives). Prior slot filling methods assume that each
intent type cannot occur more than once within a message, however this is often
not a valid assumption for real-world settings. In this work, we generalize
slot filling by removing the constraint of unique intents in a message. We cast
this as a JSON generation task and approach it using a language model. We
create a pre-training dataset by combining DBpedia and existing slot filling
datasets that we convert for JSON generation. We also generate an in-domain
dataset using GPT-3. We train T5 models for this task (with and without
exemplars in the prompt) and find that both training datasets improve
performance, and that the model is able to generalize to intent types not seen
during training.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 15:04:52 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Shah",
"Harshil",
""
],
[
"Wilcke",
"Arthur",
""
],
[
"Cobzarenco",
"Marius",
""
],
[
"Cobzarenco",
"Cristi",
""
],
[
"Challis",
"Edward",
""
],
[
"Barber",
"David",
""
]
] | TITLE: Generalized Multiple Intent Conditioned Slot Filling
ABSTRACT: Natural language understanding includes the tasks of intent detection
(identifying a user's objectives) and slot filling (extracting the entities
relevant to those objectives). Prior slot filling methods assume that each
intent type cannot occur more than once within a message, however this is often
not a valid assumption for real-world settings. In this work, we generalize
slot filling by removing the constraint of unique intents in a message. We cast
this as a JSON generation task and approach it using a language model. We
create a pre-training dataset by combining DBpedia and existing slot filling
datasets that we convert for JSON generation. We also generate an in-domain
dataset using GPT-3. We train T5 models for this task (with and without
exemplars in the prompt) and find that both training datasets improve
performance, and that the model is able to generalize to intent types not seen
during training.
| new_dataset | 0.71282 |
2305.11031 | Shoukang Hu | Shoukang Hu and Kaichen Zhou and Kaiyu Li and Longhui Yu and Lanqing
Hong and Tianyang Hu and Zhenguo Li and Gim Hee Lee and Ziwei Liu | ConsistentNeRF: Enhancing Neural Radiance Fields with 3D Consistency for
Sparse View Synthesis | https://github.com/skhu101/ConsistentNeRF | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural Radiance Fields (NeRF) has demonstrated remarkable 3D reconstruction
capabilities with dense view images. However, its performance significantly
deteriorates under sparse view settings. We observe that learning the 3D
consistency of pixels among different views is crucial for improving
reconstruction quality in such cases. In this paper, we propose ConsistentNeRF,
a method that leverages depth information to regularize both multi-view and
single-view 3D consistency among pixels. Specifically, ConsistentNeRF employs
depth-derived geometry information and a depth-invariant loss to concentrate on
pixels that exhibit 3D correspondence and maintain consistent depth
relationships. Extensive experiments on recent representative works reveal that
our approach can considerably enhance model performance in sparse view
conditions, achieving improvements of up to 94% in PSNR, 76% in SSIM, and 31%
in LPIPS compared to the vanilla baselines across various benchmarks, including
DTU, NeRF Synthetic, and LLFF.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 15:18:01 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Hu",
"Shoukang",
""
],
[
"Zhou",
"Kaichen",
""
],
[
"Li",
"Kaiyu",
""
],
[
"Yu",
"Longhui",
""
],
[
"Hong",
"Lanqing",
""
],
[
"Hu",
"Tianyang",
""
],
[
"Li",
"Zhenguo",
""
],
[
"Lee",
"Gim Hee",
""
],
[
"Liu",
"Ziwei",
""
]
] | TITLE: ConsistentNeRF: Enhancing Neural Radiance Fields with 3D Consistency for
Sparse View Synthesis
ABSTRACT: Neural Radiance Fields (NeRF) has demonstrated remarkable 3D reconstruction
capabilities with dense view images. However, its performance significantly
deteriorates under sparse view settings. We observe that learning the 3D
consistency of pixels among different views is crucial for improving
reconstruction quality in such cases. In this paper, we propose ConsistentNeRF,
a method that leverages depth information to regularize both multi-view and
single-view 3D consistency among pixels. Specifically, ConsistentNeRF employs
depth-derived geometry information and a depth-invariant loss to concentrate on
pixels that exhibit 3D correspondence and maintain consistent depth
relationships. Extensive experiments on recent representative works reveal that
our approach can considerably enhance model performance in sparse view
conditions, achieving improvements of up to 94% in PSNR, 76% in SSIM, and 31%
in LPIPS compared to the vanilla baselines across various benchmarks, including
DTU, NeRF Synthetic, and LLFF.
| no_new_dataset | 0.711825 |
2305.11034 | Samuel Mensah | Samuel Mensah, Kai Sun, Nikolaos Aletras | Trading Syntax Trees for Wordpieces: Target-oriented Opinion Words
Extraction with Wordpieces and Aspect Enhancement | Accepted at ACL 2023 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | State-of-the-art target-oriented opinion word extraction (TOWE) models
typically use BERT-based text encoders that operate on the word level, along
with graph convolutional networks (GCNs) that incorporate syntactic information
extracted from syntax trees. These methods achieve limited gains with GCNs and
have difficulty using BERT wordpieces. Meanwhile, BERT wordpieces are known to
be effective at representing rare words or words with insufficient context
information. To address this issue, this work trades syntax trees for BERT
wordpieces by entirely removing the GCN component from the methods'
architectures. To enhance TOWE performance, we tackle the issue of aspect
representation loss during encoding. Instead of solely utilizing a sentence as
the input, we use a sentence-aspect pair. Our relatively simple approach
achieves state-of-the-art results on benchmark datasets and should serve as a
strong baseline for further research.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 15:22:00 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Mensah",
"Samuel",
""
],
[
"Sun",
"Kai",
""
],
[
"Aletras",
"Nikolaos",
""
]
] | TITLE: Trading Syntax Trees for Wordpieces: Target-oriented Opinion Words
Extraction with Wordpieces and Aspect Enhancement
ABSTRACT: State-of-the-art target-oriented opinion word extraction (TOWE) models
typically use BERT-based text encoders that operate on the word level, along
with graph convolutional networks (GCNs) that incorporate syntactic information
extracted from syntax trees. These methods achieve limited gains with GCNs and
have difficulty using BERT wordpieces. Meanwhile, BERT wordpieces are known to
be effective at representing rare words or words with insufficient context
information. To address this issue, this work trades syntax trees for BERT
wordpieces by entirely removing the GCN component from the methods'
architectures. To enhance TOWE performance, we tackle the issue of aspect
representation loss during encoding. Instead of solely utilizing a sentence as
the input, we use a sentence-aspect pair. Our relatively simple approach
achieves state-of-the-art results on benchmark datasets and should serve as a
strong baseline for further research.
| no_new_dataset | 0.71227 |
2305.11044 | Ryan Boldi | Ryan Boldi, Aadam Lokhandwala, Edward Annatone, Yuval Schechter,
Alexander Lavrenenko, Cooper Sigrist | Improving Recommendation System Serendipity Through Lexicase Selection | 9 pages | null | null | null | cs.IR cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recommender systems influence almost every aspect of our digital lives.
Unfortunately, in striving to give us what we want, they end up restricting our
open-mindedness. Current recommender systems promote echo chambers, where
people only see the information they want to see, and homophily, where users of
similar background see similar content. We propose a new serendipity metric to
measure the presence of echo chambers and homophily in recommendation systems
using cluster analysis. We then attempt to improve the diversity-preservation
qualities of well known recommendation techniques by adopting a parent
selection algorithm from the evolutionary computation literature known as
lexicase selection. Our results show that lexicase selection, or a mixture of
lexicase selection and ranking, outperforms its purely ranked counterparts in
terms of personalization, coverage and our specifically designed serendipity
benchmark, while only slightly under-performing in terms of accuracy (hit
rate). We verify these results across a variety of recommendation list sizes.
In this work we show that lexicase selection is able to maintain multiple
diverse clusters of item recommendations that are each relevant for the
specific user, while still maintaining a high hit-rate accuracy, a trade off
that is not achieved by other methods.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 15:37:38 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Boldi",
"Ryan",
""
],
[
"Lokhandwala",
"Aadam",
""
],
[
"Annatone",
"Edward",
""
],
[
"Schechter",
"Yuval",
""
],
[
"Lavrenenko",
"Alexander",
""
],
[
"Sigrist",
"Cooper",
""
]
] | TITLE: Improving Recommendation System Serendipity Through Lexicase Selection
ABSTRACT: Recommender systems influence almost every aspect of our digital lives.
Unfortunately, in striving to give us what we want, they end up restricting our
open-mindedness. Current recommender systems promote echo chambers, where
people only see the information they want to see, and homophily, where users of
similar background see similar content. We propose a new serendipity metric to
measure the presence of echo chambers and homophily in recommendation systems
using cluster analysis. We then attempt to improve the diversity-preservation
qualities of well known recommendation techniques by adopting a parent
selection algorithm from the evolutionary computation literature known as
lexicase selection. Our results show that lexicase selection, or a mixture of
lexicase selection and ranking, outperforms its purely ranked counterparts in
terms of personalization, coverage and our specifically designed serendipity
benchmark, while only slightly under-performing in terms of accuracy (hit
rate). We verify these results across a variety of recommendation list sizes.
In this work we show that lexicase selection is able to maintain multiple
diverse clusters of item recommendations that are each relevant for the
specific user, while still maintaining a high hit-rate accuracy, a trade off
that is not achieved by other methods.
| no_new_dataset | 0.701112 |
2305.11052 | Shicheng Xu | Shicheng Xu, Liang Pang, Huawei Shen, Xueqi Cheng | BERM: Training the Balanced and Extractable Representation for Matching
to Improve Generalization Ability of Dense Retrieval | Accepted by ACL 2023 Main | null | null | null | cs.IR cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dense retrieval has shown promise in the first-stage retrieval process when
trained on in-domain labeled datasets. However, previous studies have found
that dense retrieval is hard to generalize to unseen domains due to its weak
modeling of domain-invariant and interpretable feature (i.e., matching signal
between two texts, which is the essence of information retrieval). In this
paper, we propose a novel method to improve the generalization of dense
retrieval via capturing matching signal called BERM. Fully fine-grained
expression and query-oriented saliency are two properties of the matching
signal. Thus, in BERM, a single passage is segmented into multiple units and
two unit-level requirements are proposed for representation as the constraint
in training to obtain the effective matching signal. One is semantic unit
balance and the other is essential matching unit extractability. Unit-level
view and balanced semantics make representation express the text in a
fine-grained manner. Essential matching unit extractability makes passage
representation sensitive to the given query to extract the pure matching
information from the passage containing complex context. Experiments on BEIR
show that our method can be effectively combined with different dense retrieval
training methods (vanilla, hard negatives mining and knowledge distillation) to
improve its generalization ability without any additional inference overhead
and target domain data.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 15:43:09 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Xu",
"Shicheng",
""
],
[
"Pang",
"Liang",
""
],
[
"Shen",
"Huawei",
""
],
[
"Cheng",
"Xueqi",
""
]
] | TITLE: BERM: Training the Balanced and Extractable Representation for Matching
to Improve Generalization Ability of Dense Retrieval
ABSTRACT: Dense retrieval has shown promise in the first-stage retrieval process when
trained on in-domain labeled datasets. However, previous studies have found
that dense retrieval is hard to generalize to unseen domains due to its weak
modeling of domain-invariant and interpretable feature (i.e., matching signal
between two texts, which is the essence of information retrieval). In this
paper, we propose a novel method to improve the generalization of dense
retrieval via capturing matching signal called BERM. Fully fine-grained
expression and query-oriented saliency are two properties of the matching
signal. Thus, in BERM, a single passage is segmented into multiple units and
two unit-level requirements are proposed for representation as the constraint
in training to obtain the effective matching signal. One is semantic unit
balance and the other is essential matching unit extractability. Unit-level
view and balanced semantics make representation express the text in a
fine-grained manner. Essential matching unit extractability makes passage
representation sensitive to the given query to extract the pure matching
information from the passage containing complex context. Experiments on BEIR
show that our method can be effectively combined with different dense retrieval
training methods (vanilla, hard negatives mining and knowledge distillation) to
improve its generalization ability without any additional inference overhead
and target domain data.
| no_new_dataset | 0.711644 |
2305.11061 | Han Jiang | Ran Shen, Gang Sun, Hao Shen, Yiling Li, Liangfeng Jin and Han Jiang | SPSQL: Step-by-step Parsing Based Framework for Text-to-SQL Generation | 8 pages, 6 figures | null | null | null | cs.CL cs.AI cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Converting text into the structured query language (Text2SQL) is a research
hotspot in the field of natural language processing (NLP), which has broad
application prospects. In the era of big data, the use of databases has
penetrated all walks of life, in which the collected data is large in scale,
diverse in variety, and wide in scope, making the data query cumbersome and
inefficient, and putting forward higher requirements for the Text2SQL model. In
practical applications, the current mainstream end-to-end Text2SQL model is not
only difficult to build due to its complex structure and high requirements for
training data, but also difficult to adjust due to massive parameters. In
addition, the accuracy of the model is hard to achieve the desired result.
Based on this, this paper proposes a pipelined Text2SQL method: SPSQL. This
method disassembles the Text2SQL task into four subtasks--table selection,
column selection, SQL generation, and value filling, which can be converted
into a text classification problem, a sequence labeling problem, and two text
generation problems, respectively. Then, we construct data formats of different
subtasks based on existing data and improve the accuracy of the overall model
by improving the accuracy of each submodel. We also use the named entity
recognition module and data augmentation to optimize the overall model. We
construct the dataset based on the marketing business data of the State Grid
Corporation of China. Experiments demonstrate our proposed method achieves the
best performance compared with the end-to-end method and other pipeline
methods.
| [
{
"version": "v1",
"created": "Wed, 10 May 2023 10:01:36 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Shen",
"Ran",
""
],
[
"Sun",
"Gang",
""
],
[
"Shen",
"Hao",
""
],
[
"Li",
"Yiling",
""
],
[
"Jin",
"Liangfeng",
""
],
[
"Jiang",
"Han",
""
]
] | TITLE: SPSQL: Step-by-step Parsing Based Framework for Text-to-SQL Generation
ABSTRACT: Converting text into the structured query language (Text2SQL) is a research
hotspot in the field of natural language processing (NLP), which has broad
application prospects. In the era of big data, the use of databases has
penetrated all walks of life, in which the collected data is large in scale,
diverse in variety, and wide in scope, making the data query cumbersome and
inefficient, and putting forward higher requirements for the Text2SQL model. In
practical applications, the current mainstream end-to-end Text2SQL model is not
only difficult to build due to its complex structure and high requirements for
training data, but also difficult to adjust due to massive parameters. In
addition, the accuracy of the model is hard to achieve the desired result.
Based on this, this paper proposes a pipelined Text2SQL method: SPSQL. This
method disassembles the Text2SQL task into four subtasks--table selection,
column selection, SQL generation, and value filling, which can be converted
into a text classification problem, a sequence labeling problem, and two text
generation problems, respectively. Then, we construct data formats of different
subtasks based on existing data and improve the accuracy of the overall model
by improving the accuracy of each submodel. We also use the named entity
recognition module and data augmentation to optimize the overall model. We
construct the dataset based on the marketing business data of the State Grid
Corporation of China. Experiments demonstrate our proposed method achieves the
best performance compared with the end-to-end method and other pipeline
methods.
| no_new_dataset | 0.709416 |
2305.11068 | Salomon Kabongo Kabenamualu | Salomon Kabongo, Jennifer D'Souza and S\"oren Auer | ORKG-Leaderboards: A Systematic Workflow for Mining Leaderboards as a
Knowledge Graph | NA. arXiv admin note: text overlap with arXiv:2109.13089 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The purpose of this work is to describe the Orkg-Leaderboard software
designed to extract leaderboards defined as Task-Dataset-Metric tuples
automatically from large collections of empirical research papers in Artificial
Intelligence (AI). The software can support both the main workflows of
scholarly publishing, viz. as LaTeX files or as PDF files. Furthermore, the
system is integrated with the Open Research Knowledge Graph (ORKG) platform,
which fosters the machine-actionable publishing of scholarly findings. Thus the
system output, when integrated within the ORKG's supported Semantic Web
infrastructure of representing machine-actionable 'resources' on the Web,
enables: 1) broadly, the integration of empirical results of researchers across
the world, thus enabling transparency in empirical research with the potential
to also being complete contingent on the underlying data source(s) of
publications; and 2) specifically, enables researchers to track the progress in
AI with an overview of the state-of-the-art (SOTA) across the most common AI
tasks and their corresponding datasets via dynamic ORKG frontend views
leveraging tables and visualization charts over the machine-actionable data.
Our best model achieves performances above 90% F1 on the \textit{leaderboard}
extraction task, thus proving Orkg-Leaderboards a practically viable tool for
real-world usage. Going forward, in a sense, Orkg-Leaderboards transforms the
leaderboard extraction task to an automated digitalization task, which has
been, for a long time in the community, a crowdsourced endeavor.
| [
{
"version": "v1",
"created": "Wed, 10 May 2023 13:19:18 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Kabongo",
"Salomon",
""
],
[
"D'Souza",
"Jennifer",
""
],
[
"Auer",
"Sören",
""
]
] | TITLE: ORKG-Leaderboards: A Systematic Workflow for Mining Leaderboards as a
Knowledge Graph
ABSTRACT: The purpose of this work is to describe the Orkg-Leaderboard software
designed to extract leaderboards defined as Task-Dataset-Metric tuples
automatically from large collections of empirical research papers in Artificial
Intelligence (AI). The software can support both the main workflows of
scholarly publishing, viz. as LaTeX files or as PDF files. Furthermore, the
system is integrated with the Open Research Knowledge Graph (ORKG) platform,
which fosters the machine-actionable publishing of scholarly findings. Thus the
system output, when integrated within the ORKG's supported Semantic Web
infrastructure of representing machine-actionable 'resources' on the Web,
enables: 1) broadly, the integration of empirical results of researchers across
the world, thus enabling transparency in empirical research with the potential
to also being complete contingent on the underlying data source(s) of
publications; and 2) specifically, enables researchers to track the progress in
AI with an overview of the state-of-the-art (SOTA) across the most common AI
tasks and their corresponding datasets via dynamic ORKG frontend views
leveraging tables and visualization charts over the machine-actionable data.
Our best model achieves performances above 90% F1 on the \textit{leaderboard}
extraction task, thus proving Orkg-Leaderboards a practically viable tool for
real-world usage. Going forward, in a sense, Orkg-Leaderboards transforms the
leaderboard extraction task to an automated digitalization task, which has
been, for a long time in the community, a crowdsourced endeavor.
| no_new_dataset | 0.709598 |
2305.11070 | Anna Wr\'oblewska | Albert Roethel, Maria Ganzha, Anna Wr\'oblewska | Enriching language models with graph-based context information to better
understand textual data | 12 pages, 5 figures | null | null | null | cs.CL cs.AI cs.LG cs.NE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A considerable number of texts encountered daily are somehow connected with
each other. For example, Wikipedia articles refer to other articles via
hyperlinks, scientific papers relate to others via citations or (co)authors,
while tweets relate via users that follow each other or reshare content. Hence,
a graph-like structure can represent existing connections and be seen as
capturing the "context" of the texts. The question thus arises if extracting
and integrating such context information into a language model might help
facilitate a better automated understanding of the text. In this study, we
experimentally demonstrate that incorporating graph-based contextualization
into BERT model enhances its performance on an example of a classification
task. Specifically, on Pubmed dataset, we observed a reduction in error from
8.51% to 7.96%, while increasing the number of parameters just by 1.6%.
Our source code: https://github.com/tryptofanik/gc-bert
| [
{
"version": "v1",
"created": "Wed, 10 May 2023 10:57:21 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Roethel",
"Albert",
""
],
[
"Ganzha",
"Maria",
""
],
[
"Wróblewska",
"Anna",
""
]
] | TITLE: Enriching language models with graph-based context information to better
understand textual data
ABSTRACT: A considerable number of texts encountered daily are somehow connected with
each other. For example, Wikipedia articles refer to other articles via
hyperlinks, scientific papers relate to others via citations or (co)authors,
while tweets relate via users that follow each other or reshare content. Hence,
a graph-like structure can represent existing connections and be seen as
capturing the "context" of the texts. The question thus arises if extracting
and integrating such context information into a language model might help
facilitate a better automated understanding of the text. In this study, we
experimentally demonstrate that incorporating graph-based contextualization
into BERT model enhances its performance on an example of a classification
task. Specifically, on Pubmed dataset, we observed a reduction in error from
8.51% to 7.96%, while increasing the number of parameters just by 1.6%.
Our source code: https://github.com/tryptofanik/gc-bert
| no_new_dataset | 0.708584 |
2305.11073 | Yifan Peng | Yifan Peng, Kwangyoun Kim, Felix Wu, Brian Yan, Siddhant Arora,
William Chen, Jiyang Tang, Suwon Shon, Prashant Sridhar, Shinji Watanabe | A Comparative Study on E-Branchformer vs Conformer in Speech
Recognition, Translation, and Understanding Tasks | Accepted at INTERSPEECH 2023. Code: https://github.com/espnet/espnet | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Conformer, a convolution-augmented Transformer variant, has become the de
facto encoder architecture for speech processing due to its superior
performance in various tasks, including automatic speech recognition (ASR),
speech translation (ST) and spoken language understanding (SLU). Recently, a
new encoder called E-Branchformer has outperformed Conformer in the LibriSpeech
ASR benchmark, making it promising for more general speech applications. This
work compares E-Branchformer and Conformer through extensive experiments using
different types of end-to-end sequence-to-sequence models. Results demonstrate
that E-Branchformer achieves comparable or better performance than Conformer in
almost all evaluation sets across 15 ASR, 2 ST, and 3 SLU benchmarks, while
being more stable during training. We will release our training configurations
and pre-trained models for reproducibility, which can benefit the speech
community.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 16:00:48 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Peng",
"Yifan",
""
],
[
"Kim",
"Kwangyoun",
""
],
[
"Wu",
"Felix",
""
],
[
"Yan",
"Brian",
""
],
[
"Arora",
"Siddhant",
""
],
[
"Chen",
"William",
""
],
[
"Tang",
"Jiyang",
""
],
[
"Shon",
"Suwon",
""
],
[
"Sridhar",
"Prashant",
""
],
[
"Watanabe",
"Shinji",
""
]
] | TITLE: A Comparative Study on E-Branchformer vs Conformer in Speech
Recognition, Translation, and Understanding Tasks
ABSTRACT: Conformer, a convolution-augmented Transformer variant, has become the de
facto encoder architecture for speech processing due to its superior
performance in various tasks, including automatic speech recognition (ASR),
speech translation (ST) and spoken language understanding (SLU). Recently, a
new encoder called E-Branchformer has outperformed Conformer in the LibriSpeech
ASR benchmark, making it promising for more general speech applications. This
work compares E-Branchformer and Conformer through extensive experiments using
different types of end-to-end sequence-to-sequence models. Results demonstrate
that E-Branchformer achieves comparable or better performance than Conformer in
almost all evaluation sets across 15 ASR, 2 ST, and 3 SLU benchmarks, while
being more stable during training. We will release our training configurations
and pre-trained models for reproducibility, which can benefit the speech
community.
| no_new_dataset | 0.709202 |
2305.11077 | Danush Shekar | Sahel Mohammad Iqbal, Danush Shekar, Subhankar Mishra | A Comparative Study of Face Detection Algorithms for Masked Face
Detection | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Contemporary face detection algorithms have to deal with many challenges such
as variations in pose, illumination, and scale. A subclass of the face
detection problem that has recently gained increasing attention is occluded
face detection, or more specifically, the detection of masked faces. Three
years on since the advent of the COVID-19 pandemic, there is still a complete
lack of evidence regarding how well existing face detection algorithms perform
on masked faces. This article first offers a brief review of state-of-the-art
face detectors and detectors made for the masked face problem, along with a
review of the existing masked face datasets. We evaluate and compare the
performances of a well-representative set of face detectors at masked face
detection and conclude with a discussion on the possible contributing factors
to their performance.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 16:03:37 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Iqbal",
"Sahel Mohammad",
""
],
[
"Shekar",
"Danush",
""
],
[
"Mishra",
"Subhankar",
""
]
] | TITLE: A Comparative Study of Face Detection Algorithms for Masked Face
Detection
ABSTRACT: Contemporary face detection algorithms have to deal with many challenges such
as variations in pose, illumination, and scale. A subclass of the face
detection problem that has recently gained increasing attention is occluded
face detection, or more specifically, the detection of masked faces. Three
years on since the advent of the COVID-19 pandemic, there is still a complete
lack of evidence regarding how well existing face detection algorithms perform
on masked faces. This article first offers a brief review of state-of-the-art
face detectors and detectors made for the masked face problem, along with a
review of the existing masked face datasets. We evaluate and compare the
performances of a well-representative set of face detectors at masked face
detection and conclude with a discussion on the possible contributing factors
to their performance.
| no_new_dataset | 0.707834 |
2305.11081 | Na Huang | Zhaochun Ren, Na Huang, Yidan Wang, Pengjie Ren, Jun Ma, Jiahuan Lei,
Xinlei Shi, Hengliang Luo, Joemon M Jose, Xin Xin | Contrastive State Augmentations for Reinforcement Learning-Based
Recommender Systems | null | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Learning reinforcement learning (RL)-based recommenders from historical
user-item interaction sequences is vital to generate high-reward
recommendations and improve long-term cumulative benefits. However, existing RL
recommendation methods encounter difficulties (i) to estimate the value
functions for states which are not contained in the offline training data, and
(ii) to learn effective state representations from user implicit feedback due
to the lack of contrastive signals. In this work, we propose contrastive state
augmentations (CSA) for the training of RL-based recommender systems. To tackle
the first issue, we propose four state augmentation strategies to enlarge the
state space of the offline data. The proposed method improves the
generalization capability of the recommender by making the RL agent visit the
local state regions and ensuring the learned value functions are similar
between the original and augmented states. For the second issue, we propose
introducing contrastive signals between augmented states and the state randomly
sampled from other sessions to improve the state representation learning
further. To verify the effectiveness of the proposed CSA, we conduct extensive
experiments on two publicly accessible datasets and one dataset collected from
a real-life e-commerce platform. We also conduct experiments on a simulated
environment as the online evaluation setting. Experimental results demonstrate
that CSA can effectively improve recommendation performance.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 16:08:34 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Ren",
"Zhaochun",
""
],
[
"Huang",
"Na",
""
],
[
"Wang",
"Yidan",
""
],
[
"Ren",
"Pengjie",
""
],
[
"Ma",
"Jun",
""
],
[
"Lei",
"Jiahuan",
""
],
[
"Shi",
"Xinlei",
""
],
[
"Luo",
"Hengliang",
""
],
[
"Jose",
"Joemon M",
""
],
[
"Xin",
"Xin",
""
]
] | TITLE: Contrastive State Augmentations for Reinforcement Learning-Based
Recommender Systems
ABSTRACT: Learning reinforcement learning (RL)-based recommenders from historical
user-item interaction sequences is vital to generate high-reward
recommendations and improve long-term cumulative benefits. However, existing RL
recommendation methods encounter difficulties (i) to estimate the value
functions for states which are not contained in the offline training data, and
(ii) to learn effective state representations from user implicit feedback due
to the lack of contrastive signals. In this work, we propose contrastive state
augmentations (CSA) for the training of RL-based recommender systems. To tackle
the first issue, we propose four state augmentation strategies to enlarge the
state space of the offline data. The proposed method improves the
generalization capability of the recommender by making the RL agent visit the
local state regions and ensuring the learned value functions are similar
between the original and augmented states. For the second issue, we propose
introducing contrastive signals between augmented states and the state randomly
sampled from other sessions to improve the state representation learning
further. To verify the effectiveness of the proposed CSA, we conduct extensive
experiments on two publicly accessible datasets and one dataset collected from
a real-life e-commerce platform. We also conduct experiments on a simulated
environment as the online evaluation setting. Experimental results demonstrate
that CSA can effectively improve recommendation performance.
| no_new_dataset | 0.709818 |
2305.11084 | Chao Wang | Chao Wang, Hengshu Zhu, Dazhong Shen, Wei wu, Hui Xiong | Preference or Intent? Double Disentangled Collaborative Filtering | null | null | null | null | cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | People usually have different intents for choosing items, while their
preferences under the same intent may also different. In traditional
collaborative filtering approaches, both intent and preference factors are
usually entangled in the modeling process, which significantly limits the
robustness and interpretability of recommendation performances. For example,
the low-rating items are always treated as negative feedback while they
actually could provide positive information about user intent. To this end, in
this paper, we propose a two-fold representation learning approach, namely
Double Disentangled Collaborative Filtering (DDCF), for personalized
recommendations. The first-level disentanglement is for separating the
influence factors of intent and preference, while the second-level
disentanglement is performed to build independent sparse preference
representations under individual intent with limited computational complexity.
Specifically, we employ two variational autoencoder networks, intent
recognition network and preference decomposition network, to learn the intent
and preference factors, respectively. In this way, the low-rating items will be
treated as positive samples for modeling intents while the negative samples for
modeling preferences. Finally, extensive experiments on three real-world
datasets and four evaluation metrics clearly validate the effectiveness and the
interpretability of DDCF.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 16:13:41 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Wang",
"Chao",
""
],
[
"Zhu",
"Hengshu",
""
],
[
"Shen",
"Dazhong",
""
],
[
"wu",
"Wei",
""
],
[
"Xiong",
"Hui",
""
]
] | TITLE: Preference or Intent? Double Disentangled Collaborative Filtering
ABSTRACT: People usually have different intents for choosing items, while their
preferences under the same intent may also different. In traditional
collaborative filtering approaches, both intent and preference factors are
usually entangled in the modeling process, which significantly limits the
robustness and interpretability of recommendation performances. For example,
the low-rating items are always treated as negative feedback while they
actually could provide positive information about user intent. To this end, in
this paper, we propose a two-fold representation learning approach, namely
Double Disentangled Collaborative Filtering (DDCF), for personalized
recommendations. The first-level disentanglement is for separating the
influence factors of intent and preference, while the second-level
disentanglement is performed to build independent sparse preference
representations under individual intent with limited computational complexity.
Specifically, we employ two variational autoencoder networks, intent
recognition network and preference decomposition network, to learn the intent
and preference factors, respectively. In this way, the low-rating items will be
treated as positive samples for modeling intents while the negative samples for
modeling preferences. Finally, extensive experiments on three real-world
datasets and four evaluation metrics clearly validate the effectiveness and the
interpretability of DDCF.
| no_new_dataset | 0.711212 |
2305.11089 | Yen Ting Lin | Javier E Santos, Zachary R. Fox, Nicholas Lubbers, Yen Ting Lin | Blackout Diffusion: Generative Diffusion Models in Discrete-State Spaces | 29 pages, 13 figures, 2 tables. Accepted by the 40th International
Conference on Machine Learning, Honolulu, Hawaii, USA | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Typical generative diffusion models rely on a Gaussian diffusion process for
training the backward transformations, which can then be used to generate
samples from Gaussian noise. However, real world data often takes place in
discrete-state spaces, including many scientific applications. Here, we develop
a theoretical formulation for arbitrary discrete-state Markov processes in the
forward diffusion process using exact (as opposed to variational) analysis. We
relate the theory to the existing continuous-state Gaussian diffusion as well
as other approaches to discrete diffusion, and identify the corresponding
reverse-time stochastic process and score function in the continuous-time
setting, and the reverse-time mapping in the discrete-time setting. As an
example of this framework, we introduce ``Blackout Diffusion'', which learns to
produce samples from an empty image instead of from noise. Numerical
experiments on the CIFAR-10, Binarized MNIST, and CelebA datasets confirm the
feasibility of our approach. Generalizing from specific (Gaussian) forward
processes to discrete-state processes without a variational approximation sheds
light on how to interpret diffusion models, which we discuss.
| [
{
"version": "v1",
"created": "Thu, 18 May 2023 16:24:12 GMT"
}
] | 2023-05-19T00:00:00 | [
[
"Santos",
"Javier E",
""
],
[
"Fox",
"Zachary R.",
""
],
[
"Lubbers",
"Nicholas",
""
],
[
"Lin",
"Yen Ting",
""
]
] | TITLE: Blackout Diffusion: Generative Diffusion Models in Discrete-State Spaces
ABSTRACT: Typical generative diffusion models rely on a Gaussian diffusion process for
training the backward transformations, which can then be used to generate
samples from Gaussian noise. However, real world data often takes place in
discrete-state spaces, including many scientific applications. Here, we develop
a theoretical formulation for arbitrary discrete-state Markov processes in the
forward diffusion process using exact (as opposed to variational) analysis. We
relate the theory to the existing continuous-state Gaussian diffusion as well
as other approaches to discrete diffusion, and identify the corresponding
reverse-time stochastic process and score function in the continuous-time
setting, and the reverse-time mapping in the discrete-time setting. As an
example of this framework, we introduce ``Blackout Diffusion'', which learns to
produce samples from an empty image instead of from noise. Numerical
experiments on the CIFAR-10, Binarized MNIST, and CelebA datasets confirm the
feasibility of our approach. Generalizing from specific (Gaussian) forward
processes to discrete-state processes without a variational approximation sheds
light on how to interpret diffusion models, which we discuss.
| no_new_dataset | 0.711443 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.