https://openreview.net/forum?id=VLdqHdp4j1H
Entropy Weighted Adversarial Training | OpenReview
We propose an instance-wise entropy-weighted adversarial training to focus on more uncertain examples.
adversarial trainingentropyweightedopenreview
https://openreview.net/forum?id=H1lZJpVFvr
Robust Local Features for Improving the Generalization of Adversarial Training | OpenReview
We propose a new stream of adversarial training approach called Robust Local Features for Adversarial Training (RLFAT) that significantly improves both the...
local featuresadversarial trainingrobustimproving
https://www.frontiersin.org/journals/neurorobotics/articles/10.3389/fnbot.2022.859610/full
Frontiers | Generative Adversarial Training for Supervised and Semi-supervised Learning
Neural networks have played critical roles in many research fields. The recently proposed adversarial training (AT) can improve the generalization ability of...
adversarial trainingfrontiersgenerativesupervisedsemi
https://openreview.net/forum?id=BygzbyHFvB&ref=ruder.io
FreeLB: Enhanced Adversarial Training for Natural Language Understanding | OpenReview
Adversarial training, which minimizes the maximal risk for label-preserving input perturbations, has proved to be effective for improving the generalization of...
training for natural languageenhancedadversarialunderstandingopenreview
https://deepai.org/publication/on-the-generalization-properties-of-adversarial-training
On the Generalization Properties of Adversarial Training | DeepAI
Aug 15, 2020 - 08/15/20 - Modern machine learning and deep learning models are shown to be vulnerable when testing data are slightly perturbed. Theoretical ...
on theadversarial traininggeneralizationpropertiesdeepai
https://deepai.org/publication/interpolated-joint-space-adversarial-training-for-robust-and-generalizable-defenses
Interpolated Joint Space Adversarial Training for Robust and Generalizable Defenses | DeepAI
Dec 12, 2021 - 12/12/21 - Adversarial training (AT) is considered to be one of the most reliable defenses against adversarial attacks. However, models train...
joint spaceadversarial traininginterpolated
https://openreview.net/forum?id=BJE-4xW0W
CausalGAN: Learning Causal Implicit Generative Models with Adversarial Training | OpenReview
We introduce causal implicit generative models, which can sample from conditional and interventional distributions and also propose two new conditional GANs...
generative modelsadversarial traininglearningcausalimplicit
https://openreview.net/forum?id=SJxSDxrKDr
Adversarial Training and Provable Defenses: Bridging the Gap | OpenReview
We propose a novel combination of adversarial training and provable defenses which produces a model with state-of-the-art accuracy and certified robustness on...
bridging the gapadversarial trainingprovabledefensesopenreview
https://arxiv.org/html/2504.15457v4
Improving Human-AI Coordination through Online Adversarial Training and Generative Models
human aiadversarial trainingimprovingcoordination
https://openreview.net/forum?id=vWS3gcvztz&referrer=%5Bthe%20profile%20of%20Baoyuan%20Wu%5D(%2Fprofile%3Fid%3D~Baoyuan_Wu1)
Improving Fast Adversarial Training With Prior-Guided Knowledge | OpenReview
Fast adversarial training (FAT) is an efficient method to improve robustness in white-box attack scenarios. However, the original FAT suffers from catastrophic...
adversarial trainingimprovingfastpriorguided
https://openreview.net/forum?id=1EWPr0ks8I
Better Diffusion Models Further Improve Adversarial Training | OpenReview
It has been recognized that the data generated by the denoising diffusion probabilistic model (DDPM) improves adversarial training. After two years of rapid...
diffusion modelsadversarial trainingbetterimproveopenreview
https://deepai.org/publication/adversarial-training-for-multi-context-joint-entity-and-relation-extraction
Adversarial training for multi-context joint entity and relation extraction | DeepAI
Aug 21, 2018 - 08/21/18 - Adversarial training (AT) is a regularization method that can be used to improve the robustness of neural network methods by addin...
adversarial trainingrelation extractionmulticontext
https://openreview.net/forum?id=BygzbyHFvB
FreeLB: Enhanced Adversarial Training for Natural Language Understanding | OpenReview
Adversarial training, which minimizes the maximal risk for label-preserving input perturbations, has proved to be effective for improving the generalization of...
training for natural languageenhancedadversarialunderstandingopenreview
https://deepai.org/publication/you-only-propagate-once-painless-adversarial-training-using-maximal-principle
You Only Propagate Once: Painless Adversarial Training Using Maximal Principle | DeepAI
May 2, 2019 - 05/02/19 - Deep learning achieves state-of-the-art results in many areas. However recent works have shown that deep networks can be vulnerabl...
adversarial trainingmaximal principlepropagatepainless
https://deepai.org/publication/mutual-adversarial-training-learning-together-is-better-than-going-alone
Mutual Adversarial Training: Learning together is better than going alone | DeepAI
Dec 9, 2021 - 12/09/21 - Recent studies have shown that robustness to adversarial attacks can be transferred across networks. In other words, we can make a...
together is betteradversarial traininggoing alonemutuallearning
https://deepai.org/publication/a3t-adversarially-augmented-adversarial-training
A3T: Adversarially Augmented Adversarial Training | DeepAI
Jan 12, 2018 - 01/12/18 - Recent research showed that deep neural networks are highly sensitive to so-called adversarial perturbations, which are tiny pertu...
adversarial trainingaugmenteddeepai
https://deepai.org/publication/adaptive-smoothness-weighted-adversarial-training-for-multiple-perturbations-with-its-stability-analysis
Adaptive Smoothness-weighted Adversarial Training for Multiple Perturbations with Its Stability...
Oct 2, 2022 - 10/02/22 - Adversarial Training (AT) has been demonstrated as one of the most effective methods against adversarial examples. While most exis...
adversarial trainingadaptivesmoothnessweighted
https://arxiv.org/abs/2205.00637v1
[2205.00637v1] Enhancing Adversarial Training with Feature Separability
Abstract page for arXiv paper 2205.00637v1: Enhancing Adversarial Training with Feature Separability
adversarial training2205enhancingfeatureseparability
https://openreview.net/forum?id=U0TCTe68s41
Demystifying Adversarial Training via A Unified Probabilistic Framework | OpenReview
Adversarial Training (AT) is known as an effective approach to enhance the robustness of deep neural networks. Recently researchers notice that robust models...
adversarial trainingdemystifyingviaunifiedprobabilistic
https://openreview.net/forum?id=jy292MSaSRd&referrer=%5Bthe%20profile%20of%20Sumukh%20K%20Aithal%5D(%2Fprofile%3Fid%3D~Sumukh_K_Aithal1)
A Closer Look at Smoothness in Domain Adversarial Training | OpenReview
Domain adversarial training has been ubiquitous for achieving invariant representations and is used widely for various domain adaptation tasks. In recent...
a closer lookin domainadversarial trainingsmoothnessopenreview
https://openreview.net/forum?id=BRdEBlwUW6
DAFA: Distance-Aware Fair Adversarial Training | OpenReview
The disparity in accuracy between classes in standard training is amplified during adversarial training, a phenomenon termed the robust fairness problem....
adversarial trainingdafadistanceawarefair
https://arxiv.org/abs/2303.05758
[2303.05758] MIXPGD: Hybrid Adversarial Training for Speech Recognition Systems
Abstract page for arXiv paper 2303.05758: MIXPGD: Hybrid Adversarial Training for Speech Recognition Systems
adversarial trainingspeech recognition2303hybridsystems
https://deepai.org/publication/adversarial-training-with-generated-data-in-high-dimensional-regression-an-asymptotic-study
Adversarial Training with Generated Data in High-Dimensional Regression: An Asymptotic Study |...
Jun 21, 2023 - 06/21/23 - In recent years, studies such as have demonstrated that incorporating additional real or generated data with pseudo-labels ...
adversarial training
https://openreview.net/forum?id=34yGUOcocD
Adversarial Training Should Be Cast as a Non-Zero-Sum Game | OpenReview
One prominent approach toward resolving the adversarial vulnerability of deep neural networks is the two-player zero-sum paradigm of adversarial training, in...
non zero sum gameadversarial trainingshould be
https://deepai.org/publication/combining-recurrent-neural-networks-and-adversarial-training-for-human-motion-synthesis-and-control
Combining Recurrent Neural Networks and Adversarial Training for Human Motion Synthesis and Control...
Jun 21, 2018 - 06/21/18 - This paper introduces a new generative deep learning network for human motion synthesis and control. Our key idea is to combine re...
recurrent neural networksadversarial training
https://openreview.net/forum?id=kdPcLdJbt1
Vulnerability-Aware Instance Reweighting For Adversarial Training | OpenReview
Adversarial Training (AT) has been found to substantially improve the robustness of deep learning classifiers against adversarial attacks. AT involves...
adversarial trainingvulnerabilityawareinstanceopenreview
https://openreview.net/forum?id=HPdxC1THU8T
Revisiting adapters with adversarial training | OpenReview
While adversarial training is generally used as a defense mechanism, recent works show that it can also act as a regularizer. By co-training a neural network...
adversarial trainingrevisitingadaptersopenreview
https://www.amazon.science/blog/adversarial-training-improves-product-discovery
Adversarial training improves product discovery - Amazon Science
Mar 6, 2024 - Method automatically generates negative training examples for deep-learning model.
adversarial trainingproduct discoveryimprovesamazonscience
https://openreview.net/forum?id=6LxMeRlkWl
Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs | OpenReview
Large language models (LLMs) can often be made to behave in undesirable ways that they are explicitly fine-tuned not to. For example, the LLM red-teaming...
adversarial training
https://openreview.net/forum?id=UDJ5tOYHot&referrer=%5Bthe%20profile%20of%20Yuntian%20Chen%5D(%2Fprofile%3Fid%3D~Yuntian_Chen1)
Focus on Hiders: Exploring Hidden Threats for Enhancing Adversarial Training | OpenReview
Adversarial training is often formulated as a min-max problem however concentrating only on the worst adversarial examples causes alternating repetitive...
focus onadversarial traininghidersexploringhidden
https://openreview.net/forum?id=w38OtHu4Ha
Adversarial Training with Generated Data in High-Dimensional Regression: An Asymptotic Study |...
adversarial training
https://aclanthology.org/2021.acl-long.401/
Data Augmentation with Adversarial Training for Cross-Lingual NLI - ACL Anthology
Xin Dong, Yaxin Zhu, Zuohui Fu, Dongkuan Xu, Gerard de Melo. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the...
data augmentationadversarial trainingfor cross
https://openreview.net/forum?id=NENo__bExYu
Make Some Noise: Reliable and Efficient Single-Step Adversarial Training | OpenReview
We introduce a novel single-step attack for adversarial training that can prevent catastrophic overfitting while obtaining a 3x speed-up.
make some noisesingle stepadversarial trainingreliable
https://deepai.org/publication/ganspeech-adversarial-training-for-high-fidelity-multi-speaker-speech-synthesis
GANSpeech: Adversarial Training for High-Fidelity Multi-Speaker Speech Synthesis | DeepAI
Jun 29, 2021 - 06/29/21 - Recent advances in neural multi-speaker text-to-speech (TTS) models have enabled the generation of reasonably good speech quality ...
adversarial traininghigh fidelityspeech synthesis
https://openreview.net/forum?id=ODkBI1d3phW
Efficient and Effective Augmentation Strategy for Adversarial Training | OpenReview
We propose an effective augmentation strategy for Adversarial Training that can be integrated with several Adversarial Training algorithms and data...
adversarial trainingefficienteffectiveaugmentationstrategy
https://deepai.org/publication/to-be-robust-or-to-be-fair-towards-fairness-in-adversarial-training
To be Robust or to be Fair: Towards Fairness in Adversarial Training | DeepAI
Oct 13, 2020 - 10/13/20 - Adversarial training algorithms have been proven to be reliable to improve machine learning models' robustness against adversarial...
to beadversarial trainingrobustfair
https://openreview.net/forum?id=HkgrZ0EYwB
Unpaired Point Cloud Completion on Real Scans using Adversarial Training | OpenReview
As 3D scanning solutions become increasingly popular, several deep learning setups have been developed for the task of scan completion, i.e., plausibly filling...
point cloudadversarial trainingcompletion
https://www.amazon.science/publications/smoothing-model-predictions-using-adversarial-training-procedures-for-speech-based-emotion-recognition
Smoothing model predictions using adversarial training procedures for speech based emotion...
Training discriminative classifiers involves learning a conditional distribution p(yi|xi), given a set of feature vectors xi and the corresponding labels yi, i...
adversarial trainingsmoothingmodelpredictionsusing
https://openreview.net/forum?id=I3HCE7Ro78H
Finding Actual Descent Directions for Adversarial Training | OpenReview
There is a subtle bug in the theory behind PGD. We show how to correct it and that it matters in practice
adversarial trainingfindingactualdescentdirections