Improving Transferability of Adversarial Examples With Input Diversity [PDF]
Computer Vision and Pattern Recognition, 2018 Though CNNs have achieved the state-of-the-art performance on various vision tasks, they are vulnerable to adversarial examples --- crafted by adding human-imperceptible perturbations to clean images.
Cihang Xie +5 more
semanticscholar +1 more source
Balancing Discriminability and Transferability for Source-Free Domain Adaptation [PDF]
International Conference on Machine Learning, 2022 Conventional domain adaptation (DA) techniques aim to improve domain transferability by learning domain-invariant representations; while concurrently preserving the task-discriminability knowledge gathered from the labeled source data.
Jogendra Nath Kundu +6 more
semanticscholar +1 more source
Rethinking the Backward Propagation for Adversarial Transferability [PDF]
Neural Information Processing Systems, 2023 Transfer-based attacks generate adversarial examples on the surrogate model, which can mislead other black-box models without access, making it promising to attack real-world applications.
Xiaosen Wang, Kangheng Tong, Kun He
semanticscholar +1 more source
Improving the Transferability of Adversarial Examples with Arbitrary Style Transfer [PDF]
ACM Multimedia, 2023 Deep neural networks are vulnerable to adversarial examples crafted by applying human-imperceptible perturbations on clean inputs. Although many attack methods can achieve high success rates in the white-box setting, they also exhibit weak ...
Zhijin Ge +6 more
semanticscholar +1 more source
Diversifying the High-level Features for better Adversarial Transferability [PDF]
British Machine Vision Conference, 2023 Given the great threat of adversarial attacks against Deep Neural Networks (DNNs), numerous works have been proposed to boost transferability to attack real-world applications.
Zhiyuan Wang +3 more
semanticscholar +1 more source
StyLess: Boosting the Transferability of Adversarial Examples [PDF]
Computer Vision and Pattern Recognition, 2023 Adversarial attacks can mislead deep neural networks (DNNs) by adding imperceptible perturbations to benign examples. The attack transferability enables adversarial examples to attack blackbox DNNs with unknown architectures or parameters, which poses ...
Kaisheng Liang, Bin Xiao
semanticscholar +1 more source
Linguistic Knowledge and Transferability of Contextual Representations [PDF]
North American Chapter of the Association for Computational Linguistics, 2019 Contextual word representations derived from large-scale neural language models are successful across a diverse set of NLP tasks, suggesting that they encode useful and transferable features of language.
Nelson F. Liu +4 more
semanticscholar +1 more source
Harmonizing Transferability and Discriminability for Adapting Object Detectors [PDF]
Computer Vision and Pattern Recognition, 2020 Recent advances in adaptive object detection have achieved compelling results in virtue of adversarial feature adaptation to mitigate the distributional shifts along the detection pipeline.
Chaoqi Chen +4 more
semanticscholar +1 more source
Improving Transferability of Adversarial Patches on Face Recognition with Generative Models [PDF]
Computer Vision and Pattern Recognition, 2021 Face recognition is greatly improved by deep convolutional neural networks (CNNs). Recently, these face recognition models have been used for identity authentication in security sensitive applications.
Zihao Xiao +7 more
semanticscholar +1 more source
On the Cross-lingual Transferability of Monolingual Representations [PDF]
Annual Meeting of the Association for Computational Linguistics, 2019 State-of-the-art unsupervised multilingual models (e.g., multilingual BERT) have been shown to generalize in a zero-shot cross-lingual setting. This generalization ability has been attributed to the use of a shared subword vocabulary and joint training ...
Mikel Artetxe +2 more
semanticscholar +1 more source