On Meaning-Preserving Adversarial Perturbations for Sequence-to-Sequence Models
Adversarial examples have been shown to be an effective way of assessing the robustness of neural sequence-to-sequence (seq2seq) models, by applying perturbations to the input of a model leading to large degradation in performance. However, these perturbations are only indicative of a weakness in the model if they do not change the semantics of the input in a way that would change the expected output. Using the example of machine translation (MT), we propose a new evaluation framework for adversarial attacks on seq2seq models taking meaning preservation into account and demonstrate that existing methods may not preserve meaning in general. Based on these findings, we propose new constraints for attacks on word-based MT systems and show, via human and automatic evaluation, that they produce more semantically similar adversarial inputs. Furthermore, we show that performing adversarial training with meaning-preserving attacks is beneficial to the model in terms of adversarial robustness without hurting test performance.
Code (0)
등록된 구현이 없습니다.
Tasks
Adversarial RobustnessMachine TranslationTranslationMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
On Evaluation of Adversarial Perturbations for Sequence-to-Sequence Models
Adversarial examples --- perturbations to the input of a model that elicit large changes in the output --- have been shown to be an effective way of assessing the robustness of sequence-to-sequence (seq2seq) models. Howe…
Adversarial RobustnessMachine TranslationRobust Tracking against Adversarial Attacks
While deep convolutional neural networks (CNNs) are vulnerable to adversarial attacks, considerably few efforts have been paid to construct robust deep tracking algorithms against adversarial attacks. Current studies on …
Adversarial AttackSemantic Robustness of Models of Source Code
Deep neural networks are vulnerable to adversarial examples - small input perturbations that result in incorrect predictions. We study this problem for models of source code, where we want the network to be robust to sou…
Protein Folding Neural Networks Are Not Robust
Deep neural networks such as AlphaFold and RoseTTAFold predict remarkably accurate structures of proteins compared to other algorithmic approaches. It is known that biologically small perturbations in the protein sequenc…
Adversarial AttackProtein FoldingSAR-GPA: SAR Generation Perturbation Algorithm
The deep learning is widely used in optical image and synthetic aperture radar (SAR) image. Current academic research shows that adversarial perturbation can effectively attack the deep learning network in optical ima…