paper-with-me

홈 › Papers

Once-for-All Adversarial Training: In-Situ Tradeoff between Robustness and Accuracy for Free

2020-10-22 · NeurIPS 2020 12 · Haotao Wang, Tianlong Chen, Shupeng Gui, Ting-Kuei Hu, Ji Liu, Zhangyang Wang

Adversarial training and its many variants substantially improve deep network robustness, yet at the cost of compromising standard accuracy. Moreover, the training process is heavy and hence it becomes impractical to thoroughly explore the trade-off between accuracy and robustness. This paper asks this new question: how to quickly calibrate a trained model in-situ, to examine the achievable trade-offs between its standard and robust accuracies, without (re-)training it many times? Our proposed framework, Once-for-all Adversarial Training (OAT), is built on an innovative model-conditional training framework, with a controlling hyper-parameter as the input. The trained model could be adjusted among different standard and robust accuracies "for free" at testing time. As an important knob, we exploit dual batch normalization to separate standard and adversarial feature statistics, so that they can be learned in one model without degrading performance. We further extend OAT to a Once-for-all Adversarial Training and Slimming (OATS) framework, that allows for the joint trade-off among accuracy, robustness and runtime efficiency. Experiments show that, without any re-training nor ensembling, OAT/OATS achieve similar or even superior performance compared to dedicatedly trained models at various configurations. Our codes and pretrained models are available at: https://github.com/VITA-Group/Once-for-All-Adversarial-Training.

📄 PDF Abstract BibTeX arXiv:2010.11828

Code (1)

VITA-Group/Once-for-All-Adversarial-Training 공식 구현 pytorch

Tasks

All

Methods 이 논문이 사용한 방법론

Batch Normalization 설명 없음

Similar Papers 제목 키워드 기반

Privacy-Preserving Adversarial Networks

2017-12-19 · Ardhendu Tripathy, Ye Wang, Prakash Ishwar

We propose a data-driven framework for optimizing privacy-preserving data release mechanisms to attain the information-theoretically optimal tradeoff between minimizing distortion of useful data and concealing specific s…

Privacy Preserving

Adversarial Training Can Hurt Generalization

2019-06-14 · ICML Workshop Deep_Phenomen 2019 6 · Aditi Raghunathan, Sang Michael Xie, Fanny Yang, John C. Duchi 외

While adversarial training can improve robust accuracy (against an adversary), it sometimes hurts standard accuracy (when there is no adversary). Previous work has studied this tradeoff between standard and robust accura…

Measuring Adversarial Robustness using a Voronoi-Epsilon Adversary

2020-05-06 · Hyeongji Kim, Pekka Parviainen, Ketil Malde

Previous studies on robustness have argued that there is a tradeoff between accuracy and adversarial accuracy. The tradeoff can be inevitable even when we neglect generalization. We argue that the tradeoff is inherent to…

Adversarial Robustness

Precise Tradeoffs in Adversarial Training for Linear Regression

2020-02-24 · Adel Javanmard, Mahdi Soltanolkotabi, Hamed Hassani

Despite breakthrough performance, modern learning models are known to be highly vulnerable to small adversarial perturbations in their inputs. While a wide variety of recent \emph{adversarial training} methods have been …

regression

Fundamental Tradeoffs in Distributionally Adversarial Training

2021-01-15 · Mohammad Mehrabi, Adel Javanmard, Ryan A. Rossi, Anup Rao 외

Adversarial training is among the most effective techniques to improve the robustness of models against adversarial perturbations. However, the full effect of this approach on models is not well understood. For example, …

Binary Classificationregression