CaltechAUTHORS
  A Caltech Library Service

Stochastic Activation Pruning for Robust Adversarial Defense

Dhillon, Guneet S. and Azizzadenesheli, Kamyar and Lipton, Zachary C. and Bernstein, Jeremy and Kossaifi, Jean and Khanna, Aran and Anandkumar, Anima (2018) Stochastic Activation Pruning for Robust Adversarial Defense. In: 6th International Conference on Learning Representations (ICLR 2018), 30 April-3 May 2018, Vancouver, Canada. https://resolver.caltech.edu/CaltechAUTHORS:20190327-085749625

[img] PDF - Published Version
See Usage Policy.

1168Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20190327-085749625

Abstract

Neural networks are known to be vulnerable to adversarial examples. Carefully chosen perturbations to real images, while imperceptible to humans, induce misclassification and threaten the reliability of deep learning systems in the wild. To guard against adversarial examples, we take inspiration from game theory and cast the problem as a minimax zero-sum game between the adversary and the model. In general, for such games, the optimal strategy for both players requires a stochastic policy, also known as a mixed strategy. In this light, we propose Stochastic Activation Pruning (SAP), a mixed strategy for adversarial defense. SAP prunes a random subset of activations (preferentially pruning those with smaller magnitude) and scales up the survivors to compensate. We can apply SAP to pretrained networks, including adversarially trained models, without fine-tuning, providing robustness against adversarial examples. Experiments demonstrate that SAP confers robustness against attacks, increasing accuracy and preserving calibration.


Item Type:Conference or Workshop Item (Poster)
Related URLs:
URLURL TypeDescription
https://openreview.net/forum?id=H1uR4GZRZPublisherArticle
http://arxiv.org/abs/1803.01442arXivArticle
ORCID:
AuthorORCID
Azizzadenesheli, Kamyar0000-0001-8507-1868
Bernstein, Jeremy0000-0001-9110-7476
Record Number:CaltechAUTHORS:20190327-085749625
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20190327-085749625
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:94174
Collection:CaltechAUTHORS
Deposited By: George Porter
Deposited On:29 Mar 2019 20:08
Last Modified:11 Nov 2020 00:49

Repository Staff Only: item control page