Efficient Neural Network Robustness Certification with General Activation Functions
Name
7742-efficient-neural-network-robustness-certification-with-general-activation-functions.pdf
Description
Published version
Size
626.71 KB
Format
Adobe PDF
Checksum (MD5)
decc2858f857de53dc1e9d515a4b0120
Author(s)
Zhang, Huan
Weng, Tsui-Wei
Chen, Pin-Yu
Hsieh, Cho-Jui
Daniel, Luca
Date Issued
2018
Citation
Zhang, Huan, Weng, Tsui-Wei, Chen, Pin-Yu, Hsieh, Cho-Jui and Daniel, Luca. 2018. "Efficient Neural Network Robustness Certification with General Activation Functions."
Version
Final published version
Abstract
© 2018 Curran Associates Inc..All rights reserved. Finding minimum distortion of adversarial examples and thus certifying robustness in neural network classifiers for given data points is known to be a challenging problem. Nevertheless, recently it has been shown to be possible to give a nontrivial certified lower bound of minimum adversarial distortion, and some recent progress has been made towards this direction by exploiting the piece-wise linear nature of ReLU activations. However, a generic robustness certification for general activation functions still remains largely unexplored. To address this issue, in this paper we introduce CROWN, a general framework to certify robustness of neural networks with general activation functions for given input data points. The novelty in our algorithm consists of bounding a given activation function with linear and quadratic functions, hence allowing it to tackle general activation functions including but not limited to four popular choices: ReLU, tanh, sigmoid and arctan. In addition, we facilitate the search for a tighter certified lower bound by adaptively selecting appropriate surrogates for each neuron activation. Experimental results show that CROWN on ReLU networks can notably improve the certified lower bounds compared to the current state-of-the-art algorithm Fast-Lin, while having comparable computational efficiency. Furthermore, CROWN also demonstrates its effectiveness and flexibility on networks with general activation functions, including tanh, sigmoid and arctan.
MIT Department
MIT-IBM Watson AI Lab
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://papers.nips.cc/paper/7742-efficient-neural-network-robustness-certification-with-general-activation-functions
