Understanding Adversarial Training: Increasing Local Stability of Neural Nets through Robust Optimization

11/17/2015
by   Uri Shaham, et al.
0

We propose a general framework for increasing local stability of Artificial Neural Nets (ANNs) using Robust Optimization (RO). We achieve this through an alternating minimization-maximization procedure, in which the loss of the network is minimized over perturbed examples that are generated at each parameter update. We show that adversarial training of ANNs is in fact robustification of the network optimization, and that our proposed framework generalizes previous approaches for increasing local stability of ANNs. Experimental results reveal that our approach increases the robustness of the network to existing adversarial examples, while making it harder to generate new ones. Furthermore, our algorithm improves the accuracy of the network also on the original test data.

READ FULL TEXT

page 7

page 8

research
12/15/2021

On the Convergence and Robustness of Adversarial Training

Improving the robustness of deep neural networks (DNNs) to adversarial e...
research
06/19/2019

Convergence of Adversarial Training in Overparametrized Networks

Neural networks are vulnerable to adversarial examples, i.e. inputs that...
research
11/15/2019

On Model Robustness Against Adversarial Examples

We study the model robustness against adversarial examples, referred to ...
research
10/03/2020

Efficient Robust Training via Backward Smoothing

Adversarial training is so far the most effective strategy in defending ...
research
01/26/2016

Unifying Adversarial Training Algorithms with Flexible Deep Data Gradient Regularization

Many previous proposals for adversarial training of deep neural nets hav...
research
05/30/2023

It begins with a boundary: A geometric view on probabilistically robust learning

Although deep neural networks have achieved super-human performance on m...
research
03/16/2021

A new framework for the stability analysis of perturbed saddle-point problems and applications in poromechanics

This paper provides a new abstract stability result for perturbed saddle...

Please sign up or login with your details

Forgot password? Click here to reset