Towards Proving the Adversarial Robustness of Deep Neural Networks

Towards Proving the Adversarial Robustness of Deep Neural Networks” by Guy Katz, Clark Barrett, David L. Dill, Kyle Julian, and Mykel J. Kochenderfer. In Proceedings of the First Workshop on Formal Verification of Autonomous Vehicles (FVAV '17), (Lukas Bulwahn, Maryam Kamali, and Sven Linker, eds.), Sep. 2017, pp. 19-26. Turin, Italy.

Abstract

Autonomous vehicles are highly complex systems, required to function reliably in a wide variety of situations. Manually crafting software controllers for these vehicles is difficult, but there has been some success in using deep neural networks generated using machine-learning. However, deep neural networks are opaque to human engineers, rendering their correctness very difficult to prove manually; and existing automated techniques, which were not designed to operate on neural networks, fail to scale to large systems. This paper focuses on proving the adversarial robustness of deep neural networks, i.e. proving that small perturbations to a correctly-classified input to the network cannot cause it to be misclassified. We describe some of our recent and ongoing work on verifying the adversarial robustness of networks, and discuss some of the open questions we have encountered and how they might be addressed.

BibTeX entry:

@inproceedings{KBD+17-FVAV,
   author = {Guy Katz and Clark Barrett and David L. Dill and Kyle Julian
	and Mykel J. Kochenderfer},
   editor = {Lukas Bulwahn and Maryam Kamali and Sven Linker},
   title = {Towards Proving the Adversarial Robustness of Deep Neural
	Networks},
   booktitle = {Proceedings of the First Workshop on Formal Verification
	of Autonomous Vehicles (FVAV '17)},
   series = {Electronic Proceedings in Theoretical Computer Science},
   volume = {257},
   pages = {19--26},
   month = sep,
   year = {2017},
   note = {Turin, Italy},
   url = {http://eptcs.web.cse.unsw.edu.au/paper.cgi?FVAV2017.3}
}

(This webpage was created with bibtex2web.)