Verifying neural networks against backdoor attacks

Neural networks have achieved state-of-the-art performance in solving many problems, including many applications in safety/security-critical systems. Researchers also discovered multiple security issues associated with neural networks. One of them is backdoor attacks, i.e., a neural network may be e...

Full description

Saved in:
Bibliographic Details
Main Authors: PHAM, Long Hong, SUN, Jun
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2022
Subjects:
Online Access:https://ink.library.smu.edu.sg/sis_research/7279
https://ink.library.smu.edu.sg/context/sis_research/article/8282/viewcontent/Verifying_neural_networks_against_backdoor_attacks.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
Description
Summary:Neural networks have achieved state-of-the-art performance in solving many problems, including many applications in safety/security-critical systems. Researchers also discovered multiple security issues associated with neural networks. One of them is backdoor attacks, i.e., a neural network may be embedded with a backdoor such that a target output is almost always generated in the presence of a trigger. Existing defense approaches mostly focus on detecting whether a neural network is ‘backdoored’ based on heuristics, e.g., activation patterns. To the best of our knowledge, the only line of work which certifies the absence of backdoor is based on randomized smoothing, which is known to significantly reduce neural network performance. In this work, we propose an approach to verify whether a given neural network is free of backdoor with a certain level of success rate. Our approach integrates statistical sampling as well as abstract interpretation. The experiment results show that our approach effectively verifies the absence of backdoor or generates backdoor triggers.