FreeEagle: Detecting Complex Neural Trojans in Data-Free Cases

TL;DR

FreeEagle is a novel data-free method for detecting complex neural Trojans in deep neural networks, effective even without access to clean or triggered samples, outperforming some existing methods.

Contribution

It introduces the first data-free backdoor detection approach capable of identifying complex Trojans without relying on any clean or triggered data.

Findings

Effective against various complex backdoor attacks

Outperforms some state-of-the-art non-data-free methods

Works across diverse datasets and model architectures

Abstract

Trojan attack on deep neural networks, also known as backdoor attack, is a typical threat to artificial intelligence. A trojaned neural network behaves normally with clean inputs. However, if the input contains a particular trigger, the trojaned model will have attacker-chosen abnormal behavior. Although many backdoor detection methods exist, most of them assume that the defender has access to a set of clean validation samples or samples with the trigger, which may not hold in some crucial real-world cases, e.g., the case where the defender is the maintainer of model-sharing platforms. Thus, in this paper, we propose FreeEagle, the first data-free backdoor detection method that can effectively detect complex backdoor attacks on deep neural networks, without relying on the access to any clean samples or samples with the trigger. The evaluation results on diverse datasets and model architectures show that FreeEagle is effective against various complex backdoor attacks, even outperforming some state-of-the-art non-data-free backdoor detection methods.