Topological Detection of Trojaned Neural Networks

Abstract

Deep neural networks are known to have security issues. One particular threat is the Trojan attack. It occurs when the attackers stealthily manipulate the model's behavior through Trojaned training samples, which can later be exploited. Guided by basic neuroscientific principles we discover subtle -- yet critical -- structural deviation characterizing Trojaned models. In our analysis we use topological tools. They allow us to model high-order dependencies in the networks, robustly compare different networks, and localize structural abnormalities. One interesting observation is that Trojaned models develop short-cuts from input to output layers. Inspired by these observations, we devise a strategy for robust detection of Trojaned models. Compared to standard baselines it displays better performance on multiple benchmarks.

References in corpus (7)

• Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning
• Poisoning Attacks against Support Vector Machines
• Sliced Wasserstein Kernel for Persistence Diagrams
• Weight Poisoning Attacks on Pre-trained Models
• Connectivity-Optimized Representation Learning via Persistent Homology
• TrojDRL: Trojan Attacks on Deep Reinforcement Learning Agents
• Towards Interaction Detection Using Topological Analysis on Neural Networks