Apr 18, 2019 · We intentionally inject trapdoors, honeypot weaknesses in the classification manifold that attract attackers searching for adversarial examples.
Apr 18, 2019 · Deep neural networks are vulnerable to adversarial attacks. Numerous efforts have focused on defenses that either try to patch `holes' in ...
This work introduces trapdoors and describes an implementation of trapdoors using similar strategies to backdoor/Trojan attacks, and shows that by ...
This repository contains code implementation of the paper "Gotta Catch'Em All: Using Honeypots to Catch Adversarial Attacks on Neural Networks", at ACM CCS ...
Missing: Concealed | Show results with:Concealed
Each trapdoor has minimal impact on classification of normal inputs, but leads attackers to produce adversarial inputs whose similarity to the trapdoor makes ...
Missing: Concealed | Show results with:Concealed
Mar 9, 2022 · Bibliographic details on Gotta Catch 'Em All: Using Concealed Trapdoors to Detect Adversarial Attacks on Neural Networks.
Deep neural networks are vulnerable to adversarial attacks. Numerous efforts have focused on defenses that either try to patch `holes' in trained models or ...
Deep neural networks are vulnerable to adversarial attacks. Numerous effortshave focused on defenses that either try to patch ...
People also ask
What are adversarial attacks on neural networks?
Gotta Catch 'Em All: Using Concealed Trapdoors to Detect Adversarial Attacks on Neural Networks. S. Shan, E. Willson, B. Wang, B. Li, H. Zheng, and B. Zhao.
Nov 2, 2020 · We intentionally inject trapdoors, honeypot weaknesses in the classification manifold that attract attackers searching for adversarial examples.