AI Paper Atlas

Research map / Adversarial robustness and security / Backdoor and poisoning

Trigger Hunting with a Topological Prior for Trojan Detection

Xiaoling Hu, Xiao Lin, Michael Cogswell, Yi Yao, Susmit Jha, Chao Chen

ICLR 2022 · Adversarial robustness and security › Backdoor and poisoning · approach: Trigger-based backdoor attacks · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use topological and diversity losses with three trigger candidatesTrojan detector · Trojaned-MNIST and Trojaned-CIFAR10AUC: 0.88 $\pm$ 0.04 vs 0.65 $\pm$ 0.07 (MNIST); 0.91 $\pm$ 0.05 vs 0.75 $\pm$ 0.07 (CIFAR10)Improved
Use topological and diversity losses with three trigger candidatesTrojan detector · TrojAI-Round1 to Round4AUC: 0.90 $\pm$ 0.02 vs 0.71 $\pm$ 0.04; 0.87 $\pm$ 0.05 vs 0.66 $\pm$ 0.07; 0.89 $\pm$ 0.04 vs 0.62 $\pm$ 0.07; 0.92 $\pm$ 0Improved
Add the topological loss (ablation)Trigger search and Trojan detector · TrojAI-Round4AUC: 0.89 $\pm$ 0.04 -> 0.92 $\pm$ 0.06Improved
Add the diversity loss and generate multiple candidates (N_T=3) (ablation)Trigger search and Trojan detector · TrojAI-Round4AUC: 0.85 $\pm$ 0.02 -> 0.92 $\pm$ 0.06Improved
Use three rather than two trigger candidates (ablation)Trigger search and Trojan detector · TrojAI-Round4AUC: 0.89 $\pm$ 0.05 -> 0.92 $\pm$ 0.06Improved
Add the topological loss (ablation)Trigger search and Trojan detector with 25 training samples · TrojAI-Round4AUC: 0.73 $\pm$ 0.03 -> 0.77 $\pm$ 0.04Improved
Add the diversity loss (ablation)Trigger search and Trojan detector with 25 training samples · TrojAI-Round4AUC: 0.68 $\pm$ 0.04 -> 0.77 $\pm$ 0.04Improved
Use the topological loss (ablation)Trigger search · Trigger reconstructioniterations to converge: ≈50 -> ≈30Improved
Vary the topological-loss weight (ablation)Trigger search and Trojan detector · TrojAI-Round4AUCNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy