Research map / Adversarial robustness and security / Backdoor and poisoning
Event Trojan: Asynchronous Event-based Backdoor Attacks
Ruofei Wang, Qing Guo, Haoliang Li, Renjie Wan
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use mutable (learned timestamp) trigger instead of immutable trigger (ablation) | Event Trojan · N-Caltech101 (VGG-16) | ASR: 18.12 -> 97.65 | Improved |
| use mutable trigger instead of immutable trigger (ablation) | Event Trojan · N-Caltech101 (ViT-S) | ASR: 14.74 -> 87.73 | Improved |
| use mutable trigger instead of immutable trigger (ablation) | Event Trojan · N-Cars (Swin-S) | ASR: 50.91 -> 100.0 | Improved |
| add cosine-similarity term to trigger loss (ablation) | Mutable trigger · N-Caltech101 (ResNet-18) | ASR: 11.93 -> 99.71 | Improved |
| add expectation/variance term psi to trigger loss (ablation) | Mutable trigger · N-Caltech101 (ResNet-18) | CDA: 85.67 -> 86.21 CDA; ASR 100.00 -> 99.71 | Improved |
| poison all dimensions (x,y,t,p) instead of one dimension (ablation) | Event Trojan · N-Caltech101 (ResNet-18) | ASR: 9.71 -> 99.71 (t only); 8.56 -> 99.71 (p only) | Improved |
| increase trigger size from 5x5 to 10x10 (ablation) | Event Trojan · N-Caltech101 (ResNet-18) | ASR: 0.0694 -> 0.9673 | Improved |
| increase trigger size from 10x10 to 30x30 (ablation) | Event Trojan · N-Caltech101 (ResNet-18) | ASR: 0.9673 -> 0.9994 | No clear effect |
| inject trigger into raw events rather than representation, against Neural Polarizer defense (ablation) | Event Trojan · N-Caltech101 (ResNet-18), Neural Polarizer defense | ASR: 1.03 -> 64.11 (BadNets vs mutable) | Improved |
| apply STC denoising filter to triggered events (ablation) | Mutable trigger · N-Caltech101 (ResNet-18) | ASR: 0.9971 -> 0.9738 | No clear effect |
| switch event representation from EST to Event Frame or Time Surface (ablation) | Event Trojan · N-Caltech101 (ResNet-18) | ASR: 0.9673 -> 0.8830 (EF), 0.8635 (TS), 0.9977 (VG) | Mixed |
| use representation-level BadNets trigger on ResNet-18 | BadNets · N-Caltech101 (ResNet-18) | ASR: 0.0 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FIBA — FIBA: Frequency-Injection Based Backdoor Attack in Medical Image Analysis (CVPR 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Rethinking Backdoor Attacks on Dataset Distillation: A Kernel Method Perspective (ICLR 2024 · similarity 0.84)
- Invisible Backdoor Attack against Self-supervised Learning (CVPR 2025 · similarity 0.84)
- Revisiting Backdoor Attacks against Large Vision-Language Models from Domain Shift (CVPR 2025 · similarity 0.84)
- Temporal-Distributed Backdoor Attack against Video Based Action Recognition (AAAI 2024 · similarity 0.84)
- Trapping Attacker in Dilemma: Examining Internal Correlations and External Influences of Trigger for Defending GNN Backdoors (NeurIPS 2026 · similarity 0.83)
- Demystifying Poisoning Backdoor Attacks from a Statistical Perspective (ICLR 2024 · similarity 0.83)
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.
