Research map / Adversarial robustness and security / Black-box and transfer attacks
Timestep-Compressed Attack on Spiking Neural Networks Through Timestep-Level Backpropagation
Donghwa Kang, Doohyun Kim, Sang-Ki Ko, Jinkyu Lee, Hyeongboo Baek, Brent ByungHoon Kang
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 |
|---|---|---|---|
| add class-specific membrane image with Ladv and Lmem (full A-MPR) (ablation) | TCA with STBP · VGG-11 CIFAR-10 (w=1, n=1) | ASR (%): 84.7 -> 87.8 | Improved |
| add full A-MPR (ablation) | TCA with HART · VGG-11 CIFAR-10 (w=2, n=1) | ASR (%): 92.5 -> 94.6 | Improved |
| add full A-MPR (ablation) | TCA with STBP · ResNet-17 CIFAR-10 (w=1, n=1) | ASR (%): 74.2 -> 74.3 | No clear effect |
| add Ladv alone (without Lmem) (ablation) | TCA with STBP · VGG-11 CIFAR-10 (w=2, n=1) | ASR (%): 89.1 -> 88.0 | Hurt |
| clamp membrane potentials to [0.2, 0.5] (ablation) | TCA with STBP · ResNet-17 CIFAR-10 (w=1, n=1) | ASR (%): 25.4 -> 74.3 | Improved |
| clamp membrane potentials to [0.2, 0.5] (ablation) | TCA with STBP · VGG-11 CIFAR-10 (w=1, n=1) | ASR (%): 14.8 -> 87.8 | Improved |
| clamp (0.2, 0.6) instead of (0.2, 0.5) (ablation) | TCA with STBP · CIFAR-10 w=1 (VGG-11 and ResNet-17) | ASR (%): VGG-11 87.8 -> 89.6; ResNet-17 74.3 -> 71.7 | Mixed |
| timestep-level backprop without temporal information (no BPTT) (ablation) | STBP FGSM attack · CIFAR-10 VGG-11 / ResNet-17 | final ASR | No clear effect |
| TCA instead of FGSM/PGD for latency | STBP attack on VGG-11 CIFAR-10 · white-box | attack latency: up to 56.6% reduction at same ASR | Improved |
| TCA instead of FGSM/PGD for latency | STBP attack on VGG-11 CIFAR-10 · black-box | attack latency: up to 57.1% reduction | Improved |
| TCA on VGG-11 CIFAR-100 | SNN attacks · VGG-11 CIFAR-100 | latency-ASR trade-off | No clear effect |
Closest papers by meaning
- Resolving the Timestep Scaling Paradox in Spiking Neural Networks with a Timestep-Scalable Neuron Model (ICML 2026 · similarity 0.88)
- Towards Effective and Sparse Adversarial Attack on Spiking Neural Networks via Breaking Invisible Surrogate Gradients (CVPR 2025 · similarity 0.86)
- Temporal Flexibility in Spiking Neural Networks: Towards Generalization Across Time Steps and Deployment Friendliness (ICLR 2025 · similarity 0.86)
- TS-SNN: Temporal Shift Module for Spiking Neural Networks (ICML 2025 · similarity 0.85)
- Efficient Logit-based Knowledge Distillation of Deep Spiking Neural Networks for Full-Range Timestep Deployment (ICML 2025 · similarity 0.85)
- Advancing Training Efficiency of Deep Spiking Neural Networks through Rate-based Backpropagation (NeurIPS 2024 · similarity 0.85)
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.
