Research map / Architectures and efficiency / Spiking networks
Optimal ANN-SNN Conversion for High-accuracy and Ultra-low-latency Spiking Neural Networks
Tong Bu, Wei Fang, Jianhao Ding, PENGLIN DAI, Zhaofei Yu, Tiejun Huang
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 |
|---|---|---|---|
| replace ReLU with quantization clip-floor-shift activation | ANN-SNN conversion · CIFAR-10, VGG-16, T=4 | accuracy: 93.96% | Improved |
| replace ReLU with quantization clip-floor-shift activation | ANN-SNN conversion · CIFAR-100, VGG-16, T=4 | accuracy: 69.62% | Improved |
| add the shift term to the clip-floor activation (ablation) | VGG-16 on CIFAR-100, L=4 · CIFAR-100, T=4 | accuracy: 21.57% -> 69.62% | Improved |
| use a trainable dynamic threshold (ablation) | VGG-16 on CIFAR-100, L=4, with shift · CIFAR-100, T=4 | accuracy: 1.00% -> 69.62% | Improved |
| increase quantization steps L from 2 to 4 (ablation) | VGG-16 on CIFAR-10 · CIFAR-10, T=4 | accuracy: 93.00% -> 93.96% | Improved |
| increase quantization steps L from 4 to 8 (ablation) | VGG-16 on CIFAR-10 · CIFAR-10, T=4 | accuracy: 93.96% -> 91.77% | Hurt |
| increase quantization steps L from 4 to 8 (ablation) | VGG-16 on CIFAR-10 · CIFAR-10, T=128 | accuracy: 95.59% -> 95.79% | Improved |
| use quantization clip-floor-shift rather than clip-floor without shift (ablation) | converted SNN · different time-steps with L=4 | accuracy | Improved |
| set quantization steps L too low (ablation) | QCFS source ANN and converted SNN · CIFAR-10/100 | accuracy | Hurt |
Closest papers by meaning
- Inference-Scale Complexity in ANN-SNN Conversion for High-Performance and Low-Power Applications (CVPR 2025 · similarity 0.94)
- Faster and Stronger: When ANN-SNN Conversion Meets Parallel Spiking Calculation (ICML 2025 · similarity 0.94)
- Efficient ANN-SNN Conversion with Error Compensation Learning (ICML 2025 · similarity 0.93)
- Temporal Misalignment in ANN-SNN Conversion and its Mitigation via Probabilistic Spiking Neurons (ICML 2025 · similarity 0.91)
- Adaptive Calibration: A Unified Conversion Framework of Spiking Neural Networks (AAAI 2025 · similarity 0.91)
- EnOF-SNN: Training Accurate Spiking Neural Networks via Enhancing the Output Feature (NeurIPS 2024 · similarity 0.91)
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.
