Research map / Label-efficient and robust learning / Knowledge distillation
A Closer Look at Knowledge Distillation in Spiking Neural Network Training
Xu Liu, Na Xia, Jinxing Zhou, Jingyuan Xu, Dan Guo
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 softmax scaling of CAM and SAM instead of no scaling (ablation) | SAMD · CIFAR-100 (ResNet-19, T=1) | accuracy: 75.56 -> 79.11 | Improved |
| use softmax instead of L2-norm scaling (ablation) | SAMD · CIFAR-100 (ResNet-19, T=1) | accuracy: 76.48 -> 79.11 | Improved |
| use softmax instead of Z-score scaling (ablation) | SAMD · CIFAR-100 (ResNet-19, T=1) | accuracy: 74.78 -> 79.11 | Improved |
| add SAMD and NLD distillation (vs without either) (ablation) | CKDSNN · CIFAR-100 (ResNet-19) | accuracy | Improved |
| replace gradient-based CAM-CAM distillation (e2KD, CATKD) with CAM-SAM (ablation) | SAMD · CIFAR-100 (ResNet-19) | accuracy | Improved |
| use adaptive Gaussian noise instead of fixed-std noise (ablation) | NLD · CIFAR-100 (ResNet-19, T=1) | accuracy | Improved |
| use CKDSNN instead of BKDSNN | SEW-ResNet-34 SNN · ImageNet-1K (T=4) | top-1 accuracy: 71.24 -> 73.05 | Improved |
| use CKDSNN instead of EnOFSNN | ResNet-20 SNN · CIFAR10-DVS (T=10) | accuracy: 80.50 -> 81.55 | Improved |
| use CKDSNN instead of EnOFSNN | ResNet-19 SNN · CIFAR-100 (T=1) | accuracy: 77.08 -> 79.11 | Improved |
| use CKDSNN instead of BKDSNN | SNN energy · ImageNet-1K (ResNet34, T=4) | fire rate (%): 15.0 -> 13.0 | Improved |
| increase time steps from 1 to 4 | CKDSNN ResNet-19 · CIFAR-100 | accuracy: 79.11 -> 83.88 | Improved |
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.
- BKDSNN — BKDSNN: Enhancing the Performance of Learning-based Spiking Neural Networks Training with Blurred Knowledge Distillation (ECCV 2024) · better on 11 of 11 shared results
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
- Temporal Separation with Entropy Regularization for Knowledge Distillation in Spiking Neural Networks (CVPR 2025 · similarity 0.93)
- Synergy Between the Strong and the Weak: Spiking Neural Networks are Inherently Self-Distillers (NeurIPS 2025 · similarity 0.93)
- Enhanced Self-Distillation Framework for Efficient Spiking Neural Network Training (NeurIPS 2025 · similarity 0.92)
- Efficient Logit-based Knowledge Distillation of Deep Spiking Neural Networks for Full-Range Timestep Deployment (ICML 2025 · similarity 0.92)
- Efficient ANN-Guided Distillation: Aligning Rate-based Features of Spiking Neural Networks through Hybrid Block-wise Replacement (CVPR 2025 · similarity 0.91)
- EnOF-SNN: Training Accurate Spiking Neural Networks via Enhancing the Output Feature (NeurIPS 2024 · similarity 0.90)
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.
