Research map / Architectures and efficiency / Spiking networks
Narrowing the ANN–SNN Gap for Continuous 1D Temporal Signal Classification with Multi-Scale Temporal Encoding and Sparsity-Regularized Transform Encoding
Qi Sun, Yulei Huang, Zhen Cao, Biao Hou
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 MTE and STE to directly trained SNN | SNN baselines (short horizon T=4) · 10 1D datasets, avg over backbones | accuracy: 0.738 -> 0.775 | Improved |
| add MTE and STE to SNN | RNN backbone · 10 1D datasets | accuracy: .725 -> .785 | Improved |
| add MTE and STE to SNN | Spikformer-1D · 10 1D datasets | accuracy: .805 -> .813 | Improved |
| add MTE and STE to ResNet18-1D SNN | ResNet18-1D · CWRU4 | Rating: .227 -> .240 | Hurt |
| add MTE (multi-scale encoding) without STE (ablation) | SLSTM · SEED, MIT-BIH, CWRU10 | spike rate: SEED R .553 -> .619 | Hurt |
| add STE alone (ablation) | SLSTM · SEED | accuracy: .896 -> .951 | Improved |
| add STE on top of MTE (ablation) | SLSTM with MTE · SOPs | SOPs (M): 7.221 -> 5.722 | Improved |
| add STE alone (ablation) | SLSTM · SOPs | SOPs (M): 4.615 -> 5.324 | Hurt |
| add MTE x STE | SLeNet · inference energy proxy | relative energy (SNN = 1.00): 1.00 -> 1.21 | Hurt |
| increase simulation horizon T in MTE (ablation) | SLSTM · SEED, CWRU10 | accuracy and spike rate | Mixed |
| use scale-aligned multi-bit schedule 4321 vs alternatives (ablation) | MTE · MTE schedule ablation (SLSTM) | accuracy and spike rate | Mixed |
| increase STE dictionary width K (ablation) | STE · SRNN on CWRU4/SEED | accuracy and spike rate | Mixed |
Closest papers by meaning
- Inference-Scale Complexity in ANN-SNN Conversion for High-Performance and Low-Power Applications (CVPR 2025 · similarity 0.89)
- Temporal Flexibility in Spiking Neural Networks: Towards Generalization Across Time Steps and Deployment Friendliness (ICLR 2025 · similarity 0.89)
- Training High Performance Spiking Neural Network by Temporal Model Calibration (ICML 2025 · similarity 0.88)
- EnOF-SNN: Training Accurate Spiking Neural Networks via Enhancing the Output Feature (NeurIPS 2024 · similarity 0.88)
- TS-SNN: Temporal Shift Module for Spiking Neural Networks (ICML 2025 · similarity 0.88)
- Temporal Misalignment in ANN-SNN Conversion and its Mitigation via Probabilistic Spiking Neurons (ICML 2025 · similarity 0.88)
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.
