Research map / Architectures and efficiency / Spiking networks
Training High-Performance Low-Latency Spiking Neural Networks by Differentiation on Spike Representation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 DSR with IF neurons instead of the strongest listed direct-training baseline | PreAct-ResNet-18 · CIFAR-10 | accuracy: 92.52%±0.17% -> 95.24%±0.17% | Helped |
| Use DSR with LIF neurons instead of the strongest listed direct-training baseline | PreAct-ResNet-18 · CIFAR-100 | accuracy: 73.07%±0.21 % -> 78.50%±0.12% | Helped |
| Use DSR with IF neurons instead of the strongest listed direct-training baseline | PreAct-ResNet-18 · ImageNet | accuracy: 67.04% -> 67.74% | Helped |
| Use DSR with LIF neurons instead of the strongest listed baseline | VGG-11 · DVS-CIFAR10 | accuracy: 74.80% -> 77.27%±0.24% | Helped |
| Add firing mechanism modification (ablation) | DSR, PreAct-ResNet-18 (IF), init. V_th=6 · CIFAR-10 | accuracy: 92.88%±0.25% -> 95.24%±0.17% | Helped |
| Add trainable spike thresholds (ablation) | DSR, PreAct-ResNet-18 (IF), V_th=6 · CIFAR-10 | accuracy: 90.45%±1.84% -> 95.24%±0.17% | Helped |
| Add trainable spike thresholds (ablation) | DSR, PreAct-ResNet-18 (IF), V_th=2 · CIFAR-10 | accuracy: 90.47%±0.12% -> 95.24%±0.17% | Helped |
| Add both firing mechanism modification and trainable thresholds (ablation) | DSR, PreAct-ResNet-18 (IF), init. V_th=6 · CIFAR-10 | accuracy: 92.59%±0.81% -> 95.24%±0.17% | Helped |
| Use firing mechanism modification alone rather than neither error-reduction method (ablation) | DSR, PreAct-ResNet-18 (IF), V_th=6 · CIFAR-10 | accuracy: 92.59%±0.81% -> 92.88%±0.25% | Hurt |
Closest papers by meaning
- Direct Training for Spiking Neural Networks: Faster, Larger, Better (AAAI 2019 · similarity 0.93)
- Going Deeper With Directly-Trained Larger Spiking Neural Networks (AAAI 2021 · similarity 0.92)
- Enhancing Training of Spiking Neural Network with Stochastic Latency (AAAI 2024 · similarity 0.91)
- EnOF-SNN: Training Accurate Spiking Neural Networks via Enhancing the Output Feature (NeurIPS 2024 · similarity 0.90)
- Enhanced Self-Distillation Framework for Efficient Spiking Neural Network Training (NeurIPS 2025 · similarity 0.90)
- Faster and Stronger: When ANN-SNN Conversion Meets Parallel Spiking Calculation (ICML 2025 · similarity 0.89)
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.