Research map / Probabilistic and scientific modeling / Variational and generative sampling
RecDis-SNN: Rectifying Membrane Potential Distribution for Directly Training Spiking Neural Networks
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 |
|---|---|---|---|
| Add MPD-Loss, combining degeneration, saturation, and gradient-mismatch penalties (ablation) | CIFARNet · CIFAR10 | accuracy: 89.83% -> 92.08% | Helped |
| Add MPD-Loss, combining degeneration, saturation, and gradient-mismatch penalties (ablation) | ResNet-19 · CIFAR10 | accuracy: 91.23% -> 94.34% | Helped |
| Add degeneration loss only (ablation) | CIFARNet · CIFAR10 | accuracy: 89.83% -> 91.29% | Helped |
| Add saturation loss only (ablation) | CIFARNet · CIFAR10 | accuracy: 89.83% -> 91.01% | Helped |
| Add gradient-mismatch loss only (ablation) | CIFARNet · CIFAR10 | accuracy: 89.83% -> 91.72% | Helped |
| Combine tdBN with MPD-Loss (ablation) | CIFARNet · CIFAR10 | accuracy: 90.69% -> 92.20% | Helped |
| Combine tdBN with MPD-Loss (ablation) | ResNet-19 · CIFAR10 | accuracy: 92.92% -> 95.53% | Helped |
| Use MPD-Loss instead of tdBN (ablation) | ResNet-19 · CIFAR10 | average quantization error: 0.78 -> 0.66 | Helped |
| Use MPD-Loss instead of tdBN (ablation) | ResNet-19 · CIFAR100 | average quantization error: 0.77 -> 0.64 | Helped |
| Use MPD-Loss instead of tdBN (ablation) | ResNet-19 · CIFAR10-DVS | average quantization error: 0.83 -> 0.61 | Helped |
| Add MPD-Loss during training (ablation) | CIFARNet · CIFAR10 | training time (min.): 2524.82 -> 2948.88 | Hurt |
| Add MPD-Loss to directly trained SNNs | RecDis-SNN · CIFAR10 | accuracy: 93.16% -> 95.55% ±0.05 | Helped |
Closest papers by meaning
- Rethinking the Membrane Dynamics and Optimization Objectives of Spiking Neural Networks (NeurIPS 2024 · similarity 0.87)
- Enhanced Self-Distillation Framework for Efficient Spiking Neural Network Training (NeurIPS 2025 · similarity 0.86)
- Efficient ANN-Guided Distillation: Aligning Rate-based Features of Spiking Neural Networks through Hybrid Block-wise Replacement (CVPR 2025 · similarity 0.85)
- Rethinking Spiking Neural Networks from an Ensemble Learning Perspective (ICLR 2025 · similarity 0.85)
- Synergy Between the Strong and the Weak: Spiking Neural Networks are Inherently Self-Distillers (NeurIPS 2025 · similarity 0.85)
- EnOF-SNN: Training Accurate Spiking Neural Networks via Enhancing the Output Feature (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.