Research map / Architectures and efficiency / Spiking networks
Cannistraci-Hebb Training on Ultra-Sparse Spiking Neural Networks
Yuan Hua, Jilin Zhang, Yingtao Zhang, Leyi You, Baobo Xiong, Carlo Vittorio Cannistraci, Hong Chen
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 |
|---|---|---|---|
| CH-SNN vs SD-SNN/DPAP/Grad R at higher link sparsity | 6Conv 2FC / 2FC · MNIST, CIFAR-10 | link sparsity: MNIST 2FC 77.36% (DPAP) -> 97.75% | Improved |
| CH-SNN vs SD-SNN accuracy on CIFAR-100 | 6Conv 2FC · CIFAR-100 | accuracy: 75.33% -> 75.22% | No clear effect |
| CH-SNN vs SD-SNN and Grad R on DVS-Gesture | 2Conv 2FC · DVS-Gesture | accuracy: 96.21% -> 95.45% (SD-SNN) | Hurt |
| CH-SNN sparse network vs fully connected on S-TP | S-TP 3FC · DVS-Gesture | accuracy: 89.02% -> 91.29% | Improved |
| CH-SNN sparse network vs fully connected on S-TP | S-TP 3FC · N-MNIST | accuracy: 96.38% -> 96.20% | Hurt |
| CH-SNN sparse network vs fully connected on S-TP | S-TP 3FC · DVS-Gesture | energy: 78mJ -> 0.8mJ | Improved |
| add SSCTI (spike-correlated topological initialization) (ablation) | CH-SNN · CIFAR-10 99% sparsity | accuracy: 10.00% -> 34.55% | Improved |
| add SSWI (sparse spike weight initialization) (ablation) | CH-SNN · CIFAR-10 99% sparsity | accuracy: 10.00% -> 77.40% | Improved |
| add both SSCTI and SSWI (ablation) | CH-SNN · MNIST 99% sparsity | accuracy: 9.80% -> 96.88% | Improved |
| vary batch size (ablation) | CH-SNN · sensitivity analysis | accuracy | No clear effect |
Closest papers by meaning
- Improving the Sparse Structure Learning of Spiking Neural Networks from the View of Compression Efficiency (ICLR 2025 · similarity 0.92)
- Sparse Spiking Neural Network: Exploiting Heterogeneity in Timescales for Pruning Recurrent SNN (ICLR 2024 · similarity 0.89)
- QP-SNN: Quantized and Pruned Spiking Neural Networks (ICLR 2025 · similarity 0.88)
- SpikingSSMs: Learning Long Sequences with Sparse and Parallel Spiking State Space Models (AAAI 2025 · similarity 0.88)
- Brain network science modelling of sparse neural networks enables Transformers and LLMs to perform as fully connected (NeurIPS 2025 · similarity 0.88)
- Take A Shortcut Back: Mitigating the Gradient Vanishing for Training Spiking Neural Networks (NeurIPS 2024 · 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.
