Research map / Architectures and efficiency / Spiking networks
EICIL: Joint Excitatory Inhibitory Cycle Iteration Learning for Deep Spiking Neural Networks
Zihang Shao, Xuanye Fang, Yaxin Li, Chaoran Feng, Jiangrong Shen, Qi Xu
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 GSI cycle iteration (GS + STDP-BW) instead of GS only | Spiking ResNet18 · CIFAR100 | test accuracy (%): 56.28 -> 63.03 | Improved |
| add GSGI cycle iteration (GS + STDP-BW-GS) instead of GS only | Spiking ResNet18 · CIFAR100 | test accuracy (%): 56.28 -> 63.47 | Improved |
| add GSGI instead of GS only | Spiking ResNet18 · CIFAR10 | test accuracy (%): 89.39 -> 90.34 | Improved |
| add GSGI instead of GS only | Spiking VGG11 · CIFAR10 | test accuracy (%): 83.96 -> 85.31 | Improved |
| add GSGI instead of GS only | Spiking VGG11 · CIFAR100 | test accuracy (%): 54.01 -> 55.71 | Improved |
| add GSGI instead of GS only | Spiking VGG11 · MNIST | test accuracy (%): 98.2 -> 98.77 | Improved |
| add GSGI instead of GS only | Excitatory inhibition Net · CIFAR100 | test accuracy (%): 53.61 -> 53.86 | Improved |
| add GSGI instead of GS only | Excitatory inhibition Net · MNIST | test accuracy (%): 99.33 -> 99.35 | No clear effect |
| replace GS with STDP-BW (backprop added to multi-layer STDP) (ablation) | Spiking ResNet18 · CIFAR10 | test accuracy (%): 89.39 -> 90.39 | Improved |
| replace GS with STDP-BW (ablation) | Spiking ResNet18 · CIFAR100 | test accuracy (%): 56.28 -> 56.20 | No clear effect |
| replace GS with STDP-BW-GS (ablation) | Spiking VGG11 · MNIST | test accuracy (%): 98.2 -> 95.85 | Hurt |
| use STDP-BW-GS instead of STDP-BW in the cycle iteration (GSGI vs GSI) (ablation) | Spiking ResNet18 · CIFAR10 | test accuracy (%): 90.05 -> 90.34 | Improved |
Closest papers by meaning
- Adaptive deep spiking neural network with global-local learning via balanced excitatory and inhibitory mechanism (ICLR 2024 · similarity 0.89)
- Balanced Resonate-and-Fire Neurons (ICML 2024 · similarity 0.85)
- NDOT: Neuronal Dynamics-based Online Training for Spiking Neural Networks (ICML 2024 · similarity 0.85)
- Autaptic Synaptic Circuit Enhances Spatio-temporal Predictive Learning of Spiking Neural Networks (ICML 2024 · similarity 0.85)
- TIDE: Asymmetric Neural Circuits for Stabilized Temporal Inhibitory-Excitatory Dynamics (NeurIPS 2026 · similarity 0.84)
- Take A Shortcut Back: Mitigating the Gradient Vanishing for Training Spiking Neural Networks (NeurIPS 2024 · similarity 0.84)
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.
