Research map / Architectures and efficiency / Spiking networks
Addressing the speed-accuracy simulation trade-off for adaptive spiking neurons
Luke Taylor, Andrew J King, Nicol Spencer Harper
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 |
|---|---|---|---|
| use Blocks reformulation instead of standard ALIF | ALIF SNN · synthetic benchmark T=2048 | training speedup: 53 | Improved |
| shorten simulation length from 2048 to 512 (ablation) | Blocks ALIF · synthetic benchmark | training speedup: 53x -> 36x | Hurt |
| use Blocks reformulation (ARP=10ms) | ALIF SNN · N-MNIST | accuracy: 97.71% -> 96.97% | Hurt |
| use Blocks reformulation (ARP=20ms) | ALIF SNN · SHD | accuracy: 87.48% -> 86.07% | Hurt |
| use Blocks reformulation with largest ARP | ALIF SNN · N-MNIST | training epoch time: 2034s -> 181s | Improved |
| use Blocks reformulation with largest ARP | ALIF SNN · SHD | training epoch time: 268s -> 20s | Improved |
| increase ARP to the largest tested values (ablation) | Blocks ALIF · N-MNIST and SHD | accuracy | Hurt |
| decrease DT from 4ms to 0.1ms (ablation) | Blocks ALIF · Allen V1 neuron fits | ETV: 0.66 -> 0.80 | Improved |
| use Blocks instead of standard ALIF for fitting | ALIF SNN · Allen V1 neuron fits | fit time per neuron: 108.4s -> 15.5s | Improved |
| use Blocks instead of standard ALIF for fitting | ALIF SNN · Allen V1 neuron fits | ETV: ~0.8 | No clear effect |
| increase ARP at fixed DT=0.1ms (ablation) | Blocks ALIF · Allen V1 neuron fits | fit accuracy | Mixed |
| only let surrogate gradients flow through non-recurrent connectivity (ablation) | Blocks ALIF · spiking classification | accuracy | Improved |
Closest papers by meaning
- CLIF: Complementary Leaky Integrate-and-Fire Neuron for Spiking Neural Networks (ICML 2024 · similarity 0.87)
- Adaptive Calibration: A Unified Conversion Framework of Spiking Neural Networks (AAAI 2025 · similarity 0.86)
- TC-LIF: A Two-Compartment Spiking Neuron Model for Long-Term Sequential Modelling (AAAI 2024 · similarity 0.86)
- Efficient Parallel Training Methods for Spiking Neural Networks with Constant Time Complexity (ICML 2025 · similarity 0.86)
- Neuronal Self-Adaptation Enhances Capacity and Robustness of Representation in Spiking Neural Networks (NeurIPS 2026 · similarity 0.86)
- P-SPIKESSM: HARNESSING PROBABILISTIC SPIKING STATE SPACE MODELS FOR LONG-RANGE DEPENDENCY TASKS (ICLR 2025 · 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.
