Reproduce “The Dormant Spiking Neuron: A State-Driven Mechanism for Efficient Spiking Neural Networks” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ruichen Ma, Yicai Chen, Pujun Zhou, Yue Zuo, Chongyang Li, Shuang Liu, Shaogang Hu, Guanchao Qiao
The authors introduce DSN, a gate that stops a spiking neuron’s computation for an input when its membrane voltage falls below a negative threshold. Unlike input-driven sparsity, this also skips membrane updates for dormant neurons. In ImageNet ResNet-18, DSN BPTT reached 64.44% accuracy with 1.17 mJ estimated energy, versus 63.10% and 4.29 mJ for the baseline; these are analytical estimates, not device measurements.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.