[2603.24692] Reconstructing Spiking Neural Networks Using a Single Neuron with Autapses

Computer Science > Neural and Evolutionary Computing arXiv:2603.24692 (cs) [Submitted on 25 Mar 2026] Title:Reconstructing Spiking Neural Networks Using a Single Neuron with Autapses Authors:Wuque Cai, Hongze Sun, Quan Tang, Shifeng Mao, Zhenxing Wang, Jiayi He, Duo Chen, Dezhong Yao, Daqing Guo View a PDF of the paper titled Reconstructing Spiking Neural Networks Using a Single Neuron with Autapses, by Wuque Cai and Hongze Sun and Quan Tang and Shifeng Mao and Zhenxing Wang and Jiayi He and Duo Chen and Dezhong Yao and Daqing Guo View PDF HTML (experimental) Abstract:Spiking neural networks (SNNs) are promising for neuromorphic computing, but high-performing models still rely on dense multilayer architectures with substantial communication and state-storage costs. Inspired by autapses, we propose time-delayed autapse SNN (TDA-SNN), a framework that reconstructs SNNs with a single leaky integrate-and-fire neuron and a prototype-learning-based training strategy. By reorganizing internal temporal states, TDA-SNN can realize reservoir, multilayer perceptron, and convolution-like spiking architectures within a unified framework. Experiments on sequential, event-based, and image benchmarks show competitive performance in reservoir and MLP settings, while convolutional results reveal a clear space--time trade-off. Compared with standard SNNs, TDA-SNN greatly reduces neuron count and state memory while increasing per-neuron information capacity, at the cost of additional temporal...