S$^2$NN: Sub-bit Spiking Neural Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wei, Wenjie, Zhang, Malu, Zhang, Jieyuan, Belatreche, Ammar, Wang, Shuai, Shan, Yimeng, Liu, Hanwen, Cao, Honglin, Wang, Guoqing, Yang, Yang, Li, Haizhou
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917038450016256
author Wei, Wenjie
Zhang, Malu
Zhang, Jieyuan
Belatreche, Ammar
Wang, Shuai
Shan, Yimeng
Liu, Hanwen
Cao, Honglin
Wang, Guoqing
Yang, Yang
Li, Haizhou
author_facet Wei, Wenjie
Zhang, Malu
Zhang, Jieyuan
Belatreche, Ammar
Wang, Shuai
Shan, Yimeng
Liu, Hanwen
Cao, Honglin
Wang, Guoqing
Yang, Yang
Li, Haizhou
contents Spiking Neural Networks (SNNs) offer an energy-efficient paradigm for machine intelligence, but their continued scaling poses challenges for resource-limited deployment. Despite recent advances in binary SNNs, the storage and computational demands remain substantial for large-scale networks. To further explore the compression and acceleration potential of SNNs, we propose Sub-bit Spiking Neural Networks (S$^2$NNs) that represent weights with less than one bit. Specifically, we first establish an S$^2$NN baseline by leveraging the clustering patterns of kernels in well-trained binary SNNs. This baseline is highly efficient but suffers from \textit{outlier-induced codeword selection bias} during training. To mitigate this issue, we propose an \textit{outlier-aware sub-bit weight quantization} (OS-Quant) method, which optimizes codeword selection by identifying and adaptively scaling outliers. Furthermore, we propose a \textit{membrane potential-based feature distillation} (MPFD) method, improving the performance of highly compressed S$^2$NN via more precise guidance from a teacher model. Extensive results on vision tasks reveal that S$^2$NN outperforms existing quantized SNNs in both performance and efficiency, making it promising for edge computing applications.
format Preprint
id arxiv_https___arxiv_org_abs_2509_24266
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle S$^2$NN: Sub-bit Spiking Neural Networks
Wei, Wenjie
Zhang, Malu
Zhang, Jieyuan
Belatreche, Ammar
Wang, Shuai
Shan, Yimeng
Liu, Hanwen
Cao, Honglin
Wang, Guoqing
Yang, Yang
Li, Haizhou
Computer Vision and Pattern Recognition
Spiking Neural Networks (SNNs) offer an energy-efficient paradigm for machine intelligence, but their continued scaling poses challenges for resource-limited deployment. Despite recent advances in binary SNNs, the storage and computational demands remain substantial for large-scale networks. To further explore the compression and acceleration potential of SNNs, we propose Sub-bit Spiking Neural Networks (S$^2$NNs) that represent weights with less than one bit. Specifically, we first establish an S$^2$NN baseline by leveraging the clustering patterns of kernels in well-trained binary SNNs. This baseline is highly efficient but suffers from \textit{outlier-induced codeword selection bias} during training. To mitigate this issue, we propose an \textit{outlier-aware sub-bit weight quantization} (OS-Quant) method, which optimizes codeword selection by identifying and adaptively scaling outliers. Furthermore, we propose a \textit{membrane potential-based feature distillation} (MPFD) method, improving the performance of highly compressed S$^2$NN via more precise guidance from a teacher model. Extensive results on vision tasks reveal that S$^2$NN outperforms existing quantized SNNs in both performance and efficiency, making it promising for edge computing applications.
title S$^2$NN: Sub-bit Spiking Neural Networks
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.24266