BASC : Behavior-Aligned Quantization and Pruning for Low-Bit Spiking Neural Networks

2026-08-11

AI Summary

为解决低比特脉冲神经网络的计算和存储成本问题，提出BASC框架，通过时间行为校正和边界级通道校正方法优化量化和剪枝。

Abstract

Spiking Neural Networks (SNNs) encode information through binary spikes and compute in an event-driven manner, offering an energy-efficient paradigm for machine intelligence. However, high-performance SNNs incur substantial memory and timestep-wise computation costs that hinder deployment on resource-constrained devices. Quantization and pruning provide complementary routes to reducing these costs, yet both make their decisions with local criteria that overlook temporal task feedback in quantization and inter-channel dependencies in pruning. Consequently, optimizing either criterion can still yield suboptimal compression performance. We refer to this discrepancy as criterion-behavior mismatch and propose Behavior-Aligned SNN Compression (BASC), a unified framework with two lightweight modules. For quantization, the scale is applied to synaptic current at every timestep and therefore shifts spike timing. Temporal-Behavior Scale Correction (TSC) makes the scale learnable under a temporal loss, allowing firing behavior to inform scale optimization. For pruning, channel importance depends on how channels jointly drive the membrane potential across the firing threshold. Boundary-Level Inter-Channel Correction (BIC) uses channelwise importance scores for initial selection and inter-channel information to re-evaluate only channels near the pruning threshold. Extensive experiments on static and neuromorphic benchmarks show that lower-bit BASC models match or outperform higher-bit baselines and retain this accuracy advantage after structured pruning, while further reducing model storage and synaptic operations.

Problem

• Spiking Neural Networks
• Quantization
• Pruning
• Memory Cost
• Computation Cost

Innovation

• Behavior-Aligned SNN Compression
• Temporal-Behavior Scale Correction
• Boundary-Level Inter-Channel Correction