比较神经网络能效需考虑表达能力，否则结论不靠谱。
Towards an Expressivity-Normalized Energy-Demand Comparison of ANNs and SNNs
- 用理论模型对比相同表达力的ANN与SNN能效
- 发现仅在特定宽度、稀疏度和深度下SNN才更省电
- 为时序网络设计提供能效与容量平衡的指导
脉冲神经网络（SNNs）常被视为人工神经网络（ANNs）的节能替代方案，但其优势依赖于网络结构和数据特性。本文建立了一个分析框架，针对时间序列数据，比较全连接ReLU ANNs与积分-放电SNNs在匹配表达能力下的理论能效。通过将推理能耗模型与表征表达力的理论边界关联，推导出表达力归一化的效率比及网络宽度、脉冲稀疏度和ANN深度缩放的明确阈值。分析揭示了事件驱动计算如何抵消SNN的时间开销，明确了SNN能效优势的适用范围，提供了面向容量感知的时序网络节能设计原则。结果表明，仅在特定条件下ANN的能效才优于SNN。
原文摘要 · Abstract (English)
Spiking neural networks (SNNs) are often regarded as energy-efficient alternatives to artificial neural networks (ANNs), yet their advantage depends critically on both network architecture and data properties. We develop an analytical framework to compare fully-connected ReLU ANNs and integrate-and-fire SNNs for time-series data with respect to their theoretical energy efficiency at matched expressive capacity. By relating an inference-energy model to theoretical bounds on representational expressivity, we derive an expressivity-normalized efficiency ratio and explicit thresholds in network width, spike sparsity, and ANN depth scaling. Our analysis characterizes the regimes in which event-driven computation offsets the temporal overhead of SNNs, providing capacity-aware principles for designing energy-efficient temporal networks. It shows that ANNs exceed SNNs in expressivity-normalized efficiency only in specific regimes.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。