用脉冲神经网络实现低功耗时序预测，精度更高且省电超96%
SpikySpace: A Spiking State Space Model for Energy-Efficient Time Series Forecasting
- 用脉冲选择性扫描替代注意力，将计算复杂度从二次降至线性
- 在4个数据集上精度比现有SNN高3.0%，能耗降低96.1%以上
- 首个全脉冲状态空间模型，适合边缘设备部署
交通管理与工业监控等领域的时序预测常需在资源受限的边缘设备上实现实时、低功耗处理。脉冲神经网络（SNN）具备事件驱动与超低功耗特性，适于该场景。但现有SNN时序预测模型多依赖复杂变压器模块。为此，本文提出SpikySpace，一种脉冲状态空间模型（SSM），通过脉冲选择性扫描将注意力模块的二次计算成本降至线性。进一步提出PTsoftplus和PTSiLU两种高效近似函数，用简单位移操作替代昂贵的指数与除法运算。在四个多变量时间序列基准上评估，SpikySpace相比领先SNN在精度上提升最高达3.0%，同时能耗降低超过96.1%。作为首个完全脉冲的状态空间模型，SpikySpace实现了类脑效率与现代序列建模的融合，为高效时序预测系统提供了可行路径。代码已开源。
原文摘要 · Abstract (English)
Time-series forecasting in domains like traffic management and industrial monitoring often requires real-time, energy-efficient processing on edge devices with limited resources. Spiking neural networks (SNNs) offer event-driven computation and ultra-low power and have been proposed for use in this space. Unfortunately, existing SNN-based time-series forecasters often use complex transformer blocks. To address this issue, we propose SpikySpace, a spiking state-space model (SSM) that reduces the quadratic cost in the attention block to linear time via spiking selective scanning. Further, we introduce PTsoftplus and PTSiLU, two efficient approximations of SiLU and Softplus that replace costly exponential and division operations with simple bit-shifts. Evaluated on four multivariate time-series benchmarks, SpikySpace outperforms the leading SNN in terms of accuracy by up to 3.0% while reducing energy consumption by over 96.1%. As the first fully spiking state-space model, SpikySpace bridges neuromorphic efficiency with modern sequence modeling, opening a practical path toward efficient time series forecasting systems. Our code is available at https://anonymous.4open.science/r/SpikySpace.
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。