提出新训练方法让脉冲神经网络适应不同时间步，提升部署灵活性。
Temporal Flexibility in Spiking Neural Networks: Towards Generalization Across Time Steps and Deployment Friendliness
- 采用随机时间步混合训练，使网络跨时间结构自适应。
- 在N-MNIST和CIFAR10-DVS上实现近乎无损的性能表现，优于传统方法10.1%。
- 首次实现大规模脉冲网络在全事件驱动芯片上的部署，适合边缘计算场景。
脉冲神经网络（SNN）因其类脑机制支持在神经形态硬件上能效高效运行，但现有直接训练方法导致其仅限于固定时间步，造成‘时间僵化’：既阻碍在无时钟事件驱动芯片上的部署，也难以根据动态推理需求调节能耗与性能。本文首次探索了跨时间步泛化的可行性，并提出混合时间步训练（MTT）方法。该方法在每轮训练中为SNN各阶段随机分配不同时间步，通过通信模块传递脉冲信号。训练完成后，在时钟驱动与事件驱动平台均进行部署与评估。实验表明，经MTT训练的模型具备显著的时间灵活性，能在事件驱动与时钟驱动环境下保持近似无损性能（如N-MNIST），并在CIFAR10-DVS上比标准方法高出10.1%；同时提升了网络泛化能力，接近当前最优水平。据我们所知，这是首个在全事件驱动场景下实现大规模SNN部署的研究。
原文摘要 · Abstract (English)
Spiking Neural Networks (SNNs), models inspired by neural mechanisms in the brain, allow for energy-efficient implementation on neuromorphic hardware. However, SNNs trained with current direct training approaches are constrained to a specific time step. This "temporal inflexibility" 1) hinders SNNs' deployment on time-step-free fully event-driven chips and 2) prevents energy-performance balance based on dynamic inference time steps. In this study, we first explore the feasibility of training SNNs that generalize across different time steps. We then introduce Mixed Time-step Training (MTT), a novel method that improves the temporal flexibility of SNNs, making SNNs adaptive to diverse temporal structures. During each iteration of MTT, random time steps are assigned to different SNN stages, with spikes transmitted between stages via communication modules. After training, the weights are deployed and evaluated on both time-stepped and fully event-driven platforms. Experimental results show that models trained by MTT gain remarkable temporal flexibility, friendliness for both event-driven and clock-driven deployment (nearly lossless on N-MNIST and 10.1% higher than standard methods on CIFAR10-DVS), enhanced network generalization, and near SOTA performance. To the best of our knowledge, this is the first work to report the results of large-scale SNN deployment on fully event-driven scenarios.
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。