提升传感器端脉冲神经网络的部署精度，解决训练与实际运行的差距问题。
Sharpness-Aware Surrogate Training for On-Sensor Spiking Neural Networks
- 在脉冲网络训练中引入平滑感知优化，使梯度更准确。
- 在N-MNIST上硬阈值精度从65.7%提升至94.7%，DVS手势数据集从31.8%升至63.3%。
- 适合资源受限的边缘设备，尤其对硬件量化和低功耗部署有显著帮助。
脉冲神经网络（SNN）是传感器端视觉计算的理想模型，但传统代理梯度训练方法在部署时使用硬阈值脉冲后性能会急剧下降，形成代理到硬脉冲的差距，直接限制了传感器端的准确率。本文提出尖锐感知代理训练（SAST），将尖锐感知最小化（SAM）应用于代理前向传播的SNN，使训练目标平滑且梯度精确，作为缩小该差距的有效策略。在双事件相机基准测试中，仅替换脉冲方式的硬脉冲精度在N-MNIST上从65.7%提升至94.7%，在DVS Gesture上从31.8%提升至63.3%。在硬件感知推理模拟下（INT8/INT4权重量化、定点膜电位、离散泄漏因子），SAST仍表现优异：在N-MNIST上硬脉冲精度从47.6%升至96.9%（INT8）和43.2%升至81.0%（INT4），DVS Gesture上从25.3%升至47.6%（INT8）和26.0%升至43.8%（INT4）。同时，计算量也显著降低，如N-MNIST（INT8）从1734k降至1315k，DVS Gesture（INT8）从86221k降至4323k。结果表明SAST在特定条件下是实现高效传感器端脉冲推理的重要工具。
原文摘要 · Abstract (English)
Spiking neural networks (SNNs) are a natural computational model for on-sensor and near-sensor vision, where event driven processors must operate under strict power budgets with hard binary spikes. However, models trained with surrogate gradients often degrade sharply when the smooth surrogate nonlinearity is replaced by a hard threshold at deployment; a surrogate-to-hard transfer gap that directly limits on-sensor accuracy. We study Sharpness-Aware Surrogate Training (SAST), which applies Sharpness-Aware Minimization (SAM) to a surrogate-forward SNN so that the training objective is smooth and the gradient is exact, and position it as one gap-reduction strategy under the tested settings rather than the only viable mechanism. Under explicit contraction assumptions we provide state-stability, input-Lipschitz, and smoothness bounds, together with a corresponding nonconvex convergence result. On two event-camera benchmarks, swap-only hard-spike accuracy improves from 65.7\% to 94.7\% on N-MNIST and from 31.8\% to 63.3\% on DVS Gesture. Under a hardware-aware inference simulation (INT8/INT4 weight quantization, fixed-point membrane potentials, discrete leak factors), SAST remains strong: on N-MNIST, hard-spike accuracy improves from 47.6\% to 96.9\% (INT8) and from 43.2\% to 81.0\% (INT4), while on DVS Gesture it improves from 25.3\% to 47.6\% (INT8) and from 26.0\% to 43.8\% (INT4). SynOps also decrease under the same hardware-aware setting, including 1734k$\rightarrow$1315k (N-MNIST, INT8) and 86221k$\rightarrow$4323k (DVS Gesture, INT8). These results suggest that SAST is a promising component in a broader toolbox for on-sensor spiking inference under the tested settings.
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。