提出可全局最优训练脉冲神经网络的新方法，解决传统训练中的误差累积问题。
Globally Optimal Training of Spiking Neural Networks via Parameter Reconstruction
- 基于参数重构的理论框架，实现脉冲神经网络的全局最优训练
- 在多个任务中表现显著优于传统方法，且与代理梯度法兼容
- 具备良好数据扩展性与模型鲁棒性，适合大规模脉冲神经网络训练
脉冲神经网络（SNN）作为生物合理且能效更高的人工神经网络替代方案备受关注。然而，由于脉冲函数不可导，其训练通常依赖代理梯度，导致多层累积的近似误差。本文将平行前馈阈值网络的凸化理论扩展至平行循环阈值网络，该类网络可包含平行SNN作为结构特例。基于此理论框架，我们提出一种用于SNN训练的参数重构算法，在多种任务中均展现出一致且显著的优势，既可独立使用，也可与代理梯度训练结合。消融实验进一步验证了该算法在数据规模上的可扩展性及对模型配置的鲁棒性，表明其在大规模SNN训练中的潜力。
原文摘要 · Abstract (English)
Spiking Neural Networks (SNNs) have been proposed as biologically plausible and energy-efficient alternatives to conventional Artificial Neural Networks (ANNs). However, the training of SNN usually relies on surrogate gradients due to the non-differentiability of the spike function, introducing approximation errors that accumulate across layers. To address this challenge, we extend the work on convexification of parallel feedforward threshold networks to parallel recurrent threshold networks, which subsume parallel SNNs as a structured special case. Building on this theoretical framework, we propose a parameter reconstruction algorithm for SNN training that demonstrates consistent and significant advantages across various tasks, both as a standalone method and in combination with surrogate-gradient training. The ablations further demonstrate the data scalability and robustness to model configurations of our training algorithm, pointing toward its potential in large-scale SNN training.
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。