用随机扰动模拟梯度，让物理神经网络也能训练。
Perturbative Gradient Training: A novel training paradigm for bridging the gap between deep neural networks and physical reservoir computing
- 通过参数空间扰动近似梯度，仅靠前向传播实现训练。
- 在仿真与硬件实验中性能接近传统反向传播。
- 适合想用物理器件做高效神经网络的工程师。
我们提出扰动梯度训练（PGT），一种新型训练范式，解决了物理储层计算的核心难题：因物理储层为黑箱而无法进行反向传播。受物理微扰理论启发，PGT 在网络参数空间施加随机扰动，仅通过前向传播即可近似梯度更新。我们在模拟神经网络架构（包括全连接网络和带储层的Transformer模型）以及基于磁振子自振环的实验硬件上验证了该方法的可行性。结果表明，在反向传播不可行或不可能的情况下，PGT 可实现与标准反向传播相当的性能。PGT 为将物理储层融入深层神经网络架构提供了可能，并有望在人工智能训练中实现显著能效提升。
原文摘要 · Abstract (English)
We introduce Perturbative Gradient Training (PGT), a novel training paradigm that overcomes a critical limitation of physical reservoir computing: the inability to perform backpropagation due to the black-box nature of physical reservoirs. Drawing inspiration from perturbation theory in physics, PGT uses random perturbations in the network's parameter space to approximate gradient updates using only forward passes. We demonstrate the feasibility of this approach on both simulated neural network architectures, including a dense network and a transformer model with a reservoir layer, and on experimental hardware using a magnonic auto-oscillation ring as the physical reservoir. Our results show that PGT can achieve performance comparable to that of standard backpropagation methods in cases where backpropagation is impractical or impossible. PGT represents a promising step toward integrating physical reservoirs into deeper neural network architectures and achieving significant energy efficiency gains in AI 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。