近日，美国康奈尔大学McMahon, Peter L.团队提出了机器学习中多模波传播的任意控制。相关论文于2025年12月8日发表在《自然—物理学》杂志上。

受控的多模波传播可以实现比基于单模波导连接的离散元件的架构更节省空间的光子处理器。代替定义离散元素，人们可以雕刻光子处理器的连续衬底，通过二维多模干涉来执行计算。

研究组设计并展示了一个折射率的设备，可以在空间上快速重新编程，允许任意控制波的传播。该器件是一种二维可编程波导，使用具有约104个可编程空间自由度的平板波导折射率的平行电光调制。研究组在没有数字预处理或后处理的情况下，在单次通过多达49维向量的基准任务上实现了神经网络推理。理论和数值分析进一步表明，二维可编程波导不仅可以提供恒定因子的器件面积减少，而且还可以提供缩放效益，所需面积以N1.5而不是N2增长。

附：英文原文

Title: Arbitrary control over multimode wave propagation for machine learning

Author: Onodera, Tatsuhiro, Stein, Martin M., Ash, Benjamin A., Sohoni, Mandar M., Bosch, Melissa, Yanagimoto, Ryotatsu, Jankowski, Marc, McKenna, Timothy P., Wang, Tianyu, Shvets, Gennady, Shcherbakov, Maxim R., Wright, Logan G., McMahon, Peter L.

Issue&Volume: 2025-12-08

Abstract: Controlled multimode wave propagation can enable more space-efficient photonic processors than architectures based on discrete components connected by single-mode waveguides. Instead of defining discrete elements, one can sculpt the continuous substrate of a photonic processor to perform computations through multimode interference in two dimensions. Here we designed and demonstrated a device with a refractive index that can be rapidly reprogrammed across space, allowing arbitrary control of wave propagation. The device, a two-dimensional programmable waveguide, uses parallel electro-optic modulation of the refractive index of a slab waveguide with about 104 programmable spatial degrees of freedom. We implemented neural network inference on benchmark tasks with up to 49-dimensional vectors in a single pass, without digital pre-processing or post-processing. Theoretical and numerical analyses further indicated that two-dimensional programmable waveguides may offer not only a constant-factor reduction in device area but also a scaling benefit, with the area required growing as N1.5 rather than N2.

DOI: 10.1038/s41567-025-03094-2

Source: https://www.nature.com/articles/s41567-025-03094-2