ReDON: Recurrent Diffractive Optical Neural Processor with Reconfigurable Self-Modulated Nonlinearity

Diffractive optical neural networks (DONNs) have demonstrated unparalleled energy efficiency and parallelism by processing information directly in the optical domain. However, their computational expressivity is constrained by static, passive diffractive phase masks that lack efficient nonlinear responses and reprogrammability. To address these limitations, we introduce Recurrent Diffractive Optical Neural Processor (

), a novel architecture featuring reconfigurable, recurrent self-modulated nonlinearity. This mechanism enables dynamic, input-dependent optical transmission through in-situ electro-optic self-modulation, providing a highly efficient and reprogrammable approach to optical computation. Inspired by the gated linear unit (GLU) in large language models, senses a fraction of the propagating optical field and modulates its phase or intensity via a lightweight, parametric function, enabling effective nonlinearity with minimal inference overhead. As a non-von Neumann architecture with the main weighting units (metasurfaces) being fixed, we substantially extend the DONN’s nonlinear representational capacity and task adaptability via recurrent optical hardware reuse and dynamically tunable nonlinearity. We systematically investigate various self-modulation configurations to uncover the trade-offs between hardware efficiency and expressivity. On image recognition and segmentation tasks, improves test accuracy and mIoU by up to 20% over prior DONNs with optical or digital nonlinearities at comparable complexity and negligible power overhead. This work establishes a new paradigm for reconfigurable nonlinear optical computing, uniting the benefits of recurrence and self-modulation in non-von Neumann analog processors.