High-performance real-world optical computing trained by in situ gradient-based model-free optimization
This work addresses the challenge of efficient training for optical computing systems, potentially accelerating their transition to real-world applications, though it appears incremental as it builds on existing gradient estimation techniques.
The authors tackled the problem of computationally demanding training and simulation-to-reality gaps in optical computing systems by proposing a gradient-based model-free optimization method, which outperformed hybrid training on MNIST and FMNIST datasets and enabled high-speed classification of cells from phase maps.
Optical computing systems provide high-speed and low-energy data processing but face deficiencies in computationally demanding training and simulation-to-reality gaps. We propose a gradient-based model-free optimization (G-MFO) method based on a Monte Carlo gradient estimation algorithm for computationally efficient in situ training of optical computing systems. This approach treats an optical computing system as a black box and back-propagates the loss directly to the optical computing weights' probability distributions, circumventing the need for a computationally heavy and biased system simulation. Our experiments on diffractive optical computing systems show that G-MFO outperforms hybrid training on the MNIST and FMNIST datasets. Furthermore, we demonstrate image-free and high-speed classification of cells from their marker-free phase maps. Our method's model-free and high-performance nature, combined with its low demand for computational resources, paves the way for accelerating the transition of optical computing from laboratory demonstrations to practical, real-world applications.