The proposed all-optical autoencoder (OAE) framework can simultaneously achieve image reconstruction, representation, and generation by leveraging the non-reciprocal property of diffractive deep neural networks (D2NNs).
Optical next generation reservoir computing (NGRC) driven by time-delay inputs can generate implicit polynomial features in the speckle patterns, enabling superior performance in chaotic time series forecasting and unmeasured state variable prediction compared to conventional optical reservoir computing.
The core message of this paper is that a gradient-based model-free optimization (G-MFO) method can efficiently train optical computing systems in situ without relying on computationally heavy and biased system simulations, overcoming the simulation-to-reality gaps faced by conventional training approaches.
Optical processors can efficiently perform high-dimensional linear operations at the speed of light, but limitations in precision require a quantization-aware training framework to optimize performance.
Optical random neural networks can achieve high performance through genetically programmable random projection kernels.
Optical signal processing enhances noise resilience in visual perception tasks.