Spatial photonic Ising machine: an emerging free-space optical computing architecture [Invited]

Free-space optical computing is a domain-specific architecture that performs information processing based on optical modulation, control, and free-space light propagation. It inherently offers useful features such as spatial parallelism, high-throughput processing using passive devices, and lossless information transmission. Fueled by the explosive growth and rapid advancement of artificial intelligence in recent years, attention to free-space optical computing has been revived over the past decade. In such circumstances, the spatial photonic Ising machine (SPIM) has emerged as a novel free-space optical computing architecture. SPIM is expected to serve as a dedicated solver for combinatorial optimization problems with high speed and efficiency, motivating efforts to understand its characteristics and drive innovations in performance and functionality. This review provides a detailed overview of the principles and current statuses of free-space optical computing to highlight its potential. We then summarize recent advances in SPIM models, implementation techniques, and the exploration of new functionalities, and offer insights into future directions for this rapidly evolving field.