All-optical temporal integration mediated by subwavelength heat antennas
This work addresses the dimensional constraints in optical computing for AI workloads, representing a significant leap rather than an incremental improvement.
The paper tackled the problem of integrated optical computing systems being limited to low-dimensional tensor operations, and demonstrated an all-optical neuromorphic system capable of processing input vectors exceeding 250,000 elements using time division multiplexing and thermo-optic modulation.
Optical computing systems deliver unrivalled processing speeds for scalar operations. Yet, integrated implementations have been constrained to low-dimensional tensor operations that fall short of the vector dimensions required for modern artificial intelligence. We demonstrate an all-optical neuromorphic computing system based on time division multiplexing, capable of processing input vectors exceeding 250,000 elements within a unified framework. The platform harnesses optically driven thermo-optic modulation in standing wave optical fields, with titanium nano-antennas functioning as wavelength-selective absorbers. Counterintuitively, the thermal time dynamics of the system enable simultaneous time integration of ultra-fast (50GHz) signals and the application of programmable, non-linear activation functions, entirely within the optical domain. This unified framework constitutes a leap towards large-scale photonic computing that satisfies the dimensional requirements of AI workloads.