Sound waves enable faster, efficient photonic AI computing

Researchers have demonstrated an all-optical activation function using sound waves, a key step toward more energy-efficient photonic computing for AI. This advancement could lead to faster and more efficient AI systems. The new activation function, developed by researchers at the Max Planck Institute, Leibniz University Hannover, and MIT, utilizes stimulated Brillouin scattering to introduce nonlinearity in optical neural networks. This allows for all-optical processing, avoiding the need for electronic conversions. This research is significant because photonic computing offers potential advantages in speed and energy efficiency compared to traditional electronic systems, particularly for complex AI tasks. The team demonstrated the function's versatility by implementing different activation functions.