An international research consortium, spearheaded by Aalto University in Finland, has unveiled a groundbreaking method. This approach allows for the execution of intricate tensor operations through a single instance of light propagation, thereby performing crucial computational tasks in deep - learning at the blistering speed of light.

Traditional computing methods often rely on electronic circuits and active regulation. In contrast, this innovative method harnesses the natural propagation of light waves. The crux of this technique is that it encodes digital data into both the amplitude and phase of light. When light fields interact, matrix and tensor operations can be carried out spontaneously.

This remarkable breakthrough opens up an entirely new avenue for the advancement of general AI hardware. It also offers a promising solution to overcome the performance limitations that currently plague existing computing platforms.