Return
Multiplying matrices in a single pass with light
DOI:10.1038/s41566-025-01828-5.png)
Abstract
En 中文
Optical computing has been limited to vector–matrix multiplications, with matrix–matrix operations requiring wavelength- or time-division multiplexing, reducing energy efficiency and speed. Now, researchers have demonstrated a free-space optical approach that overcomes these limitations, enabling parallel matrix–matrix and tensor–matrix multiplications in a single optical operation.
Keywords:
optical computing
matrix multiplication
tensor multiplication
free-space optics
parallel processing
Journal
IF:
32.9
Papers:
4.3K
Citations:
6.1W

