Patent · US Active

Photonic tensor accelerators for artificial neural networks

US12387094B2 · kind B2 · utility

0Cited by
6References
20Claims
0Family size

Assignee

Inventors

Key dates

Filing dateOct 15, 2019
Grant dateAug 12, 2025
Priority date
Expiry dateApr 30, 2042

Classification

  • Technology area (CPC H)Electricity
  • CPC primaryH04B10/58
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

Photonic units for vector-vector multiplication, matrix-vector multiplication, matrix-matrix multiplication, batch matrix-matrix multiplication, and tensor-tensor multiplication are described. Multiplications are through coherent mixing and square-law detection. There are many dimensions—wavelength, vector mode, quadrature, and three dimensions of space—that can be used to construct photonic accelerators. The encoded input vector or input matrix is fanned out into a desired number of copies and mixed with the corresponding encoded local oscillators containing the weight vectors comprising the weight matrix. Any subset of two (three) dimensions can be used to construct photonic accelerators for matrix-vector (matrix-matrix) multiplications. Multiple dimensions can be combined into a hyperdimension to increase the scalability. Each dimension, each non-overlapping subset of a dimension, or each non-overlapping subset of a hyperdimension, can be used independently to construct a photonic tensor accelerator (PTA) for batch matrix multiplication, or tensor multiplication operations.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.