US 12,153,975 B2
Hybrid analog-digital matrix processors
Tyler J. Kenney, Boston, MA (US); Martin B. Z. Forsythe, Jamaica Plain, MA (US); Tomo Lazovich, Cambridge, MA (US); and Darius Bunandar, Boston, MA (US)
Assigned to Lightmatter, Inc., Boston, MA (US)
Filed by Lightmatter, Inc., Boston, MA (US)
Filed on Dec. 13, 2023, as Appl. No. 18/538,172.
Application 18/538,172 is a continuation of application No. 17/545,137, filed on Dec. 8, 2021, granted, now 11,886,942.
Application 17/545,137 is a continuation of application No. 17/246,892, filed on May 3, 2021, granted, now 11,775,779, issued on Oct. 3, 2023.
Application 17/246,892 is a continuation of application No. 16/995,674, filed on Aug. 17, 2020, granted, now 11,023,691, issued on Jun. 1, 2021.
Application 16/995,674 is a continuation of application No. 16/800,998, filed on Feb. 25, 2020, granted, now 10,803,258, issued on Oct. 13, 2020.
Claims priority of provisional application 62/810,851, filed on Feb. 26, 2019.
Prior Publication US 2024/0193379 A1, Jun. 13, 2024
Int. Cl. G06J 1/02 (2006.01); G06F 17/16 (2006.01)
CPC G06J 1/02 (2013.01) [G06F 17/16 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A hybrid analog-digital processor comprising:
circuitry comprising an analog processor, wherein the circuitry is configured to perform a multi-layer neural network, the performing comprising:
obtaining a first scaling factor for a first matrix associated with a first layer of the multi-layer neural network;
scaling at least a portion of the first matrix using the first scaling factor to obtain a first scaled matrix;
performing the first layer of the multi-layer neural network by controlling the analog processor to perform matrix multiplication using the first scaled matrix;
obtaining a second scaling factor for a second matrix associated with a second layer of the multi-layer neural network, wherein the first scaling factor is different from the second scaling factor;
scaling at least a portion of the second matrix using the second scaling factor to obtain a second scaled matrix; and
performing the second layer of the multi-layer neural network by controlling the analog processor to perform matrix multiplication using the second scaled matrix.