Invention Grant
- Patent Title: Learning a truncation rank of singular value decomposed matrices representing weight tensors in neural networks
-
Application No.: US15962996Application Date: 2018-04-25
-
Publication No.: US11238346B2Publication Date: 2022-02-01
- Inventor: Regan Blythe Towal , Raghuraman Krishnamoorthi
- Applicant: QUALCOMM Incorporated
- Applicant Address: US CA San Diego
- Assignee: QUALCOMM Incorporated
- Current Assignee: QUALCOMM Incorporated
- Current Assignee Address: US CA San Diego
- Agency: Arent Fox LLP
- Main IPC: H04L12/24
- IPC: H04L12/24 ; G06Q10/04 ; G06N3/08 ; G06N3/04 ; G06N3/063
Abstract:
An apparatus for learning a rank of an artificial neural network is configured to decompose a weight tensor into a first weight tensor and a second weight tensor. A set of rank selection parameters are applied to the first weight tensor and the second weight tensor to truncate the rank of the first weight tensor and the second weight tensor. The set of rank selection parameters are updated simultaneously with the weight tensors by averaging updates calculated for each rank selection parameter of the set of rank selection parameters.
Public/Granted literature
Information query
