Title:
METHOD AND APPARATUS FOR TRAINING AND APPLYING A NEURAL NETWORK
Document Type and Number:
WIPO Patent Application WO/2020/220191
Kind Code:
A1
Abstract:
A neural network training method comprises: determining parameters for a weight matrix of each layer of a network, based on the number of shared elements in a structured implementation with shared elements of each layer; determining the weight matrix of each layer based on the determined parameter; decompressing the determined weight matrix to derived an integrated matrix; iterating over forward propagation and backward propagation on all layers of the network based on the integrated matrix until a preset condition is satisfied, wherein the integrated matrix is updated in each iteration; and storing the iterated parameters of the network.
Inventors:
ZENG TIEYONG (CN)
YANG HONGFEI (CN)
DING XIAOFENG (CN)
HU HUI (CN)
YANG HONGFEI (CN)
DING XIAOFENG (CN)
HU HUI (CN)
Application Number:
PCT/CN2019/084969
Publication Date:
November 05, 2020
Filing Date:
April 29, 2019
Export Citation:
Assignee:
HUAWEI TECH CO LTD (CN)
International Classes:
G06N3/08
Foreign References:
| CN104899641A | 2015-09-09 | |||
| US20170103308A1 | 2017-04-13 |
Download PDF: