HLSTM MODEL-BASED ACOUSTIC MODELING METHOD AND DEVICE, AND STORAGE MEDIUM

Title:

HLSTM MODEL-BASED ACOUSTIC MODELING METHOD AND DEVICE, AND STORAGE MEDIUM

Document Type and Number:

WIPO Patent Application WO/2018/153200

Kind Code:

A1

Abstract:

A highway long short time memory (HLSTM) model-based acoustic modeling method and device, and a storage medium, the method comprising: training a randomly initialized HLSTM model on the basis of a preset function and optimizing a training result (101); carrying out forward calculation on training data by means of the optimized HLSTM model (102); and training a randomly initialized long short time memory (LSTM) model on the basis of the result of the forward calculation and the preset function so as to obtain a model that is an acoustic model of a speech recognition system (103), the HLSTM model and the LSTM model having the same network parameters.

More Like This:

WO/2015/184525	SYSTEMS AND METHODS OF INTERPRETING SPEECH DATA
JP5377157	A communication karaoke system that automatically creates a singing history for each customer based on the characteristics of the singing voice.
JP2002135673	INFORMATION PROCESSOR

Inventors:

ZHANG PENGYUAN (CN)
DONG ZHENJIANG (CN)
ZHANG YU (CN)
JIA XIA (CN)
LI JIE (CN)
ZHANG HENGSHENG (CN)

Application Number:

PCT/CN2018/073887

Publication Date:

August 30, 2018

Filing Date:

January 23, 2018

Export Citation:

Click for automatic bibliography generation Help

Assignee:

ZTE CORP (CN)

International Classes:

G10L15/08

Foreign References:

CN104538028A	2015-04-22
CN105810193A	2016-07-27
CN106098059A	2016-11-09
CN106328122A	2017-01-11
CN106170800A	2016-11-30
CN105529023A	2016-04-27

Attorney, Agent or Firm:

CHINA PAT INTELLECTUAL PROPERTY OFFICE (CN)

Download PDF:

View/Download PDF PDF Help

Previous Patent: METHOD AND DEVICE FOR ADJUSTING SENSITIVITY OF TOUCH SCREEN

Next Patent: DEEP LEARNING TRAINING METHOD AND APPARATUS