Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
TRANSLITERATION BASED DATA AUGMENTATION FOR TRAINING MULTILINGUAL ASR ACOUSTIC MODELS IN LOW RESOURCE SETTINGS
Document Type and Number:
WIPO Patent Application WO/2022/078506
Kind Code:
A1
Abstract:
A computer-implemented method of building a multilingual acoustic model for automatic speech recognition in a low resource setting includes training a multilingual network on a set of training languages with an original transcribed training data to create a baseline multilingual acoustic model. Transliteration of transcribed training data is performed by processing through the multilingual network a plurality of multilingual data types from the set of languages, and outputting a pool of transliterated data. A filtering metric is applied to the pool of transliterated data output to select one or more portions of the transliterated data for retraining of the acoustic model. Data augmentation is performed by adding one or more selected portions of the output transliterated data back to the original transcribed training data to update training data. The training of a new multilingual acoustic model through the multilingual network is performed using the updated training data.

Inventors:
THOMAS SAMUEL (US)
AUDHKHASI KARTIK (US)
KINGSBURY BRIAN E D (US)
Application Number:
PCT/CN2021/124149
Publication Date:
April 21, 2022
Filing Date:
October 15, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
IBM (US)
IBM CHINA CO LTD (CN)
International Classes:
G10L15/00; G06F40/58
Foreign References:
US20180307679A12018-10-25
CN111507114A2020-08-07
US20200098351A12020-03-26
CN111557029A2020-08-18
CN108920473A2018-11-30
CN111261144A2020-06-09
Attorney, Agent or Firm:
ZHONGZI LAW OFFICE (CN)
Download PDF: