Title:
HYPER-PARAMETER OPTIMIZATION SYSTEM, METHOD, AND PROGRAM
Document Type and Number:
WIPO Patent Application WO/2022/079848
Kind Code:
A1
Abstract:
A speech enhancement means 81 determines an enhancement mask generated based on a mask for speech enhancement, when a test utterance is input as speech data. A first hyper-parameter optimization means 82 determines, when the test utterance is input, a first hyper-parameter which is a hyper-parameter representing the degree to which the signal representing the test utterance is kept using the mask, and the first hyper-parameter which is set to take into account a downstream task that is processed using an enhanced test utterance. A mask generation means 83 generates an adaptive mask from the determined enhancement mask and the first hyper-parameter that enhances the test utterance for the downstream task. The mask generation means 83 generates the adaptive mask in which the first hyper-parameter is a power of the mask.
Inventors:
WANG QIONGQIONG (JP)
KOSHINAKA TAKAFUMI (JP)
KOSHINAKA TAKAFUMI (JP)
Application Number:
PCT/JP2020/038860
Publication Date:
April 21, 2022
Filing Date:
October 15, 2020
Export Citation:
Assignee:
NEC CORP (JP)
International Classes:
G10L25/30; G10L21/0272
Domestic Patent References:
| WO2020045313A1 | 2020-03-05 |
Other References:
RAVANELLI MIRCO; BRAKEL PHILEMON; OMOLOGO MAURIZIO; BENGIO YOSHUA: "A network of deep neural networks for Distant Speech Recognition", 2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), IEEE, 5 March 2017 (2017-03-05), pages 4880 - 4884, XP033259338, DOI: 10.1109/ICASSP.2017.7953084
Attorney, Agent or Firm:
IWAKABE Fuyuki et al. (JP)
Download PDF: