MULTIPLE REPRESENTATION TRANSFER FROM LARGE LANGUAGE MODELS TO END-TO-END ASR SYSTEMS
- Takuma Udagawa
- Masayuki Suzuki
- et al.
- 2024
- ICASSP 2024
Bio
Masayasu Muraoka is a researcher of Natural Language Processing at IBM Research - Tokyo. He mainly focuses on multi-modal research in vision and language and large language models. Please visit the Google Scholar page for his publications.
Publications
Incorporating Syntactic Knowledge into Pre-trained Language Model using Optimization for Overcoming Catastrophic Forgetting
- Ran Iwamoto
- Issei Yoshida
- et al.
- 2023
- EMNLP 2023
A Simple Yet Strong Domain-Agnostic De-bias Method for Zero-Shot Sentiment Classification
- Yang Zhao
- Tetsuya Nasukawa
- et al.
- 2023
- ACL 2023
Visual Objects As Context: Exploiting Visual Objects for Lexical Entailment
- Masayasu Muraoka
- Tetsuya Nasukawa
- et al.
- 2020
- EMNLP 2020
Patents
- 23 Oct 2023
- US
- 11797425
- 24 Apr 2023
- US
- 11636338
- 27 Sep 2021
- US
- 11132393
- 05 Apr 2021
- US
- 10970488
Projects
Top collaborators
HK
Hiroshi Kanayama
Senior Technical Staff Member, Knowledge Infrastructure, AI Technologies, IBM Research - Tokyo
RI
GS
George Saon
Speech strategy lead, distinguished research scientist
GK
Gakuto Kurata
Distinguished Engineer and Chief Scientist for Spoken Conversational Systems