SLP-L18.1

PRODUCTION-SCALE DYNAMIC VOCABULARY ASR BIASING WITH WORD-LEVEL FST AND ROBUST TRAINING

José Enrique Garcia Lainez, Tianyang Sun, Shaoshi Ling, Yifan Gong, Huaming Wang, Microsoft, Spain
Session:
SLP-L18: ASR Model Adaptation and Customization Oral
Track:
Speech and Language Processing [SL]
Location:
Room 114
Presentation Time:
Fri, 8 May, 09:00 - 09:20
Presentation
Discussion
Resources
No resources available.
Session SLP-L18
SLP-L18.1: PRODUCTION-SCALE DYNAMIC VOCABULARY ASR BIASING WITH WORD-LEVEL FST AND ROBUST TRAINING
José Enrique Garcia Lainez, Tianyang Sun, Shaoshi Ling, Yifan Gong, Huaming Wang, Microsoft, Spain
SLP-L18.2: DO WE REALLY NEED SELF-ATTENTION FOR STREAMING AUTOMATIC SPEECH RECOGNITION?
Youness Dkhissi, Valentin Vielzeuf, Elys Allesiardo, Orange Innovation, France; Anthony Larcher, Le Mans Université, France
SLP-L18.3: SYNTHETIC DATA DOMAIN ADAPTATION FOR ASR VIA LLM-BASED TEXT AND PHONETIC RESPELLING AUGMENTATION
Natsuo Yamashita, Koichi Nagatsuka, Hiroaki Kokubo, Kota Dohi, Tuan Vu Ho, Hitachi, Ltd., Japan
SLP-L18.4: SSVD-O: PARAMETER-EFFICIENT FINE-TUNING WITH STRUCTURED SVD FOR SPEECH RECOGNITION
Pu Wang, KU Leuven, Belgium; Shinji Watanabe, Carnegie Mellon University, United States of America; Hugo Van hamme, KU Leuven, Belgium
SLP-L18.5: THREE SECONDS IS SUFFICIENT: A MULTI-PRONGED FRAMEWORK FOR MODEL-BASED SPEAKER ADAPTATION IN ASR UNDER DATA-SCARCE CONDITIONS
Jiajun Deng, Huawei, China; Guinan Li, Chunyat Wu, The Chinese University of Hong Kong, China; Tristan Tsoi, Huawei, China; Huimeng Wang, Tao Zhong, Zhaoqing Li, Chengxi Deng, Youjun Chen, Shujie Hu, Xunying Liu, The Chinese University of Hong Kong, China; Simon Lui, Huawei, China
SLP-L18.6: IN-SYNC: ADAPTATION OF SPEECH AWARE LARGE LANGUAGE MODELS FOR ASR WITH WORD LEVEL TIMESTAMP PREDICTIONS
Xulin Fan, University of Illinois Urbana-Champaign, United States of America; Vishal Sunder, Samuel Thomas, IBM Research, United States of America; Mark Hasegawa-Johnson, University of Illinois Urbana-Champaign, United States of America; Brian Kingsbury, George Saon, IBM Research, United States of America
Contacts