Paper ID | HLT-1.3 | ||
Paper Title | SPEECH RECOGNITION BY SIMPLY FINE-TUNING BERT | ||
Authors | Wen-Chin Huang, Nagoya University, Japan; Chia-Hua Wu, Shang-Bao Luo, Academia Sinica, Taiwan; Kuan-Yu Chen, National Taiwan University of Science and Technology, Taiwan; Hsin-Min Wang, Academia Sinica, Taiwan; Tomoki Toda, Nagoya University, Japan | ||
Session | HLT-1: Language Modeling 1: Fusion and Training for End-to-End ASR | ||
Location | Gather.Town | ||
Session Time: | Tuesday, 08 June, 13:00 - 13:45 | ||
Presentation Time: | Tuesday, 08 June, 13:00 - 13:45 | ||
Presentation | Poster | ||
Topic | Speech Processing: [SPE-GASR] General Topics in Speech Recognition | ||
IEEE Xplore Open Preview | Click here to view in IEEE Xplore | ||
Abstract | We propose a simple method for automatic speech recognition (ASR) by fine-tuning BERT, which is a language model (LM) trained on large-scale unlabeled text data and can generate rich contextual representations. Our assumption is that given a history context sequence, a powerful LM can narrow the range of possible choices and the speech signal can be used as a simple clue. Hence, comparing to conventional ASR systems that train a powerful acoustic model (AM) from scratch, we believe that speech recognition is possible by simply fine-tuning a BERT model. As an initial study, we demonstrate the effectiveness of the proposed idea on the AISHELL dataset and show that stacking a very simple AM on top of BERT can yield reasonable performance. |