Massively Multilingual ASR on 70 Languages: Tokenization, Architecture, and Generalization Capabilities

Bibliographic Details
Title: Massively Multilingual ASR on 70 Languages: Tokenization, Architecture, and Generalization Capabilities
Authors: Tjandra, Andros, Singhal, Nayan, Zhang, David, Kalinli, Ozlem, Mohamed, Abdelrahman, Le, Duc, Seltzer, Michael L.
Publication Year: 2022
Collection: Computer Science
Subject Terms: Computer Science - Computation and Language, Computer Science - Sound, Electrical Engineering and Systems Science - Audio and Speech Processing
More Details: End-to-end multilingual ASR has become more appealing because of several reasons such as simplifying the training and deployment process and positive performance transfer from high-resource to low-resource languages. However, scaling up the number of languages, total hours, and number of unique tokens is not a trivial task. This paper explores large-scale multilingual ASR models on 70 languages. We inspect two architectures: (1) Shared embedding and output and (2) Multiple embedding and output model. In the shared model experiments, we show the importance of tokenization strategy across different languages. Later, we use our optimal tokenization strategy to train multiple embedding and output model to further improve our result. Our multilingual ASR achieves 13.9%-15.6% average WER relative improvement compared to monolingual models. We show that our multilingual ASR generalizes well on an unseen dataset and domain, achieving 9.5% and 7.5% WER on Multilingual Librispeech (MLS) with zero-shot and finetuning, respectively.
Comment: Submitted to ICASSP 2023
Document Type: Working Paper
Access URL: http://arxiv.org/abs/2211.05756
Accession Number: edsarx.2211.05756
Database: arXiv
More Details
Description not available.