LoRA-Whisper: Parameter-Efficient and Extensible Multilingual ASR

TL;DR

LoRA-Whisper introduces a parameter-efficient method for multilingual ASR that reduces language interference and improves performance on new languages by integrating LoRA matrices into the Whisper model.

Contribution

The paper presents LoRA-Whisper, a novel approach that incorporates LoRA matrices into Whisper to enhance multilingual ASR and facilitate language expansion without performance loss.

Findings

Achieves 18.5% relative gain over baseline in multilingual ASR

Achieves 23.0% relative gain in language expansion tasks

Effectively mitigates language interference in multilingual models

Abstract

Recent years have witnessed significant progress in multilingual automatic speech recognition (ASR), driven by the emergence of end-to-end (E2E) models and the scaling of multilingual datasets. Despite that, two main challenges persist in multilingual ASR: language interference and the incorporation of new languages without degrading the performance of the existing ones. This paper proposes LoRA-Whisper, which incorporates LoRA matrix into Whisper for multilingual ASR, effectively mitigating language interference. Furthermore, by leveraging LoRA and the similarities between languages, we can achieve better performance on new languages while upholding consistent performance on original ones. Experiments on a real-world task across eight languages demonstrate that our proposed LoRA-Whisper yields a relative gain of 18.5% and 23.0% over the baseline system for multilingual ASR and language expansion respectively.