Multilingual Automatic Speech Recognition for Scandinavian Languages

Rafal Cerniavski, Sara Stymne

Keywords: Scandinavian ASR, Multilingual ASR, Language Classification, Language Models

TL;DR: We train a multilingual ASR model to transcribe Swedish, Danish and Norwegian

Abstract: We investigate the effectiveness of multilingual automatic speech recognition models for Scandinavian languages by further fine-tuning a Swedish model on Swedish, Danish, and Norwegian. We first explore zero-shot models, which perform poorly across the three languages. However, we show that a multilingual model based on a strong Swedish model, further fine-tuned on all three languages, performs well for Norwegian and Danish, with a relatively low decrease in the performance for Swedish. With a language classification module, we improve the performance of the multilingual model even further.

Student Paper: Yes, the first author is a student