eduzhai > Applied Sciences > Engineering >

End-to-End Code Switching Language Models for Automatic Speech Recognition

  • Save

... pages left unread,continue reading

Document pages: 5 pages

Abstract: In this paper, we particularly work on the code-switched text, one of themost common occurrences in the bilingual communities across the world. Due tothe discrepancies in the extraction of code-switched text from an AutomatedSpeech Recognition(ASR) module, and thereby extracting the monolingual textfrom the code-switched text, we propose an approach for extracting monolingualtext using Deep Bi-directional Language Models(LM) such as BERT and otherMachine Translation models, and also explore different ways of extractingcode-switched text from the ASR model. We also explain the robustness of themodel by comparing the results of Perplexity and other different metrics likeWER, to the standard bi-lingual text output without any external information.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×