To What Degree Can Language Borders Be Blurred In BERT-based Multilingual Spoken Language Understanding?

  • 2020-11-10 09:59:24
  • Quynh Do, Judith Gaspers, Tobias Roding, Melanie Bradford
  • 5

Abstract

This paper addresses the question as to what degree a BERT-based multilingualSpoken Language Understanding (SLU) model can transfer knowledge acrosslanguages. Through experiments we will show that, although it workssubstantially well even on distant language groups, there is still a gap to theideal multilingual performance. In addition, we propose a novel BERT-basedadversarial model architecture to learn language-shared and language-specificrepresentations for multilingual SLU. Our experimental results prove that theproposed model is capable of narrowing the gap to the ideal multilingualperformance.

 

Quick Read (beta)

loading the full paper ...