CHALLENGES IN AUTOMATIC SPEECH RECOGNITION IN VOICE ASSISTANTS AND HOW TO MITIGATE THEM

Abstract

 Automatic Speech Recognition (ASR) has increasingly been finding more and more applications, 
especially because of the growth seen in smart speaker awareness, adoption and engagement, and ASR being a 
critical early step in the end-to-end voice assistant usage. However, despite advancements, there continue to be 
gaps in ASR. This article explores those gaps and provides strategies to overcome those challenges. The gaps 
and challenges have been articulated specifically from the point of view of the use case of music and music- 
related queries on voice assistants.