Automatic Speech Recognition (ASR) systems generalize poorly on accented speech. The phonetic and linguistic variability of accents present hard challenges for ASR systems today in both data collection and modeling strategies. The resulting bias in ASR performance across accents comes at a cost to both users and providers of ASR. We present a survey of current promising approaches to accented speech recognition and highlight the key challenges in the space. Approaches mostly focus on single model generalization and accent feature engineering. Among the challenges, lack of a standard benchmark makes research and comparison especially difficult.
@article{arxiv.2104.10747,
title = {Accented Speech Recognition: A Survey},
author = {Arthur Hinsvark and Natalie Delworth and Miguel Del Rio and Quinten McNamara and Joshua Dong and Ryan Westerman and Michelle Huang and Joseph Palakapilly and Jennifer Drexler and Ilya Pirkin and Nishchal Bhandari and Miguel Jette},
journal= {arXiv preprint arXiv:2104.10747},
year = {2021}
}