Accented Speech Recognition: A Survey
This addresses the bias in ASR performance across accents, which affects users and providers, but it is incremental as it reviews existing methods without introducing new solutions.
The paper surveys approaches to improve automatic speech recognition (ASR) for accented speech, highlighting that current systems generalize poorly due to phonetic and linguistic variability, and it identifies key challenges such as the lack of a standard benchmark.
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.