Addressing racial bias in automatic speech recognition is an area of concern fields associated with human-computer interaction. Research to date suggests that sociolinguistic variation, namely systematic sources sociophonetic has yet be extensively exploited acoustic model architectures. This paper reports a study evaluates the performance one ASR system for multi-ethnic sample speakers from Am...

Recently, the speech community is seeing a significant trend of moving from deep neural network based hybrid modeling to end-to-end (E2E) for automatic recognition (ASR). While E2E models achieve state-of-the-art results in most benchmarks terms ASR accuracy, are still used large proportion commercial systems at current time. There lots practical factors that affect production model deployment ...