Target Accent

Target accent research focuses on improving speech technologies' ability to accurately process and generate speech with diverse accents, aiming to overcome limitations in data availability and model generalizability. Current research emphasizes developing robust and adaptable models, often employing generative frameworks and leveraging techniques like data augmentation, transfer learning, and the exploitation of linguistic relationships between accents to improve automatic speech recognition (ASR) and voice conversion (VC) performance, even with limited data. These advancements have significant implications for improving accessibility in healthcare, enhancing human-computer interaction, and fostering inclusivity in speech technology applications.

Papers