@inproceedings{b64f5c768e0c4a99b061a63d15ebdac2,
title = "Multilingual TTS Accent Impressions for Accented ASR",
abstract = "Automatic Speech Recognition (ASR) for high-resource languages like English is often considered a solved problem. However, most high-resource ASR systems favor socioeconomically advantaged dialects. In the case of English, this leaves behind many L2 speakers and speakers of low-resource accents (a majority of English speakers). One way to mitigate this is to fine-tune a pre-trained English ASR model for a desired low-resource accent. However, collecting transcribed accented audio is costly and time-consuming. In this work, we present a method to produce synthetic L2-English speech via pre-trained text-to-speech (TTS) in an L1 language (target accent). This can be produced at a much larger scale and lower cost than authentic speech collection. We present initial experiments applying this augmentation method. Our results suggest that success of TTS augmentation relies on access to more than one hour of authentic training data and a diversity of target-domain prompts for speech synthesis.",
keywords = "accented speech recognition, data augmentation, low-resource speech technologies, speech synthesis",
author = "Georgios Karakasidis and Nathaniel Robinson and Yaroslav Getman and Atieno Ogayo and Ragheb Al-Ghezi and Ananya Ayasi and Shinji Watanabe and Mortensen, {David R.} and Mikko Kurimo",
note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; International Conference on Text, Speech, and Dialogue, TSD ; Conference date: 04-09-2023 Through 06-09-2023",
year = "2023",
doi = "10.1007/978-3-031-40498-6_28",
language = "English",
isbn = "978-3-031-40497-9",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "317--327",
editor = "Kamil Ek{\v s}tein and Franti{\v s}ek P{\'a}rtl and Miloslav Konop{\'i}k",
booktitle = "Text, Speech, and Dialogue - 26th International Conference, TSD 2023, Proceedings",
address = "Germany",
}