speech language model for multiple domains
Dnr:
NAISS 2024/22-915
Type:
NAISS Small Compute
Principal Investigator:
Siyang Wang
Affiliation:
Kungliga Tekniska högskolan
Start Date:
2024-06-26
End Date:
2025-07-01
Primary Classification:
10208: Language Technology (Computational Linguistics)
Webpage:
Allocation
Abstract
This project explores new methods for speech language models, a new family of speech foundation models that enable various speech technology applications. We are especially interested in the Text-to-speech (TTS) application. We intend to construct a speech language model for English and Swedish.