What's in a speech token?
Priberam Machine Learning Lunch Seminar
Abstract:
Discrete speech tokenization is a surprisingly effective technique to adapt text language models to speech. But despite the simplicity of the method, the nature of the units it produces is not widely understood. In this talk, I ask some questions (and hopefully answer a few) about how, when, and why discrete speech tokens work.
Bio:
Ben Peters is a postdoctoral researcher at INESC-ID. He works at the intersection of NLP and speech.
Priberam Machine Learning Lunch Seminar
Abstract:
Discrete speech tokenization is a surprisingly effective technique to adapt text language models to speech. But despite the simplicity of the method, the nature of the units it produces is not widely understood. In this talk, I ask some questions (and hopefully answer a few) about how, when, and why discrete speech tokens work.
Bio:
Ben Peters is a postdoctoral researcher at INESC-ID. He works at the intersection of NLP and speech.
Good to know
Highlights
- 1 hour
- In-person
Location
Instituto Superior Técnico, Anfiteatro PA2
1 Avenida Rovisco Pais
1049-001 Lisboa
How would you like to get there?
