Date: 24 March
Time: 1 p.m.
Venue: PA2 Amphitheatre, Mathematics Building, Alameda Campus
Speaker: Ben Peters (INESC-ID)
Title: “What’s in a speech token?”
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.
Speaker bio:
Ben Peters is a postdoctoral researcher at INESC-ID. He works at the intersection of NLP and speech.
Priberam is a member of the IST Spin-Off Community®.
The Priberam Machine Learning Lunch Seminars are free of charge. Prior registration is required.
