Seminar from Anthony Larcher, Professor at LIUM

Date: 02/07/2024
Time: 10h00
Localization: IC2, Boardroom
Speaker: Anthony Larcher
 
 

Trustworthy, explainable, interpretable… what can we expect for speech processing?

Improvements in the performance of automated systems have led to the widespread commercial deployment of applications that are still poorly understood. To ensure the long-term viability of these applications and the security of those that are not, the field of operability or interpretability, or the development of trustworthy AI, has seen a major boom in recent years.

While the definition of these concepts has not yet been established and the vast majority of
approaches have been developed for the processing of textual natural language and images, this presentation will review existing methods and discuss their potential application to speech.