Explainability and evolutive systems for Speaker Verification and Anti-Spoofing an event organized by the EXTENSOR project (ANR-19-CE23-0001-01)

In the context of the EXTENSOR project, LIUM and EURECOM organise two sessions of talk on two topics addressed in this project: explainability and interpretability for speech processing and evolutive systems and end-to-end speech processing.

The talks will take place on Thursday 2 September and and Friday 3 September .

To join online: https://univ-lemans-fr.zoom.us/j/92143709904?pwd=Qy9QTXUydExNSlFnS0pWY0ZaNXpzUT09

THURSDAY 2 SEPTEMBER

• 3:30pm (CEST): Evolutive network architectures for speech deepfake detection (Wanying Ge – EURECOM)
• 4:15pm (CEST): Measuring biases for speaker recognition (Ambuj Mehrish – LIUM)

FRIDAY 3 SEPTEMBER

• 2:00pm (CEST) Jean-François Bonastre from LIA will present his work on explainability and interpretability in speech processing.