I am a second year PhD student at IRIT (Toulouse, France) and I work on interactive approaches for speaker diarization, my goal is to find how to make the most out of human time to improve speaker diarization performance. I work under the supervision of Hervé Bredin.
Recent work
- On the calibration of powerset speaker diarization models (Alexis Plaquet, Hervé Bredin) - Proc. INTERSPEECH 2024 (📃 🌐 🐍)
- Powerset multi-class cross entropy loss for neural speaker diarization (Alexis Plaquet, Hervé Bredin) - Proc. INTERSPEECH 2023 (📃🌐)
- Participation in JSALT 2023 : xdiar team - confidence measures, calibration, semi-supervised learning
- Semi and self-supervised learning for speaker diarization - 5 months internship from feb-jun 2022, supervised by Hervé Bredin (📃)
Education
- Ecole Normale Supérieure de Lyon - Master 2 Informatique Fondamentale - 2021-2022
- Université Lyon 1 Claude Bernard - Master 1 Informatique, rank 2/125 - 2020-2021
- Université Picardie Jules Verne - Licence d’informatique, mention Très bien - 2017-2020
Other
- 💾 datasets-pyannote : a Python package to download and prepare datasets for
pyannote.audio