Pierre Champion

  • PostDoc, INRIA
  • Open-source , Maintainer
Currenlty working as a postdoc at INRIA’s Mission Sécurité-Défense, focusing on; automatic retrieval of incorrectly ASR-transcribed proper nouns/last names (KWS) , detection of audio deep-fakes , evaluation of voice privacy systems and contributing to SpeechBrain.
Completed a PhD, in 2023, in collaboration with both the MULTISPEECH and LIUM teams, under the co-supervision of Dr. Denis Jouvet and Dr. Anthony Larcher. My thesis, titled “Anonymizing Speech: Evaluating and Designing Speaker Anonymization Techniques," focuses on privacy-preserving deep learning for speech processing, supported by the Deep-Privacy ANR project.
Formerly, I received an MSc (Research) in Natural Language Processing from Le Mans University, France in 2019, and worked as an Data Scientist & Android Developer in an SME for tow years.
Interests. Deep Neural Networks, Speech & Language Processing, Privacy Preservation, ASR/VC/ASV.

Projects

Relevant Publications

For more papers, see google scholar.
2020
2021
2022
2023
2024