Paul-Gauthier Noé
Présentation
I am a researcher in Computer Science (chargé de recherche CNRS) in Marseille.
In 2025, I joined the team QARMA in the Laboratoire d'Informatique et des Systèmes. My main research interests are Calibration of probabilities, Information theory, Machine learning and Privacy in Speech technologies. In 2024, I worked as a postdoctoral researcher at Statify, a joint team of Inria and Laboratoire Jean Kuntzmann in Grenoble. In 2023, I received a PhD in Computer Science in Avignon Université under the supervision of Jean-François Bonastre and Driss Matrouf. I was working for the international JST-ANR VoicePersonae project. My thesis is named Representing evidence for attribute privacy: bayesian updating, compositional evidence, and calibration and can be found here.Education:
Research visits:
Awards:
Teaching Experience:
Publications
Publications
|
|
Representing evidence for attribute privacy : bayesian updating, compositional evidence and calibrationOther [cs.OH]. Université d'Avignon, 2023. English. ⟨NNT : 2023AVIG0113⟩ |
|
|
Explaining a probabilistic prediction on the simplex with Shapley compositionsECAI 2024, Oct 2024, Santiago de Compostela, Spain. ⟨10.3233/FAIA240605⟩ |
|
|
Hiding Speaker’s Sex in Speech Using Zero-Evidence Speaker Representation in an Analysis/Synthesis PipelineICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Jun 2023, Rhodes Island, Greece. pp.1-5, ⟨10.1109/ICASSP49357.2023.10096749⟩ |
|
|
Faire le pont entre l'observation et la preuve : Application au respect de la vie privéeJournées d’Etudes sur la Parole - JEP2022, Jun 2022, Île de Noirmoutier, France |
|
|
A bridge between features and evidence for binary attribute-driven perfect privacyICASSP 2022, May 2022, Singapore, Singapore |
|
|
Benchmarking and challenges in security and privacy for voice biometricsSPSC 2021, 1st ISCA Symposium on Security and Privacy in Speech Communication, ISCA, Nov 2021, Magdeburg, Germany. ⟨10.21437/SPSC.2021-11⟩ |
|
|
Adversarial Disentanglement of Speaker Representation for Attribute-Driven Privacy PreservationInterspeech 2021, Aug 2021, Brno, Czech Republic |
|
|
The Privacy ZEBRA: Zero Evidence Biometric Recognition AssessmentInterspeech 2020, Oct 2021, Shanghai, China. pp.1698-1702, ⟨10.21437/Interspeech.2020-1815⟩ |
|
|
Speech Pseudonymisation Assessment Using Voice Similarity MatricesInterspeech 2020, Oct 2020, Shanghai, China |
|
|
Introducing the VoicePrivacy initiativeINTERSPEECH 2020, Oct 2020, Shanghai, China |
|
|
CGCNN: COMPLEX GABOR CONVOLUTIONAL NEURAL NETWORK ON RAW SPEECHICASSP 2020, May 2020, Barcelona, Spain |
|
|
Denoising x-vectors for Robust Speaker RecognitionOdyssey 2020 The Speaker and Language Recognition Workshop, Nov 2020, Tokyo, Japan. pp.75-80, ⟨10.21437/Odyssey.2020-11⟩ |
|
|
Towards a unified assessment framework of speech pseudonymisationComputer Speech and Language, 2022, 72, pp.101299. ⟨10.1016/j.csl.2021.101299⟩ |
|
|
The VoicePrivacy 2020 Challenge: Results and findingsComputer Speech and Language, 2022, 74, pp.101362. ⟨10.1016/j.csl.2022.101362⟩ |
|
|
Compositional discriminant analysis through calibrated evidence functionsThe 10th International Workshop on Compositional Data Analysis (CoDaWork2024), Jun 2024, Girona, Catalonia, Spain |
|
|
The distribution of calibrated likelihood functions on the probability-likelihood Aitchison simplex2025 |
|
|
Supplementary material to the paper The VoicePrivacy 2020 Challenge: Results and findings2022 |
|
|
The VoicePrivacy 2020 Challenge Evaluation Plan[0] LIA - Laboratoire Informatique d'Avignon; MULTISPEECH - Speech Modeling for Facilitating Oral-Based Communication Inria Nancy - Grand Est, LORIA - NLPKD - Department of Natural Language Processing & Knowledge Discovery; Eurecom [Sophia Antipolis]; University of Edinburgh. 2020 |