Contrat postdoctoral: Evolutive end-to-end neural networks for speaker recognition

 

Laboratoire d’accueil : LIUM, équipe LST, https://lium.univ-lemans.fr/lium/lst/
Site : Le Mans Université
Encadrement : Anthony Larcher (LIUM)
Durée : Contrat post-doctoral d’un an, prise de fonction : dès que possible, au plus tard fin juin 2020

Mots clef : Deep Learning, evolutive networks, genetic algorithms, explainability, speaker recognition, end-to-end

Contexte

The LST team from LIUM (Le Mans University) is focusing on evolutive end-to-end neural networks for speaker recognition. The Extensor project (French ANR funded) aims at developing novel architectures for end-to-end speaker recognition as well as explaining the behavior of those networks. The focus of Extensor is threefold:

1. get rid of the legacy of bayesian system’s architecture and explore wider opportunities offered in deep learning;
2. explore real end-to-end architectures exploiting the tax signal instead of classical features (such as MFCC of filterbanks);
3. Develop tools for explainability in speaker recognition.

 

Missions

1. Develop end-to-end speaker recognition system based on state-of-the-art approaches (xvectors, sincnet…)

2. Develop evolutive architectures making use of existing genetic algorithms and study their behavior.

3. Participate to the three hackathons organized by the Extensor project in order to develop tools for evolutive neural network architecture and explainability for speaker recognition.

4. Dissemination: the research will be published in the major conferences and journals

 

Profil et compétences recherchées

• Doctorat portant sur l’apprentissage automatique (Machine learning et Deep Learning)
• Des travaux dans le domaine du traitement de la parole sont un plus
• Maîtrise de python et d’un framework de deep learning (Pytorch, TensorFlow)

 

Salaire : 2 600€ net (après taxes)

Postuler : Contacter anthony.larcher@univ-lemans.fr, marie.tahon@univ-lemans.fr
Joindre un CV et une lettre de motivation.