{"id":24175,"date":"2019-11-28T17:13:19","date_gmt":"2019-11-28T16:13:19","guid":{"rendered":"https:\/\/lium.univ-lemans.fr\/?p=24175"},"modified":"2020-01-08T11:01:00","modified_gmt":"2020-01-08T10:01:00","slug":"postdoc-position-evolutive-end-to-end-neural-networks-for-speaker-recognition","status":"publish","type":"post","link":"https:\/\/lium.univ-lemans.fr\/en\/postdoc-position-evolutive-end-to-end-neural-networks-for-speaker-recognition\/","title":{"rendered":"PostDoc position: evolutive end-to-end neural networks for speaker recognition"},"content":{"rendered":"

Contrat postdoctoral: Evolutive end-to-end neural networks for speaker recognition<\/h2>\n

 <\/p>\n

Host Laboratory <\/strong> : LIUM, \u00e9quipe LST, http:\/\/lium.univ-lemans.fr\/lium\/lst\/<\/a>
\nSite<\/strong> : Le Mans University, France
\nSupervisor<\/strong> : Anthony Larcher (LIUM)
\nDuration<\/strong> : 24 months, Starting date <\/strong> : asap, end of Jaune 2020 at the latest<\/br><\/p>\n

Mots clef<\/strong> : Deep Learning, evolutive networks, genetic algorithms, explainability, speaker recognition, end-to-end<\/p>\n

Contexte<\/H3><\/p>\n

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:<\/p>\n

    \n
  1. get rid of the legacy of bayesian system\u2019s architecture and explore wider opportunities offered in deep learning;<\/li>\n
  2. explore real end-to-end architectures exploiting the tax signal instead of classical features (such as MFCC of filterbanks);<\/li>\n
  3. Develop tools for explainability in speaker recognition.<\/li>\n<\/ol>\n

    Missions<\/H3><\/p>\n

    1. Develop end-to-end speaker recognition system based on state-of-the-art approaches (xvectors, sincnet\u2026)<\/p>\n

    2. Develop evolutive architectures making use of existing genetic algorithms and study their behavior.<\/p>\n

    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.<\/p>\n

    4. Dissemination: the research will be published in the major conferences and journals<\/p>\n

     
    \n 
    \n

    Expected competences:<\/H3><\/p>\n