19/11/2021

DDimensioning AI based algorithms : Bounds, information criterion and applications to times series detection problems.

This job offer has expired


  • ORGANISATION/COMPANY
    Université Grenoble Alpes
  • RESEARCH FIELD
    Computer scienceProgramming
  • RESEARCHER PROFILE
    Recognised Researcher (R2)
  • APPLICATION DEADLINE
    19/12/2021 22:00 - Europe/Athens
  • LOCATION
    France › Saint Martin d'Hères
  • TYPE OF CONTRACT
    Temporary
  • JOB STATUS
    Full-time
  • HOURS PER WEEK
    35
  • OFFER STARTING DATE
    01/01/2022

OFFER DESCRIPTION

The Institute of Earth Sciences (ISTerre) is a joint research unit (UGA / CNRS / USMB / IRD / Univ Gustave Eiffel) of nearly 300 people. This unit is part of the Observatoire des Sciences de l'Univers de Grenoble (OSUG) and the PAGE research cluster of the Université Grenoble Alpes (UGA).

 

The work will be led at GIPSA by O. Michel, Professor at UGA/Grenoble-INP. Within the GAIA cluster, the work will be carried out in close collaboration with O. Michel, P. Comon (DR CNRS) and S. Elbouch (PhD student). The other researchers of the pole are involved to a lesser extent but will interact with the post-doctoral fellow on theoretical developments related to the project. A senior researcher (R. Cabral-Farias, I3S/Univ Coté d'Azur) is to join the laboratory for 1 year, with the aim of becoming involved in the theme. The applications and validation of the proposed methods in the context of real data will be pursued at ISTerre under the direction of M. Campillo, and in collaboration with the staff involved in the MIAI chair (S. Mouaoued -doctoral student- and H. Zhu -engineer) and more broadly with the researchers in this group (L Seydoux,...).

The proposed research track is about model assessment and sparse modeling or model complexity in IA representation of time series. This issues, usually adressed via cross validation procedures are requiring large data sets. This project proposes to investigate the effectiveness of analytical approaches from a signal processing perspective , i.e to evaluate their performances on real data sets recorded in the context of geophysical experiments. To assess the relevance of these approaches, some classical multivariate algorithms (ranging from VARMA models to Settering Networks will serve as benchmark for dimensioning e.g LSTM networks designed for anomaly detection. The project is twofold : evaluate and extend analytical information theoretic criteria for dimensioning network, and assess their relevance for real seismic large data arrays. 

Project within the framework of MIAI (Chair " Geophysical applications of AI for natural hazard and georesources " M. Campillo , O. Michel)

Activities :

- Construct and synthetize state of the art bibliographic research

- Implement codes to assessing methods of AI algorithm dimension (hyperparameter tuning) settings

- Apply and evaluate performances on geophysical data, collaborating with Isterre researchers

 

More Information

Offer Requirements

Skills/Qualifications

PhD

Autonomy in research. Synthesis of scientific publications. Algorithmic development. Python or Matlab programming. Writing of publications

know how to be : Scientific communication, teamwork.

Work location(s)
1 position(s) available at
GIPSA LAB
France
Saint Martin d'Hères
38400
11 rue des Mathématiques

EURAXESS offer ID: 710321

Disclaimer:

The responsibility for the jobs published on this website, including the job description, lies entirely with the publishing institutions. The application is handled uniquely by the employer, who is also fully responsible for the recruitment and selection processes.

 

Please contact support@euraxess.org if you wish to download all jobs in XML.