3 post-doctoral positions in the fields of

nonstationary signal processing &

statistical decision and learning theories

The ANR project "StaRAC" and its three groups — ENS Lyon (SISYPHE group, P. Flandrin, P. Borgnat), GIPSA-lab Grenoble (SIGMA-PHY group, P.-O. Amblard), ICD-UTT Troyes (C. Richard) — are looking for 3 post-doctoral researchers (16 months each, starting fall 2008). Prospective candidates are encouraged to get in touch by e-mail with the contact persons mentioned below.

Context

The scientific objectives of the StaRAC project are:

Description of the positions

- propose and develop operational (i.e., interpretable, relative and testable) approaches to the concept of stationarity, with the purpose of filling existing gaps between theory and practice;
- extend the concept of stationarity to general groups of transformations;
- develop new methods to measure, test and model departures from stationarity.

1) post-doctoral researcher at ENS Lyon

The work to be developed with the Lyon group is mostly concerned with revisiting the concept of stationarity within a time-frequency framework (at large). Theoretical studies will be conducted for contributing to the formalization of an operational concept of relative stationarity, together with its assessment and implementation.

This work will be done in collaboration with ICD-UTT Troyes.

Requirements: excellent knowledge of nonstationary, statistical, digital signal processing.

Open to PhD in signal processing, theoretical physics or applied mathematics.

Contact: Patrick.Flandrin (at) ens-lyon.fr or Pierre.Borgnat (at) ens-lyon.fr

2) post-doctoral researcher at GIPSA-lab

The work to be developed in Grenoble concerns the study of generalized stationarities, especially in connection with the notion of self-similarity with stationary increments and their extensions (e.g., processes with finite size scale invariance and stationary increments). To this aim, the well-known notion of self-similar signal with stationary increments will have to be recast using the general work of Hannan and specialized to the case of multiparameter groups. Extensions using warpings will then be performed, and synthesis algorithm will be studied.

This work will be done in collaboration with ENS Lyon.

Requirements: excellent knowledge of nonstationary, statistical, digital signal processing; knowledge of group theory will be appreciated.

Open to PhD in signal processing, theoretical physics or applied mathematics.

Contact: bidou.amblard (at) gipsa-lab.inpg.fr

3) post-doctoral researcher at ICD-UTT Troyes

The main research theme to be developed in Troyes concerns the design of tests of stationarity with statistical-based approaches.

This work will be done in collaboration with ENS Lyon.

Requirements: excellent knowledge of nonstationary signal processing, statistical signal processing and learning theory, preferably on kernel-based methods.

Open to PhD in signal processing, theoretical physics or applied mathematics.

Contact: Cedric.Richard (at) utt.fr