Subject : Dealing with Misspecifications in Signal Processing
A motivated post-doctoral candidate is solicited in the LEME laboratory of Paris Nanterre University at Ville d’Avray (50 Rue de Sèvres, 92410 Ville-d'Avray) and the L2S laboratory of CentraleSupelec.
The candidate must possess a doctoral degree in signal processing, applied mathematics (optimization, statistics...), or information theory. Good mathematical background is required. Solid programming skills (matlab/python) will be appreciated. Above all, the applicants must be motivated to learn quickly and work effectively on challenging research problems.
Please send your CV (containing contact information of at least two referees), statement of research experience and interests and representative publications (maximum 3 articles) in attachment to M. N. El Korso (firstname.lastname@example.org), A. Breloy (email@example.com) and F. Pascal (firstname.lastname@example.org).
Context and research project:
Parametric inference and information recovery methods in signal processing are strongly dependent on the statistical model and the observed data. Nevertheless, in practice, statistical models are seldom fully or correctly specified. This leads to a dramatic performance loss when applying classical information recovery methods assuming perfectly specified models in misspecified scenario.
For this post-doc position, we aim at studying new basis for information recovery and performance prediction for non regular statistical models. By non-regular, we mean misspecified models with possibly under-determined, partial, incomplete, censured or truncated data. Specifically, by using advanced optimization tools, we will combine the advantages of both likelihood based robust parametric and empirical non-parametric methodologies for signal processing applications. Finally, it should be noted that our primary focus, in terms of application, will be related to calibration in the context of very large radio-interferometer and/or MIMO radar.
The research grant is awarded for 12 months and its monthly net salary is about 2000€.
Expected starting: Sept/Oct 2019.
- A. M. Zoubir, V Koivunen, Y Chakhchoukh, M Muma, Robust estimation in signal processing: A tutorial-style treatment of fundamental concept, IEEE Signal Process. Mag. Vol. 29 (4), 61-80, 2014.
- A. B. Owen, Empirical Likelihood. Chapman & Hall/CRC, Boca Raton, 2001.
- F. Pascal, H. Harari-Kermadec, and P. Larzabal, The Empirical Likelihood method applied to covariance matrix estimation, Signal Process., vol. 90, no. 2, pp. 566-578, Feb. 2010.
- H. White, Maximum likelihood estimation of misspeciffied models, Econometrica, vol. 50, no. 1, pp. 1-25, 1982.
- C. D. Richmond and L. L. Horowitz, Parameter bounds on estimation accuracy under model misspeciffication, IEEE Trans. on Signal Process., vol. 63, no. 9, pp. 2263-2278, 2015.
- A. Mennad, S. Fortunati, M. N. El Korso, A. Younsi, A. M. Zoubir and A. Renaux, Slepian-Bangstype formulas and the related Misspecified Cramér-Rao Bounds for Complex Elliptically Symmetric distributions, Elsevier Signal Process., Vol. 142, Jan. 2018, Pages 320-329
- A. Breloy, G. Ginolhac, F. Pascal, P. Forster, Robust Covariance Matrix estimation in Low-Rank Heterogeneous Context, IEEE Trans. Signal Process., vol. 64, no. 22, pp. 5794-5806, 2016.
- V. Ollier, M. N. El Korso, R. Boyer, P. Larzabal and M. Pesavento, Robust Calibration of Radio Interferometers in Non-Gaussian Environment, IEEE Trans. on Signal Proc., Vol. 65, Nov. 2017, pp. 5649-5660.
- Y. Sun, A. Breloy, P. Babu, D.P. Palomar, F. Pascal, G. Ginolhac, Low-Complexity Algorithms for Low Rank Clutter Parameters Estimation in Radar Systems, IEEE Trans. Signal Process., vol. 64, no. 8, pp. 1986-1998, 2016.
(c) GdR 720 ISIS - CNRS - 2011-2019.