Guillaume Lecué

lecue@latp.univ-mrs.fr

Contact

• Mail Adress:
  • Laboratoire d'Analyse, Topologie et Probabilités UMR 6632
  • Centre de Mathématiques et Informatique (CMI)
  • Université de Provence
  • Technopôle Château-Gombert
  • 39 rue F. Joliot Curie
  • 13453 Marseille Cedex 13
  • France
• Phone: +33 (0) 4 91 11 35 58
• Fax: +33 (0) 4 91 11 35 34

Current situation

I am a CNRS researcher since the 1st of October 2007. I did a Post-doc with Pr. Shahar Mendelson at the Australian National University in Canberra and at the Technion in Haïfa. I did my Phd at the Laboratoire de Probabilités et Modèles Aléatoires of Université Paris 6. My PhD supervisor was Pr. Alexandre Tsybakov.

Publications and preprints

1. S. Gaïffas and G. Lecué, "Aggregation of penalized empirical risk minimizers in regression." Submitted.
   
2. G. Lecué and S. Mendelson, "Sharper lower bounds on the performance of the Empirical Risk Minimization Algorithm." Submitted.
   
3. K. Bertin and G. Lecué, "Selection of variables and dimension reduction in high-dimensional non-parametric regression." Electronic Journal of Statistics, Vol. 2, pp. 1224-1241. 2008.
   
4. G. Lecué and S. Mendelson, "Aggregation via Empirical Risk Minimization." To appear in Probability theory and related fields.
   
5. C. Chesneau and G. Lecué, "Adapting to Unknown Smoothness by Aggregation of Thresholded Wavelet Estimators." To appear in Statistica Sinica.
   
6. G. Lecué, "Classification with minimax fast rates for Classes of Bayes Rules with Sparse Representation." Electronic Journal of Statistics, Vol. 2, pp. 741-773. 2008
7. S. Gaïffas and G. Lecué, "Optimal rates and adaptation in the Single-Index Model using aggregation." Electronic Journal of Statistics, Vol. 1, pp. 538-573. 2007
   
8. G. Lecué, "Suboptimality of Penalized Empirical Risk Minimization In Classification."
   20th Annual Conference On Learning Theory, COLT07. Proceedings. Bshouty, Gentile (Eds.). Springer. LNAI 4539, 142-156. Springer.
9. G. Lecué, "Optimal rates of aggregation in classification under low noise assumption."
   Bernoulli 13(4), 2007, 1000-1022.
10. G. Lecué, "Simultaneous adaptation to the margin and to complexity in classification."
    Annals of Statistics, Vol. 35, No. 4. pp. 1698-1721. August 2007.
11. G. Lecué, "Optimal oracle inequality for aggregation of classifiers under low noise condition."
    19th Annual Conference On Learning Theory, COLT06. Proceedings. Gabor Lugosi, Hans Ulrich Simon (Eds.). Springer. LNAI 2006, 364-378. "Mark Fulk award" for the best student paper.
12. G. Lecué, "Lower Bounds and Aggregation in Density Estimation."
    Journal of Machine Learning Research, 7(Jun):971-981, 2006.

PhD

I have defended my PhD on the 18th of May 2007 in front of the following jury:
• Alexandre Tsybakov, Professor at Université Paris 6. PhD advisor.
• Vladimir Koltchinskii, Professor at Georgia Institute of Technology, USA. Referee.
• Gábor Lugosi, Professor at Universitat Pompeu Fabra, espagne. Referee.
• Pascal Massart, Professor at Université Paris 11. President.
• Olivier Catoni, Professor at Université Paris 6. Member.
• Arkadi Nemirovski, Professor at Georgia Institute of Technology, USA. Member.
• Stéphane Boucheron, Professor at Université Paris 6.Member.
• Lucien Birgé, Professor at Université Paris 6. Member
My PhD manuscript is available here and the slides of my PhD defense are available here.

Invitations, presentations

• An aggregation procedure in classification. Applications to adaptivity.
• Suboptimality of penalized empirical risk minimization in classification.
• Aggregation methods in classification: optimality and adaptation.