Organized by :
Chair of Applied Statistics at Conservatoire National des Arts
et Métiers (CNAM), Paris, France
Invited speakers :
Professor, Department of Statistics, Stanford University, and
Leader, Computation Research Group, Stanford Linear Accelerator Center[bio]
Professor of Statistics in the Department of Mathematics at
Imperial College [bio]
Director of MPI for Biological Cybernetics, Member of KXEN
Scientific Committee [bio]
Senior Research Scientist [bio]
[Top of page]
Program Scientific committee :
- G. D’Aubigny (Grenoble)
- M. Bera (KXEN)
- P. Bertail (CREST-ENSAE)
- P. Deheuvels (Académie des Sciences)
- G. Hébrail (EDF DER)
- G. Saporta (CNAM)
- A.J. Valleron (INSERM)
[Top of page]
Objectives :
- Statistical Learning Theory (SLT) brings a new vision to prediction modelization problems, the predicted target being qualitative or quantitative. It is a paradox but SLT is better known in the AI/Computer Science community (artificial intelligence, neural networks, machine learning) than in the Statistics community. SLT has brought significant results in fields such as pattern recognition and decision making algorithms. Some recent SLT developments are very close to rather old works in Multivariate Data Analysis but SLT efficiency, in terms of computing complexity, now allows huge scalability in database and problem sizes.
- The purpose of this conference is to provide a good summary of state
of the art Statistical Learning Theory, as well as the future perspective
for SLT development presented by specialist mathematicians, with worldwide
reputation, who all have written fundamental books on SLT subject.
- This conference is intended to address research and operational individuals, in both Statistics and Computer Science fields.
- Suggestions for conference papers of 30 minutes can be sent, in English,
to the conference Scientific Committee (saporta@cnam.fr)
before October 10, 2002, on subjects relating to Statistical Learning
Theory: SVM, regularization, model validation, model sorting, VC dimension,
reproducing kernel-based embeddings into Hilbert spaces, specific algorithms,
etc.).
[Top of page]
Publications :
A selected list of forthcoming conference communications will be published in a special issue of Applied Stochastic Models in Business and Industry, or in the Journal de la Société Française de Statistique.
Organizing committee :
[Top of page]
Access map :
CNAM
Auditorium C - Abbe Grégoire
292 rue Saint Martin
75003 Paris - France
Metro : Reaumur Sebastopol or Strasbourg Saint Denis
http://www.cnam.fr/home/Infos_pratiques/plan.html
Registration on November will start at 8:30 am.
Registration fees :
€ 125 per participant for academics (including lunch
and refreshments)
€ 250 per participant from businesses (including lunch and refreshments)
€ 25 per student (lunch excluded)
[Top of page]
