Inria
Inria.fr
Inria Channel
- Presentation
- HAL publications
- Activity reports
TAO Research team
Machine Learning and Optimisation
- Leader : Marc Schoenauer
- Type : Project team
- Research center(s) : Saclay
- Field : Applied Mathematics, Computation and Simulation
- Theme : Optimization, Learning and Statistical Methods
- Université Paris-Sud (Paris 11), CNRS, Laboratoire de recherche en informatique (LRI) (UMR8623)
Team presentation
- Machine Learning can be viewed as an optimization problem. A critical issue, that lead to the development of the Statistical Learning Theory, is that the objective function (the generalization error) is not exactly known.
- A common practive of several sub-fields of Evolutionary Computation (e.g. Multi-Objective Optimization, Constraint Handling, Co-evolution) is to gather and later exploit some archive of previously encountered solutions. An efficient use of such archive naturally appeals to Machine Learning techniques. The project TAO, result of a crossover between part of the Fractales group at INRIA Rocquencourt and part of the Inférence et Apprentissage group at LRI, thus aims at exploiting efficiently such a synergy between Machine Learning techniques and Evolutionary Algorithms.
- CLASSIC - Computational Learning, Aggregation, Supervised Statistical, Inference, and Classification
- DOLPHIN - Parallel Cooperative Multi-criteria Optimization
- GEOSTAT - Geometry and Statistics in acquisition data
- MISTIS - Modelling and Inference of Complex and Structured Stochastic Systems
- MODAL - MOdel for Data Analysis and Learning
- REALOPT - Reformulations based algorithms for Combinatorial Optimization
- SELECT - Model selection in statistical learning
- SEQUEL - Sequential Learning
- SIERRA - Statistical Machine Learning and Parsimony
Research themes
The main application areas are process control (after having caracterized the defects, the issue of minimizing their frequency naturally arises), some medical applications (like the caracterization of the multi-admitted persons), and intelligent control in robotics.International and industrial relations
TAO members are deeply involved in several European Network of Excellence: Evonet and KDNet (FP5) and the PASCAL project (FP6).Keywords: Machine Learning Evolutionary Optimization Data Mining
Research teams of the same theme :
Contact
Team leader
Marc Schoenauer
Tel.: +33 1 69 15 66 26
Secretariat
Tel.: +33 1 69 15 34 71
Find out more
See also
