SVM Viability Controller Active Learning: Application to Bike Control

Abstract : It was shown recently that SVMs are particularly adequate to define action policies to keep a dynamical system inside a given constraint set (in the framework of viability theory). However, the training set size of the SVMs faces the dimensionality curse, because it is based on a regular grid of the state space. In this paper, we propose an active learning approach, aiming at decreasing dramatically the training set size, keeping it as close as possible to the final number of support vectors. We use a virtual multi-resolution grid, and some particularities of the problem, to choose very efficient examples to add to the training set. To illustrate the performances of the algorithm, we solve a six-dimensional problem, controlling a bike on a track, problem usually solved using reinforcement learning techniques.
Type de document :
Communication dans un congrès
Approximate Dynamic Programming and Reinforcement Learning, 2007. ADPRL 2007. IEEE International Symposium on, Apr 2007, United States. pp.193-200, 2007
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-00616856
Contributeur : Laetitia Chapel <>
Soumis le : mercredi 24 août 2011 - 16:35:48
Dernière modification le : mercredi 24 août 2011 - 16:35:48

Identifiants

  • HAL Id : hal-00616856, version 1

Collections

Citation

Laetitia Chapel, Guillaume Deffuant. SVM Viability Controller Active Learning: Application to Bike Control. Approximate Dynamic Programming and Reinforcement Learning, 2007. ADPRL 2007. IEEE International Symposium on, Apr 2007, United States. pp.193-200, 2007. <hal-00616856>

Partager

Métriques

Consultations de la notice

39