Skip to Main content Skip to Navigation
Habilitation à diriger des recherches

Apprentissage : Paradigmes, Structures et abstractions

Abstract : The work presented here concerns Machine Learning, and particularly supervised learning, together with knowledge discovery through data mining. The approach is logical and learning is considered as a search problem in a partially ordered space. The hidden structure of the data relates the statements of the representation language and their extension in the universe of instances, in an intension/extension lattice. I first discuss how using the hidden structure allows to reduce the search in supervised concentp learning problems. Then I present a form of abstraction that reduces the intension/exension lattice by changing the granlularity of the representation language or of the universe of instances. Some work concerning the extraction of repeated sequential motifs, in particular in bioInformatics, are also discussed from this intension-extension-abstraction point of view. Finally two recnt paradigms in supervised learning are explored. The first one addresses the ambiguity of the instances, and the second one extends supervised learning to a multi-agent collaborative setting.
Document type :
Habilitation à diriger des recherches
Complete list of metadata

Cited literature [140 references]  Display  Hide  Download
Contributor : Henry Soldano Connect in order to contact the contributor
Submitted on : Wednesday, September 1, 2010 - 2:39:25 PM
Last modification on : Saturday, June 25, 2022 - 8:49:46 PM
Long-term archiving on: : Tuesday, October 23, 2012 - 3:21:21 PM


  • HAL Id : tel-00514160, version 1


Henry Soldano. Apprentissage : Paradigmes, Structures et abstractions. Informatique [cs]. Université Paris-Nord - Paris XIII, 2009. ⟨tel-00514160⟩



Record views


Files downloads