My research work, present activities and projects.
Fast is fine, but accuracy is everything.
Wyatt Earp.
I successfully defended a PhD thesis entitled
Language bias in the generate-and-test algorithm
and Contributions to disjunctive learning.
This work was supervised by
Céline Rouveirol
at the Laboratory of Computer Science (LRI, Paris-Saclay University).
My main interests are:
voting methods like boosting and applications to various
frameworks in which a least general generalization algorithm exists ;
in this task, I developped
volata,
a java library that implements ensemble methods and least general generalization algorithms ;
volata
allows such boosting simulations ;
application of machine learning to psychology
with Romina Manoli,
and problems involved by medical data like the INDANA database ;
multi-task learning with Jean-Baptiste Faddoul ;
information extraction with Patrick Marty ;
unsupervised learning and text clustering
with Laurent Candillier ;