Research activities

I work on machine learning, set constraints, tree automata, tree rewriting. See the list of my publications .

Member of the Grappa Team, in the LIFL (Laboratoire d'informatique fondamentale de Lille.

Member of the Mostare research group of the french research institure INRIA.

Projects

• Lampada (2009-2013)
• Crotal (2007-2009) on conditional random fields for natural language.
• Marmota (2005-2009) on statistical machine learning and tree structured data. (Scientific coordination)

past projects

• ATASH on document transformations
• WebContent on information extraction
• Member of the ACI group Masses de données.

Students

• Edouard Gilbert with Rémi Gilleron: on probabilistic models from tree structured data.

Past

• Patrick Marty: with Rémi Gilleron and Fabien Torre: on information extraction and supervised learning. (Defense on December 4, 2007)
• Florent Jousse: with Isabelle Tellier and Rémi Gilleron: on machine learning and question answering. (Defense on October 31, 2007)
• Julien Carme: with Rémi Gilleron, Aurélien Lemay and Joachim Niehren: on information extraction from tree structured data and grammatical inference. (Defense on September 23, 2005)
• Daniela Dudau: with Rémi Gilleron and Isabelle Tellier: Machine learning and categorial grammars. (Defense on April 13, 2004).
• Franck Seynhaeve: with Sophie Tison Set constraints and term rewriting. (Defense on May 12, 1999).

TATA

TATA (Tree Automata Techniques and Applications) is an online-book on tree automata writen with H. Comon, M. Dauchet, R. Gilleron, F Jacquemard, Christof Löding, D. Lugiez et S. Tison.

Programme Committee

• ECML/PKDD 2010,
• LATA 2010,
• ATANLP workshop on tree automata and natural languages in ACL 2010.
• TAL journal of the ATALA for the Machine Learning for NLP, Volume 50 num 3.

Softs

• posunlbayes: from Text Classification from Positive and Unlabeled Examples François Denis and Rémi Gilleron and M. Tommasi. (A for ever prototype ...)
• movieXML: A slight modification of the freely avalaible code to request the Imdb movie database. Produces XML output and lets you apply stylesheets to get benchmarks for information extraction, segmentation of tree structured data.

Homepage