In the first part,we shall consider several classes of categorial grammars and discuss theirlearnability. We consider learning as a symbolic issue in an unsupervisedsetting, from raw or from structured data, for some variants of Lambek grammarsand for categorial dependency grammars. In that perspective, we discuss forthese frameworks different type constructors and structures, some limitations(negative results) but also some algorithms (positive results) under somehypothesis.

The second partwill describe some works in the team LIS (Logical Information Systems); formalmodels have been developed for information systems that allow for navigation,querying, updating, and analysis of heterogeneous data collections where dataare given (logical) descriptors. While the LIS approach is rather general, and generalpurpose tools are provided, it has been applied for example to geographic data;categorial grammars can also be seen as a particular case of LogicalInformation System; current research in the LIS team also includes a semanticweb oriented version. If time permits, some tool demonstrations are alsopossible.

Annie Foret is an associate-professor (Maître de Conférences) of computerscience in Rennes University, France. She is a member of the LogicalInformation System (LIS) team in the Data and Knowledge Management departmentat the Irisa laboratory in Rennes; she also works in collaboration with the NLPteam (TALN) at LINA in Nantes.

She is a visitor to ILCC in 2013-2014, duringher sabbatical year at French CNRS. Her general research interests are in logic,language and computation. Her current research interests include grammarlearning and categorial grammars.

Previously, she studied mathematics andcomputer science at École normale supérieure, and was a PHD student at Inria onnon-classical logics and rewriting, under the supervision of G. Huet. She thenjoined Irisa and Rennes University where she recently completed her habilitation(HDR) on “some classes of type-logical grammars that model syntax”.