Damien Nouvel Damien Nouvel

Associate Professor, Inalco, Paris, France

"The question whether machines can think is about as relevant as the question whether submarines can swim" (E. Dijkstra)

> Site fran├žais ...

I work on various topics within Natural Language Processing, such as:

My PhD focused on how to combine knowledge-based and data-driven learning approaches for Named Entity Recognition over oral transcripts. For this purpose, we propose to discover hierarchical sequential patterns by mining large annotated corpora in an exhaustive and objective fashion. We are then able to induce automata (or, more precisely, transducers) called annotation rules that may be used to recognize named entities. The originality of our proposal is that we do not focus on categorizing words, but instead on individually marking boundaries of entities (discriminating a single annotation tag as a local instruction). In this context, the system does not have to categorize words of an entity, but instead aims at inserting beginning and ending tags for recognizing named entities. You may find recent information in the Publications section.

About implementation, have a look at the mXS page for the French version of this system or for information about adaptions to other languages.

For Named Entity Recgnition, if you are looking for references, here are some I'd suggest as a starting point: