The event will be recorded and available on YouTube later!
L'évènement sera filmé et diffusé sur YouTube plus tard !
Seating is on a first come, first served basis whether you have RSVPed or not, so we suggest arriving early. We can host 100 people. For security reasons, please have your real name on Meetup and bring an ID card.
La salle permet d'accueillir 100 personnes. L'inscription est obligatoire mais ne garantit pas que vous pourrez entrer, nous vous recommandons donc d'arriver un peu en avance. Pour des raisons de sécurité, prière d'apporter un document d'identité qui doit correspondre à votre nom réel sur Meetup.
• Piotr Bojanowski (Facebook AI Research [FAIR] (https://research.facebook.com/ai))
FastText, a library for efficient text classification and word representation learning.
Piotr is a Post Doctoral Researcher at Facebook AI Research. Before Facebook, he worked on his Ph.D. at Willow team at INRIA Paris. He graduated from Ecole polytechnique in 2013 and received a Masters Degree in Mathematics, Machine Learning and Computer Vision (MVA). His work focuses on automatic video understanding and related learning problems. More recently, he also started working on sequence modeling and recurrent networks, with applications to language modeling.
• Éric de la Clergerie (INRIA Alpage (https://www.rocq.inria.fr/alpage-wiki/tiki-index.php?page=Accueil))
What can we do with a parser? Learn!
"There are actually many potential applications for parsers. But one of them concerns the acquisition of knowledge about the world and specialized domains. More precisely, starting from Harris Distributional Hypothesis that "words occurring in similar contexts tend to be close semantically", and using large parsed corpora (such as Wikipedia), I will present two methods for building "semantic spaces" (as graphs or word embeddings). I will also advocate for the coupling of language and knowledge."
Eric is a Research Officer working at INRIA in the field on Natural Language Processing (NLP), more specifically on Parsing and Syntax. He is the main developer of FRMG, a large coverage French grammar that may be tried online at http://alpage.inria.fr/frmgwiki . FRMG has been used to parse all kinds of corpora (including French wikipedia), in particular to perform knowledge acquisition (terminology and semantic networks). It has also been used for information extraction. More recently, he has also worked on statistical transition-based parsers, on hybridation of FRMG with statistical information, and on the coupling of symbolic parsers (such as FRMG) with statistical ones.
Most of his work relies on the DyALog system, a Prolog-like system implementing tabulation and computation sharing, initially developed during his PhD.
• Sylvain Utard (Algolia (https://www.algolia.com/))
Inside the Algolia engine: the indexing challenge of instant search
Founded in 2012, Algolia is a privately held company created by veterans in the fields of algorithms, search engines and text mining. They help users deliver an intuitive search-as-you-type experience on their websites and mobile apps. Basically, they make finding things easier. They are a customer-focused, dev-centric company dedicated to changing the way people think about search.