Mardi 16 Février
Heure: 
14:00  17:00 
Lieu: 
Salle B107, bâtiment B, Université de Villetaneuse 
Résumé: 
TBC 
Description: 
Benjamin Hellouin 
Jeudi 18 Février
Heure: 
14:00  15:00 
Lieu: 
Salle B107, bâtiment B, Université de Villetaneuse 
Résumé: 
Sampled Weighted MinHashing for LargeScale Topic Mining 
Description: 
Ivan Vladimir MEZA Sampled Weighted MinHashing (SWMH) is a randomized approach to automatically mine topics from largescale corpora. SWMH generates multiple random partitions of the corpus vocabulary based on term cooccurrence and agglomerates highly overlapping interpartition cells to produce the mined topics. While alternative approaches define a topic as a probabilistic distribution over the complete vocabulary, SWMH topics are subsets of such vocabulary. Interestingly, the topics mined by SWMH underlie themes from the corpus at different levels of granularity. We extensively evaluate the meaningfulness of the mined topics both qualitatively and quantitatively on the NIPS (1.7K documents), 20 Newsgroup (20K), Reuters (800K) and Wikipedia (4M) corpora. 
Vendredi 19 Février
Heure: 
11:00  12:00 
Lieu: 
Salle B107, bâtiment B, Université de Villetaneuse 
Résumé: 
Relational typechecking of connected proofstructures 
Description: 
Luc Pellissier It is possible to define a typing system for Multiplicative Exponential Linear Logic (MELL): in such a system, typing judgments are of the form ? R : x : ?, where R is a MELL proofstructure, ? is the list of types of the conclusions of R, and x an element of the relational interpretation of ?, meaning that x is an element of the relational interpretation of R (of type ?). As relational semantics can be used to infer execution properties of the proofstructure, these judgment can be considered as forms of quantitative typing. We provide an abstract machine that decides, if R satisfies a geometric condition, whether the judgment ? R : x : ? is valid. Also, the machine halts in bilinear time in the sizes of R and x. 

