Programme détailléCyril Labbé, Univ Grenoble, LIGFlagging suspect scientific publications for post-publication reassessments. Natural Language Generation has been used to produce fake scientific papers for more than a decade. Now more complex AI-powered generation techniques produce texts indistinguishable from that of humans and the generation of scientific texts starting from a few keywords used as input has been documented. Among other things, our study introduces the concept of tortured phrases: unexpected weird phrases in lieu of established ones, such as ‘counterfeit consciousness’ instead of ‘artificial intelligence.’ Hypothesising the use of advanced language models we ran a detector on the abstracts of recent articles and on several control sets. This will lead us to discuss various research avenues related to the processing of scientific texts: detection of meaning less texts, spotting factual errors, flagging mis-citations,...
Anastasia Bezerianos, Univ. Paris-Saclay, Inria, LISNVisual Text Analytics in Data Journalism Visual text analytics combines natural language processing with interactive visualizations, revealing patterns in individual text documents or document collections. I will briefly present the field of visual text analytics from a visualization perspective, with new and historical visualization examples. Next, I will then discuss three visual text analytics systems we designed within the iCODA project, in collaboration with journalists from OuestFrance and the LINKMEDIA team in Rennes. First, the Storyfier project, that focused on helping journalists conduct close reading of large document collections. And the HyperStorylines and GeoStorylines projects, providing high-level overviews of relationships between entities extracted from news articles.
Patrick Paroubek, CNRS, LISNDu traitement automatique du langage naturel et de l'accès aux données expérimentales L'exposé arbordera la problème de l'accès aux données d'expérimentation par le biais des publications scientifiques et
Géraldine Damnati, Orange InnovationExtraction d'information et gestion de la connaissance au sein d'une organisation. Comment mener des travaux de recherche ouverte sur des données "fermées" ?
Liste des Posters et démos :
|
Personnes connectées : 2 | Flux RSS | Vie privée |