Linking entities for enriching ands structuring social media content

Raphael TRONCY - Maite de conférences
Data Science

Date: -
Location: Eurecom

Abstract: Social media platforms such as Twitter, Facebook or LinkedIn become a reliable source of news and play a key role for being aware of events around the world. Using social media to recognize, enrich or summarize events is however very challenging. In the first part of this talk, we will present ADEL, a novel hybrid architecture for an adaptive entity linking system, that combines methods from the natural language processing, information retrieval and semantic fields. We will show how ADEL can outperform the state-of-the-art systems on the reference NEEL challenge that happens in the yearly #Micropost workshop (2014-2016). In the second part of this talk, we will present a novel method that enables to create so-called news semantic snapshots, that are composed of a set of named entities that best summarize and contextualize the content of a news item. We will show how those algorithms are used in applications that enrich news videos or scientific talks through the HyperTED demonstrator. Speaker's bio: Dr. Raphaël Troncy (http://www.eurecom.fr/~troncy) is an Associate Professor in the Data Science Department of EURECOM leading the Multimedia Semantics group. He is also co-chair of the W3C Media Fragments Working Group and W3C Incubator Group on Multimedia Semantics, contributes to the W3C Web Annotations Working Group and numerous W3C Community Groups such as Schema.org or Open Linked Education. He is working on information extraction (project ASRAEL, http://asrael.limsi.fr/) and in particular named entity linking (framework NERD, http://nerd.eurecom.fr/), ontology modeling in the multimedia domain (project DOREMUS, http://doremus.org/) and large scale data integration. Applications range from smart cities (project 3cixty, http://3cixty.com/) to social TV and second screen (projects LinkedTV, http://www.linkedtv.eu/ and NexGen-TV, http://nexgentv.fr/) where semantic web, information extraction and multimedia analysis technologies are used together.