Fact-checking is the task of assessing the factual accuracy of claims, typically prior to publication. Modern fact-checking is faced with a triple revolution in terms of scale, complexity, and visibility, as claims and background knowledge are increasingly digital.
Our project brings together experts in data management, natural language processing, automated reasoning and data mining from Inria, LIMSI/CNRS and U. Paris Saclay, U. Rennes 1, U. Lyon 1 (full participant list below), and the fact-checking team “Les Décodeurs” from Le Monde, the leading French national newspaper. We work to establish fact-checking as a data management problem, endow it with sound foundations, design and deploy novel algorithms for automating fact-checking, and validate them by close interaction with the journalists.
ContentCheck talks, presentations and general-audience publications:
- Ioana Manolescu in the Inria Alumni session “Fake news and post-truth: straight to the facts!”, CNAM, October 2017
- “Les chercheurs en informatique au service du fact-checking”, in the DIXIT newsletter of Ecole Polytechnique, June 2017
- Ioana Manolescu participated to a fact-checking panel at Web2Day 2017, a 3000-strong IT and digital conference in Nantes
- “La vérité, rien que la vérité” article by Ioana Manolescu in the “Blog Binaire” of Le Monde (computer science topics presented to the general audience), April 2017
- Les Décodeurs published an article analyzing the Twitter content produced during the 2016 primaries of the right, in view of the 2017 French election campaign. The article is based on a collection of public tweets, stored and analyzed as part of the ContentCheck project.
- Michaël Thomazo presented FactMinder, our system for archiving, annotating, and querying semantic-rich Web conten at the Tech & Check Conference of the Duke University in North Carolina
- Ioana Manolescu and Samuel Laurent gave invited presentations at the Computation and Journalism workshop on March 15 in Rennes: http://compjournalism2016.irisa.fr/ co-organized by Xavier Tannier
- ContentCheck in the news: Journal du CNRS, Ouest France, Le Devoir (Canada), INRIA news, Le Monde, Rue 89, NLTO, Sciences pour tous, interstices.info
ContentCheck scientific publications:
- Flash points: Discovering exceptional pairwise behaviors in vote or rating data, Adnene Belfodil, Sylvie Cazalens, Philippe Lamarre, Marc Plantevit. ECML/PKDD conference, 2017.
- Associated material available at: https://github.com/Adnene93/DiscoveringSimilarityChanges/
- Extracting Linked Data from statistic spreadsheets Tien Duc Cao, Ioana Manolescu, Xavier Tannier, Workshop on Semantic Big Data (SBD), next to the SIGMOD conference, 2017.
- Mixed-instance querying: a lightweight integration architecture for data journalism Raphaël Bonaque, Tien Cao, Bogdan Cautis, François Goasdoué, Javier Letelier, Ioana Manolescu, Oscar Mendoza, Swen Ribeiro, Xavier Tannier, Michaël Thomazo VLDB, Sep 2016, New Delhi, India. VLDB, <http://vldb2016.persistent.com/>
- Creation, Visualization and Edition of Timelines for Journalistic Use by Xavier Tannier and Fréderic Vernier, Natural Language Processing Meets Journalism workshop next to IJCAI 2016 http://nlpj2016.fbk.eu
- Cedar — Le Monde — LIMSI meeting: Aug 28, 2017 @Skype meeting
- Cedar — Le Monde — LIMSI meeting: Jun 28, 2017 @ Le Monde
- Cedar — Le Monde — LIMSI meeting: Apr 11, 2017 @ Le Monde
- Third project meeting: Mar 07, 2017 @ Le Monde
- CEDAR — Le Monde — LIMSI meeting: Nov 11, 2016 @ Skype meeting
- CEDAR — Le Monde — LIMSI meeting: Oct 11, 2016 @ Skype meeting
- CEDAR — Le Monde — LIMSI meeting: Oct 18, 2016 @ Skype meeting
- CEDAR — Le Monde — LIMSI meeting: Sep 27, 2016 @ Le Monde
- Second project meeting: July 5, 2016 @ Le Monde
- Presentation of “Creation, Visualization and Edition of Timelines for Journalistic Use” by Xavier Tannier and Frédéric Vernier
- Presentation of “Towards Automatic Topic Assignment for News Articles” by Tien-Duc Cao, Ioana Manolescu and Xavier Tannier
- (Brief) presentation of “Tatooine, a lightweight integration architecture for data journalism” by Raphaël Bonaque, Tien Cao, Bogdan Cautis, François Goasdoué, Javier Letelier, Ioana Manolescu, Oscar Mendoza, Swen Ribeiro, Xavier Tannier, Michaël Thomazo
- Discussion on fact-checking procedures and datasets
- CEDAR — Le Monde — LIMSI meeting: May 30, 2016 @ INRIA, Palaiseau
- First project meeting: January 29, 2016 @ Le Monde in Paris
- INRIA – CEDAR team: models, languages, and efficient database techniques for complex, semantic-rich data
- IRISA, Université de Rennes 1 – SHAMAN team: automated reasoning, representing and querying knowledge, and integrating heterogeneous information sources
- LIMSI, CNRS – ILES team: natural language processing, information retrieval and extraction, event extraction and information
Xavier Tannier, Brigitte Grau, Patrick Paroubek
- LIRIS, CNRS – DB & DM2L teams: data management and mining, in particular declarative and logical approaches, query evaluation, heterogeneous data integration in large-scale distributed systems, and constraint-based pattern mining in particular for spatio-temporal data
Philippe Lamarre, Sylvie Cazalens, J.-Marc Petit, Marc Plantevit, Céline Robardet
- Le Monde, Les Décodeurs: major leading French newspaper and fact-checking pioneer in the French news industry; provides application knowledge, helps define a typology of needs, builds a reference set for evaluation, and proof-tests the ideas and tools produced by the project
Samuel Laurent, Aline Rouyer, Ludovic Werwinski