Semantic Enrichment of Twitter Microposts Helps Understand Post-Brexit Reactions

Ontotext is a global leader and innovator in the area of semantic graph databases (triplestores), Linked Open Data, Semantic Web, linguistic analysis, cognitive computing and artificial intelligence. The company develops key solutions to data challenges in various industries such as medical, pharmaceutical, publishing, financial risk management and compliance, legal as well as in the fields of text mining and text analytics, data integration and master data management.

At SEMANTiCS 2016, Ontotext will present PHEME – an ongoing project funded by Seventh Framework Programme (FP7) of the EU for research in the field of ICT, coordinated by the University of Sheffield, and in partnership with nine European organizations from England, Germany, Austria, Spain, Switzerland and Kenya. The goal of the project is to create a software for automatic discovery and verification of information at scale and fast. The benefits of applying such computational framework will be shown in the context of social media analytics - better understanding the impact of rumours in social media – where they originate, how they spread, who and why spreads them, with multiple uses in many domains. 

Ontotext is the developer of GraphDB – a semantic graph database that serves enterprises to store, organize and manage content and data in the form of semantically enriched Linked Data. 

The project aims to develop ways of verifying information disseminated through social media networks, as well as debunk false and abusive information (rumorous memes) that may be generated by inauthentic users. 


Ontotext Booth at the SEMANTiCS 2016 Marketplace

Come talk to us at the booth in the exhibition area to find out more about how we tackle big data challenges.

Presentation during the Main Conference

Vladimir Alexiev, Lead of the Data & Ontology Team at Ontotext will briefly present some of the underlying technologies of the PHEME project that make it possible for social network analysis and information visualization to take place eventually. That is, on the combination of natural language processing and text analytics methods that break down a piece of information (e.g. a tweet) to its lexical, semantic and syntactic components, and then additional algorithms perform a series of cross-reference tasks and semantic enrichment by harnessing knowledge from Linked Open Data via Ontotext’s scalable semantic graph database (GraphDB). To better illustrate how LOD annotations can be a powerful tool for semantic enrichment, he will focus on the process of linking a series of Social Media microposts to information available in other knowledge bases (i.e. not directly transmitted through Social Media channels) in order to provide better and richer context for further analysis of these microposts.

Time & Date of Presentation: Wednesday, Sep 14th, 16:15 - 16:45, WiFa SR 3, Track: Industry 4.0

Presentation during the 7th DBpedia Community Meeting

In addition to the PHEME presentation during the Main Conference, Vladimir Alexiev will give a presentation on Multisensor Linked Open Data at the DBpedia & NLP Session. The Multisensor project analyzes and extracts data from mass- and social media documents, including text, images and video, across several languages. It uses a number of ontologies for representing that data: NIF and OLIA for linguistic info, ITSRDF for NER, DBpedia and Babelnet for entities and concepts, MARL for sentiment, OA for image and cross-article annotations, etc. He will talk about the challenges of fitting all these ontologies together, and about some innovations like embedding FrameNet in NIF.

Vlado will also be in the committee for the DBpedia Citation Challenge.

Time & Date of Presentation: Thursday, Sep 15th, 14:30 - 15:45, WiFa SR4, room: Parallel Session 2 


Ontotext AD
Polygraphia Office Center, fl. 4
47A Tsarigradsko Shosse
Sofia 1124