This Meetup is past

32 people went

University of Maryland, Baltimore County

1000 Hilltop Circle, Technology and Engineering Building, Room 456 · Baltimore, MD

How to find us

Directions: http://www.umbc.edu/aboutumbc/directions.php Greg Milbank (M) 201-232-9195

Location image of event venue

Details

6:00-6:45 Pizza Social

6:45-8:00 Featured Speakers, Tim Finan, Professor of Computer Science at UMBC and the Research Team in the Advanced Research Lab ;

Topic: Collaboration, Automatic Linked Data, and Graph based Queries – Semantic Web Research at UMBC

Tim starts off with a broad overview of the lab and then provides details on three current research activities.

1)
Laura Zavela; Mobile, Collaborative, Context-Aware Systems:

The Semantic Web provides the technology and knowledge constructs to create a rich notion of context that goes beyond current networking applications focusing mostly on location. The context model includes location and surroundings, the presence of people and devices, inferred activities and the roles people fill in them.

2) Varish Mulwad; Automatically Generating Linked Data:

Evidence for a table's meaning can be found in its metadata but currently requires human interpretation. We describe techniques grounded in graphical models and probabilistic reasoning to infer meaning associated with a table. Using background knowledge from the Linked Open Data cloud, we automatically infer the semantics of column headers, table cell values (e.g., strings and numbers) and relations between columns and represent the inferred meaning as graph of RDF triples.

3) Lushan Han; GoRelations - A Question Answering System on DBpedia:

We developed an intuitive { semantic graph} notation allowing one to pose queries by annotating graphs with natural language terms denoting entities and relations. The query is automatically translated into SPARQL to produce an answer. Key contributions are the robust techniques mapping user terms to the most appropriate classes and properties in the ontologies used in the linked data. Our approach combines a statistical analysis of the underlying data and lexical semantic similarity metrics derived from a large text corpus and WordNet.

Level: Intermediate