NICOP - Ontology-Based Big Data Information Extraction and Integration
Abstract
This research focuse on Statistical Ontology-Based Big Data Information 1) Extraction & 2) Integration - via the Knowledge Graph. An ontology is a formal, explicit, shared conceptualization of a domain, which defines the concepts and vocabulary used within a community of interest. The formal representation must be both machine readable and " machine understandable" in order for the ontology to be used for logical inference. The researchers will investigate designs and build ontologies that will drive the text mining and information extraction from big data sets containing difficult to find, vital information for the warfighter. Southampton Uiversity is well known for its cobntibutions to symantic web, vocbulary buiding and information retrieval and The PI and Co-Pi will be building on their prviously published work work where they used different vocabularies to model a number of data instances and datasets, including the building of effective grammars and linguistic annotation components as part of anentity recognition system. They will use Probabilistic Soft Logic (PSL) to resolve the data entities and produce the knowledge graph of these entities. PSL allows users to specify rich probabilistic models over continuous- valued random variables. The research will produce tools and methodst for extracting information from various sources, static, dynamic and streaming, and then creating a networked integrated infrastructure, i.e. a knowledge graph, of the entities and their relationships in these environments. b) This project is well aligned with the vision of ONR of providing a globally networked and integrated intelligent enterprise.Warfighter missions now rely on a virtual net of sensors and communications systems for battlefield awareness. There is also a proliferation of information being derived from social media, where the intelligence community is increasingly focusing its attention. Harvesting, analyzing, and rapidly converting these information sources into actionable intelligence for military decision makers are the challenges investigated in this research. Institutional cost sharing is stated to be: $88,866.00 c) Code 31, Code 30 d) A testable, networked integrated platform, which can be shown to collect and itegrate valuable information as support to ONR s mission of providing collaborative analytic tools for the US military and intelligence communities.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 08, 2016
- Source ID
- N629091612056
Entities
People
- Gary Wills
Organizations
- Office of Naval Research
- United States Navy
- University of Southampton