Bryan Thompson

Bryan B. Thompson

SYSTAP, LLC

Company URL: http://www.systap.com

Bigdata Project URL: http://www.bigdata.com/blog

email: bryan-at-systap-com

Mr. Bryan B. Thompson has 30 years experience in small business, government, and the private sector with a focus in applied research and development and emerging technology. His technical background includes expertise in cloud computing; the semantic web; web architecture; relational, object, and RDF database architectures; knowledge management and collaboration; artificial intelligence and connectionist models; natural language processing; metrics, scalability studies, benchmarks and performance tuning; decision support systems; and usability design. His work for the last several years has focused on assessing and applying Semantic Web technologies to support semantics-based federation of disparate data sources for the Intelligence Community. Significant developments include:

• The creation of bigdata®, an open source distributed database holding several world records, including the fastest and most scalable semantic web platform. Bigdata® is commercial licensed through OEMs and is used within in the intelligence community, telecommunications industry, managed storage networks, heath care, bioinformatics, etc.;
• The synthesis of a cognitive model of human reasoning about risk and uncertainty ("Recognition / Metacognition theory": Cohen, Freeman & Thompson, 1998) with a connectionist model of human reflexive inference ("Shruti": Shastri & Ajjanagadde 1993; Shastri, 1996) that provides a cognitive and computational explanation and model of expertise and critical thinking behaviors;
• Novel extensions to the IBIS architecture include support for reasoning with positive and negative evidence, support for planning, and support for engaging individuals in critical thinking dialogs that identify the sources of uncertainty in their expressed arguments. IBIS systems provide a framework for expressing structured arguments that link evidence (text, images, etc.) to conclusions.
• A class of optimal incremental planners (Thompson and Cohen, 1999) and their application to sophisticated knowledge models within an architecture for a connectionist reasoner ("Shruti"; Shastri, http://www.icsi.berkeley.edu/~shastri/shruti);
• An adaptive control methodology known as Action Dependent Heuristic Dynamic Programming (Lukes, Thompson and Werbos, Jan 1990). This method was first and most extensively published by Watkins as Q-Learning in his doctoral thesis (Watkins, 1989).