As a key contributor, project leader, and manager, Dr. Frawley has conducted and directed leading-edge applied research and development in knowledge-based systems; distributed, adaptive systems; knowledge discovery in databases; and network information agents.

PROFESSIONAL BACKGROUND: At the NASA Lewis Research Center Dr. Frawley did early work in the analysis and computation of multi-group neutron diffusion and single-group neutron transport. For two years, he was an instructor of mathematics at John Carroll University, introducing courses in numerical analysis and programming. In two periods of employment at the Schlumberger-Doll Research Center (SDRC) he headed the mathematics section and founded one of the first industrial informatics research groups. His own research dealt with the mathematical physics of low-frequency electromagnetics and real-time computation and control of nuclear spectroscopic

measurements.

Since 1977 he has worked in the fields of knowledge based systems and artificial intelligence. At SDRC he established the first industrial artificial intelligence research team and was responsible for Gamma, a knowledge based system for the interpretation of gamma ray spectra, and the Dipmeter Advisor, an expert system for subsurface geological interpretation. At the Mitre Corporation he held a staff position with and later was leader of the Knowledge Based Systems Group which produced Knobs, a resource allocation and natural language system for tactical air mission planning, which was the research prototype for the system which successfully handled that task in the Gulf War.

Joining GTE Labs in 1983 as a research manager, he established and directed a research department focused on adaptive systems and distributed intelligence. The department installed and maintained an adaptive system for quality control in Sylvania light bulb plants. Individually, he worked on models of learning by example, and their application to telephone service provisioning.

From 1987 through 1996, he was a Principal MTS at GTE. As project leader for Learning In Expert Domains (through 1993), he was responsible for the Integrated Learning System (ILS), a multi-agent, platform-independent adaptive telephone network controller, and for ARMS, an inductively-defined automated maintenance scheduler for airborne radios. The ILS was the first practical demonstration of multistrategy learning. ARMS reduced maintenance costs for Airphone by $3M per year. During this period, Dr. Frawley worked on methods by which domain knowledge can direct data-driven inductive techniques. His system for function-based induction (Fbi) was used in the ILS and ARMS, and to identify failing components by discovering associated patterns in Mobilnet call data.

In 1994 he established a Network Information Agents project, targeting services provided over the information superhighway, and was its principal investigator through 1996. The main result was a prototype internet news reader and filter, Gina, soon to be available to GTE's ISP customers. For Gina he provided the personalization and customization module, clustering users based on their usage histories and presenting to them options common to their behavioral counterparts.

OTHER: He is a member of the American Mathematical Society, IEEE, Sigma Xi, and AAAI, has been an NSF Research Fellow and a NASA Fellow, and holds two patents. He served as co-chair of the first international workshop on knowledge discovery in databases (IJCAI, 1989), was an organizer of subsequent knowledge discovery workshops at national conferences, and was a co-author of the paper selected as Best Paper at the 1990 European AI conference. In 1992 he lectured at four universities in Brazil, sponsored by the Council for the Development of Science and Technology of Brazil. At GTE, he received that company's highest award for technical achievement, the Leslie H. Warner Award For Research, in 1992, and in 1993 received a Superior Performance award.

EDUCATION:

B.S, M.S. - Physics, John Carroll University. Theses: Ultrasonic Absorption in Fluids, The Fundamentals of Digital Computers. Advisor: E. F. Carome.

Ph. D. - Mathematics, University of Oklahoma. Thesis: Locally Disconjugate Families of Continuous Functions. Advisor: W. T. Reid.

Return to Bud's Home Page