Ph.D., Brown University, Computer Science, 1991
M.Sc. Computer Science, Brown University, 1988
B.A. Philosophy, Yale University, Magna Cum Laude, 1983
Dr. Goldman specializes in Artificial Intelligence. His particular research interests center around planning, in particular the boundary between planning and control theory; and reasoning and action under uncertainty. Dr. Goldman is currently leading SIFT's activity on the BBN SMITE team on DARPA's Scalable Network Monitoring (SNM) program, developing technology to drastically reduce the false-positive rate of network security systems.
Dr. Goldman has led numerous major projects at SIFT. Most recently, he guided SIFT's successful contributions over the multi-phase Integrated Learning program, providing infrastructures in which learning agents may combine their knowledge. He earlier executed SIFT’s DARPA-sponsored Phase 2 SBIR contract " Playbook-enhanced Variable Autonomy Control System (PVACS)" (DAAH01-03-C-R177). He also contributed to SIFT’s effort on DARPA’s Mixed-initiative Control of Automa-teams (MICA) program, which is a subcontract from Cornell University.
Prior to joining SIFT, Dr. Goldman was a senior principal research scientist in the Automated Reasoning group at the Honeywell Technology Center (HTC) from 1994 through the summer of 2002. In 2001, Dr. Goldman won two technical achievement awards: one for his work on the Circa hard real-time AI system and one for his work on the Argus/Scyllarus system for information fusion in computer intrusion detection. He also received two peer-nominated awards for specific technical contributions to these programs, and was co-author of two patent applications. Dr. Goldman worked on closed-loop hard real-time control, in the context of the Circa system, for many years while at Honeywell Technology Center. His first contribution to the program was the introduction of dynamic abstraction planning, which replaced the previous forward search of the Circa system with a more sophisticated divide-and-conquer approach featuring context-sensitive abstraction. Dr. Goldman also pioneered the use of timed automata to provide semantics for Circa controllers, and the use of timed automata verifiers in Circa planning. With his colleagues, Dr. Goldman has submitted a patent application concerning incremental verification techniques for controller synthesis. While at Honeywell, Dr. Goldman managed Honeywell contracts to NASA (Thinking Space Systems) and DARPA (Advanced Logistics Program).