Mark Boddy's Personal Homepage
My training is in Computer Science, with a specialization in Artificial Intelligence. Over the years, most of my work has been on adapting and extending techniques from a broad variety of areas for application to a range of problems, primarly in planning and scheduling domains. This work has drawn from and in some cases contributed to research in constraint satisfaction, heuristic search, mathematical optimization, classical planning and extensions thereto, temporal reasoning, multi-agent cooperative negotiation, resource-bounded reasoning, and a number of other areas.
In general, I am drawn to problems that involve the application of theoretical concepts to particular classes or structures of problems, rather than to either purely theoretical investigations or the detailed implementation of product-ready applications, though I have done both of those as well.
For a full Curriculum Vitae, including a complete list of publications, click here.
Here is a selection, intermittently updated, of current and recent projects:
- We are currently working on a DARPA Seedling on the topic of learning abstract features.
- SLATE (System-Level Autonomy Trust Enabler) is an ongoing project in compositional verification for complex autonomous systems.
- We had a role on Honeywell's Coordinators team, focussed primarily on the "Change Evaluation" and "Coordination" modules.
- We had a role on a project in NASA's Human and Robotics Technology program, primed by USRA, with the objective of providing robust, reconfigurable autonomous systems for manned space operations. Here is a brief description.
- "Closing the Loop on Net-Centric Defense" was a project in AFRL's CDOT program.
- BAMS, the "Behavioral Adversary Modeling System," is software we developed under ARDA's ACIT program. BAMS applies classical AI planning to vulnerability analysis for cyber defense. Here are the slides for a talk describing the system. This domain was used in the International Planning Competition at ICAPS-09.
- We generated an automation study for the Department of Energy, focussed on identifying opportunities to use automation to achieve energy and cost savings in the DOE's "Industries of the Future" (IOF).
- Under a grant from NASA, we produced a report surveying requirements, methods, and opportunities for using compilation methods to improve the implementation of autonomous systems for NASA missions.
- Under a grant from NASA, we produced a report surveying NASA's requirements, current projects, and strategic directions in planning and scheduling.
I earned a Ph.D. from Brown University in early 1991, and promptly relocated from Rhode Island to Minnesota, where I have lived since. I moved here to take a job in the "Knowledge-Based Systems" section at what was then Honeywell's Systems and Research Center. There, I worked on a variety of research projects, both internally and externally funded, eventually reaching the rank of Research Fellow.
In that capacity, I led teams of up to a dozen engineers in solving technical problems and applying the results to Honeywell products, including the first-ever implementation of an effective scheduling system for an entire petroleum refinery. I was also involved in delivering an avionics processor and communications scheduling tool for the Boeing 777, which required solving a scheduling problem of a scale and complexity far beyond the current state of the art at that time. While at Honeywell, I won the H.W. Sweatt Award (Honeywell's highest engineering honor) in 2000.
In late 2002, I moved to Adventium Labs, which was then just starting up. Adventium Labs promised to be (and has turned out to be) an opportunity to build a new organization free of the constraints of working in a large corporation. We have the freedom to pursue a broader range of intellectual interests, working for a broader range of customers, both government and commercial. In addition, having new colleagues brings new opportunities to collaborate. Among the most fruitful of these interactions has been our investigations of the application of AI techniques to cyber security, for example in the BAMS project, in which we applied classical planning techniques to computer network vulnerability analysis.