• View
• Edit
• History
• Print

Human-Machine Collaborative Optimization Via Apprenticeship Scheduling

Abstract:

Health care, manufacturing, and military operations require the careful choreography of resources - people, robots, and machinery - to effectively fulfill the responsibilities of the profession. Poor resource utilization has been shown to have drastic health, safety, and economic consequences. However, coordinating a heterogeneous team of agents to complete a set of tasks related through upper- and lower-bound temporal constraints is NP-Hard. Further, the process of modeling the multi-faceted aspects of real-world scenarios is labor-intensive and leaves much to be desired. Yet, there is hope. In practice, we know there are a rare breed of human experts who effectively reason about complex resource optimization problems every day. The question then becomes: How can we autonomously learn the rules-of-thumb and heuristics from these domain experts to support real-time decision-support and autonomously coordinate human-robot teams? In this talk, I will present a novel computational technique, known as Apprenticeship Scheduling, which 1) learns the heuristics and implicit rules-of-thumb developed by domain experts from years of experience, 2) embeds and leverages this knowledge within a scalable resource optimization framework, and 3) provides decision support in a way that engages users and benefits them in their decision-making process. By intelligently leveraging the ability of humans to learn heuristics and the speed of modern computation, we can improve the ability to coordinate resources in these time- and safety-critical domains.

Bio:

Matthew Gombolay is a PhD candidate and NSF Graduate Research Fellow in the Interactive Robotics Group at the Massachusetts Institute of Technology (MIT). He received his S.M. (2013) from the department of Aeronautics and Astronautics at MIT and his B.S (2011) from the department of Mechanical Engineering at Johns Hopkins University. Matthew studies the interaction of humans and automation and is developing computational methods for real-time and collaborative resource optimization. Matthew focuses on harnessing the strengths of human domain experts and sophisticated computational techniques to form collaborative human-machine teams for manufacturing, healthcare, and military operations. Matthew has worked for MIT Lincoln Laboratory and the Johns Hopkins University Applied Physics Laboratory developing cutting-edge planning and scheduling algorithms for ballistic and anti-ship missile defense with the US Navy and Missile Defense Agency. Matthew has received a Best Technical Paper Award from the AIAA Intelligent Systems Committee, and his work has been highlighted in media outlets such as PBS, NBC, Harvard Business Review, and public radio.