One of the most fundamental challenges of building a human-multiagent team is adjustable autonomy, a process in which the autonomy over decisions is dynamically transferred between humans and agents. This talk will focus on adjustable autonomy in the context of a human interacting with a team of agents and focuses on four issues that arise when addressing this team-level adjustable autonomy problem in real-time uncertain domains.
My approach makes four contributions to the field in order to address these challenges. First, I have included, in the adjustable autonomy framework, the modeling the duration of the resolution of inconsistencies between the human and agent view. Second, in order to address the challenges brought about by dealing with time, I have modeled these new adjustable autonomy strategies using continuous time planning. Third, I have introduced a new model for Interruptible Time dependent Markov Decision Problems (ITMDPs) that allows for an action to be interrupted at any point in continuous time. This results in a more accurate modeling of actions and produces additional time-dependent policies that guide interruption during the execution of an action. Fourth, I have created a hybrid approach that decomposes the team level adjustable autonomy problem in a separate ITMDP for each team decision.
Dr Nathan Schurr
Nathan Schurr has just defended his Ph.D. thesis in Artificial Intelligence at the University of Southern California and will be joining Aptima in January. His interests are in allowing humans and artificially intelligent entities to collaborate and interact. He has worked on projects ranging in domains from software personal assistants to human-multirobot teams. He has developed an Incident Command training system, DEFACTO, which allows human incident commanders to coordinate with fire fighter agents during a large-scale disaster response.