What's my teammate going to do? Reasoning about probable plans for successful teamwork

Опубликовано 7 сентября 2016, 18:00

As people are increasingly connected to other people and computer agents, forming mixed networks, collaborative teamwork offers great promise for transforming the way people perform their everyday activities and interact with computer agents. This talk presents new representations and algorithms, developed to enable computer systems to function as effective team members in settings characterized by uncertainty and partial information. For a collaboration to succeed in such settings, participants need to reason about the possible plans of others, adapt their plans accordingly, and support each other when needed. The probabilistic representation of agentsΓÇÖ beliefs about the probable plans of others presented in this talk, enables efficient decision-theoretic reasoning on general teamwork models. The talk will describe empirical studies that demonstrate the usefulness of this representation in designing agent strategies for deciding when and how to help teammates. Other studies presented in this talk, show that to collaborate successfully, computer agents need to consider how people themselves perceive the joint utility of helpful behavior to the group, and make decisions accordingly. The talk will conclude with a real-world application where collaborative plans deliver value for users and the environment.