Bidirectional Communication for Effective Human-Agent Teaming

Abstract

The recent proliferation of artificial intelligence research is reaching a point where machines are able to learn and adapt to dynamically make decisions independently or in collaboration with human team members. With such technological advancements on the horizon, there will come a mandate to develop techniques to deploy effective human-agent teams. One key challenge to the development of effective teaming has been enabling a shared, dynamic understanding of mission space, and a basic knowledge about the states and intents other teammates. Bidirectional communication is an approach that fosters communication between human and intelligent agents to improve mutual understanding and enable effective task coordination. This session focuses on current research and scientific gaps in three areas necessary to advance the field of bidirectional communication between human and intelligent agent team members. First, intelligent agents must be capable of understanding the state and intent of the human team member. Second, human team members must be capable of understanding the capabilities and intent of the intelligent agent. Finally, in order for the entire system to work, systems must effectively integrate information from and coordinate behaviors across all team members. The combination of these three areas will enable future human-agent teams to develop a shared understanding of the environment as well as a mutual understanding of each other, thereby enabling truly collaborative human-agent teams.