We are pleased to have Dr. David Pynadath as the invited speaker at MSDM-2014.
Sequential Decision-Making in Human-Agent Interactions
Abstract: Advances in decision-making algorithms have made it possible for agents to interact with people in more complex domains than ever before. However, interacting with a human being raises unique challenges when compared to interacting with another agent. In this talk, I will present some of the features of human-agent interaction that distinguish the sequential decision-making required in such domains from that required in software multiagent interaction. I will illustrate these features using the example of a training system developed by USC ICT for teaching people cross-cultural negotiation through interactions with different agents. Examining these features illuminates opportunities for MSDM researchers to contribute in novel ways to existing domains of human-agent interaction. Beyond the benefit that such agents would derive from algorithmic research, agent *designers* could also benefit from further contributions from the MSDM community. By observing the obstacles that currently prevent many researchers in human-agent interaction from exploiting MSDM algorithms, we can identify additional avenues of investigation that could potentially expand the impact of those algorithms.
Dr. David Pynadath is a research scientist at USC Institute for Creative Technologies. His research interest is in decision-making in multi-agent systems, with a specific focus on how agents model others. He has developed multi-agent systems for applications in social simulation, virtual training environments, automated personal assistants and Unmanned Aerial Vehicle (UAV) coordination and has published papers on social simulation, teamwork, plan recognition, and adjustable autonomy. Dr. Pynadath received his Ph.D. from the University of Michigan, Ann Arbor in 1999.