Home → News and Events → Events Calendar → Seminar
Effective Computer Agents for Interacting With People
Event Type: Seminar Date: March 07, 2008 Time: 10:00AM - 11:30AM Venue: CC-2-2540 Abstract: Technological advancements have created opportunities for computers to interact with people in diverse settings, whether as autonomous actors or as proxies for individual people or organizations. The range and complexity of human behavior raises particular challenges to the design of successful computer agents for these mixed human-computer settings. This talk presents work that addresses these challenges by synthesizing techniques from computer science with insights from the behavioral and social sciences. I will present a language that makes explicit the different mental models that agents use to make their decisions, described as nodes in a graphical network. The language defines an equilibrium that makes a distinction between agents' optimal strategies and the way they actually behave in reality. I will show how this language facilities the design of computer agents that represent and learn the social factors that affect people's negotiation behavior in settings of varying complexity. In empirical investigations, agents using this language to negotiate with other people were able to outperform traditional game-theoretic equilibria. Finally, I will show that for best performance, computers participating in mixed human-computer settings must model human behavior in a way that reflects the contextual setting in which the decision is presented to people. Speaker: Yakov Gal Speaker Bio: Yakov GalComputer Science and Artificial Intelligence Laboratory
Massachusetts Institute of Technology School of Engineering and Applied Sciences
Harvard University
