Please note: This PhD seminar will take place online.
David
Radke,
PhD
candidate
David
R.
Cheriton
School
of
Computer
Science
Supervisors: Professors Kate Larson, Tim Brecht
For problems requiring cooperation, many multiagent systems implement solutions among either individual agents or across an entire population towards a common goal. Multiagent teams are primarily studied when in conflict; however, organizational psychology (OP) highlights the benefits of teams among human populations for learning how to coordinate and cooperate.
In this talk, I will present a model of multiagent teams for reinforcement learning (RL) agents inspired by OP and early work on teams in artificial intelligence. I will then relax assumptions about common interest among teammates when presenting credo, a model of how agents optimize their behavior for various groups they belong to. Our empirical results show how some amount of mixed incentives can increase global reward beyond a fully cooperative population while maintaining high equality and supporting emergent role specialization.
To join this PhD seminar on Zoom, please go to https://uwaterloo.zoom.us/j/93175380339.