“A mechanism is essentially a game,” says Professor Amy Greenwald of Brown CS, “and the mechanism design problem is to design a game such that strategic play leads to a desirable outcome.”
In a recent Keynote address ("Learning Equilibria in Simulation-Based Games ... and the Ensuing Empirical Design of Mechanisms") at the IntelliSys 2020 conference, a conference focusing on areas of intelligent systems and AI as well as applications to the real world, she put forth a methodology to design these mechanisms. It applies under two key conditions:
- The mechanisms induce games that can be simulated, but that do not afford an analytic description, and
- The agents play approximate equilibria in these simulation-based games.
"Under these conditions," Amy says, "we use the probably approximately correct learning framework to construct algorithms that learn equilibria. We show experimentally that our methodology can be used to design effective mechanisms that capture stylized, but rich multiagent systems, such as advertisement exchanges, which are not generally amenable to analytical mechanism design."
Amy's research focuses on game-theoretic and economic interactions among computational agents, applied to areas like autonomous bidding in wireless spectrum auctions and ad exchanges. She recently moderated an AAAI Panel, “Advancing AI By Playing Games”, won second place in The International Automated Negotiation Agents Competition’s (ANAC) Supply Chain Management League, and delivered a keynote address at UEMCOM in 2019.
A link to the keynote address can be found here.
For more information, click the link that follows to contact Brown CS Communications Outreach Specialist Jesse Polhemus.