DCL Lecture Series: Lillian Ratliff "Learning in Stackelberg Games: An Implicit Approach with Provable Guarantees"

Decision and Control Laboratory

Coordinated Science Laboratory

Learning in Stackelberg Games: An Implicit Approach with Provable Guarantees

Dr. Lillian Ratliff

University of Washington

Wednesday, November 6, 2019

3:00pm – 4:00pm

CSL Auditorium (B02)

__________________________________________________________________________________________________________________________________________________

Abstract:

Algorithms are being used to supplement or even replace human decision makers in a variety of applications from intelligent infrastructure to online marketplaces. Drawing on several real-world examples, I will provide some motivation for conducting research at the interface of economics/game theory and machine learning.  Focusing on learning in hierarchical decision problems, I will present new results on an implicit approach to finding equilibria in both the zero-sum and general-sum regimes. The technical results include guarantees on convergence in deterministic and stochastic settings, as well as new analytical tools for understanding the impact of algorithm synthesis on learning trajectories and near-equilibrium behavior. Accompanying the technical results, will be several illustrative examples from adversarial learning and statistics which underscore a dual perspective: game theoretic ideas can be used to expose new insights into the optimization landscape and synthesize robust learning algorithms, and conversely, machine learning tools may provide new technical approaches to solving complex multi-party decision problems emerging in a diverse set of applications. Time permitting, I will conclude the talk by discussing some interesting new application domains we are currently working on with the goal highlighting salient features that require a new set of tools combining economics and decision theory with optimization and machine learning.

Bio: Lillian Ratliff is an Assistant Professor in the Department of Electrical and Computer Engineering at the University of Washington. Prior to joining UW, she also obtained her PhD (2015) in EECS at UC Berkeley. Lillian holds a MS (UNLV 2010) and BS (UNLV 2008) in Electrical Engineering as well as a BS (UNLV 2008) in Mathematics. Her research interests lie at the intersection of game theory, learning, and optimization. She draws on theory from these areas to develop analysis tools for studying algorithmic competition, cooperation and collusion and synthesis tools for designing algorithms with performance guarantees. In addition, she is interested in developing new theoretical models of human decision-making in consideration of behavioral factors in societal-scale systems (e.g., intelligent infrastructure, platform-based markets and e-commerce, etc.) and computational schemes to shape the outcome of competitive interactions. Lillian is the recipient of an NSF Graduate Research Fellowship (2009), NSF CISE Research Initiation Initiative award (2017), and an NSF CAREER award (2019).