CSE Colloquia Series-Phebe Vayanos

Nov 22, 2019
11:00 a.m.
Lopata Hall, Room 101

Robust Active Preference Elicitation to Learn the Priorities of Policy-Makers at the Los Angeles Homeless Services Authority


Motivated by our collaboration with the Los Angeles Homeless Services Authority, the authority in charge of allocating housing resources and services to those experiencing homeless in LA, we consider the problem faced by a recommender system which seeks to offer a user with unknown preferences their favorite item among a potentially infinite collection. Before making a recommendation, the system has the opportunity to elicit the user's preferences by making a moderate number of queries. We take the point of view of a risk-averse recommendation system which only possesses limited, set-based, information on the user utility function and investigate two complementary settings. In the first setting, each query corresponds to a pairwise comparison, in the spirit of choice-based conjoint analysis. In the second setting, each query asks the user to rate an item on a scale from 0 to 1. We show that these problems can be formulated as multi-stage robust optimization problems with decision-dependent information discovery and propose reformulations in the form of mixed-binary linear problems that, combined with decomposition techniques, can be solved efficiently with off-the-shelf solvers. We evaluate the performance of our approaches on both synthetic and real-world data from the Homeless Management Information System where we learn the preferences of policy-makers in terms of characteristics (fairness-efficiency-interpretability trade-offs) of a policy for allocating housing resources. Our results illustrate that our framework outperforms state-of-the-art techniques from the literature.

This work is based on two papers, one joint with Duncan McElfresh, Yingxiao Ye, John Dickerson, and Eric Rice and the other with Angelos Georghiou and Han Yu.


Phebe Vayanos is an Assistant Professor of Industrial & Systems Engineering and Computer Science at the University of Southern California. She is also an Associate Director of the CAIS Center for Artificial Intelligence in Society at USC. Her research aims to address fundamental questions arising in data-driven optimization (a.k.a. prescriptive analytics) with aim to tackle real-world decision- and policy-making problems in uncertain and adversarial environments. Her work is motivated by resource allocation problems that are important for social good, such as those arising in public health, public safety and security, public housing, biodiversity preservation, and education. She is also interested in issues surrounding fairness, efficiency, and interpretability in resource allocation. Prior to joining USC, she was lecturer in the Operations Research and Statistics Group at the MIT Sloan School of Management, and a postdoctoral research associate in the Operations Research Center at MIT. She holds a PhD degree in Operations Research and an MEng degree in Electrical & Electronic Engineering, both from Imperial College London.