HMI DAIS 15 - Public online seminar, 9am 10 June 2021 AEST
Angela Zhou is a graduating PhD from Cornell University/Cornell Tech in Operations Research and Information Engineering. She works at the intersection of statistical machine learning and operations research in order to inform reliable data-driven decision-making in view of fundamental practical challenges that arise from realistic information environments. In particular, her research has focused on robust causal inference for decision-making, and credible performance evaluation for algorithmic fairness and disparity assessment.
Seminar Title: Credible Evaluation for Algorithmic Fairness
Abstract: Credible performance evaluation is necessary in algorithmic fairness in order to align the performance benchmarks, by which we measure machine learning progress and ultimately justify the deployment of algorithms, with the real-world actual operating conditions and impacts of algorithms in consequential settings. In particular, these settings generate specific and common challenges.
I will present results for credible evaluation when the protected class is not observed, e.g. Assessing Algorithmic Fairness with Unobserved Protected Class Using Data Combination. In this paper we study a fundamental challenge to assessing disparate impacts in practice: protected class membership is often not observed in the data. This is particularly a problem in lending and healthcare. We consider the use of an auxiliary dataset, such as the US census, to construct models that predict the protected class from proxy variables, such as surname and geolocation. We provide exact characterizations of the tightest-possible set of all possible true disparities that are consistent with the data (and possibly any assumptions). We further provide optimization-based algorithms for computing and visualizing these sets and statistical tools to assess sampling uncertainty. Together, these enable reliable and robust assessments of disparities – an important tool when disparity assessment can have far-reaching policy implications. We demonstrate this in two case studies with real data: mortgage lending and personalized medicine dosing.
This is joint work with Xiaojie Mao and Nathan Kallus.