Skip to main content

Bio: Amanda Coston is a PhD student in Machine Learning and Public Policy at Carnegie Mellon University. Her research investigates how to make algorithmic decision-making more reliable and more equitable using causal inference and machine learning. She is advised by Alexandra Chouldechova and Edward H. Kennedy. She completed a B.S.E from Princeton where she was advised by Robert Schapire. Prior to her PhD, she worked at Microsoft, the consultancy Teneo, and the Nairobi-based startup HiviSasa. She is a Meta Research PhD Fellow (2022), K & L Gates Presidential Fellow in Ethics and Computational Technologies (2020), and NSF GRFP Fellow (2019).

Talk Title: Validity, equity, and oversight in societally consequential machine learning

Talk Abstract: Machine learning algorithms are widely used for decision-making in societally high-stakes settings from child welfare and criminal justice to healthcare and consumer lending. Recent history has illuminated numerous examples where these algorithms proved unreliable or inequitable. We take a principled approach to the use of machine learning in societally high-stakes settings, guided by three pillars: validity, equity, and oversight. We address data problems that challenge the validity of algorithmic decision support systems by developing methods for algorithmic risk assessments that account for selection bias, confounding, and bandit feedback. We conduct audits for bias throughout the systems in which algorithms are used to inform decision-making. Throughout we propose novel methods that use doubly-robust techniques for bias correction. We apply our methods to the consumer lending context where we find that use of our methods could extend lending opportunities to applicants from historically underbanked regions without an increase in the average rate of default. We present additional findings in the child welfare, criminal justice and healthcare settings.