AI Seminar

Machine Learning and Causality: Building Efficient, Reliable Models for Decision-Making

Maggie MakarPresidential Postdoctoral FellowU-M CSE
3725 Beyster BuildingMap

Location: BBB 3725 and Zoom (Zoom link; Password if needed: MichiganAI)

Abstract: Increasingly, practitioners are turning to ML to build causal models, and predictive models that perform well under distribution shifts. However, current techniques for causal inference typically rely on having access to large amounts of data, limiting their applicability to data-constrained settings. In addition, empirical evidence has shown that most predictive models are insufficiently robust with respect to shifts at test time. This talk presents novel techniques addressing both of these problems.

Much of the causal literature focuses on learning accurate individual treatment effects, which can be complex and hard to estimate from small samples. However, it is often sufficient for the decision maker to have estimates of upper and lower bounds on the potential outcomes of decision alternatives to assess risks and benefits. I will show that in such cases we can improve sample efficiency by estimating simple functions that bound these outcomes instead of estimating their conditional expectations. I will present a novel algorithm that leverages these theoretical insights.

I will also present an approach to deal with distribution shifts. I will discuss how to use causal knowledge and auxiliary data to design regularization schemes that encourage robust prediction. I will present a causally-motivated regularization scheme that enables prediction of the target label with high accuracy even if the training data is collected in biased settings.

About the speaker: Maggie Makar’s research interests lie at the intersection of machine learning and causal inference. Her work leverages causal ideas to make ML models robust to distributional shifts, and utilizes ideas from machine learning to make causal inference more statistically efficient. She focuses on developing ML and causal models to guide decision making, particularly in the field of healthcare. Her work has appeared in ICML, AAAI, JSM, JAMA, Health Affairs, and Epidemiology.



AI Lab