Reliable Decision-Support using Counterfactual Models

Abstract

In real-time environments, decision-makers are faced with the challenge of quickly integrating high-dimensional data to inform their actions. For instance, physicians in a hospital's intensive care unit must process long histories of vital signals, laboratory test results, and treatments in order to decide whether a patient is at risk and what future interventions might be necessary to prevent undesirable outcomes. To aid in this process, machine learning practitioners commonly use supervised learning algorithms to fit models that predict outcomes given the high-dimensional history, but this approach can produce biased models that may lead to dangerous downstream decisions. The key issue is that supervised learning algorithms are highly sensitive to the policy used to choose actions in the training data, which causes the model to capture relationships that do not generalize. We propose using a different learning objective that predicts counterfactuals instead of predicting outcomes under an existing action policy as in supervised learning. To support decision-making in temporal settings, we introduce the Counterfactual Gaussian Process (CGP) to predict the counterfactual future progression of continuous-time trajectories under sequences of future actions.

Bio

Peter Schulam is a PhD candidate in the Computer Science Department at Johns Hopkins University where he is working with Professor Suchi Saria. His research interests lie at the intersection of machine learning, statistical inference, and healthcare with an emphasis on developing methods to support the personalized medicine initiative. Before coming to JHU, he obtained his MS from Carnegie Mellon’s School of Computer Science and his BA from Princeton University. He has been awarded a National Science Foundation Graduate Research Fellowship and the Dean's Centennial Fellowship within Johns Hopkins' Whiting School of Engineering.