CHDS Seminar with Michael Lingzhi Li: “The Cram Method for Efficient Simultaneous Learning and Evaluation”
Abstract:
We introduce the ‘cram’ method, a general, rigorous, and efficient approach to simultaneously developing and evaluating an individualized treatment rule (ITR). In a single pass of data, the proposed method repeatedly trains an machine learning algorithm and tests its empirical performance. Because it utilizes the entire sample for both learning and evaluation, cramming is significantly more data-efficient than traditional sample-splitting and reduces the evaluation standard error by more than 40% when compared to sample-splitting, while improving the performance of learned policy.
Bio:
Michael Lingzhi Li is an Assistant Professor in the Technology and Operations Management unit at Harvard Business School. His research focuses on the end-to-end development of decision algorithms based on machine learning, causal inference and operations research. He also examines the implementation of such algorithms in hospitals, pharmaceutical companies, and public health organizations. He is the recipient of awards including the INFORMS Edelman Finalist, the INFORMS Pierskalla Award, and the Innovative Applications in Analytics Award.