Michael Cooper
Ph.D. Student, Computer Science
University of Toronto
Welcome to my website! I'm Michael - I'm a Ph.D. student in Computer Science at the University of Toronto, advised by Rahul G. Krishnan, and Michael Brudno. My research focuses on designing clinical machine learning systems for implementation in national-scale, high-stakes decision-making settings like liver transplant prioritization. In this same vein, I also design and study algorithms to make modern machine learning methods more reliable and interpretable.
Previously, I earned B.S. and M.S. degrees in Computer Science from Stanford University, where I had the privilege of investigating the influence of indoor built space design on human wellbeing with James Landay and Sarah Billington, building an augmented reality application to align patients and medical images in 3D space with Bruce Daniel, and constructing a computer vision dataset comprising complex multi-object multi-actor scenes with Ehsan Adeli and Fei-Fei Li.
In my free time, I enjoy scuba diving, long-distance running, alpine skiing, cheering on the Vancouver Canucks, and reading science fiction and historical non-fiction (a few favorites: Liu Cixin's Remembrance of Earth's Past Trilogy, Kim Stanley Robinson's Mars Trilogy, and Ben Rich and Leo James' Skunk Works).
( June 2025) Our preprint, The Curious Language Model: Strategic Test-Time Information Acquisition, presents a simple test-time policy for language models to acquire novel information from external sources to solve prediction tasks. Grateful to the folks at Abridge for the wonderful internship experience!
( June 2025 ) Honoured to be named a Swartz-Reisman Institute for Technology and Society graduate fellow! The support of this fellowship and community will significantly advance my research on equitable liver transplant prioritization.
( May 2025 ) Our paper, Red Teaming Large Language Models for Healthcare, summarizes the design process and findings of our 2024 workshop at MLHC. Our identified vulnerabilities—hallucination, syncophany, misprioritization, and more—motivate future study and cautious deployment of language models in high-stakes healthcare settings.
- M. Cooper, X. Gao, X. Zhao, D. Khoroshchuk, et al. “DynaMELD: A Dynamic Model of End-Stage Liver Disease for Equitable Prioritization,” medRxiv preprint, 2024.
- A. Gharari*, M. Cooper*, R. Greiner, RG. Krishnan. “Copula-based Deep Survival Models for Dependent Censoring,” Uncertainty in Artificial Intelligence (UAI), 2023.
- M. Cooper, R. Wadhawan, J. Giorgi, C. Tan, D. Liang. “The Curious Language Model: Strategic Test-Time Information Acquisition,” arXiv preprint, 2025.