Ann Lee
Also known as Ann B. Lee
Ann Lee develops statistical and machine learning methods for science, including predictive distributions and uncertainty estimates. She completed a physics PhD at Brown University in 2002 and joined Carnegie Mellon University in 2005 after research at Brown and a mathematics position at Yale. Her coauthored papers introduced FlexCode in 2017 for estimating predictive distributions and WALDO in 2023 for constructing confidence regions. She helped found STAMPS, which became a CMU research center in 2024, and joined the editorial leadership of PRX Intelligence in 2026. As of October 2026, she is a professor in CMU's Department of Statistics & Data Science with a joint appointment in the Machine Learning Department.
- Works
- 53
- Citations
- 2,924
- h-index
- 16
- Badges
Publication numbers from OpenAlex, as of 27 October 2026.
Timeline
-
2002
Physics studies at Brown
She completes a physics PhD at Brown University in 2002, studying with David Mumford. Her education also includes MSc/BSc degrees in engineering physics at Chalmers University of Technology in Sweden. She serves as a visiting research associate in Brown's Department of Applied Mathematics before moving to Yale.
-
2005
Joining Carnegie Mellon
In 2005, she joins Carnegie Mellon University. Before CMU, she serves as the J.W. Gibbs Assistant Professor in Yale University's Department of Mathematics.
-
2006
Diffusion maps and coarse-graining
With Stéphane Lafon, she coauthors "Diffusion Maps and Coarse-Graining" in 2006 in IEEE Transactions on Pattern Analysis and Machine Intelligence. The paper addresses dimensionality reduction, graph partitioning and data set parameterization.
-
2011
The McCandless Professorship
In 2011, CMU awards her the Estella Loomis McCandless Professorship while she is an associate professor of statistics in its Dietrich College of Humanities and Social Sciences.
-
2017
FlexCode for predictive distributions
With Rafael Izbicki, she introduces FlexCode in 2017, a method for estimating conditional probability distributions from data. It uses regression to estimate the coefficients of a mathematical series and can handle complex, high-dimensional data.
-
2023
Confidence regions with WALDO
In 2023, she coauthors the WALDO paper with Luca Masserano, Tommaso Dorigo, Rafael Izbicki and Mikael Kuusela at the International Conference on Artificial Intelligence and Statistics. The method constructs confidence regions with statistical coverage guarantees using prediction algorithms or estimates of posterior distributions.
-
2024
The STAMPS research center
In 2024, STAMPS becomes the STAtistical Methods for the Physical Sciences Research Center at CMU. Lee and Mikael Kuusela are its founders and co-directors; the center develops statistical methods for fundamental physics, environmental science and climate science.
-
2026
Deputy editor of PRX Intelligence
In 2026, she joins the editorial leadership of the American Physical Society's PRX Intelligence as a deputy editor. The journal covers AI and machine learning for the physical sciences and related fields.
-
As of October 2026
Professor at Carnegie Mellon
As of October 2026, she is a professor in CMU's Department of Statistics & Data Science with a joint appointment in the Machine Learning Department. Her work includes neural simulation-based inference, calibrated predictive inference and probabilistic forecasting.
Career
- NowCarnegie Mellon UniversityProfessor, Co-director, STAMPS Research Center
- 2000Brown UniversityVisiting research associate
- Yale UniversityJ.W. Gibbs Assistant Professor
Education
- 2002Brown UniversityPhD in physics
- Chalmers University of TechnologyMSc in engineering physics
- Chalmers University of TechnologyBSc in engineering physics
Sources and credits
Photo: We have no published photo of Ann Lee yet. Send us a free-licensed photo.
- Ann B Lee
- Vision Introductory Comments
- Department of Statistics and Data Science < Carnegie Mellon University
- PRX Intelligence - Editors of PRX Intelligence
- Diffusion Maps and Coarse-Graining: A Unified Framework for Dimensionality Reduction, Graph Partitioning, and Data Set Parameterization
- McCandless Chairs Awarded to Two Young and Dynamic Researchers - The Piper - Carnegie Mellon University
- [1704.08095] Converting High-Dimensional Regression to High-Dimensional Conditional Density Estimation
- Simulator-Based Inference with WALDO: Confidence Regions by Leveraging Prediction Algorithms and Posterior Estimators for Inverse Problems
- About STAMPS - STAMPS@CMU - Carnegie Mellon University
- American Physical Society to launch new open access journal on AI and machine learning in scientific research


