- Published papers at NeurIPS 2022, KDD 2017, NeurIPS 2019, ICML 2020, AAAI 2022.
- PaCMAP algorithm won two software awards from the American Statistical Association.
- Work on optimal scoring systems won the 2019 INFORMS Innovative Applications in Analytics Award.
- Maintenance of an underground electrical distribution network won the 2013 INFORMS Innovative Applications in Analytics Award.
- Solved a well-known theoretical problem in machine learning and earned a prize for solving a COLT open problem.
- Delivered invited and keynote talks at INFORMS, KDD, SDM, AISTATS, ECML-PKDD, and other venues.
- Teams have won awards in several data science competitions, including the ASA Data Challenge Expo in 2022.
Research Experience
- Developed practical code for sparse models such as decision lists, decision trees, and additive models.
- Introduced the Rashomon Set paradigm, allowing users to choose among many good models.
- Developed theory for why simpler models often perform well.
- PaCMAP algorithm is widely used in bioinformatics, biology, and ecology.
- Optimal scoring systems (sparse linear models with integer coefficients) applied to healthcare and criminal justice.
- Led a team using machine learning to maintain an underground electrical distribution network.
- Developed methods for detecting crime series in cities.
Background
Gilbert, Louis, and Edward Lehrman Distinguished Professor of Computer Science; Departments of Computer Science, Electrical and Computer Engineering, Statistical Science, Mathematics, and Biostatistics & Bioinformatics at Duke University; PI, Interpretable Machine Learning Lab; Research focuses on interpretable machine learning and its applications.
Miscellany
Enjoys competing in data science competitions and coaching student teams.