Hayden McTavish

Research

I work on optimization problems to facilitate interpretable machine learning. My algorithms find accurate machine learning models that succinctly and transparently communicate their entire reasoning process. Recently I have been focused on optimal decision trees, generalized additive models, and prototypical parts models. When optimized well, these model classes are competitive with complex black-box models while providing transparency.

I also work on approximating Rashomon sets. These are sets of all near-optimal models in a model class. They allow for rapid and transparent incorporation of expert input as to which model best fits a domain. Rashomon sets also provide a degree of freedom for addressing additional constraints with negligible loss in accuracy, and allow for robust characterization of properties of all near-optimal models in a model class (such as reliance on a missing value).

My CV can be found here.

Publications

Leveraging Predictive Equivalence in Decision Trees, ICML, 2025. Hayden McTavish*, Zachery Boner*, Jon Donnelly*, Margo Seltzer, Cynthia Rudin.

Near-Optimal Decision Trees in a SPLIT Second, ICML, 2025. Spotlight. Varun Babbar*, Hayden McTavish*, Margo Seltzer, Cynthia Rudin.

Interpretable Generalized Additive Models for Datasets with Missing Values, Neurips, 2024. Hayden McTavish*, Jon Donnelly*, Margo Seltzer, Cynthia Rudin.

Rashomon Sets for Prototypical-Part Networks: Editing Interpretable Models in Real-Time, CVPR, 2025. Jon Donnelly, Zhicheng Guo, Alina Jade Barnett, Hayden McTavish, Chaofan Chen, Cynthia Rudin

Selective Preference Aggregation, ICML, 2025. Shreyas Kadekodi*, Hayden McTavish*, Berk Ustun. Link to Neurips 2024 Workshop version.

Fast Sparse Decision Tree Optimization via Reference Ensembles, AAAI, 2022. (video teaser) (talk) (code) Hayden McTavish*, Chudi Zhong*, Reto Achermann, Ilias Karimalis, Jacques Chen, Cynthia Rudin, Margo Seltzer

Engagement and Anonymity in Online Computer Science Course Forums, ICER, 2023. (slides) Mrinal Sharma*, Hayden McTavish*, Zimo Peng, Anshul Shah, Vardhan Agarwal, Caroline Sih, Emma Hogan, Ismael Villegas Molina, Adalbert Gerald Soosai Raj, Kristen Vaccaro

(*co-first authors)