This project will develop interpretable machine-learning models that combine programmatic features with simple predictors for scientific modeling tasks. The project will extend the LeaPR framework to climate science, materials science and neuroscience, with the goal of automatically discovering human-inspectable features that can support prediction, hypothesis generation and scientific understanding.
2026
Learning Interpretable Models of Scientific Data
Learned Programmatic Representation models combine programmatic features, synthesized by LLMs as code, and decision tree predictors, yielding interpretable models.
Other Researchers
Gabriel Poesia Reis e Silva, Assistant Professor, Computer Science and Engineering