$\texttt{LucidAtlas}$: Learning Uncertainty-Aware, Covariate-Disentangled, Individualized Atlas Representations

TMLR Paper6127 Authors

Abstract: Interpreting how covariates influence spatially structured biological variation remains a key challenge in developing models suitable for clinical application. We present $\texttt{LucidAtlas}$, a versatile framework for modeling and interpreting spatially varying information with associated covariates. To address the limitations of neural additive models when analyzing dependent covariates, we introduce a marginalization approach that enables accurate explanations of how combinations of covariates shape the learned atlas. $\texttt{LucidAtlas}$ integrates covariate interpretation, spatial representation, individualized prediction, population distribution analysis, and out-of-distribution detection into a single interpretable model. We validate its effectiveness on one synthetic spatiotemporal dataset and two real-world medical datasets. Our findings underscore the critical role of by-construction interpretable models in advancing scientific discovery.

Submission Length: Regular submission (no more than 12 pages of main content)

Previous TMLR Submission Url: https://openreview.net/forum?id=R0UcKQhWw5

Changes Since Last Submission: Our submission was desk-rejected for “Modified template.” We have now checked and corrected the file to conform fully to the official template.

Assigned Action Editor: ~Sameer_Deshpande1

Submission Number: 6127