Understanding Structured Health Data through Interaction-Aware Mixture-of-Experts
Machine Learning
2026-07-14 v1
Abstract
We study interaction-aware mixture-of-experts for post-stroke rigidity prediction using multi-level views of structured health records. Despite minimal performance gains, routing attribution reveals systematic importance differences across views, underscoring view construction as key to interpretability.
Cite
@article{arxiv.2607.12255,
title = {Understanding Structured Health Data through Interaction-Aware Mixture-of-Experts},
author = {Ji Hwan Park and Ying Ding and Tianjin Guo},
journal= {arXiv preprint arXiv:2607.12255},
year = {2026}
}