aiqbench analysis
This research demonstrates a novel pipeline that extracts interpretable decision rules from the latent space of transformer models and translates them back into clinically measurable features. By leveraging gradient-based attribution, the method successfully uncovers complex variable interactions that raw-feature rules miss, though it highlights inherent limitations when high-performing latent patterns cannot be perfectly mapped to simple clinical conditions. The findings offer a practical pathway for deploying transparent AI in healthcare without sacrificing predictive accuracy derived from deep learning architectures.
From Latent Biomarkers to Clinical Rules: Embedding-Guided Rule Mining and Attribution-Based Translation for Interpretable Tabular Learning
arXiv:2609.22155v1 Announce Type: new Abstract: Clinical decision support tools are most useful when accurate predictions are accompanied by understandable explanations. Rule-based models provide transparency, but rules derived directly from raw clinical measurements may miss patterns arising from interactions between multiple variables. We present a four-step pipeline that mines decision rules in the latent space of an FT-Transformer and translates them back into measurable clinical features. Embedding dimensions that consistently separate patient groups are treated as latent biomarkers, rules are mined using small decision trees, and selected rules are translated using gradient-input saliency and CLS attention attribution. We evaluate the framework on six public clinical and population health datasets at four embedding dimensions. Translated rules outperformed raw-feature rules in five of six datasets, with mean AUROC gains ranging from 0.04 to 0.23. On the heart disease dataset, embedding-space rules reached 0.98 AUROC, but translation reduced this to 0.72, showing that high-performing latent rules cannot always be represented by simple raw-feature conditions. These results show that latent-space rule discovery can uncover predictive patterns while translating them into clinically measurable features that can be evaluated by clinicians.
Abstract:Clinical decision support tools are most useful when accurate predictions are accompanied by understandable explanations. Rule-based models provide transparency, but rules derived directly from raw clinical measurements may miss patterns arising from interactions between multiple variables. We present a four-step pipeline that mines decision rules in the latent space of an FT-Transformer and translates them back into measurable clinical features. Embedding dimensions that consistently separate patient groups are treated as latent biomarkers, rules are mined using small decision trees, and selected rules are translated using gradient-input saliency and CLS attention attribution. We evaluate the framework on six public clinical and population health datasets at four embedding dimensions. Translated rules outperformed raw-feature rules in five of six datasets, with mean AUROC gains ranging from 0.04 to 0.23. On the heart disease dataset, embedding-space rules reached 0.98 AUROC, but translation reduced this to 0.72, showing that high-performing latent rules cannot always be represented by simple raw-feature conditions. These results show that latent-space rule discovery can uncover predictive patterns while translating them into clinically measurable features that can be evaluated by clinicians.
Machine Learning (cs.LG)
arXiv:2609.22155 [cs.LG]
(or arXiv:2609.22155v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2609.22155 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 271 words
Related coverage
Bayesian Networks with Latent Time Embedding for Stage-Aware Causal Modeling of Alzheimer's Disease Progression
arXiv · Jun 16, 2026
Tech to Replace Animal Testing Is Almost Ready. Scientists Are Not
IEEE Spectrum · Sep 29, 2026
CoPoE: Multimodal Fusion via Decomposable Disease-Coordinate Product-of-Experts for Missing-Modality Alzheimer's Diagnosis
arXiv · Oct 9, 2026
CARDINAL Predicts Cardiovascular Risk From Non-contrast Cardiac CT
arXiv · Sep 1, 2026