aiqbench analysis
This study addresses critical bottlenecks in deploying tabular in-context learning for clinical risk prediction, demonstrating that naive retrieval methods degrade under real-world data heterogeneity. The proposed AWARE framework significantly improves performance by aligning retrieval tasks with inference goals through supervised embedding learning. These findings offer essential guidance for researchers aiming to scale AI models within the complex constraints of electronic health records.
Retrieval-aligned Tabular Foundation Models Enable Robust Clinical Risk Prediction in Electronic Health Records Under Real-world Constraints
arXiv:2604.01841v3 Announce Type: replace Abstract: Clinical prediction from structured electronic health records (EHRs) is challenging due to high dimensionality, heterogeneity, class imbalance, and distribution shift. While tabular in-context learning (TICL) and retrieval-augmented methods perform well on generic benchmarks, their behavior in clinical settings remains unclear. We present a multi-cohort EHR benchmark comparing classical, deep tabular, and TICL models across varying data scale, feature dimensionality, outcome rarity, and cross-cohort generalization. PFN-based TICL models are sample-efficient in low-data regimes but degrade under naive distance-based retrieval as heterogeneity and imbalance increase. We propose AWARE, a task-aligned retrieval framework using supervised embedding learning and lightweight adapters. AWARE improves AUPRC by up to 12.2% under extreme imbalance, with gains increasing with data complexity. Our results identify retrieval quality and retrieval-inference alignment as key bottlenecks for deploying tabular in-context learning in clinical prediction.
Authors:Minh-Khoi Pham, Thang-Long Nguyen Ho, Thao Thi Phuong Dao, Tai Tan Mai, Minh-Triet Tran, Marie E. Ward, Una Geary, Rob Brennan, Nick McDonald, Martin Crane, Marija Bezbradica
Abstract:Clinical prediction from structured electronic health records (EHRs) is challenging due to high dimensionality, heterogeneity, class imbalance, and distribution shift. While tabular in-context learning (TICL) and retrieval-augmented methods perform well on generic benchmarks, their behavior in clinical settings remains unclear. We present a multi-cohort EHR benchmark comparing classical, deep tabular, and TICL models across varying data scale, feature dimensionality, outcome rarity, and cross-cohort generalization. PFN-based TICL models are sample-efficient in low-data regimes but degrade under naive distance-based retrieval as heterogeneity and imbalance increase. We propose AWARE, a task-aligned retrieval framework using supervised embedding learning and lightweight adapters. AWARE improves AUPRC by up to 12.2% under extreme imbalance, with gains increasing with data complexity. Our results identify retrieval quality and retrieval-inference alignment as key bottlenecks for deploying tabular in-context learning in clinical prediction.
Not peer-reviewed. Under revision
Artificial Intelligence (cs.AI)
arXiv:2604.01841 [cs.AI]
(or arXiv:2604.01841v3 [cs.AI] for this version)
https://doi.org/10.21203/rs.3.rs-9085469/v1 DOI(s) linking to related resources
Full text under the publisher's open licence · 180 words
Related coverage
Amazon is investing in the Lean Focused Research Organization
Amazon Science · Jul 26, 2026
34 Amazon Research Awards Build on Trainium recipients announced
Amazon Science · Aug 5, 2026
Amazon Research Awards recipients announced
Amazon Science · May 27, 2026
MMAP: Multimodal Missing-Aware Pretraining for Longitudinal Alzheimer's Prediction
arXiv · Sep 23, 2026