Generating interpretable context-aware medical concept embedding implicating expert knowledge using LLMs

Bong Gyun Kang

doi:10.1016/j.artmed.2026.103543

Generating interpretable context-aware medical concept embedding implicating expert knowledge using LLMs

2026-10-07

0

PRE

AI

DOI：10.1016/j.artmed.2026.103543

Original

Original request for help

Save

Abstract

En 中文

• Using LLMs by integrating column names and clinical knowledge into each cell value. • ICAME consistently outperforms baselines, even with just a dozen samples. • Offering interpretable embeddings for users to confirm cell values and clinical knowledge. • Protecting data privacy by not sharing patient data online, reducing LLM API costs.

Keywords:

Medical concept embedding

Large language model

Tabular learning

Learning from prior knowledge

Generating interpretable context-aware medical concept embedding implicating expert knowledge using LLMs

Abstract

Journal

Organization

Cited Papers