AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Expert-guided Clinical Text Augmentation via Query-Based Model Collaboration

Dongkyu Cho, Miao Zhang, Gregory D Lyng, Rumi Chunara

ICML 2026 · Large language models › Retrieval and knowledge · approach: Domain-specific QA datasets

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add weak-expert (biomedical NER) guidance to LLM augmentationQwen-3 (Naive augmentation) · MIMIC-III readmissionaccuracy: 0.520 -> 0.599Improved
add weak-expert guidance to LLM augmentationBERT (Naive augmentation) · MIMIC-III readmissionaccuracy: 0.736 -> 0.757Improved
add weak-expert guidance to LLM augmentationQwen-3 (Naive augmentation) · MIMIC-III length-of-stayRMSE: 16.357 -> 15.563Improved
apply naive LLM augmentation instead of noneQwen-3 · MIMIC-III readmissionaccuracy: 0.526 -> 0.520Hurt
apply CATO augmentation instead of noneQwen-3 · MIMIC-III length-of-stayRMSE: 17.835 -> 18.677Hurt
add weak-expert guidance to naive augmentationLlama-3.2-3B generalist · synthetic notes token levelPR: 0.51 -> 0.79Improved
add weak-expert guidance to naive augmentationLlama-3.2-3B generalist · synthetic notes token levelHR: 0.59 -> 0.33Improved
use biomedical-ner-all weak expert rather than general-expert (ablation)augmentation pipeline · synthetic notes token levelPR: 0.53 -> 0.79Improved
use general-expert NER instead of no expert (ablation)augmentation pipeline · synthetic notes token levelHR: 0.59 -> 0.50Improved
use a larger strong generalist (Llama-3.2-3B instead of 1B) (ablation)augmentation pipeline · synthetic notesPR: 0.66 -> 0.79Improved
replace expert collaboration with DPO-trained single Strong Expert (ablation)Llama-3.2-3B · MIMIC-III readmissionaccuracy: 0.583 -> 0.560Hurt
replace expert collaboration with DPO-trained single Strong Expert (ablation)Qwen-3 · MIMIC-III readmissionaccuracy: 0.599 -> 0.582Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy