AI Paper Atlas

Research map / Large language models / Reasoning and chain of thought

Learning Self-Critiquing Mechanisms for Region-Guided Chest X-Ray Report Generation

Sixing Yan, Ziao Wang, Kejing Yin, William K. Cheung, Ka Chun Cheung, Simon See

ICLR 2026 · Large language models › Reasoning and chain of thought · approach: Process reward models

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a self-critiquing learning mechanism for region-guided chest X-ray report generation so that models ground diagnoses in abnormality regions rather than superficial image-report correlations.

Problem

Automatic radiology report generation from chest X-rays with supervised learning tends to learn superficial statistical correlations between images and reports and lacks grounding of findings in abnormality locations.

Approach

Introduces a self-critiquing mechanism that lets the model reason about and critique the relevant image regions when generating reports, so that abnormality grounding is learned rather than inferred from superficial correlations. Only the abstract was available, so finer details are not stated.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add all three self-critiquing mechanisms (train and test) (ablation)RadSCR · MIMIC-CXRCheXbert F-1: 0.561 -> 0.610Improved
add self-critiquing at inference (ablation)RadSCR · MIMIC-CXRCE-Abn: 0.379 -> 0.572Improved
add alternative-abnormality critique C(Abn) (ablation)RadSCR · MIMIC-CXRCE-Abn: 0.523 -> 0.572Improved
add false-negative critique C(Neg) (ablation)RadSCR · MIMIC-CXRCheXbert F-1: 0.602 -> 0.610Improved
add alternative-image critique C(Img) (ablation)RadSCR · MIMIC-CXRRadGraph-F1 (Complete): 0.371 -> 0.367No clear effect
add LLM decoder (ablation)RadSCR · MIMIC-CXRRadNLI F1: 0.326 -> 0.408Improved
add LLM decoder (ablation)RadSCR · MIMIC-CXRCheXbert F-1: 0.611 -> 0.610No clear effect
add abnormality prototypes (ablation)RadSCR · MIMIC-CXRCE-Abn: 0.377 -> 0.572Improved
add abnormality prototypes (ablation)RadSCR · MIMIC-CXRCE-Organ: 0.751 -> 0.744Mixed
increase sampled alternatives to Np=2RadSCR · MIMIC-CXRclinical metricsImproved
sample only hard alternatives instead of randomRadSCR · MIMIC-CXRCheXbert F-1: 0.610 -> 0.577Hurt
use region-aware RadSCR instead of MLP headSwin · MIMIC-CXR abnormality predictionROC-AUC: 0.756 -> 0.900Improved

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