Hidden-RAD2

Introduction

Welcome to the Hidden-RAD2 challenge, a Core Task of NTCIR-19!

Building on the NTCIR-18 Hidden-RAD task, Hidden-RAD2 investigates whether AI systems can predict structured radiological reasoning from chest X-ray images and verify the factual and logical correctness of generated medical explanations. 

The challenge consists of two tasks:

- Task 1 — Structured Radiological Reasoning Prediction: Predict radiologist-validated A1–A5 structured reasoning annotations from a chest X-ray image.

- Task 2 — Causal Verification and Correction: Detect and correct factual or logical errors in AI-generated medical explanations.

Hidden-RAD2 aims to support the development of safer, more transparent, and more trustworthy medical AI.

Contact us: hidden-rad2@googlegroups.com