Papers by Yunseok Han

GenGO

Omakase Read Later

Search

Papers by Yunseok Han

Findings

2026.findings-acl

stage-wise diagnostic probing
accuracy
disease evolution
VQA
generation
Temporal Event Localization
Interval-wise Change Reasoning
Global Trajectory Summarization
vision–language models
longitudinal medical reasoning

2026

Multimodality

1 papers

MI-CXR: A Benchmark for Longitudinal Reasoning over Multi-Interval Chest X-rays (2026.findings-acl)

Copied to clipboard

Sunghwan Steve Cho, Yunseok Han, Jaeyoung Do

Challenge:	Existing medical VQA benchmarks focus on single images or short-horizon image pairs.
Approach:	They propose a benchmark for standardized evaluation of longitudinal reasoning over multi-visit sequences.
Outcome:	The proposed benchmark shows low overall performance (29.3% accuracy) and is only modestly above random guessing.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations