publications

2026

1. TMLR

   

   Is the Geometry Doing the Work? An Operating-Point Audit of Hierarchy in Hyperbolic Vision-Language Models

   Jaeyoung Kim, Eunseok Kim, and Dongsuk Jang

   Transactions on Machine Learning Research, 2026

   Abs PDF

   Auditing whether trained hyperbolic vision-language models (MERU, HyCoCLIP, PHyCLIP) actually use hyperbolic geometry and entailment cones to encode concept hierarchy.

2025

1. EMNLP

   

   MedTutor: A Retrieval-Augmented LLM System for Case-Based Medical Education

   Dongsuk Jang, Ziyao Shangguan, Kyle Tegtmeyer, and 4 more authors

   In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, 2025

   Abs PDF Video Code

   A novel system designed to augment resident training by automatically generating evidence-based educational content from clinical case reports.

2. NAACL

   

   Multi-Perspective Integration via Mixture-of-Agents for Enhanced Healthcare QA Summarization

   Dongsuk Jang, Alan Li, and Arman Cohan

   In Proceedings of the Second Workshop on Patient-Oriented Language Processing (CL4Health), 2025

   Abs PDF Code

   Solving enhanced healthcare QA summarization issues with MoA and other approaches.

2024

1. arXiv

   

   LoraMap: Harnessing the Power of LoRA Connections

   Hyeryun Park, Jeongwon Kwak, Dongsuk Jang, and 2 more authors

   Under review at IEEE Journal of Biomedical and Health Informatics, 2024

   Abs PDF

   Suggesting advanced LoRA connection method.

2. AMIA

   

   Automated Information Extraction from Thyroid Operation Narrative: A Comparative Study of GPT-4 and Fine-tuned KoELECTRA

   Dongsuk Jang, Hyeryun Park, Jiye Son, and 3 more authors

   In AMIA Informatics Summit, 2024

   Abs PDF Video

   Automated Information Extraction from Thyroid Operation Narrative with LLMs and ELECTRA model.