Joohyung Yun

Hi! I'm Joohyung Yun, a 4th year integrated M.S./Ph.D. student (2nd year Ph.D.) in Computer Science at POSTECH, advised by Professor Wook-Shin Han in the Data Systems Lab. My research sits at the intersection of Natural Language Processing and Information Retrieval, with a focus on Retrieval-Augmented Generation (RAG) for complex, real-world documents.

I'm particularly interested in multimodal document RAG: collections that mix text, tables, and images. My recent projects explore multihop reasoning across these modalities and how to combine retrieval with LLM reasoning effectively. Currently, I'm extending this work toward Agentic RAG—building autonomous, data-centric systems that go beyond simple retrieval into complex reasoning with history-aware backtracking and adaptive traversal strategies.

Looking ahead, I aim to scale these ideas to Enterprise-level RAG: high-velocity, high-trust systems powered by Knowledge Graphs and Multi-Modal RAG that deliver verifiable, multi-modal insights. My long-term vision is to contribute to cost-efficient, data-centric platforms designed for AGI-level reasoning and agentic autonomy across enterprise ecosystems.

Contact: jhyun@dblab.postech.ac.kr · joohyung00@postech.ac.kr

Research Focus

I'm currently working on methods and systems for:

• Multimodal & multihop retrieval over documents composed of text, tables, and images
• Agentic RAG with history-aware backtracking and adaptive traversal
• Enterprise-level RAG powered by Knowledge Graphs and Multi-Modal RAG

News

Publications

• Conference
  🔄 Failure is Feedback: History-Aware Backtracking for Agentic Traversal in Multimodal Graphs
  Yun, J., Lee, D., Han, W.
  ICML, 2026.
• Conference
  🌸 LILaC: Late Interacting in Layered Component Graph for Open-domain Multimodal Multihop Retrieval
  Yun, J., Lee, D., Han, W.
  EMNLP, 2025.
• Conference
  ☀️ HELIOS: Harmonizing Early Fusion, Late Fusion, and LLM Reasoning for Multi-Granular Table-Text Retrieval
  Park, S., Yun, J., Lee, J., Han, W.
  ACL, 2025.
• Conference
  🧩 ReCG: Bottom-Up JSON Schema Discovery Using a Repetitive Cluster-and-Generalize Framework
  Yun, J., Tak, B., Han, W.
  PVLDB, 2024.
• Workshop
  🏆 KDD Cup Meta CRAG 2024 Technical Report: Three-Step Question-Answering Framework
  Park, S., Seok, J., Lee, J., Yun, J., Lee, W.
  KDD Cup Workshop on Retrieval-Augmented Generation, 2024.