Portrait of Junhyuk Kwon

Junhyuk Kwon

I am an M.S. student at the Graduate School of Artificial Intelligence, POSTECH, where I am a member of the Machine Learning Laboratory advised by Prof. Jungseul Ok. Previously, I received B.S. degrees in Electrical Engineering and Artificial Intelligence from Inha University.

My research interests lie in agentic multimodal AI and vision-language reasoning. I build interactive and verifiable AI agents that understand ambiguous human requests, reason over visual environments, and reliably execute multi-step tasks.

If you are interested in my research, please feel free to contact me at tree.jhk@gmail.com.

News

ProCompNav is accepted to NeurIPS 2026 as an Oral presentation (Top 0.36% - 112/30,709 submissions).
Completed an AI Research Internship at KT Corp.
VIRO is accepted to CVPR 2026.
RecKG is accepted to ACM SAC 2024 as an Oral presentation.

Education

Mar, 2025 – Feb, 2027 Pohang University of Science and Technology (POSTECH), Pohang, South Korea
M.S. in Graduate School of Artificial Intelligence
Advisor: Prof. Jungseul Ok
Mar, 2018 – Feb, 2025 Inha University, Incheon, South Korea
B.S. in Electrical Engineering and Artificial Intelligence (Double Major)
Undergraduate Research Advisor: Prof. Young-Duk Seo (KDD Lab)

Publications

  1. ProCompNav: Proactive Instance Navigation with Comparative Judgment for Ambiguous User Queries
    Junhyuk Kwon, Seungjoon Lee, Hyejin Park, Kyle Min, and Jungseul Ok
    Neural Information Processing Systems (NeurIPS), 2026 Oral (Top 0.36% - 112/30,709 submissions)
    arXiv Code Project Page
  2. VIRO: Robust and Efficient Neuro-Symbolic Reasoning with Verification for Referring Expression Comprehension
    Hyejin Park, Junhyuk Kwon, Suha Kwak, and Jungseul Ok
    Computer Vision and Pattern Recognition (CVPR), 2026
    arXiv Code
  3. RecKG: Knowledge Graph for Recommender Systems
    Junhyuk Kwon, Seokho Ahn, and Young-Duk Seo
    ACM/SIGAPP Symposium on Applied Computing (SAC), 2024 Oral
    arXiv Code

Preprint

  1. MMTB: Evaluating Terminal Agents on Multimedia-File Tasks
    Chiyeong Heo, Jaechang Kim, Junhyuk Kwon, Hoyoung Kim, Dongmin Park, Jonghyun Lee, and Jungseul Ok
    Under review 2026
    arXiv
  2. SegmentWeave: Post-Processing Framework for Enhanced Dense Video Captioning
    Junhyuk Kwon (second author), et al.
    Under review 2026

Honors

2024Grand Prize (2nd of 12), Inha AI Challenge — LLM-based Question Answering
2024Undergraduate Paper Award, Korea Computer Congress (Top 12.8%)
2023Excellence Award (2nd of 28), Hyundai Motor Group CTO AI Competition

Experience

Research & Industry

Jul – Aug, 2026 AI Research Intern, KT Corp
One of Korea’s largest telecom and AI companies, serving 30%+ of the country’s mobile users.
Mar, 2023 – Feb, 2025 Undergraduate Researcher, Inha University KDD Lab
Knowledge Graphs, Recommender Systems, and LLMs
Jul – Sep, 2024 AI Research Intern, Growdle

Additional

Sep, 2022 – Feb, 2023 NAVER Boostcamp AI Tech, Recommender Systems
Jun, 2019 – Jan, 2021 Korean Augmentation to the U.S. Army (KATUSA)