Taekhyun Park (박택현)

I’m a Master’s student at BAELAB, Graduate School of Data Science, Pusan National University, Korea.

You can find my CV here.

📌 Research Interests

• Novel Architectures: JEPA, Deep Supervision, State Space Model with Deep Learning , Looped LLM
• Sequence Modeling: Time Series, Language, Video, Event Log
• LLM : Process Mining with LLM, RLVR , Agent system

🔥 News

• 2026.05:  🎉🎉 Design and Implementation of an LLM-Based Maritime Risk Scenario Generation System for Vessel Traffic Services was accepted at IEEE Case 2026.
• 2026.05:  🎉🎉 A new arXiv paper was published: LoopUS: Recasting Pretrained LLMs into Looped Latent Refinement Models Paper
• 2025.12:  🎉🎉 JustDense: Just using Dense instead of Sequence Mixer for Time Series Analysis was accepted as a Regular Paper at IEEE BigData 2025.
• 2025.10:  🎉🎉 A new arXiv paper was published: Reasoning-Aware GRPO using Process Mining Paper

📝 Publications

^* indicates first author.

Arxiv

LoopUS: Recasting Pretrained LLMs into Looped Latent Refinement Models

Taekhyun Park^*, Yongjae Lee, Dohee Kim and Hyerim Bae

Paper Submission, arXiv, Project Pages

ICIC

LURKER: LANGUAGE MODELS FOR UNVEILING MARITIME RISKS VIA KNOWLEDGE-DRIVEN EVENT FORECASTING AND REASONING

Seongmoon Hong^*, Taekhyun Park, Dohee Kim, Sangmin Jo and Hyerim Bae

ICIC Express Letters & ICICIC 2025

ICIC

IDENTIFYING SEVERITY OF MARITIME ACCIDENTS BASED ON WEIGHTED KERNEL DENSITY ESTIMATION

Sangmin Jo^*, Dohee Kim, Seongmoon Hong, Taekhyun Park, and Hyerim Bae

ICIC Express Letters & ICICIC 2025

CASE

Design and Implementation of an LLM-Based Maritime Risk Scenario Generation System for Vessel Traffic Services

Taekhyun Park^*, Sangmin Jo, Dohee Kim and Hyerim Bae

IEEE Case 2026, Project Pages

Arxiv

Generative AI and Machine Learning Collaboration for Container Dwell Time Prediction via Data Standardization

Minseop Kim^*, Takhyeong Kim, Taekhyun Park, Hanbyeol Park and Hyerim Bae

Paper Submission

Arxiv

Application of Large Language Models for Container Throughput Forecasting: Incorporating Contextual Information in Port Logistics

Minseop Kim^*, Jaeeun Kwon, Hanbyeol Park, Kikun Park, Taekhyun Park and Hyerim Bae

Paper Submission

Arxiv

Rewarding Structural Conformance of Reasoning using Process Mining

Yongjae Lee^*, Taekhyun Park^* and Hyerim Bae

Paper Submission, arXiv, code

IEEE Bigdata 2025

JustDense: Just using Dense instead of Sequence Mixer for Time Series analysis

Taekhyun Park^*, Yongjae Lee^*, Daesan Park^*, Dohee Kim and Hyerim Bae

The 13th IEEE International Conference on Big Data (IEEE BigData 2025)(acceptance rate: 18.7%) code

S&M

LA-RCS: LLM-agent-based Robot Control System

Taekhyun Park^*, Young-Jun Choi, Seung-Hoon Shin, Chang-Eun Lee and Kwangil Lee

Sensors and Materials & IMETI2024, Project Pages

🎖 Honors and Awards

• 2022.11 블록체인 기술을 활용한 글쓰기 하이퍼 플랫폼, AI.블록체인 아이디어 경진대회, 경남테크노파크원장상 수상
• 2022.12 이미지 생성 AI 학습을 위한 P2P 기반 데이터셋 구축 및 AI 소유권 가상자산화 플랫폼, KMOU AI 아이디어 챌린지 대회, 대상 수상, Link
• 2023.12 RAG 시스템 + 되먹임 AI 시스템 (S2I, AutoCoT… etc)를 활용한 학교 게시판 DB 기반 검색 AI 시스템, KMOU Capstone Design, 레인보우브레인상 수상
• 2025.11 해양 LLM 시스템, 2025 K-해양 AI 첼린지, 장려(부산테크노파크 원장) 수상

📖 Educations

• M.S. in Data Science, Pusan National University (2025.03 ~ )
  • Advisor: Prof. Hyerim Bae
• B.S. in Control Engineering, Korea Maritime & Ocean University (2019.03 ~ 2025.02)
  • Under graduate student (2023.09 ~ 2025.02) Advisor: Prof. Kwangil Lee
  • I served military service in ROKA. (2020.10 ~ 2022.02)

💻 Internships

• 2024.01 - 2024.03, DeepLogicChain, Seoul.