Gyusam Chang

About Me

Gyusam Chang is a Ph.D. candidate at Korea University, advised by Prof. Sangpil Kim. He previously interned at Samsung Advanced Institute of Technology (SAIT), where he worked with Principal Researcher Sujin Jang on domain generalization and adaptation for 3D object detection. He also completed a Visiting Graduate Researcher program at the University of California, Los Angeles (UCLA), working with Prof. M. Khalid Jawed on 3D reconstruction for smart farming systems. His current research focuses on Agentic AI for Neural Architecture Design, developing Multi-agent frameworks for interpretable architecture design across diverse real-world applications.

Research Interests

• Computer Vision: 3D Object Detection, Multi-view 3D Reconstruction, 3D Gaussian Splatting
• Multimodal AI: Large Language Models (LLMs), Vision-Language Models (VLMs), Retrieval-Augmented Generation (RAG), Multimodal Knowledge Graphs (MMKGs), Recommender Systems
• Machine Learning: Domain Generalization and Domain Adaptation, Neural Architecture Design (NAD)

News

• [Nov. 2025] Our paper ‘NIRSplat’ has been accepted to AAAI 2026!
• [Sep. 2024] Our paper ‘UDGA’ was accepted to NeurIPS 2024!
• [Sep. 2024] He joined SCI Lab at UCLA as a Visiting Graduate Researcher!
• [Sep. 2023] Our paper ‘CMDA’ was accepted to AAAI 2024!
• [Jun. 2023] He joined CV Lab at SAIT as a Research Internship!

Publications

Patents

Method and apparatus with object detection model training

Sujin Jang, Sangpil Kim, Jinkyu Kim, Wonjeong Ryu, Dongwook Lee, Gyusam Chang, Daehyun Ji
US Patent App. 18/897,759, 2025

Method and apparatus with object detector training

Sujin Jang, Sangpil Kim, Jinkyu Kim, Wonseok Roh, Gyusam Chang, Dongwook Lee, Daehyun Ji
US Patent App. 18/451,287, 2024

Services

Conference Reviewers

• NeurIPS 2025

• AAAI 2026

Journal Reviewers

• IEEE Transactions on Intelligent Transportation Systems (TITS)

Powered by Jekyll and Minimal Light theme.