Mar 2016 - Aug 2020
Yonsei University
Bachelor's degree in Applied Statistics
GPA: 3.72 / 4.3
PhD Candidate in Artificial Intelligence
Dahee Kwon
I am a PhD student at the KAIST Graduate School of AI advised by Prof. Jaesik Choi. My research focuses on the internal mechanisms of deep learning models, especially how representations in vision and multimodal systems support generalizable reasoning, robustness, and open-ended generation.
Email
daheekwon@kaist.ac.kr
Location
Seongnam, Korea
Phone
+82 10 6287 7674
Website
daheekwon.github.io
Research Interests
Generative AI
Representation Learning
Interpretability
Vision-Language Models
Computer Vision
Summary
My current work centers on understanding the expressive capabilities and inherent failure modes of visual generative models, with the long-term goal of improving robustness and generalization in open-ended real-world settings. I am particularly interested in how structural organization in learned representations enables scalable reasoning across vision and multimodal systems.
Education
Aug 2020 - Aug 2022
Korea Advanced Institute of Science and Technology
Master of Science in Artificial Intelligence
Advisor: Jaesik Choi | GPA: 4.3 / 4.3
Aug 2022 - Present
Korea Advanced Institute of Science and Technology
PhD Candidate in Artificial Intelligence
Advisor: Jaesik Choi | GPA: 4.2 / 4.3
Selected Publications
My recent work spans interpretability, representation learning, and generative modeling in vision and multimodal systems.
Granular Concept Circuits: Toward a Fine-Grained Circuit Discovery for Concept Representations
Dahee Kwon, Sehyun Lee and Jaesik Choi
Enhancing Creative Generation on Stable Diffusion-based Models
Jiyeon Han, Dahee Kwon and Jaesik Choi
Understanding Distributed Representations of Concepts in Deep Neural Networks without Supervision
Wonjoon Chang, Dahee Kwon and Jaesik Choi
Work Experience
Oct 2025 - Present
Research Intern, Naver Cloud
Visual Generation Team
- Architected an automated data curation pipeline to improve image editing performance.
- Implemented a Ray-based distributed framework for high-throughput large-scale data processing.
Scholarships and Awards
Insung Scholarship
Jan 2025
KAIST Breakthroughs Spring 2026
Mar 2026
Selected for KAIST Breakthroughs 50 for the paper "Granular Concept Circuits: Toward a Fine-Grained Circuit Discovery for Concept Representations."
Talks
Understanding Deep Neural Networks Decision-Making Through Exploring Learned Features
AI EXPO KOREA 2024 Workshop | May 2024
Analyzing the Attribute-relevant Featuremaps in Stable Diffusion Models
KCC XAI Workshop | Jun 2024
Understanding Diffusion-based Generative Models
KAIST XAI Tutorial Series | Nov 2024
Enhancing Creativity in Text-to-Image Generation
Samsung AI Forum | Sep 2025
Projects
Development of Reinforcement Learning Technology for Optimizing and Automating Semiconductor Design
Samsung SAIT | Feb 2021 - Jun 2022
Worked on reinforcement learning and evolutionary algorithm baselines including DDPG, PPO, GCN-RL, and NSGA3.
Creative Generation in Text-to-Image Diffusion Model
Naver AI Lab | Sep 2022 - Aug 2024
Studied internal representations and image editing behavior in text-to-image diffusion models.
AI Research Hub
IITP | Aug 2024 - Present
Leading projects on neural scaling laws and interpretation of information stored in internal model modules.
Skills
Python
PyTorch
Representation Learning
Interpretability
Generative Models
English