Kyungmin Choi
AI Researcher · GIST Data Science Lab
Abstract reasoning and analogy toward general intelligence
I am an AI researcher at the Gwangju Institute of Science and Technology (GIST) , working in the Data Science Lab under Prof. Sundong Kim. My research centers on abstract reasoning and analogy - building datasets and methods that align machine reasoning with human intuition (ARC-style tasks).
I am broadly interested in the intersection of machine learning, reasoning, and the philosophy of science, and in what it takes to move from pattern recognition toward genuine, general intelligence.
Research Interests
Abstract Reasoning & ARC
How can machines acquire the kind of fluid, compositional reasoning that ARC-style tasks demand?
Gestalt-level Reasoning Machine
Examining what should be done, not what seems working.
Reasoning & the Philosophy of Science
What the history of physics can teach us about interpreting learned representations.
Publications
See all publications →News
CV
Education
Awards
- Great Thesis Award (2026) (2026)