Yunpyo An (안윤표)

About

I am pursuing a combined M.S.–Ph.D. degree at the UNIST Graduate School of Artificial Intelligence, as a member of the IMIL Lab advised by Prof. Jooyeon Kim, after receiving a B.Sc. in Computer Science and Engineering from UNIST in 2022 (Cum Laude). My research is driven by one question: What are the fundamental limitations of machine learning, and how can we transcend them? We like to believe that machine learning solves many of our problems, yet in practice it often does not — the universal approximation theorem guarantees that neural networks can represent almost anything, but only under an unbounded budget, far removed from the situations we actually face. My goal is to draw a clear boundary of what is learnable under a limited budget: currently, I study the limitations of transformers in multi-step reasoning, combining theoretical analysis — such as learnability and circuit complexity under bounded resources — with empirical studies on reasoning benchmarks such as ARC-AGI, alongside a broader interest in world models. Previously, I worked on active learning, which led to a publication at AAAI 2024, and I am always open to research discussions and collaborations — feel free to reach out.

Publications

1. Yunpyo An, Suyeong Park, and Kwang In Kim. "Active Learning Guided by Efficient Surrogate Learners." AAAI 2024.

Education

• 2022 – present: M.S.–Ph.D., UNIST Graduate School of Artificial Intelligence
• 2018 – 2022: B.Sc. in Computer Science and Engineering, UNIST (Cum Laude)

Academic Service

• Reviewer: AAAI, CVPR

Experience

• 2024 – present: IMIL Lab, UNIST (world models, reasoning)
• 2021 – 2024: MLV Group, UNIST (active learning, federated learning)

Projects

• Bus information system for UNIST, Host by Hacking and Programming Club HeXA
• Speaki: Automatic grading system for Introduction to AI Programming 2 Course (101 for AI)

Teaching at UNIST

• 2026 Spring: Introduction to AI Programming 2 (head TA): basic AI course for first-year students
• 2023 Fall: Introduction to AI Programming 2 (TA)
• 2023 Fall: Discrete Mathematics (TA): basic mathematics course for undergraduate students
• 2023 Spring: Principles of Deep Learning (TA): deep learning course designed for graduate students
• 2022 Spring: Introduction to Machine Learning (TA): machine learning course designed for undergraduate students whos major in Computer Science and Engineering