An invited seminar titled “Structured Learning for Reliable Autonomy” was held on 21 September 2026 from 13:00 to 14:00 in Room 303, Building 301 at Seoul National University, with remote participation available via Zoom. The event was organized by the Aerospace Autonomy and Intelligence Laboratory (AAI) and was hosted by Prof. Namhoon Cho.
The seminar was delivered by Youngjae Min, a final-year Ph.D. candidate in Autonomous Systems and Statistics at the Massachusetts Institute of Technology, advised by Professor Navid Azizan in the Laboratory for Information and Decision Systems (LIDS). Min holds an S.M. in aeronautics and astronautics from MIT (2023) and a B.S. in electrical engineering and mathematical sciences from KAIST (2020). His background includes internships at Honda Research Institute (2025) and Tesla AI (2026), and his work has been supported by the MIT School of Engineering MathWorks Fellowship and a Kwanjeong Educational Foundation Scholarship.
In his presentation, Min outlined structured learning approaches that embed inductive biases across multiple levels of system design to improve the reliability of autonomous systems. He described architectural methods that enforce certifiable properties such as feasibility, stability, and safety as hard constraints; algorithmic techniques for structured sequential learning to enable continual adaptation while mitigating catastrophic forgetting; and system-level integration of these principles into structured robotic platforms. The talk emphasized moving beyond purely data-driven learning toward designs shaped by mathematical structure, control-theoretic principles, and system-level considerations.