Hee Bin Yoo

Hee Bin Yoo

he/him

Assistant Professor at the School of AI Software, Soongsil University

Sitemap

A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.

Pages

Posts

portfolio

ENS

Project construction and management application developed at R inc. from 2017 to 2018.

GDG

College recommendation platform developed at R inc. in 2018.

NINEWATT

Power consumption calculation and prediction backend system developed at R inc. in 2018.

publications

Unveiling the Significance of Toddler-Inspired Reward Transition in Goal-Oriented Reinforcement Learning

Published in Proceedings of the AAAI Conference on Artificial Intelligence, 2024

AAAI 2024 oral presentation.

Recommended citation: Junseok Park, Yoonsung Kim*, Hee bin Yoo*, Min Whoo Lee, Kibeom Kim, Won-Seok Choi, Minsu Lee, and Byoung-Tak Zhang. (2024). "Unveiling the Significance of Toddler-Inspired Reward Transition in Goal-Oriented Reinforcement Learning." Proceedings of the AAAI Conference on Artificial Intelligence, 38(1), 592-600.
Download Paper

Continuous SO(3) Equivariant Convolution for 3D Point Cloud Analysis

Published in European Conference on Computer Vision, 2024

ECCV 2024 poster.

Recommended citation: Jaein Kim, Hee Bin Yoo, Dong-Sig Han, Yeon-Ji Song, and Byoung-Tak Zhang. (2024). "Continuous SO(3) Equivariant Convolution for 3D Point Cloud Analysis." European Conference on Computer Vision, Springer, 59-75.

Confidence Controls Deep Metric Learning

Published in Machine Learning, 2026

Machine Learning, 2026.

Recommended citation: Jinhee Park, Hee Bin Yoo, Byoung-Tak Zhang, and Junseok Kwon. (2026). "Confidence Controls Deep Metric Learning." Machine Learning, 115, 77.
Download Paper

Neural Collapse-Informed Initialization with Perturbation Injection in Classification-based Metric Learning

Published in Proceedings of the AAAI Conference on Artificial Intelligence, 2026

AAAI 2026 oral presentation.

Recommended citation: Jinhee Park*, Hee Bin Yoo*, Minjun Kim, Byoung-Tak Zhang, and Junseok Kwon. (2026). "Neural Collapse-Informed Initialization with Perturbation Injection in Classification-based Metric Learning." Proceedings of the AAAI Conference on Artificial Intelligence, 40(10), 8287-8295.
Download Paper

Deep Neural Networks as Finite-Step Hopfield Dynamics

Published in New Frontiers in Associative Memories Workshop at ICLR, 2026

New Frontiers in Associative Memories Workshop at ICLR 2026.

Recommended citation: Wooyul Jung, Hochan Bang, Hee Bin Yoo, and Byoung-Tak Zhang. (2026). "Deep Neural Networks as Finite-Step Hopfield Dynamics." New Frontiers in Associative Memories Workshop at ICLR.
Download Paper

talks

Theory Seminar

Published:

Speaker and organizer for a seminar covering feature learning dynamics in deep learning, based on physics-inspired theory and the book Principles of Deep Learning Theory. Trained presenters on foundational concepts and conducted reviews before presentations.

Seminar for Freshman

Published:

Speaker and organizer for freshman seminars held each semester from 2022 to 2025. Covered mathematical foundations of random signals and functional theory, Bayesian views of time series, and applications of game theory to time series analysis.

Cognitive Dynamic Systems Seminar

Published:

Speaker and organizer for summer/winter 2022 seminars connecting cognitive dynamic systems with contemporary studies in stochastic differential equations, video and language, Gaussian-process views of neural networks, and neural tangent kernels.

Machine Learning 8

Published:

Speaker and organizer for a machine learning theory paper review seminar held from spring 2022 to autumn 2023, with participants from Kakao, Naver, KAIST, and Seoul National University.

Reviewing Seminar

Published:

Speaker and organizer for a summer 2023 reviewing seminar on recent deep learning theory, including implicit bias of SGD, transfer learning theory, interpolation, and overfitting from the Deep Learning Theory Workshop at the Simons Institute.

NeurIPS 2025 Paper Review Seminar

Published:

Organized and led a weekly paper reading seminar from October to November 2025. Presented multiple papers per session and facilitated discussions on optimization, generalization, and deep learning theory.

Brown Bag Seminar Poster Presentation

Published:

Presented the poster “How Classifier Features Transfer to Downstream: An Asymptotic Analysis in a Two-Layer Model” at the 6th Brown Bag Seminar, attended by over 150 researchers across interdisciplinary clusters.

AI Lunch Seminar

Published:

Delivered an invited talk titled “How Classifier Features Transfer to Downstream: An Asymptotic Analysis in a Two-Layer Model.” Presented theoretical insights on feature transfer in neural networks and conditions for generalization to unseen classes.

EWHA-CUHK Global Student Partnership Program

Published:

Organized and delivered a seminar on Physical AI and lab research activities for participants in the EWHA-CUHK Global Student Partnership Program. Coordinated a guided lab tour including Physical AI demonstrations.

teaching

Computing Utilization Education Advisor

Education advising, Hyper Metacognitive Lab, 2017

Researched and advised on applying computing to mathematics education. Collaborated with Dong-Hun Lee, founding member of Hana High School and former president of the National Association of Mathematics Teachers.

Yongsan High School Project Advising and Competition Judging

Advising and judging, Yongsan High School, 2018

Advised Computer Science Club projects on self-driving systems, human posture estimation, VR, and embedded systems. Taught project execution methodologies and current technologies, and judged annual research, IoT, and project competitions hosted by the school’s natural sciences department.