Stephen Baek

Stephen Baek is a faculty member at the University of Iowa, Department of Industrial and System Engineering, with a PhD from Seoul National University's School of Mechanical and Aerospace Engineering. His research spans interdisciplinary applications of machine learning in computational modeling, biomedical engineering, and geometric data processing.

• Current Affiliation: University of Iowa, Department of Industrial and System Engineering
• PhD Institution: Seoul National University

Stephen's work focuses on physics-informed machine learning, multiscale modeling, and geometric deep learning. He develops algorithms that integrate physical principles with neural networks for applications in energetic materials, human pose estimation, and medical imaging. His research also includes federated learning and interpretable AI for constraint-based synthesis and text classification.

Recent publications highlight a physics-aware deep learning framework (PARCv2) for spatiotemporal dynamics, graph convolutional networks for airway mesh smoothing, and prototype trajectory methods for explainable AI. This work bridges computer science with applied physics and healthcare domains.

Stephen actively collaborates across disciplines, evidenced by co-authors from institutions like Iowa, Seoul National University, and Samsung Electronics, with publications in venues such as CoRR, J. Mach. Learn. Res., and NeurIPS. His methodological innovations address challenges in model heterogeneity, 3D surface processing, and constraint satisfaction.