Kim, Donghwan (김동환)

Donghwan Kim is an associate professor in the Department of Mathematical Sciences at Korea Advanced Institute of Science and Technology (KAIST). His research focuses on optimization theory and applications, particularly in the areas of machine learning and generative models. Kim is affiliated with the Graduate School of AI for Math (AI4Math) and has been actively involved in academic conferences and seminars. He has contributed to various research topics, including convergence of gradient methods, loss landscapes, and implicit bias in learning algorithms. Kim's work has been recognized through several academic events and publications, and he continues to mentor students at both undergraduate and graduate levels.

Donghwan Kim is a researcher at KAIST, specializing in Optimization for Machine Learning. His work focuses on developing efficient optimization algorithms for machine learning tasks, particularly in convex and nonconvex settings. He has contributed to the design of first-order methods, proximal point methods, and momentum-based techniques for accelerating convergence in image reconstruction and restoration problems. His research also explores the application of optimization techniques in X-ray CT image reconstruction, including ordered subsets and spatially nonuniform optimization transfer. Kim's publications highlight the importance of structured nonconvex-nonconcave minimax problems and the use of inexact fixed-point iterations for solving such problems.