Hyeonbin Moon

I'm a Ph.D. candidate in Mechanical Engineering at KAIST, where I use mechanics to predict and design how materials and structural systems behave. As engineering systems grow more complex, machine learning is the natural tool to reach for — but on scarce or noisy data it readily returns predictions that violate the very physics they describe. I embed governing equations and numerical structure directly into the learning framework instead, so that physics constrains what a model is allowed to predict, making reliable prediction possible from far less data and fast enough to run in real time. Extending this with uncertainty quantification, my aim is real-time digital twins that are not only fast but dependable enough to deploy in real engineering systems. Read more about my research →