Introduction
I am a researcher with a focus on deep learning and numerical analysis.
Feel free to reach out if you're interested in my research.
Contact
Email: hgh2134@gist.ac.kr
Employment and Education
Assistant Professor at the Gwangju Institute of Science and Technology (April 2025 - Present)
Research Fellow at Center for AI and Natural Sciences, Korea Insitute for Advanced Study (March 2023 - March 2025)
Ph.D. in Mathematical Sciences, Seoul National University supervised by Myungjoo Kang (March 2017 – February 2023)
B.S. in Mathematical Sciences and B.S. in Statistics, Seoul National University (March 2013 – February 2017)
My current Research Interest
Foundations of Deep Learning
Universal Approximation Theorem
Optimization
Generalization
Preconditioning for Large Numerical Systems
Generative Models
Publishing
Floating-Point Neural Networks Are Provably Robust Universal Approximators (Link)
Geonho Hwang, Wonyeol Lee, Yeachan Park, Sejun Park, and Feras Saad
International Conference on Computer Aided Verification, 2025.
Floating-point neural networks can represent almost all floating-point functions (Link)
Geonho Hwang , Yeachan Park, Wonyeol Lee, and Sejun Park
Forty-second International Conference on Machine Learning. 2025.
Minimum width for universal approximation using squashable activation functions (Link)
Jonghyun Shin, Namjun Kim, Geonho Hwang, Sejun Park
Forty-second International Conference on Machine Learning. 2025.
Analysis of efficient preconditioner for solving Poisson equation with Dirichlet boundary condition in irregular three-dimensional domains (Link)
Geonho Hwang, Yesom Park, Yueun Lee, and Myungjoo Kang
Journal of Computational Physics, 2024.
Analyzing the latent space of GAN through local dimension estimation for disentanglement evaluation (Link)
Jaewoong Choi, Geonho Hwang, Hyunsoo Cho, and Myungjoo Kang
Pattern Recognition, 2025.
Expressive power of ReLU and step networks under floating-point operations (Link)
Yeachan Park*, Geonho Hwang*, Wonyeol Lee, and Sejun Park (*=equal contribution)
Neural Networks, 2024.
Minimal Width for Universal Property of Deep RNN (Link)
Chang hoon Song, Geonho Hwang, Jun ho Lee, and Myungjoo Kang
Journal of Machine Learning Research, 2023.
MAGANet: Achieving Combinatorial Generalization by Modeling a Group Action (Link)
Geonho Hwang, Jaewoong Choi, Hyunsoo Cho, and Myungjoo Kang
Proceedings of the 40th International Conference on Machine Learning, 2023.
Finding the Global Semantic Representation in GAN through Fréchet Mean (Link)
Jaewoong Choi, Geonho Hwang, Hyunsoo Cho, and Myungjoo Kang
The Eleventh International Conference on Learning Representations. 2023.
Disentangling the correlated continuous and discrete generative factors of data (Link)
Jaewoong Choi, Geonho Hwang and Myungjoo Kang
Pattern Recognition, 2023.
Do Not Escape From the Manifold: Discovering the Local Coordinates on the Latent Space of GANs (Link)
Jaewoong Choi, Junho Lee, Changyeon Yoon, Jung Ho Park, Geonho Hwang and Myungjoo Kang
International Conference on Learning Representations, 2022.