Taesun Yeom

I'm a first-year M.S.-Ph.D. student at EffL (Efficient Learning Lab), POSTECH (advisor: Prof. Jaeho Lee). My research primarily focuses on developing theories that closely align with empirical phenomena to gain a deeper understanding of neural nets.

Recently, I have been particularly interested in (1) understanding the inductive biases of neural nets, (2) neural fields, (3) implicit bias and learning dynamics.

Before joining EffL, I received my bachelor's degree in Mechanical Engineering from Chung-Ang University and worked closely with Prof. Minhyeok Lee and Prof. Seokwon Lee.

Email  /  GitHub  /  Google Scholar  /  LinkedIn  /  CV

Publications
project image

On the Internal Representations of Graph Metanetworks


Taesun Yeom and Jaeho Lee
ICLR Workshop on Weight Space Learning, 2025
arxiv / openreview

What do graph metanetworks learn?
project image

Fast Training of Sinusoidal Neural Fields via Scaling Initialization


Taesun Yeom*, Sangyoon Lee*, and Jaeho Lee
International Conference on Learning Representations (ICLR), 2025
arxiv / code / openreview

We propose a simple yet effective approach for accelerating neural field training.
project image

DuDGAN: Improving Class-Conditional GANs via Dual-Diffusion


Taesun Yeom, Chanhoe Gu, and Minhyeok Lee
IEEE Access, 2024
paper / code

We propose a dual-diffusion process, which aims to reduce overfitting of class-conditional GANs.
project image

Superstargan: Generative adversarial networks for image-to-image translation in large-scale domains


Kanghyeok Ko, Taesun Yeom, and Minhyeok Lee
Neural Networks, 2023
paper / code

Enabling GANs for image-to-image translation in large-scale domains.



Education
M.S.–Ph.D. in Artificial Intelligence
Pohang University of Science and Technology (POSTECH), South Korea 		2024.09 – Present
B.S. in Mechanical Engineering
Chung-Ang University, South Korea 		2018.03 – 2024.08

Design and source code from Jon Barron's website