Byeongho Yu
PhD student @ POSTECH CSE
bhyu418@postech.ac.kr · LinkedIn · GitHub · X
I am a PhD student in Computer Science and Engineering at POSTECH, advised by Prof. Eunhyeok Park at the Eco Lab (Efficient Computing Lab), part of the ML Lab.
I started my PhD in 2024 and earned my B.S. in Computer Science and Engineering from POSTECH. I was also an exchange student at Université de technologie de Troyes, France, in 2023.
I am broadly interested in improving the efficiency and performance of AI models through approaches such as model compression and efficient inference. Recently, I have been particularly interested in reasoning in LLMs and how to improve it efficiently. My interests span large language models, looped language models, and spiking neural networks.
News
| Sep 25, 2026 | Our paper “Robust and Efficient Continual Model Merging via Global Singular Subspace Separation and Restoration” has been accepted to NeurIPS 2026 🎉 |
|---|---|
| Aug 21, 2026 | Our paper “LoopCD: Loop-wise Contrastive Decoding for Improving Reasoning in Looped Language Models” has been accepted to EMNLP 2026 Main Conference 🎉 |
| Nov 11, 2025 | Our paper “Stabilizing Direct Training of Spiking Neural Networks: Membrane Potential Initialization and Threshold-robust Surrogate Gradient” has been accepted to WACV 2026 🎉 |
| Aug 21, 2025 | My first paper “PruneCD: Contrasting Pruned Self Model to Improve Decoding Factuality” has been accepted to EMNLP 2025 Main Conference 🎉 |
Publications
2026
- Robust and Efficient Continual Model Merging via Global Singular Subspace Separation and Restoration
2025
Teaching
Teaching Assistant
- Fall 2025 · CSED490F — Deep Learning Implementation
- Spring 2025 · CSED311 — Computer Architecture
Mentoring
- Fall 2025 · CSED199 — Freshman Research Participation · Mentor
- Fall 2023 · CSED211 — Introduction to Computer Software Systems · Student Mentor
- Fall 2022 · CSED211 — Introduction to Computer Software Systems · Student Mentor