Byeongho Heo

Sungkyunkwan University — South Korea

Computer Science · Physical Sciences

20h-index2.6kcitations70works2.72yr avg
OpenAlex↗ORCID↗Google Scholar↗Lab page↗

Accepting Students?

No reports yet. Know if this professor is taking students?

Research Topics

Advanced Neural Network Applications(32), Domain Adaptation and Few-Shot Learning(32), Advanced Image and Video Retrieval Techniques(24), Multimodal Machine Learning Applications(10), Adversarial Robustness in Machine Learning(8)

Publications69 total

Ten questions concerning Large Language Models (LLMs) for building applications
Building and Environment·2026Open Access
Exploring Conditions for Diffusion models in Robotic Control
arXiv (Cornell University)·2025Open Access
Token Bottleneck: One Token to Remember Dynamics
arXiv (Cornell University)·2025Open Access
Masking meets Supervision: A Strong Learning Alliance
2025
Match Me If You Can: Semi-supervised Semantic Correspondence Learning with Unpaired Images
Lecture notes in computer science·2024· 1 cited
MaskRIS: Semantic Distortion-aware Data Augmentation for Referring Image Segmentation
arXiv (Cornell University)·2024Open Access
Improving ViT interpretability with patch-level mask prediction
Pattern Recognition Letters·2024· 2 cited
Similarity of Neural Architectures Using Adversarial Attack Transferability
Lecture notes in computer science·2024· 1 cited
Rotary Position Embedding for Vision Transformer
Lecture notes in computer science·2024· 26 cited
SeiT++: Masked Token Modeling Improves Storage-Efficient Training
Lecture notes in computer science·2024· 1 cited
Learning with Unmasked Tokens Drives Stronger Vision Learners
Lecture notes in computer science·2024
Token-Supervised Value Models for Enhancing Mathematical Problem-Solving Capabilities of Large Language Models
arXiv (Cornell University)·2024Open Access
Rotary Position Embedding for Vision Transformer
arXiv (Cornell University)·2024· 1 citedOpen Access
Morphing Tokens Draw Strong Masked Image Models
arXiv (Cornell University)·2023Open Access
SeiT++: Masked Token Modeling Improves Storage-efficient Training
arXiv (Cornell University)·2023Open Access
Match me if you can: Semi-Supervised Semantic Correspondence Learning with Unpaired Images
arXiv (Cornell University)·2023Open Access
Learning with Unmasked Tokens Drives Stronger Vision Learners
arXiv (Cornell University)·2023· 1 citedOpen Access
Scratching Visual Transformer's Back with Uniform Attention
2023· 34 cited
SeiT: Storage-Efficient Vision Training with Tokens Using 1% of Pixel Storage
2023· 6 cited
Masking meets Supervision: A Strong Learning Alliance
arXiv (Cornell University)·2023Open Access
page 1 of 4Next →

Frequent Co-authors

Sangdoo Yun(27), Dongyoon Han(20), Jin Young Choi(19), Sanghyuk Chun(12), Seong Joon Oh(8), Minsik Lee(5), Junsuk Choe(5), Jung-Woo Ha(5), Jongwon Choi(5), Wonjae Kim(5), Youngjoon Yoo(4), Youngjung Uh(3), Youngmin Ro(3), Jongin Lim(3), Hawook Jeong(3), Jeesoo Kim(2), Hyojin Park(2), Nojun Kwak(2), Mingi Ji(2), Sungrae Park(2)