Scholar
Jonghyun Lee
Google Scholar ID: GPi5hw4AAAAJ
KRAFTON AI | PhD, SNU
Citations & Impact
Citations
343
H-index
7
i10-index
6
Publications
9
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- 2024: Paper 'Efficient Diffusion-Driven Corruption Editor for Test-Time Adaptation' accepted to ECCV 2024.
- 2024: Paper 'Entropy is not Enough for Test-Time Adaptation: From the Perspective of Disentangled Factors' selected as Spotlight at ICLR 2024.
- 2024: Paper 'SF(DA)^2: Source-free Domain Adaptation Through the Lens of Data Augmentation' accepted to ICLR 2024.
- 2022: Paper 'Confidence score for source-free unsupervised domain adaptation' accepted to ICML 2022 as Spotlight.
- 2021: Co-authored ICLR paper 'Removing undesirable feature contributions using out-of-distribution data'.
- 2020: Co-authored ECCV paper 'iCaps: An interpretable classifier via disentangled capsule networks'.
- 2024: Preprint 'Gradient Alignment with Prototype Feature for Fully Test-time Adaptation' posted on arXiv.
Research Experience
- Spring 2024: Research Scientist Intern at LG AI Research.
- 2024–Present: Developing techniques to reduce toxic responses in LLMs.
- 2023: Developed continual learning methods for tabular data.
- 2021–2022: Developed unsupervised domain adaptation methods.
- 2019–2020: Developed a DATA-based NVH prediction program.
- 2018–2019: Developed a smart management system for HANA.
Co-authors