Scholar
Jae Myung Kim
Google Scholar ID: eP6FHFAAAAAJ
PhD candidate, University of Tübingen
Citations & Impact
Citations
450
H-index
7
i10-index
7
Publications
16
Co-authors
6
list available
Publications
Resume (English only)
Academic Achievements
- Published multiple papers in areas including synthetic training data, few-shot learning, explainability, zero-shot learning, and bias. Notable works include 'LoFT: LoRA-Fused Training Dataset Generation with Few-shot Guidance' at BMVC 2025, 'Does Feasibility Matter? Understanding the Impact of Feasibility on Synthetic Training Data' at CVPRW on SynData4CV & FGVC12 2025 (Best Paper Award), and 'DataDream: Few-shot Guided Dataset Generation' at ECCV 2024.
Research Experience
- Closely working with Cordelia Schmid (Inria/Google) on topics such as synthetic data as training data, weak alignment, zero-shot and few-shot learning.
Education
- PhD student at the University of Tübingen, member of ELLIS and IMPRS-IS programs, advised by Zeynep Akata (TU Munich/Helmholtz Munich); previously completed B.S. and M.S. degrees at Seoul National University.
Background
- Currently seeking research scientist or applied scientist roles in the industry. Broadly interested in efficient data-centric approaches, particularly in synthetic data as training data, weak alignment, and zero-shot and few-shot learning.
