Scholar
Keivan Rezaei
Google Scholar ID: NsJKrKIAAAAJ
Ph.D. Student, University of Maryland
Citations & Impact
Citations
261
H-index
9
i10-index
8
Publications
18
Co-authors
9
list available
Publications
Resume (English only)
Academic Achievements
- Published multiple papers including 'Localizing Knowledge in Diffusion Transformers' accepted to NeurIPS 2025, 'RESTOR: Knowledge Recovery via Machine Unlearning' accepted to TMLR 2025, and more; also involved in several research works such as 'Online Advertisements with LLMs: Opportunities and Challenges', 'On Mechanistic Knowledge Localization in Text-to-Image Generative Models', etc.
Research Experience
- Worked as a Research Intern at Adobe, Student Researcher at Google, and will be joining MOSAIC at Allen Institute for AI (Ai2) as a Research Intern.
Education
- Pursuing a Ph.D. at the University of Maryland, advised by Prof. Feizi and Prof. Hajiaghayi.
Background
- Fourth-year Ph.D. student at the University of Maryland, focusing on the interpretability of generative AI models from both a model and data perspective. Research interests include localizing knowledge within models, detecting and explaining their failure modes, analyzing the impact of individual data points on a model through unlearning and data selection for language model pretraining, and integrating ads into the output of LLMs as a strategy to monetize them effectively.
Miscellany
- Currently on the job market and seeking research positions.
