|
Yajat Yadav Hi! I'm an incoming PhD student in computer science at UT Austin. My research will be supported by a NSF Graduate Research Fellowship. Previously, I completed my B.S. in EECS at UC Berkeley, where my research in robot learning was advised by Sergey Levine. I've also done research focusing at the intersection of deep learning and biology with Yun Song and Ahmed Alaa. Email / CV / Resume / Google Scholar / Github / LinkedIn / Twitter |
|
ResearchI am interested in developing reinforcement learning algorithms that enable agents like robots to continually learn generalizable skills. |
|
Robust Fine-tuning of Vision-Language-Action Robot Policies via Parameter
Merging
Yajat Yadav*, Zhiyuan Zhou*, Andrew Wagenmaker, Karl Pertsch, Sergey Levine International Conference on Learning Representations (ICLR), 2026 website / arXiv Model-merging based approach for robust, generalizable finetuning of VLA robot policies. |
|
ONG: Orthogonal Natural Gradient Descent
Yajat Yadav* , Jathin Korrapati* , Patrick Mendoza* NeurIPS 2025: Non-Euclidean Foundation Models and Geometric Learning Workshop, Class Project, 2025 arXiv / code A continual-learing algorithm based on natural gradients and their projections onto task-specific subspaces. Rotated MNIST figure on the left is from this paper. |
|
Agent-based Modeling for Patient-Centered Clinical Decision Support in
Neuro-Oncology
Eduardo Rodriguez Almaraz, Brenda Miao, Yajat Yadav , J de Groot, J Young, A Vo, A Butte, B S Mitchel, D Raleigh, N Butowski, Ahmed Alaa Neuro-Oncology, Volume 26 Supplement 5, 2024 Journal Article Using LLM agents to simulate clinical decision-making scenarios in neuro-oncology, validated by several independent clinicians. |
|
VROOM: Visual Reconstruction over Onboard Multiview
Yajat Yadav* , Varun Bharadwaj* , Tanish Baranwal* , Jathin Korrapati* Preprint (Class Project), 2025 website / arXiv / code Reconstructing 3D F1 circuits and car motion using multiple onbard camera views. |
|
Website inspired by Jonathan Barron. |