Saurabh Kumar
PhD Student at Stanford University
I am a final-year PhD student in Stanford's CS department, where I'm fortunate to be advised by Benjamin Van Roy.
The goal of my research is to design algorithms which can continually adapt to changes in dynamic environments. I have been fortunate to be supported by the Stanford Knight Hennessy Fellowship and the NSF GRFP Fellowship. Previously, I was an AI resident at Google Brain. I graduated from Georgia Tech with a Bachelors in Computer Science and minor in Applied Math. At Georgia Tech, I was fortunate to be supported by the Barry Goldwater Scholarship and the Stamps President's Scholarship.
I spend most of my free time playing the tabla. I also enjoy swimming and playing tennis.
Selected Publications (See all)
-
Continual Learning as Computationally Constrained Reinforcement LearningSaurabh Kumar*, Henrik Marklund*, Yifan Zhu, Hong Jun Jeon, Yueyang Liu, Benjamin Van Roy
Foundations and Trends in Machine Learning, 2025 (to appear) -
Learning Continually by Spectral RegularizationAlex Lewandowski, Michał Bortkiewicz, Saurabh Kumar, Dale Schuurmans, András György, Mateusz Ostaszewski, Marlos C. Machado
International Conference on Learning Representations, 2025 -
Conformal Transformations for Symmetric Power TransformersSaurabh Kumar*, Jacob Buckman*, Carles Gelada, Sean Zhang
ICLR 2025 Workshop on Scalable Optimization for Efficient and Adaptive Foundation Models -
Maintaining Plasticity in Continual Learning via Regenerative Regularization
-
The Need for a Big World Simulator: A Scientific Challenge for Continual LearningSaurabh Kumar*, Hong Jun Jeon*, Alex Lewandowski, Benjamin Van Roy
Finding the Frame: An RLC Workshop for Examining Conceptual Frameworks, 2024 (oral) -
Satisficing Exploration for Deep Reinforcement LearningDilip Arumugam, Saurabh Kumar, Ramki Gummadi, Benjamin Van Roy
Finding the Frame: An RLC Workshop for Examining Conceptual Frameworks, 2024 (oral) -
One Solution is Not All You Need: Few-shot Extrapolation via Structured MaxEnt RLSaurabh Kumar, Aviral Kumar, Sergey Levine, Chelsea Finn
Neural Information Processing Systems, 2020
Contact
You can reach me at szk at cs dot stanford.edu
Last updated: March 6, 2025
Website Template Credits