Samuel Maxwell Bateman
Robot Learning PhD Student
NSF CISE Fellow
@ Princeton
I am a Robot Learning PhD student advised by Professor Dhruv Shah in the PRISM Lab at Princeton University, supported by an NSF CISE Graduate Fellowship.
Humans get a driver’s license after only ~50 hours of instruction, practice, and relatively sparse feedback. A prep cook can learn a station in an afternoon with perhaps one demonstration of each task. Then both continuously improve at these tasks over periods of years with minimal to no additional feedback. Why do robots and AI agents need orders of magnitude more data and feedback to learn comparatively simpler tasks? Can they continue improving through self-supervised interaction and experience? What kinds of algorithms make those lessons generalizable to new tasks and environments?
My research is motivated by these questions, and broadly focuses on reinforcement learning, diffusion models, and generative world models for sample-efficient robot learning. I am interested in finding simple, theoretically grounded methods that scale beyond any individual task, dataset, or environment. I think the positive impact of robotics and AI more broadly will be predicated on broad accessibility of data efficient learning algorithms which are usefully applicable on a wide range of real world problems.
Previously, I was a Senior Machine Learning Engineer on the Localization and Mapping team at the autonomous vehicles startup Nuro, working on applied research in online mapping and perception foundation models. Before that, I did my undergrad in Applied Math and Computer Science at CU Boulder where I was a part of the Autonomous Robotics and Perception Group under Dr. Christoffer Heckman and the Autonomous Vehicle Systems Lab under Dr. Hanspeter Schaub.
news
| Sep 01, 2025 |  Started my PhD at Princeton, working with Dhruv Shah on reinforcement learning and diffusion models for robot learning. |
|---|---|
| May 19, 2025 |  We presented a paper at ICRA 2025 in Atlanta: Evaluating Global Geo-alignment for Precision Learned Autonomous Vehicle Localization using Aerial Data. |
| Jun 17, 2024 |  I presented a paper at CVPR 2024 Workshop on Autonomous Driving: Exploring Real World Map Change Generalization of Prior-Informed HD Map Prediction Models. |
