Scholar
Justin Kerr
Google Scholar ID: k6fOmQgAAAAJ
PhD Student, UC Berkeley
Citations & Impact
Citations
2,793
H-index
15
i10-index
17
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- GaussGym: An open-source real-to-sim framework for learning locomotion from pixels
- Eye, Robot: Learning to Look to Act with a BC-RL Perception-Action Loop
- Predict-Optimize-Distill: A Self-Improving Cycle for 4D Object Understanding
- Persistent Object Gaussian Splat (POGS) for Tracking Human and Robot Manipulation of Irregularly Shaped Objects
- Robot See Robot Do: Imitating Articulated Object Manipulation with Monocular 4D Reconstruction
- GARField: Group Anything with Radiance Fields
- LERF-TOGO: Language Embedded Radiance Fields for Zero-Shot Task-Oriented Grasping
- Nerfstudio: A Modular Framework for Neural Radiance Field Development
Research Experience
- Worked on warehouse bin-packing automation at Berkshire Grey and was involved in early work on a moon rover tech demo at NASA JPL.
Education
- Ph.D. (in progress): AI Research Lab at UC Berkeley, Advised by Angjoo Kanazawa and Ken Goldberg; B.S.: Carnegie Mellon University, Worked with Howie Choset on multi-robot path planning.
Background
- Research Interests: Advancing robot perception for real-world manipulation through 3D multi-modal understanding, 4D reconstruction, tactile sensing, and active vision. Maintains Nerfstudio, a large open-source repo for 3D neural reconstruction.
Miscellany
- Personal interests not explicitly mentioned.
Co-authors