About

Dinesh Jayaraman is a prominent robotics and machine learning researcher whose work sits at the intersection of robot manipulation, tactile sensing, reinforcement learning, and visual representation learning. He is perhaps best known for pioneering the integration of touch and vision in robotic grasping — his 2018 paper "More Than a Feeling" (353 citations) demonstrated how tactile feedback can dramatically improve a robot's ability to grasp and regrasp objects, a fundamental shift from purely vision-based approaches. This thread continued with his influential work on touch-based control using deep predictive models, collectively amassing nearly 150 citations. Jayaraman has also made significant contributions to large-scale robot learning, contributing to the Open X-Embodiment collaboration (119 citations) and the DROID manipulation dataset (108 citations), both of which advance generalizable, cross-platform robotic intelligence. His research extends into reward learning, with Eureka leveraging large language models for automated reward design, and value-implicit pre-training for universal visual representations. Through reproducible benchmarks like REPLAB and work on safety-critical reinforcement learning, Jayaraman consistently bridges theoretical innovation with practical, deployable robotic systems — making him a compelling figure in modern embodied AI research.

Research Focus

Key Achievements

14
H-Index
24
Papers
1,017
Total Citations
42
Avg Citations/Paper
🏆 Most Cited Paper
More Than a Feeling: Learning to Grasp and Regrasp Using Vision and Touch
353 citations · 2018
📈 Most Prolific Year: 2019 (6 Papers)
🤝 Key Collaborators: 216
🏛 Institutions: University of California, Berkeley, California University of Pennsylvania, Institute of Occupational Medicine, Berkeley College, University of Pennsylvania, Menlo School

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago