Chethan Bhateja
Papers
1
Total Citations
5
H-Index
1
About
Chethan Bhateja is a researcher pushing the boundaries of robotic reinforcement learning (RL) by bridging the gap between internet-scale data and real-world robot control. His key research areas include offline RL, representation learning, and leveraging large-scale pre-training for robotics. In his notable 2024 work, "Robotic Offline RL from Internet Videos via Value-Function Learning," Bhateja explores how to harness internet video data—without robot-specific interaction—to pre-train value functions that generalize broadly. This approach addresses a critical bottleneck: enabling robots to learn from diverse, passively observed human demonstrations rather than requiring costly, task-specific robot datasets. By showing that offline RL methods can effectively transfer knowledge from internet videos to robotic control, his work offers a scalable path toward more versatile and adaptable robots. Though early in its impact, with 5 citations to date, this contribution is already shaping conversations around foundation models for robotics. Bhateja’s research sits at the intersection of machine learning and robotics, aiming to unlock the kind of broad generalization that has transformed other AI domains.
Research Focus
Key Achievements
Top Papers
- 1Robotic Offline RL from Internet Videos via Value-Function Learning5 citations · 2024