Tushar Nagarajan
Papers
1
Total Citations
3
H-Index
1
About
Tushar Nagarajan is a leading researcher in embodied AI and computer vision, with a focus on grounding intelligent agents in real-world physical interactions. His work bridges the gap between passive video understanding and active robotic manipulation, leveraging egocentric video to teach agents about object affordances and task structure. In his highly cited paper "Shaping embodied agent behavior with activity-context priors from egocentric video," Nagarajan introduces a method to extract activity-context priors from in-the-wild egocentric footage, enabling robots to learn complex physical tasks—such as sequential object interactions—more efficiently by understanding preconditions from human demonstrations. This approach reduces the need for costly trial-and-error learning, making it a foundational contribution to data-efficient robot learning. With over three citations and growing influence, Nagarajan’s work is shaping how agents perceive and act in dynamic environments, advancing the frontier of interactive AI. His research is essential reading for anyone interested in embodied intelligence, human-robot collaboration, or learning from video.
Research Focus
Key Achievements
Top Papers
- 1