Aditya Kannan

Papers

Total Citations

H-Index

About

Aditya Kannan is a leading researcher in dexterous robotic manipulation, with a focus on bridging the gap between human hand dexterity and robotic control. His work centers on enabling robots to perform complex, real-world tasks—particularly with soft, deformable objects—by learning directly from human demonstrations. In his highly cited 2024 paper, "Learning dexterity from human hand motion in internet videos" (12 citations), Kannan proposes a novel approach to circumvent the safety and hardware limitations of unguided robot experience collection by leveraging vast internet video data of human hands. This work has significant implications for building generalist robotic agents capable of operating in diverse environments. His 2023 paper, "DEFT: Dexterous Fine-Tuning for Real-World Hand Policies" (2 citations), further advances the field by addressing the challenges of long-horizon manipulation tasks with deformable objects. Kannan’s research is distinguished by its focus on open-source platforms, as demonstrated in his 2024 demonstration paper, making dexterous hand control more accessible to the broader robotics community. His contributions are paving the way for more adaptable, human-like robotic hands.