Abhi Gupta
Papers
2
Total Citations
68
H-Index
2
About
Abhi Gupta is a leading researcher in robotic manipulation, with a focus on enabling dexterous, real-world performance through learning from limited data. His work centers on two pivotal challenges: achieving robust multi-fingered grasping and efficiently learning long-horizon tasks. In his highly cited work, "Generative Attention Learning: a 'GenerAL' framework for high-performance multi-fingered grasping in clutter" (52 citations), Gupta introduced a novel framework that combines generative modeling with attention mechanisms, dramatically improving a robot's ability to grasp objects in dense, unstructured environments. This contribution is foundational for advancing robotic dexterity beyond simple parallel-jaw grippers. Complementing this, his paper "SQUIRL: Robust and Efficient Learning from Video Demonstration of Long-Horizon Robotic Manipulation Tasks" (16 citations) tackles the data inefficiency of deep reinforcement learning. SQUIRL enables robots to learn complex, multi-step tasks directly from a single video demonstration, bypassing the need for extensive real-world trial-and-error. By bridging the gap between simulation and reality, Gupta’s work is paving the way for more autonomous, adaptable robots in manufacturing, logistics, and home assistance.
Research Focus
Key Achievements
Top Papers
- 1
- 2