Arpit Bahety
Papers
4
Total Citations
35
H-Index
3
About
Arpit Bahety is a robotics researcher whose work sits at the intersection of manipulation, learning, and explainable AI. He is best known for introducing **REFLECT** (2023, 18 citations), a pioneering framework that leverages Large Language Models to automatically detect, summarize, and correct robot failures from execution logs—a critical step toward building transparent and robust autonomous systems. In parallel, Bahety has made significant contributions to deformable object manipulation with **Bag All You Need** (2023, 11 citations), which presents a generalizable bagging strategy for heterogeneous objects, tackling the complex interplay between rigid and deformable items under partial observability. His **BaRiFlex** gripper design (2024, 4 citations) addresses a key bottleneck in robot learning by offering both versatility and collision robustness, enabling safer and more effective policy learning in human environments. With a clear focus on bridging perception, reasoning, and physical interaction, Bahety’s work is shaping how robots can understand their own mistakes and handle the messy, unstructured objects of daily life.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4