Naman Agarwal
Papers
1
Total Citations
5
H-Index
1
About
Naman Agarwal is a researcher at the forefront of differentiable control and robotics, whose work bridges the gap between simulation and learning. His key contributions center on developing tools and methods that leverage automatic differentiation for control tasks, enabling faster and more efficient training of robotic systems. Agarwal is the lead author of "Deluca -- A Differentiable Control Library," a seminal 2021 paper that introduced an open-source library of natively differentiable physics and robotics environments. This work, which has garnered 5 citations, provides a comprehensive benchmarking suite and gradient-based control methods, allowing researchers to auto-differentiate through simulation dynamics. By making these environments accessible, Agarwal has significantly accelerated research in model-based reinforcement learning and optimal control. His efforts are particularly notable for democratizing access to differentiable simulators, a critical resource for training complex robotic behaviors. Through Deluca, Agarwal has established himself as a key architect of the infrastructure that enables modern, data-efficient approaches to robot learning, making his work indispensable for students and researchers pushing the boundaries of intelligent control.
Research Focus
Key Achievements
Top Papers
- 1