Shreenabh Agrawal
Indian Institute of Science Bangalore, Carnegie Mellon University
Papers
2
Total Citations
11
H-Index
1
About
Shreenabh Agrawal is a researcher advancing the frontiers of reinforcement learning (RL) and robot learning from demonstration (LfD). His work addresses critical bottlenecks in scaling intelligent systems to real-world complexity. In his highly cited 2024 paper, "Barrier Functions Inspired Reward Shaping for Reinforcement Learning" (10 citations), Agrawal introduces a novel framework that leverages control barrier functions to shape rewards, offering a principled alternative to traditional value-function-based methods. This approach significantly accelerates training in large state spaces, a persistent challenge in modern RL. His subsequent 2025 work, "Scalable Learning of High-Dimensional Demonstrations with Composition of Linear Parameter Varying Dynamical Systems" (1 citation), tackles a core issue in LfD: enabling robots to generalize complex, high-dimensional tasks from sparse user demonstrations. By composing stable dynamical systems, Agrawal’s method eliminates the need for end-users to have coding expertise, democratizing robot programming. Together, these contributions demonstrate a clear trajectory toward more efficient, scalable, and accessible autonomous systems, positioning Agrawal as a rising voice in the intersection of control theory and machine learning.
Research Focus
Key Achievements
Top Papers
- 1Barrier Functions Inspired Reward Shaping for Reinforcement Learning10 citations · 2024
- 2