Rudra P.K. Poudel
Papers
1
Total Citations
5
H-Index
1
About
Dr. Rudra P.K. Poudel is a leading researcher in reinforcement learning (RL) and computer vision, with a focus on bridging the gap between simulated environments and real-world robotic applications. His work centers on developing sample-efficient, robust learning algorithms that can handle significant visual appearance variations—a critical challenge for deploying RL in everyday tasks like visual navigation. In his highly regarded paper "ReCoRe: Regularized Contrastive Representation Learning of World Model" (2024, 5 citations), Dr. Poudel introduces a novel framework that enhances model-based RL by learning invariant representations through contrastive objectives and regularization. This approach directly addresses the poor sample efficiency and fragility of current model-free methods, enabling agents to generalize across changing environments without exhaustive retraining. His contributions are pivotal for advancing autonomous systems that must operate reliably in the real world, where lighting, textures, and layouts constantly shift. Dr. Poudel’s work stands out for its theoretical rigor and practical impact, earning recognition among peers for pushing the boundaries of representation learning in RL. His research continues to inspire new directions in robust, generalizable artificial intelligence.
Research Focus
Key Achievements
Top Papers
- 1ReCoRe: Regularized Contrastive Representation Learning of World Model5 citations · 2024