Parameswaran Kamalaruban

École Polytechnique Fédérale de Lausanne

Papers

Total Citations

H-Index

About

Parameswaran Kamalaruban is a researcher whose work lies at the intersection of machine learning, reinforcement learning, and human-AI interaction. His most notable contribution is in the domain of **inverse reinforcement learning (IRL)**, where he addresses the critical challenge of learning from limited teacher interaction. In his highly cited 2020 paper, "Interaction-limited Inverse Reinforcement Learning," Kamalaruban proposes a novel framework that enables an agent to efficiently infer a reward function even when a helpful teacher is unavailable or has restricted access. This work is particularly impactful for real-world scenarios where continuous expert feedback is impractical, such as in autonomous systems or robotics. By tackling the bottleneck of interaction scarcity, his research paves the way for more sample-efficient and practical learning algorithms. With 3 citations on this key paper alone, his ideas are gaining traction in the reinforcement learning community. Kamalaruban’s work is essential reading for students and researchers interested in making AI systems that learn robustly from limited human guidance, bridging the gap between theoretical IRL and real-world deployment.