Gaurav Ashish

Papers

Total Citations

H-Index

About

Gaurav Ashish is a rising researcher whose work sits at the intersection of reinforcement learning (RL), inverse reinforcement learning (IRL), and safe AI systems. His primary focus is on enabling autonomous agents to learn and respect real-world constraints—such as safety limits or ethical boundaries—that are often unspoken or difficult to specify mathematically. In his highly cited 2022 paper, "Learning Soft Constraints From Constrained Expert Demonstrations" (4 citations), Ashish pioneered methods to infer these hidden constraints directly from expert behavior, moving beyond traditional IRL that assumes only reward optimization. This work is critical for deploying RL agents in physical systems where violations can be costly. He further solidified his impact with "Benchmarking Constraint Inference in Inverse Reinforcement Learning" (2 citations), where he established standardized evaluation frameworks for this emerging field. By tackling the fundamental challenge of unknown constraints, Ashish is helping to bridge the gap between theoretical RL and safe, real-world deployment. His contributions are particularly valuable for students and researchers working on AI safety, robotics, and human-in-the-loop learning, offering a clear path toward more robust and trustworthy autonomous systems.