Samarth Shukla

About

Samarth Shukla is a leading researcher in autonomous navigation and deep reinforcement learning, with a focus on developing sample-efficient, learning-based systems for robotic mapless navigation. His most influential work, "Reinforced Imitation," introduces a novel hybrid training paradigm that combines expert demonstrations, imitation learning, and reinforcement learning to train end-to-end neural networks for target-driven navigation. This approach dramatically improves sample efficiency, enabling robots to learn complex navigation policies with far fewer interactions than traditional RL methods. The paper has garnered over 198 citations, underscoring its impact on the field. Shukla’s contributions address a critical bottleneck in deploying deep RL in real-world robotics—the high cost of data collection—by leveraging prior demonstrations to bootstrap learning. His work has been widely recognized for bridging the gap between imitation learning and reinforcement learning, offering a practical pathway for mapless navigation in unstructured environments. Through this research, Shukla has advanced the state of the art in autonomous systems, making his work essential reading for students and researchers interested in efficient, learning-driven robot control.

Research Focus

Key Achievements

Top Papers

1. 1
   Reinforced Imitation: Sample Efficient Deep Reinforcement Learning for Mapless Navigation by Leveraging Prior Demonstrations

   198 citations · 2018
2. 2
   Reinforced Imitation: Sample Efficient Deep Reinforcement Learning for Map-less Navigation by Leveraging Prior Demonstrations

   16 citations · 2018

Key Collaborators

Contact & Links