Dmytro Korenkevych

About

Dmytro Korenkevych is a robotics and machine learning researcher whose work sits at the intersection of deep reinforcement learning and real-world robotic control. His research has made significant strides in bridging the gap between simulated environments and the practical challenges of deploying learning algorithms on physical robots. His most influential contribution, "Benchmarking Reinforcement Learning Algorithms on Real-World Robots" (2018, 46 citations), established a rigorous framework for evaluating model-free reinforcement learning methods in real-world settings, providing the research community with reproducible standards that accelerated progress in the field. Complementing this, his work on "Setting up a Reinforcement Learning Task with a Real-World Robot" (2018, 24 citations) directly addressed the reliability and reproducibility challenges that had long hindered physical robot adoption in RL research. Beyond benchmarking, Korenkevych has pushed the boundaries of exploration strategies in continuous control, proposing autoregressive policies as a compelling alternative to conventional Gaussian approaches, enabling smoother and more effective robot trajectories. Collectively, his contributions have helped democratize real-world reinforcement learning research, making it more accessible, reliable, and practically viable for the broader scientific community.

Research Focus

Key Achievements

Top Papers

1. 1
   Benchmarking Reinforcement Learning Algorithms on Real-World Robots

   46 citations · 2018
2. 2
   Setting up a Reinforcement Learning Task with a Real-World Robot

   24 citations · 2018
3. 3
   Autoregressive Policies for Continuous Control Deep Reinforcement Learning

   13 citations · 2019

Key Collaborators

Contact & Links