Mohammadhossein Malmir

Technical University of Munich, Politecnico di Milano

Papers

Total Citations

H-Index

About

Mohammadhossein Malmir is a robotics researcher whose work focuses on bridging the critical gap between simulated and real-world environments—a challenge known as Sim2Real transfer. His primary research areas include reinforcement learning, domain randomization, and continual learning for robotic manipulation. Malmir’s most significant contribution is his pioneering analysis of how randomization techniques affect the transfer of robotic policies from simulation to reality, a study that has garnered 11 citations and serves as a foundational reference for researchers navigating the complexities of Sim2Real. He further advanced the field by introducing "Continual Domain Randomization," a novel approach that addresses the limitations of traditional static randomization by dynamically adapting simulation parameters during training, earning 5 citations. His earlier work on "Continual Learning on Incremental Simulations" (2 citations) laid the groundwork for scalable Sim2Real methods, challenging the prevailing reliance on hyper-realistic simulations. Through these contributions, Malmir has established himself as a key voice in making robotic learning more robust and practical, directly impacting how autonomous systems are trained for complex, real-world manipulation tasks.