Achkan Salehi

Centre National de la Recherche Scientifique

Papers

Total Citations

H-Index

About

Achkan Salehi is a researcher whose work bridges the frontiers of machine learning, reinforcement learning, and robotics, with a particular focus on sample efficiency and adaptive control. His key research areas include few-shot learning, quality-diversity optimization, and meta-learning for control systems. Salehi’s major contributions are twofold: first, he pioneered the concept of "Few-Shot Quality-Diversity Optimization" (2022, 11 citations), which innovatively combines the principles of few-shot learning with quality-diversity algorithms to enable rapid adaptation and behavioral diversity in sequential decision-making tasks, a significant leap for domains like robotics where data is scarce. Second, his work on "Adaptive Asynchronous Control Using Meta-Learned Neural Ordinary Differential Equations" (2023, 2 citations) introduces a novel framework that leverages meta-learned neural ODEs to create robust, asynchronous controllers, addressing critical real-world challenges in robotics such as irregular sampling and system delays. Though his citation counts are still growing, these works represent foundational steps toward more intelligent, adaptable autonomous systems. Salehi’s research is particularly notable for its forward-looking integration of meta-learning with control theory, promising to unlock more resilient and efficient robotic behaviors in unstructured environments.