D. Patrone

Johns Hopkins University Applied Physics Laboratory

Papers

Total Citations

H-Index

About

D. Patrone is a researcher focused on advancing the frontiers of robot learning, particularly in the areas of meta-imitation and goal-directed policy generalization. Their most notable contribution, "Visual Goal-Directed Meta-Imitation Learning" (2022), tackles a fundamental challenge in robotics: enabling agents to perform entirely new tasks on the very first attempt, without any prior demonstration or example trajectory. This zero-shot generalization capability pushes beyond traditional meta-learning, which often requires at least one adaptation step. By integrating visual goal specification with meta-imitation, Patrone’s work aims to create policies that can instantly infer and execute unseen goals from a single visual cue. Though early in its impact cycle with 3 citations, this work represents a bold step toward truly autonomous and adaptable robotic systems. Patrone’s research sits at the intersection of imitation learning, meta-learning, and computer vision, addressing the critical need for sample efficiency and rapid generalization in real-world robot deployment. Their contributions are particularly relevant for students and researchers interested in building robots that learn from minimal human input and adapt on the fly.