Yiming Ding

Papers

Total Citations

H-Index

About

Yiming Ding’s research lies at the intersection of reinforcement learning, robotics, and representation learning, with a focus on enabling machines to learn complex behaviors from limited data. In their highly cited work, “Goal-Conditioned Imitation Learning” (2019, 32 citations), Ding tackled the fundamental challenge of reward design in reinforcement learning—particularly in robotics, where defining task success is often computationally expensive or ambiguous. By proposing a framework that leverages goal-conditioned imitation, they offered a practical pathway to bypass handcrafted reward functions, making RL more accessible for real-world robotic control. Building on this, Ding’s “Mutual Information Maximization for Robust Plannable Representations” (2020, 7 citations) extended their impact into the domain of high-dimensional state spaces. Here, they introduced a method to learn compressed, robust representations that improve planning and sample efficiency—a critical step toward deploying robots in unstructured environments. Ding’s work is notable for bridging the gap between model-free and model-based approaches, offering principled solutions to long-standing bottlenecks in robot learning. Their contributions continue to inspire researchers seeking to build more autonomous, data-efficient robotic systems.