Yishu Miao

Papers

Total Citations

H-Index

About

Yishu Miao’s research centers on reinforcement learning for robotics, with a particular focus on enabling autonomous navigation in complex, real-world environments. His major contribution is the development of a novel framework that addresses the critical challenges of sparse rewards and high environmental variation, which often cause standard algorithms like Deep Deterministic Policy Gradient (DDPG) to suffer from high variance and instability. By introducing stochastic guidance into the learning process, Miao’s work provides a more robust and sample-efficient approach for training robot navigation policies. His most-cited paper, "Learning With Stochastic Guidance for Robot Navigation" (2020), has garnered 46 citations, reflecting its impact on the field. This work, along with an earlier conference version (2018), demonstrates his sustained effort to bridge the gap between reinforcement learning theory and practical robotic applications. Miao’s research is particularly valuable for students and engineers working on autonomous systems, offering a principled method to improve the reliability of learning-based controllers in unpredictable settings.