Larry Yang

Papers

Total Citations

244

H-Index

About

Larry Yang is a robotics and machine learning researcher whose work sits at the intersection of deep learning, reinforcement learning, and autonomous robotic control. His most influential contribution, "End-To-End Robotic Reinforcement Learning without Reward Engineering" (2019), addresses one of the central challenges in real-world robot learning: enabling agents to acquire complex behaviors directly from raw sensory inputs like camera images, without relying on hand-crafted reward functions that are difficult to design in practice. This work has accumulated over 200 citations, reflecting its significant impact on the robotics and AI communities. Yang's earlier research, "GPLAC: Generalizing Vision-Based Robotic Skills Using Weakly Labeled Images" (2017), tackles the equally critical problem of generalization — training robotic sensorimotor policies that can transfer to visually diverse, previously unseen environments using weakly labeled data. Together, these contributions demonstrate a coherent research vision: making robotic learning more practical, scalable, and deployable in unstructured real-world settings. Yang's work is particularly valuable for researchers seeking to bridge the gap between simulation-based training and robust real-world robot performance, pushing the boundaries of what autonomous systems can learn from experience alone.