Hanjun Dai

Papers

Total Citations

H-Index

About

Hanjun Dai is a leading researcher at the intersection of machine learning, reinforcement learning, and generative AI. His work is distinguished by a bold vision: building universal agents that can generalize across diverse tasks, often by leveraging the power of large language models and generative systems. A key contribution is his pioneering work on "Learning Universal Policies via Text-Guided Video Generation," which reimagines planning as a video generation problem, enabling agents to solve novel tasks without task-specific training. This highly influential paper (29 citations) exemplifies his ability to bridge language, vision, and decision-making. Beyond this, Dai has made foundational contributions to graph neural networks and combinatorial optimization, with his work on structure-aware learning and scalable inference methods accumulating thousands of citations. His research consistently pushes the boundaries of what AI agents can learn, from mastering complex board games to generating coherent long-horizon plans. For students and researchers, Hanjun Dai represents a rare blend of theoretical depth and practical impact—a thinker who uses generative models not just to create, but to reason and act.