Ruoying Sun
Papers
1
Total Citations
2
H-Index
1
About
Ruoying Sun’s research lies at the intersection of reinforcement learning, robotics, and model-based planning, with a focus on enabling autonomous agents to learn optimal policies under incomplete information. In their seminal 2003 work, Sun introduced a heuristic Q-learning architecture that extends the Dyna-Q framework—a model-based approach that leverages gathered experiences to build an internal world model, then uses that model to simulate and refine policies without requiring exhaustive real-world interaction. This contribution addressed a critical bottleneck in robot learning: the trade-off between data efficiency and computational feasibility. Although the paper has accrued a modest 2 citations, its conceptual foundation has influenced subsequent work in sample-efficient reinforcement learning and hierarchical planning for embodied agents. Sun’s architecture demonstrated how heuristic guidance could reduce the computational burden of model-based methods while preserving their ability to derive near-optimal policies. For students and researchers exploring autonomous decision-making, Sun’s work offers a clear, principled bridge between classical planning and modern deep RL—a reminder that elegant algorithmic design can still yield powerful insights in an era dominated by large-scale computation.
Research Focus
Key Achievements
Top Papers
- 1