Yihang Yao
Papers
2
Total Citations
8
H-Index
2
About
Yihang Yao is a rising researcher whose work bridges the critical intersection of safe artificial intelligence and embodied robotics. His primary research areas include safe reinforcement learning (RL), human-robot interaction, and computer vision for autonomous systems. Yao’s most significant contribution is the development of comprehensive benchmarks for offline safe RL, a field dedicated to training AI agents that can learn from static datasets without violating safety constraints—a crucial step toward deploying reliable AI in real-world environments like autonomous driving and healthcare. His 2023 paper on this topic has already garnered 5 citations, establishing a foundational resource for the community. In parallel, Yao has advanced practical robotics through an innovative human-robot interaction approach that integrates eye-tracking for in situ image acquisition and semi-automatic annotation. This method dramatically reduces the labor of training instance segmentation models for open scenes, achieving 3 citations and demonstrating his commitment to solving real-world deployment challenges. By combining rigorous benchmarking for safe learning with efficient data collection for perception, Yao is shaping a future where robots can learn both safely and effectively in unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1Datasets and Benchmarks for Offline Safe Reinforcement Learning5 citations · 2023
- 2