Chris Xie
Papers
2
Total Citations
55
H-Index
2
About
Chris Xie is a leading researcher in robotic manipulation and model-based reinforcement learning, whose work bridges the gap between physical reasoning and autonomous decision-making. In his highly influential 2016 paper, "Model-based reinforcement learning with parametrized physical models and optimism-driven exploration" (45 citations), Xie pioneered a method that integrates model identification with model predictive control, enabling robots to learn dynamics from sparse data using a feature-based representation and a simple least-squares fitting procedure. This work introduced a principled optimism-driven exploration strategy, allowing robots to efficiently and safely interact with their environments—a foundational contribution to sample-efficient robot learning. Xie’s more recent research tackles the complex challenge of semantic object placement. In his 2021 paper, "Predicting Stable Configurations for Semantic Placement of Novel Objects" (10 citations), he developed a framework that enables robots to re-pose unseen objects in novel environments by learning both physical stability and semantic relationships. By decomposing the problem into finding physically valid locations and then predicting context-appropriate configurations, Xie’s work advances the frontier of robots operating in unstructured human spaces. His contributions are essential for building robots that can understand and interact with the world in a human-like manner, making him a key figure in modern robotics research.
Research Focus
Key Achievements
Top Papers
- 1
- 2Predicting Stable Configurations for Semantic Placement of Novel Objects10 citations · 2021