In-context learning helps robots understand and perform new tasks
In-Context Learning for Robots: Methods and Applications
RoboticsMachine Learning
Summary
Robots need to figure out what to do when given new tasks and then do them correctly. The authors review different ways robots can learn from examples or interactions without changing their internal brains during use. They group the approaches by how the robot connects examples to actions and explain how these methods handle changes in objects or environments. The review also looks at how robots remember and reuse experience to improve learning over time.
What this means in practice
- •For robot developers: Design robots that quickly adapt to new tasks by following example-based learning methods that do not require re-training internal parameters.
- •For automation engineers: Create flexible robotic systems that maintain taught task constraints in changing physical environments using context-based execution strategies.
A survey. It maps existing work.
Authors
Haojian Huang, Zexi Li, Junhao Guo, Yehang Zhang, Wenxuan Peng, Bohan Zhou, Weilin Ruan, Leyi Wu, Chenxu Wang, Jianchong Su, Binghui Xie, Wosong Chen, Yingjie Xu, Tianhao Zhou, Suzeyu Chen, Pukun Zhao, Jiaqi He, Xinyi Li, Runze Li, Peiran Dong, Shaoxiang Dang, Jing Huang, Yingbing Chen, Yifan Chang, Tianyi Zhang, Shiyuan Deng, Haozhi Wang, Yangkai Wei, Wenqian Li, Han Yang, Kaiwen Zhou, Huaping Liu, James Cheng, Rui Shao, Donglin Wang, Yaochu Jin, Jianye Hao, Ying-Cong Chen, Yinchuan Li
Abstract
General-purpose robots must infer what a new task requires and translate that understanding into appropriate physical action. In-context learning (ICL) for robots supports this process by using demonstrations and interaction to direct existing competence with neural parameters held fixed during deployment. We organize this literature review around the interfaces connecting contextual evidence to execution, distinguishing four families: context-conditioned policies, geometric demonstration transfer, world-model-based control, and skill- and agent-based execution. Comparing these interfaces clarifies their transfer assumptions and the roles of training, correspondence, and memory in making context useful. Across manipulation and navigation, we examine how these mechanisms preserve taught requirements as objects, environments, and execution conditions change. This analysis links method design to evaluation practices that distinguish responsiveness to teaching, physical transfer, and benefits from retained experience. The resulting agenda connects compositional task acquisition and faithful transfer with physical recursive self-improvement, in which experience improves the ability to learn subsequent tasks.