Robot Learning from Human Videos: A Survey
Junyi Ma, Erhang Zhang, Haoran Yang, Ditao Li, Chenyang Xu, Guangming Wang, Hesheng Wang
THE PROBLEM
This paper focuses on Imitation & Reinforcement LearningImitation Learning (IL)Teaching a robot by showing it examples of how to do a task.. Comprehensive survey covering Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. Modern Robot LearningSkillA reusable behavior like grasp, push, place, or open drawer. learning from human video data. Reviews Imitation & Reinforcement LearningPolicy learningTraining a model that maps observations to actions. foundations, human-video incorporation interfaces, and a hierarchical taxonomy of Modern Robot LearningSkillA reusable behavior like grasp, push, place, or open drawer. transfer pathways (task-oriented, observation-oriented, action-oriented). Analyzes human video datasets, video generation schemes, and provides statistical trends in Robot LearningDatasetA collection of training or evaluation data. development. Read the paper by tracking the Core ConceptsTaskThe job the robot is supposed to complete, such as pick-and-place, navigation, or drawer opening. definition, the Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or data assumptions, and the evidence that supports the claimed improvement.
HOW IT WORKS
Task framing
Core method
Data and supervision
Evaluation evidence
KEY RESULTS
This survey maps how to train robots on Manipulation & TasksManipulationUsing a robot arm or hand to move or interact with objects. skills using freely available human video data instead of expensive Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. demonstrations. It catalogs three core pathways—task-level, observation-level, and action-level transfer—with concrete design patterns for each, letting you pick the right approach for your Robot LearningDatasetA collection of training or evaluation data. and learning setup.
WHY DEVELOPERS SHOULD CARE
This survey maps how to train robots on Manipulation & TasksManipulationUsing a robot arm or hand to move or interact with objects. skills using freely available human video data instead of expensive Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. demonstrations. It catalogs three core pathways—task-level, observation-level, and action-level transfer—with concrete design patterns for each, letting you pick the right approach for your Robot LearningDatasetA collection of training or evaluation data. and learning setup.
LIMITATIONS
The main limitation to check is whether the claimed behavior holds outside the paper's reported setup. That means testing across different Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. embodiments, scenes, objects, and data distributions.
WHAT COMES NEXT
The practical next step is independent reproduction with clear baselines, ablations, and stress tests. For a developer, the useful follow-up is to map the paper's Imitation & Reinforcement LearningImitation Learning (IL)Teaching a robot by showing it examples of how to do a task. assumptions onto a concrete Core ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. stack, then test the smallest version of the method that could run end to end.