CHORROBOT: framework to learn CHallenging Operations

This lecture presents the CHORROBOT framework, which enables learning challenging operations using reactive robot control. The framework involves motivation, demonstrations, evaluation, execution, and generalization, focusing on contact points, ground truth, position, force, orientation, moments, and speed. It emphasizes the adaptability of the kinematic model, control specification, and hyperparameters. The system's predictability, flexibility, and composability allow for easy deployment of similar tasks, requiring minimal examples for learning. The lecture concludes by highlighting the efficiency of the system compared to neural networks, thanking the audience for their attention and providing additional information for those interested.