Multimodal Learning of Keypoint Predictive Models for Visual Object Manipulation | IEEE Journals & Magazine | IEEE Xplore