Fetching the paper…

SE3-Pose-Nets: Structured Deep Dynamics Models for Visuomotor Planning and Control · Around