This workshop explores new approaches to mobile manipulation with an emphasis on the relationship between machine learning and successful interaction in human environments. Autonomous manipulation in human environments is challenging because of the associated high dimensional state space and its inherent uncertainties. It requires perceptual and manipulation skills which are robust against sparse, incomplete and noisy information.
In such environments, leveraging past experience, oftentimes resulting from the robot's own interactions, promises an increased robustness and reliability. Therefore, these challenges naturally connect autonomous manipulation to machine learning. This workshop will explore these issues through a discussion of how the intersection of perception, learning, and manipulation are required to create flexible robust systems.