In this workshop, we aim to bring together machine learning and robotics researchers who work at the intersection of these fields... Dexterous manipulation is a grand challenge in robotics and a necessary step in performing many manipulation tasks such as ones involving tool use. Today, many different approaches are being explored to tackle this problem: underactuated soft hands, fully actuated rigid hands, model-based control, reinforcement learning, imitation learning, experiments with many different tactile sensors, simulators for contact-based interactions, and many others. Furthermore, there is little convergence on dexterous manipulation tasks being solved by different research groups. Due to the diversity of approaches and lack of dialogue between different sub-communities, quantifying progress and identifying challenges and opportunities is a formidable issue. We believe that big jumps in dexterous manipulation will result from interdisciplinary work spanning the subfields of..