Search

Acquiring multiple skills has commonly involved collecting a large number of expert demonstrations per task or engineering custom reward functions. Recently it has been shown that it is possible to acquire a diverse set of skills by self-supervising control on top of human teleoperated play data. Play is rich in state space coverage and a policy trained on this data can generalize to specific tasks at test time outperforming policies trained on individual expert task demonstrations. In this work, we explore the question of whether robots can learn to play to autonomously generate play data that can ultimately enhance performance. By training a behavioral cloning policy on a relatively small quantity of human play, we autonomously generate a large quantity of cloned play data that can be used as additional training. We demonstrate that a general purpose goal-conditioned policy trained on this augmented dataset substantially outperforms one trained only with the original human data on 18 difficult user-specified manipulation tasks in a simulated robotic tabletop environment. A video example of a robot imitating human play can be seen here: https://learning-to-play.github.io/videos/undirected_play1.mp4

Coherent population trapping (CPT) provides a highly sensitive means for probing the energy level structure of an atomic system. For a nitrogen vacancy center in diamond, the CPT offers an alternative to the standard optically-detected magnetic resonance method for measuring the hyperfine structure of the electronic ground states. We show that the nuclear spin dependent CPT measures directly the hyperfine splitting of these states due to the 14N nuclear spin. The CPT spectral response obtained in the presence of a strong microwave field, resonant or nearly resonant with a ground state spin transition, maps out the dynamic Stark splitting induced by the coherent spin excitation.