The reviewers agree that the paper addresses an interesting problem of active meta-learning and provides a novel solution.