Policies trained from data instead of hand-written controllers.
Papers and code on using LLMs and VLMs in robotics.
