Generate and Revise: Reinforcement Learning in Neural Poetry | IEEE Conference Publication | IEEE Xplore