AI and poetry创作诗歌对于人来说都很难，对于机器来说就更是难上加难了。虽然机器与人相比，在一些方面有着先天的优势，例如，机器有无穷无尽的词汇库可供选择，机器可以很容易的解决对仗、平仄和押韵等问题。但是，真正的诗歌是有灵魂的，诗歌本质上是在传达诗人的思想。机器自动创作在主题控制方面很弱，很难让整首诗都围绕一个统一的主题来生成。目前文本生成的模型（例如GPT-2）等，已经取得了不错的效果，例如华为已经开发了“乐府”系统等。
We at first define a unified format for formulating all types of training samples by integrating more detailed form information, present a simple form-stressed weighting method in GPT2 to strengthen the control to the form of CI
MM '18: ACM Multimedia Conference
Republic of Korea
October, 2018, (2018): 783-791
Automatic generation of natural language from images has attracted extensive attention. In this paper, we take one step further to investigate generation of poetic language (with multiple lines) to an image for automatic poetry creation. This task involves multiple challenges, in...
As the first attempt on unsupervised style disentanglement, we find Chinese poetry generation is an ideal choice for evaluation because we can better focus on stylistic generation rather than dealing with genre requirements
As an exquisite and concise literary form, poetry is a gem of human culture. Automatic poetry generation is an essential step towards computer creativity. In recent years, several neural models have been designed for this task. However, among lines of a whole poem, the coherence ...
Semantic keywords are extracted as the skeleton of a poem, where a topic memory network is proposed that can take as many keywords as possible and dynamically select the most relevant ones to use during poem generation