如何看待人工智能写诗（AI写诗）？

近来的机器人作诗系统应该都脱胎于

这篇论文。就公开的系统来说，前有

，后有题主分享的编诗姬。

算法而言，机器作诗无非是基于字符级RNN的各种变形，加之格律的约束。答主之前也稍微有玩过，贴两首自己训练的模型做出的七言诗：

用的算法只是两层的字符级RNN，看起来就可以像模像样的作诗了。

然而机器人作诗的问题，仍然是有诗感而无诗意。从题主贴出的这首诗来看，只能说乍一看像一首诗，但是诗歌想要表达什么？句与句之间有怎样的演进？终究不禁细想。RNN虽然能通过学习一些长程的字词搭配，保持一首诗风格上，或者诗意上的统一，但是仍然不能说学习到了真正的语义。

说到底，计算机能够像模像样地写诗，还是利用了诗歌本身的模糊性和普通人对诗的不熟悉。当计算机开始写作文、写故事的时候，这种语义上的缺点暴露无疑。

从去年的工作来看，最好的通用故事生成器应该是

，通过训练两组RNN，去encode小说中的句子，再decode出上句和下句。在论文最后，他们给出一个模型生成的故事是这样的：

she grabbed my hand . come on . she fluttered her bag in the air . i think we ’re at your place . i ca n’t come get you . he locked himself back up . no . she will . kyrian shook his head . we met ... that congratulations ... said no . the sweat on theirfingertips’s deeper from what had done it all of his flesh hard did n’t fade . cassie tensed between her arms suddenly grasping him as her sudden her senses returned to its big form . her chin trembled softly as she felt something unreadable in her light . it was dark . my body shook as i lost what i knew and be betrayed and i realize just how it ended . it was n’t as if i did n’t open a vein . this was all my fault , damaged me . i should have told toby before i was screaming . i should ’ve told someone that was an accident . never helped it . how can i do this , to steal my baby ’s prints ?

仍然是一个有小说感，但是nonsensical的东西。

不过基于Skip-Thought Vectors的一个看图讲故事的demo还挺像回事的，题主可以关注

：

总的来说，计算机创作在近期来看，仍然是篇幅越短、含义越模糊（越需要人脑补），越能够糊弄住人。未来只能期待DL在更好的记忆和处理语义上取得突破。