acm-header
Sign In

Communications of the ACM

ACM TechNews

Neural Network Poetry Is So Bad We Think It's Written By Humans


View as: Print Mobile App Share:
A human poet at work.

A neural network trained on thousands of lines of poetry has tried its hand at penning its own rhymes that mimic certain forms of verse. Its best efforts even fool people into thinking theyre reading the words of a human poet.

Credit: Getty Images

A neural network trained on poetry has attempted to write its own lines that mimic certain forms of verse. Its best efforts can convince people they are reading the words of a human poet.

The poetic bot was developed by Jack Hopkins while he was a researcher at the University of Cambridge. It can be programmed to write in a particular rhythm or write poems on specific themes and can be endlessly tweaked to generate various forms. For example, Hopkins says it could write about Brexit in the style of a Greek epic or rewrite parts of Romeo and Juliet while mimicking a rapper.

The AI poet was trained on more than 7 million words of 20th-century English poetry, most of it from poetry books found online.

Hopkins also instructed the neural network to keep checking to make sure some of the words in each line relate to the selected theme.

From New Scientist
View Full Article

 

Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account