Communications of the ACM
Scientific Sonnets: Duke Team Wins Competition for Poetry-Generating Algorithm
A team of Duke University researchers won the 2018 PoetiX Literary Creative Turing Test contest for an algorithm that composed the most "human-like" sonnets. The winning program was designed to select the rhyming end-words for a 14-line sonnet before filling in the rest of the syllables by writing backward. The team added more rhythmic automated punctuation positioning and fundamental rules to prevent certain part-of-speech errors.
"We collected many [part-of-speech] sequences that could not occur, then used this knowledge to ensure that word sequences did not violate these basic rules," the team wrote in a paper describing their algorithm.
The team also employed machine learning to teach the algorithm poetry as it was fed a collection of sonnet-like texts. "There's a lot of synergy between the aspects of the poetry that humans can generate and the other aspects that the computer can generate," said Duke's Cynthia Rudin.
From The Chronicle
View Full Article
Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA
No entries found