Times Haiku: when computers write poetry
Inventing an algorithm that can find haikus within the New York Times is impressive, but it also shows that – at least for now – creativity still requires the human touch.
Jacob Harris, a senior software architect at The New York Times, is the man behind Times Haiku – a blog of serendipitous poetry found by an algorithm in the publication’s website. On the blog, he explains how the algorithm finds phrases in stories that will form haikus:
“It scans each sentence looking for potential haikus by using an electronic dictionary containing syllable counts. We started with a basic rhyming lexicon, but over time we’ve added syllable counts for words like “Rihanna” or “terroir” to keep pace with the broad vocabulary of The Times.”
So that the poems retain their visual integrity when shared online, Jacob and his team post them as images. In the background of each poem is a pattern of coloured lines generated by the computer according to the metre of the poem’s first line.
Poets shouldn’t worry too much about redundancy just yet, though. Jacob acknowledges that a computer lacks one important quality: it has no aesthetic sense. It takes a team of human reporters to spot the best haikus found by the algorithm and post them to the blog.