{"id":60339,"date":"2021-11-20T03:42:59","date_gmt":"2021-11-20T08:42:59","guid":{"rendered":"https:\/\/world.edu\/?p=60339"},"modified":"2021-11-29T22:34:39","modified_gmt":"2021-11-30T03:34:39","slug":"can-artificial-intelligence-create-a-new-kind-of-poetry","status":"publish","type":"post","link":"https:\/\/world.edu\/can-artificial-intelligence-create-a-new-kind-of-poetry\/","title":{"rendered":"Can artificial intelligence create a new kind of poetry?"},"content":{"rendered":"

\"\"
\nThere are hundreds of poetry generation programs on the Internet – that is, systems capable of producing poetry automatically – but what can they be used for?\u00a0Do these programs have an interest, beyond that of satisfying their designer?<\/span><\/p>\n

We have heard a lot about GPT2 or GPT3, these huge computer programs capable of producing very realistic texts, and even poetry.\u00a0GPT2 and GPT3 are in fact “models”, species of knowledge bases, fed by billions of sentences and texts gleaned from the Internet, and “digested” in order to be able to produce new texts, inspired by old texts, but in very different times.<\/span><\/p>\n

Add constraints<\/strong><\/h4>\n

To produce poetry, it is “enough” to add constraints: monitor the rhymes and the length of the lines, respect the overall structure, the absence of repetition in rhymed position, etc.\u00a0Automatic generation systems are having some success (we find a phenomenal number of them on the Internet) because the task is fun, playful, but also complex if we want to produce texts with meaning (and even more if we want to control what is said).<\/span><\/p>\n

We therefore have here an ideal setting for experimenting (often without funding constraints): the generation of poetry is often a hobby and a hobby, for the researcher as well as for the enlightened amateur.<\/span><\/p>\n

A question arises, however.\u00a0Are these systems worthy of attention?<\/span><\/p>\n

On the literary level, most systems are still, it must be admitted, quite rudimentary and have difficulty in competing with Baudelaire or Rimbaud.\u00a0The most advanced are however impressive, and it is above all the training base that plays a crucial role (that is to say all the texts which have enabled the system to learn. We can indeed provide a system dataset reduced but specialized (works by poets of the XIX\u00a0<\/span>th<\/span><\/sup>\u00a0century for example) to “specialize” a system, adapt cost. We can then get systems who write paragraphs on the way de Balzac, or poetry in the style of Baudelaire.<\/span><\/p>\n

However, it should be noted that the results appearing in the press (whether it be the generation of prose or poetry) are often the fruit of multiple trials, or even the fruit of post-editing work on the part of the journalist.<\/span><\/p>\n

The Oupoco project<\/strong><\/h4>\n

The\u00a0<\/span>Oupoco<\/span><\/a>\u00a0project\u00a0(Ouvroir de po\u00e9sie combinatoire) that we developed with a team from\u00a0<\/span>the LATTICE laboratory<\/span><\/a>\u00a0, had a more modest goal.\u00a0Like\u00a0<\/span>Queneau’s<\/span><\/a>\u00a0experience\u00a0in\u00a0<\/span>One Hundred Thousand Billion Poems<\/span><\/em><\/a>\u00a0, our primary ambition was to produce billions of poems simply by recombining verses from a representative French poetic corpus.<\/span>
\n

\"\"

One hundred thousand billion poems , Raymond Queneau, 1961. M\u00e9dard Museum<\/p><\/div><\/p>\n

To this end, we have assembled a database of more than 4000 sonnets by authors ranging from the beginning of the 19\u00a0<\/span>th<\/span><\/sup>\u00a0to the beginning of the 20\u00a0<\/span>th<\/span><\/sup>\u00a0century.\u00a0While all of Queneau’s lines rhyme together, we had to automatically determine the rhyme of each line in order to be able to produce rhyming poetry.\u00a0From the start, the project was therefore more of an analysis project than a generation project (as shown in\u00a0<\/span>this video<\/span><\/a>\u00a0.<\/span>
\n