Skip to main content
[an error occurred while processing this directive]

[an error occurred while processing this directive]
HomeProjects >

Turing Tests in the Creative Arts

2016 turing test

 

Sponsored by The Neukom Institute for Computational Science at Dartmouth College, the Neukom Institute Prizes in Computational Arts aim to inspire innovations in computational methods that generate artistic products, such as literary, musical, and visual art.

Neukom Institute runs three different competitions: DigiLit for short stories, PoetiX for sonnets, and AlgoRhythms for dance music DJ sets. All entries are in the form of computer code that generates novel works in one of these three creative modes.

The Neukom Turing Test website is here.

 

2016 Winners

 

AlgoRhyhtms

First Prize (joint): Jaume Parera and Pritish Chandna (Music Technology Group, Upf, Barcelona, Spain) - DJ Codo Nudo, 39% of audience voted set as "Human": $2000
First Prize (joint): CJ Carr and Zack Zukowski (Medford, MA, USA) - DadaBots - Jungle Bot, 38% of audience voted set as "Human": $2000
Second Prize:Gerard Roma (Surrey, UK), Dub Life, 28% of audience voted set as "Human": $1000

 

Two evaluation rounds:

There were three Human DJs and three Algorithm Finalists - all hidden from view while audience danced and voted to 10 mins of music from each. We counted the votes: Human versus Algorithm. Winning the turing test would mean 50% or more votes for Human. One of the algorithms achieved this in the on-line vote. After the dance party, overall, two algorithms achieved about 40% Human votes, and they were highly sophisticated technical achievements, using cutting-edge methods in music information research (MIR), so they were given joint first place. 

Examples:
Try it for yourself, take the on-line poll: On-line listening poll: Human vs. Algorithm?

 

DigiLit

First prize: no first prize awarded.
Second prize:Judy Malloy: second prize for her algorithm, which creates variations on a story called “Another Party in Woodside”, each constructed by randomly arranging pre-written sentences into paragraphs; $1000. 

Judges:

  • David Cope – Algorithmic music pioneer
  • Lynn Neary - NPR Correspondent
  • David Krakauer - President of The Santa Fe Institute

Evaluation Details:
Each judge received 11 short stories – 8 “human generated” and 3 “machine generated” (one from each of the two machines)
“Passing the Turing Test” – means a given algorithm (machine having a majority of the judges rating its entries as “human”
Results: No machine won, but one submission generated by Toksu and Ibrahim on the seed “thesaurus” “fooled” one of the judges!

 

DigiLit

First prize: no first prize awarded.
Second prize:Judy Malloy: second prize for her algorithm, which creates variations on a story called “Another Party in Woodside”, each constructed by randomly arranging pre-written sentences into paragraphs; $1000. 

Judges:

  • David Cope – Algorithmic music pioneer
  • Lynn Neary - NPR Correspondent
  • David Krakauer - President of The Santa Fe Institute

Evaluation Details:
Each judge received 11 short stories – 8 “human generated” and 3 “machine generated” (one from each of the two machines)
“Passing the Turing Test” – means a given algorithm (machine having a majority of the judges rating its entries as “human”
Results: No machine won, but one submission generated by Toksu and Ibrahim on the seed “thesaurus” “fooled” one of the judges!

 

Poetix

First Prize: Marjan Ghazvini Nejad, Xing shi, Yejin Choi, and Kevin Knight (USC/ISI, USA) extraordinary programming effort that is a marriage of multiple AI techniques: $3000. See an example poem...
Second Prize: Andrea Gagliano, Emily Paul, Kyle Booten, and Marti Hearst. (UC Berkeley, USA) "Pythonic Poet": $1000

Judges:

  • Luke Menand – Pulitzer Prize-winning author and New Yorker contributor
  • Robert Siegel - NPR Correspondent
  • Ray Monk - Award-winning biographer and philosopher

Evaluation Details:
Each judge received 10 sonnets – 6 “human generated” and 4 “machine generated” (two from each of the two machines)
"Passing the Turing Test” – means a given algorithm (machine having a majority of the judges rating its entries as “human”
Results: All judges had perfect scores – so all machines were identified.

 

 

Press

New York Times: Dartmouth Contest Shows Computers Aren't Such Good Poets [New York Times, AP]

The Washington Post: Can machines create quality art and creative literature? We’re about to find out.

Dartmouth Now: Can Robot Artists Create Human-Quality Work? Not Yet

Tribtown.com: Computers are pretty good at stocking shelves and operating cars, but are not so good at writing poetry

U. S. News: Dartmouth contest shows computers aren't such good poets

ColumbusCEO: Dartmouth contest shows computers aren’t such good poets

The Christian Science Monitor: Computer or poet? Humans win this round of poetry contest

 

 

 

 

Last Updated: 5/8/17