Can a Machine Write a Sonnet That Is as Good as a Human's? We're About to Find Out

0
0
Turing Tests in Creative Arts
Dartmouth College
Since it was conceived in 1950, the Turing Test – named for Alan Turing, saint of “The Imitation Game” – has been the standard method for surveying counterfeit consciousness: Machines are judged on how well they display clever conduct, more often than not in discussion or amusement playing, that to a human audience or eyewitness would be undefined from that of a genuine individual.
The previous summer, two educators at Dartmouth College proposed an innovative variety: the Turing Tests in Creative Arts, testing members to submit calculations that can produce human-quality craftsmanship.
“In particular,” Dan Rockmore (an educator of math and software engineering) and Michael Casey (a teacher of music and software engineering) write in a paper that talks about the task, “we inquire as to whether machines are fit for creating poems, short stories, or move music that is undefined from human-produced works, however maybe not yet so progressed as Shakespeare, O. Henry or Daft Punk.”
The opposition has three sections: DigiLit, where the test is making a New Yorker-level short story; PoetiX, where the item should be a 14-line work in measured rhyming; and AlgoRhythms, where the PC needs to make a 15-minute movie set. In all cases, the product will be granted a “seed” – a verbal picture in the artistic challenge and a solitary track of music for the move. Coordinators will blend the sections in with human-created work. A board of artistic judges will be requested that make sense of which sonnets and stories were composed by machines; for the music, judges will be more understudies. A champ is any PC passage that tricks the judges into speculation its maker was still alive.
The outcomes will be declared May 18 at Dartmouth’s Digital Arts Exposition.

© 2016 The Washington Post

LEAVE A REPLY