DigiLit

The DigiLit prize competition encourages the creation of algorithms able to produce a "human-level" short story of the kind that might be intended for a short story collection produced in a well-regarded MfA program or a piece for The New Yorker. The Prize seeks to reward algorithms that could, for example, write stories for a creative writing class in which students are asked to submit a new short story each day.

Rules:

  1. Entries must be submitted via the competition Web site [TBA] by 11:59PM UTC, April 15, 2016.
  2. A contemporary computer should be able to run the code and generate a story of at most 7000 words within 24 hours using less than 16GB of RAM and less than 50GB of disk space. We will run entries using x86-compatible CPUs. Preferred platforms are Apple OSX 10.10, Ubuntu Linux 14.04, or Microsoft Windows 8.1 or 10.x. All libraries must be statically linked, and an executable “RunMe” provided.
  3. The algorithm/software should respond to a noun or noun phrase “prompt” (e.g., “hat”, “car keys”, “wedding”, “sorrow”, “violin case”) and be capable of producing an effectively unlimited number of original short stories.
  4. We prefer that submissions include source code, and that you provide a binary executable as stated in rule 2. The program must be completely self-contained and require no connectivity to the Internet. Programs must be capable of generating distinct short stories with high probability. We will test this requirement and then, should your entry make the list of finalists, after some unspecified number of iterations we will pick one or two of the generated works as your entry.

Prizes:

The DigiLit Short Story Turing Test: After an initial screening, the top entrants will have their stories mixed in with a collection of human generated stories. To keep with the spirit of a Turing Test, the computer-generated stories that are included will be generated by two noun prompts per algorithm and not pre-screened before inclusion. Two collections of mixed human- and computer-generated stories will be read separately by two panels, each with three judges. The judges will be asked to rate the stories as either human- or computer-generated. If a computer-generated story is scored as human by a majority of the judges on its panel, the creators will win a $5000 prize. At most one prize will be awarded to any competitor.

The DigiLit Short Story Competition: A first prize of $3000 will be awarded to the team that enters the “best” short story generating software. Two second prizes of $1000 will be given to the runner-ups. Computer-generated submissions are judged against each other according to criteria used by professional short story writers (e.g., instructions given to judges of the "Best American Short Stories" collection). Judging criteria will also include an evaluation of the submitted algorithm and code.

Note: Contestants might be interested in the lessons and resources gathered by the NaNoGenMo 2014 event.

Final Caveats:
Competition details are subject to change at any time.
Interpretation of these rules is subject to the sole and final decision of the "Turing Tests Team"