Minstrel remixed: Procedurally generating stories

Abstract

We are recreating, investigating, and defining new uses for one of the most influential artificial intelligence projects of the past 25 years: Scott Turner's Minstrel, which is regarded as a landmark in both the story generation and computational creativity communities. We compare our new system, Minstrel Remixed, with the implementation of the original, and discuss the various additions made during our rational reconstruction which facilitate investigations into the inner workings of the system. In conclusion, we evaluate the performance of Minstrel Remixed and determine that its results are quite close to those of the original. Copyright © 2010, Association for the Advancement of Artificial.