EIS

Team

Aaron A. Reed

Advisors

Jim Whitehead, Noah Wardrip-Fruin

 

Lab

Expressive Intelligence Studio

Website

http://almostgoodbye.textories.com/

Overview

 Almost Goodbye is an experiment in minimalist procedural content generation for interactive narratives. It does not try to generate a whole story or plot points from scratch, but instead asks what is the minimum amount of procedural generation that can be added to a hand-authored story to produce something both computationally interesting but still narratively sound. The resulting narrative, about a scientist leaving Earth forever and saying her final goodbyes, generates “satellite” sentences that color the narrator’s description and perception of her conversations based on the choices made by the player in prior conversations and other player-influenced contextual cues.

Expressive Intelligence Studio

and

Center for Games Science

 

Team

UCSC: Adam Smith

UW: Erik Andersen, Yun-En Liu, Eric Butler, and the Refraction Team

 

Videos

Early level generation prototype

 

Website

games.cs.washington.edu/refraction/

Overview

Refraction is an online puzzle game for teaching fractions developed by the Center for Game Science at the University of Washington. In the game, the player builds a network of devices that split and recombine lasers (dividing and adding their fractional power levels) in order to power disabled spaceships and save the animals inside. Along with the mathematical challenge of each level, the player must use devices to bend lasers around obstacles and to form the right beams from the right directions. Puzzles generally have many solutions, but all of them involve the target concepts. The integration of mathematical and spatial reasoning into every puzzle yields an engaging experience even for those who are competent with the game’s target fraction concepts: equal partitioning, addition, multiplication, mixed numbers, improper fractions, and common denominators.

Expressive Intelligence Studio

 

Team Members

Brandon Tearse

Peter Mawhorter

 

Advisors

Michael Mateas

Noah Wardrip-Fruin

 

Website, with the software available for download:

http://minstrel.soe.ucsc.edu/

Overview

Almost two decades ago, Scott R. Turner created MINSTREL, a program that could creatively manufacture cohesive stories without the need for human input. Although these stories contain diction akin to the writing of a child, they are nonetheless complex at many levels. MINSTREL was able to plan and achieve goals such as dramatic rise and fall, realistic human emotion, changing mental states, generation and followthrough with a final moral in mind, and generation of novel content. Unfortunately, MINSTREL is currently unavailable and was written in an old variant of LISP. Minstrel Remixed is an attempt to recreate MINSTREL using Scala. In addition to recreating MINSTREL, Minstrel Remixed attempts to surpass its predecessor in a handful of ways. Since Minstrel needs some sort of information to manipulate through its internal processes, it is limited by the size and variety of its internal libraries. We aim to allow users from the internet to generate information to enrich Minstrel's understanding of the world, allowing for rapid content generation and thus a quick growth in what Minstrel is able to come up with. In addition, since Minstrel Remixed runs much faster than the original, it is designed to accept user-generated story modifications.

This material is based upon work supported by the National Science Founcation under Grants No. 1048385 and 0747522. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.