This project explores the utility of "computational creativity" as
an organizing theme for interdisciplinary curricula in computer
science, media arts, and cognitive science. A computational creativity
system is a computational system that simulates the cognitive processes
involved in human creativity, or that implements an algorithm that
meets the criteria for creativity even if it is not cognitively based.
Many computational models of creativity have been described in the
literature, and simple versions of many of these models have been
implemented in software. One of the co-PIs of this proposal has used
the development of such models as an organizing theme for the
instruction of introductory computer science for 15 years; the present
proposal expands the application of this theme to broader,
interdisciplinary curricula in computer science, media arts, and
cognitive science. The goal is to develop, in joint work among computer
science and media arts faculty and undergraduate students,
computational creativity systems that are useful in real artistic
practice. These systems will be extended and used in computer science,
media arts, and interdisciplinary courses. The curriculum
integrates research and teaching while drawing new populations of
potentially creative and diverse students into the study of computer
science.
Highlight: Igor: An Open-Ended
Creative Variation-Generation System
Control flow of the Igor creative variation-generation system
prototype
developed in
the Hampshire College CreativeIT project, “The
Computational Creativity Curriculum.”
NSF-funded researchers Lee Spector and Chris Perry, working at
Hampshire College with undergraduate students Adam Sax and Brian
Cipriano, have developed a prototype software tool that uses
computational intelligence techniques to assist creative artists in the
development of professional-grade computer animation. Their system,
called Igor, provides artist-users with a selection of
computer-generated variations of an artist-generated element such as an
image, texture or palette. Igor uses a genetic algorithm to generate
these variations, and although human artists are ultimately in control
Igor may make contributions to the art-making process that are complex,
surprising, and arguably creative.
The Igor prototype was developed to assist users of the "Shake"
software for image compositing -- Shake is a commercial tool that is
widely used both in animation courses and in industry -- but it was
designed to be applicable, with little modification, to other
environments including Maya (for 3D modeling), Processing (for
algorithmic image generation), and Max/MSP (for sound). It is accessed
through a web interface in an ordinary browser, so that users need have
no software aside from a browser and their primary art-making tools. It
farms computationally intensive out to additional compute nodes as
needed; the prototype farms out image processing tasks to Hampshire
College's NSF/MRI-funded cluster.
Igor is the first product of an NSF CreativeIT project -- "The
Computational Creativity Curriculum" -- the objective of which is to
enhance and energize the undergraduate computer science curriculum
using methods and projects from the arts. It will simultaneously
increase participation in computer science by engaging artists and
scientists in a collective enterprise based on computational models of
creative processes. Such models, and related "computational creativity"
tools, will be developed in interdisciplinary projects and courses and
used in professional-quality media arts projects.