Creative Computers and Data Mining

(invited lecture for DS 2013)

Author: Hannu Toivonen

Affiliation: Department of Computer Science
University of Helsinki
Finland

Abstract. In the field of computational creativity, researchers aim to give computers creative skills, such as those needed in writing poetry or composing music. Obviously, an agent needs to know the field in which it operates. This is where data mining has great potential: making creative agents adaptive to various fields and genres by automatic discovery of relevant information from existing creative artifacts. We give several examples of how verbal creativity can benefit from data mining of existing text corpora.

On the other hand, computational creativity tools allow a whole new approach to data analysis. In this ``Affective Data Analysis,'' the goal is to turn data into a subjective, esthetic experience by automatic or semiautomatic creation of a novel artifact using the user's data as inspiration. This is in strong contrast with traditional data analysis methods that emphasize cold facts instead of warm feelings. We illustrate this idea with musicalization of sleep measurements and chat discussions.