Interaction-based Authoring for Scalable Co-creative Agents
Mikhail Jacob, Brian Magerko · ICCC · 2015
This article presents a novel approach to authoring cocreative systems - called interaction-based authoring that combines ideas from case-based learning and imitative learning, while emphasizing its use in open-ended co-creative application domains. This work suggests an alternative to manually authoring knowledge for computationally creative agents that relies on user interaction “in the wild” as opposed to high-effort manual authoring beforehand. The Viewpoints AI installation is described as an instantiation of the interaction-based authoring approach. Finally, the interaction-based authoring approach is evaluated within the Viewpoints AI installation and the results are discussed guiding development and further evaluation in the future.