Pharaoh: Conceptual Blending of Cognitive Scripts for Computationally Creative Agents

Improvisational acting is a creative group performance where actors co-construct stories on stage in real-time based on actors’ perceptions of the environment. The Digital Improv Project has been engaged in a multi-year study of the cognitive processes involved in improvisational acting. This better understanding of human cognition and creativity has led to formal computational models of some aspects of our findings. In this work, we consider enriching AI improv agents with the ability to improvise new nontraditional scenes based on existing social cognitive scripts. This paper shows how the use of Pharaoh -a context based structural retrieval algorithm for cognitive scriptsand simple blending rules can help digital improv agents to create new interesting scenes. The paper also provides an illustrative example at the end.