Believable Non-Player Characters (NPCs) are a crucial component of narrative-driven games. An important aspect of believable characters is their contextually-relevant reactions, which is often driven by emotion in humans. The plausibility of NPCs’ “emotions” partly depends on their psychological validity. A Computational Model of Emotion (CME), grounded in emotion theories and/or models from psychology, is an attractive solution. Play-testing believability can be expensive. Theory-independent acceptance tests offer a cheaper pre-test of a CME’s output against expected responses. We propose the first methodology for creating verifiable, replicable, and reusable test cases with known believable characters from professionally-created stories.

Citation (ACM)

Geneva M. Smith and Jacques Carette. 2023. Building Test Cases for Video Game-Focused Computational Models of Emotion presented at the Interdisciplinary Design of Emotion-Sensitive Agents Workshop (IDEA 2023). May 30, 2023. Held in conjunctions with the 22nd International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2023), May 29–June 2, 2023, London, England.