Semantic-map-based analysis of insight problem solving

Abstract

Abstract Intelligent agents and co-robots, or cobots, become increasingly popular today as assistants of creators of arts. They can be also expected to become popular in the near future as assistants in other creative work, including research and, in particular, insight problem solving. Arguably, the success of such tools is linked to their social-emotional intelligence. This work continues the previous study of the concept of a cobot-assistant of an insight problem solver (Kuznetsova & Samsonovich, Procedia Computer Science 123:258–64, 2018). The concept is based on a semantic map, that represents cognitive and emotional states and attitudes. The previously constructed semantic map is used here to study the dynamics of human cognitive and emotional states during insight problem solving. It is found here that subjects solving an insight problem become more excited and try to take more radical approaches, and also become slightly more confident, as they approach the insight moment, regardless of whether they actually succeed. This finding should help to formulate rules and cognitive schemas to be tested in future studies of the proposed insight problem solver assistant. The expected impact is on the anticipated emergence of general-purpose virtual cobots-assistants, unleashing higher creativity in their users.