A Fuzzy Reasoning Process for Conversational Agents in Cognitive Cities

Abstract

Facing the challenges in a city that is to be understood as a complex construct, this article presents a solution approach for the further development of existing conversational agents, which should be used particularly in cities, for instance, as a source of information. The proposed framework consists of a fuzzy analogical reasoning process (based on structure-mapping theory) and a network-like memory (i.e., fuzzy cognitive maps stored in graph databases) as additions to the general architecture of a chatbot. Thus, it represents a concept of a global fuzzy reasoning process, which allows conversational agents to emulate human information processing by using cognitive computing (consisting of soft computing methods and cognition and learning theories). The framework is already in the third iteration of its development. Three experiments were conducted to examine the stability of the theoretical foundation as well as the potential of the framework.