Fuzzy Analogical Reasoning

Automatically understanding the meaning of words is a powerful technique for interpreting perceptions expressed in natural language. However, in order to correctly react to the understood information, reasoning comes into play, enabling the inference of causes or consequences associated with expressed perceptions. This process holds great potential for artificial entities to better adapt interactions according to user needs and desires. Indeed, by mapping known situations (base) to unknown ones (target), artificial entities can extrapolate knowledge about expected reactions from users in specific contexts to react in analogous ways in unfamiliar situations where user expectations are unknown. In this chapter, an automated approach to reasoning by analogies, building on the theory of computing with words, is presented. This approach promotes a robust interaction, aligned with the core principles of phenotropics, emphasizing the importance of being an ever-improving guesser rather than a perfect decoder. Concretely, a framework for fuzzy analogical reasoning is introduced, followed by an implementation allowing to showcase its potential. An evaluation of the proposed fuzzy analogical reasoning framework is represented by an analysis of the Strengths, Weaknesses, Opportunities, and Threats (SWOT) identified in a workshop with experts, showcasing the great potential of this method.