Fuzzy Cognitive Maps as a Bridge between Symbolic and Sub-Symbolic Artificial Intelligence

Maikel Leon, University of Miami, USA

Abstract

The intersection of symbolic and sub-symbolic Artificial Intelligence (AI) presents a fertile ground for innovations that combine the interpretability of the former with the learning capabilities of the latter. This paper introduces Fuzzy Cognitive Maps (FCMs) as a hybrid model that encapsulates the strengths of both paradigms, proposing them as a viable solution to the challenges of explainability and interpretability in AI systems. FCMs have emerged as a compelling framework for representing causal knowledge and facilitating decision-making processes intuitively and justifiably. FCMs can handle the inherent uncertainty and vagueness seen in real-world scenarios, thus enabling a more natural and flexible approach to problem-solving. This intrinsic flexibility, combined with the capacity for learning and adaptation derived from sub-symbolic AI, positions FCMs as an ideal candidate for applications demanding high degrees of explainability and interpretability.

Keywords

Fuzzy Cognitive Maps, Symbolic AI, Sub-symbolic AI, Explainable AI, and Interpretable AI.

Full Text : https://ijcionline.com/paper/13/13424...
Abstract URL: https://ijcionline.com/abstract/13424...
Volume URL : https://ijcionline.com/volume/v13n4

#fuzzy #explainableai #artificialintelligence #machinelearning #generativeai