Computing with Words 2.0: A Machine Learning Driven Fuzzy Linguistic Framework for Human-Centric AI

To address that, this paper sets out an extended Computing with Words 2.0 (CWW 2.0) system that represents combination of fuzzy linguistic modeling and machine learning to improve the capacity of AI systems to comprehend and reason with the vagueness and ambiguity of human language. Using machine learning, the framework is capable of dynamically learning membership functions, linguistic term definitions, and fuzzy inference rules by adapting to a domain-specific data, allowing higher adaptation than in traditional static fuzzy models. The ability to model uncertainty provided by fuzzy logic can be combined with data-driven learning, which provides the proposed approach with applicability in the real world with context-specific and interpretable decision support. Its architecture combines a hybrid fuzzy-statistical engine, natural language linguistic term extraction and a human-in-the-loop contextual refinement mechanism. Experimentation on the healthcare and learning datasets shows that accuracy, versatility and explainability are significantly better than that found in both traditional fuzzy systems and pure statistical AI models. The solution to the black-box AI drawbacks consists in the fact that the output given is not only precise but also clear and correlating with the patterns of human logic. Finally, the ML-augmented CWW 2.0 framework constitutes an important contribution to the development of human-centric AI systems that will efficiently work in uncertain environments without losing trust and interpretability.