Interpretability of Fuzzy Systems: Current Research Trends and Prospects

摘要: Fuzzy systems are universally acknowledged as valuable tools to model complex phenomena while preserving a readable form of knowledge representation. The resort natural language for expressing the terms involved in fuzzy rules, fact, is key factor conjugate mathematical formalism and logical inference with human-centered interpretability . That makes specifically suitable every real-world context where people charge crucial decisions. This because self-explanatory nature rules profitably supports expert assessments. Additionally, far investigated, it appears that (a) simple adoption sets modeling not enough ensure interpretability; (b) representation must confront problem overall system accuracy, thus yielding trade-off which frequently debated. Such issues have attracted growing interest research community became assume central role current literature panorama computational intelligence. chapter gives an overview topics related interpretability, facing ambitious goal proposing some answers number open challenging questions: What interpretability? Why worth considering? How how assess (quantify) it? Finally, design interpretable models?