Generating trustworthy explanations with a language model enriched by fuzzy rule-based systems
In this work, we have developed, validated, and deployed fuzzy-grounded, trustworthy, interactive explanations by combining fuzzy systems with Large Language Models (LLMs). More precisely, we profit from fuzzy systems' ability to deal appropriately with vagueness and uncertainty in approximate reasoning and propose an efficient and effective way to combine fuzzy rule-based systems with the natural language verbalization capabilities of LLMs. The aim is to generate explanations that balance naturalness and faithfulness. The proposed software architecture is open-source and highly modular. We validated our work in the medical field. We observed significant statistical differences with related-samples Wilcoxon signed rank tests (p-value < 0.001) regarding hallucinations in the experimental scenarios under study, which hints at good faithfulness. We also produced significantly more compact narratives, which helps alleviate the usual ‘over-explanation’ issue of LLMs without a penalization in accuracy.
keywords: Trustworthy AI, Explainable Fuzzy Systems, Heart Disease
Publication: Congress
1759743168489
October 6, 2025
/research/publications/generating-trustworthy-explanations-with-a-language-model-enriched-by-fuzzy-rule-based-systems
In this work, we have developed, validated, and deployed fuzzy-grounded, trustworthy, interactive explanations by combining fuzzy systems with Large Language Models (LLMs). More precisely, we profit from fuzzy systems' ability to deal appropriately with vagueness and uncertainty in approximate reasoning and propose an efficient and effective way to combine fuzzy rule-based systems with the natural language verbalization capabilities of LLMs. The aim is to generate explanations that balance naturalness and faithfulness. The proposed software architecture is open-source and highly modular. We validated our work in the medical field. We observed significant statistical differences with related-samples Wilcoxon signed rank tests (p-value < 0.001) regarding hallucinations in the experimental scenarios under study, which hints at good faithfulness. We also produced significantly more compact narratives, which helps alleviate the usual ‘over-explanation’ issue of LLMs without a penalization in accuracy. - Pablo Miguel Perez-Ferreiro, Alejandro Catala, Alberto Bugarin-Diz, Jose M. Alonso-Moral - 10.1109/FUZZ62266.2025.11152139 - 979-8-3315-4319-8
publications_en