Fuzzy–Bayesian Modeling for Culture Medium Optimization in Cell Manufacturing

This article introduces a hybrid Fuzzy–LLM framework that integrates fuzzy logic, symbolic reasoning, and large language models (LLMs) to produce traceable, cognitively faithful explanations of sensory and cognitive data. Building upon previous works (Explainable Hybrid Modeling, Robust Explainable Modeling, and From Explainability to Prescription), it uses a new benchmark dataset, FuzzyCog-Sensory v1, which combines physico-chemical, hedonic, and linguistic information. The study quantitatively assesses how fuzzy rules can guide LLMs to generate scientifically consistent narratives. Using the Cognitive Fidelity Index (CFI), Align-Score, and Bias Rate, it evaluates the coherence between numerical models and linguistic outputs. Results demonstrate that Fuzzy–LLM systems can achieve both human-level interpretability and scientific traceability, advancing explainable AI toward auditable, human-centered cognitive modeling.