A model agnostic eXplainable AI based fuzzy framework for sensor constrained Aerospace maintenance applications

Explainability matters for practical maintenance decisions: The paper demonstrates that combining explainable AI techniques (SHAP) with machine learning helps maintenance teams understand why the system recommends prioritizing certain sensors, rather than just accepting a "black box" prediction—critical for building trust in high-stakes aerospace applications.

AI can work with limited data and sensors: The framework addresses a real-world constraint in small and medium-sized operations: companies often cannot afford extensive sensor networks or labeled datasets. By using unsupervised methods (Fuzzy C-means clustering) combined with dimensionality reduction (UMAP), the approach makes AI maintenance tools accessible to resource-constrained organizations.

Combining multiple AI techniques improves reliability: Rather than relying on a single machine learning method, the paper shows that integrating SHAP (for explanations), UMAP (for data visualization), and fuzzy clustering (for handling uncertainty) creates a more robust framework—suggesting that responsible AI implementation often requires thoughtful combination of complementary tools rather than single-solution approaches.

Responsible AI requires domain-specific design: The framework was specifically tailored for aerospace jet engine maintenance, indicating that trustworthy AI systems cannot be one-size-fits-all; they must account for industry constraints, safety requirements, and operational realities to be genuinely useful and responsible.