Explainable Federated Bayesian Causal Inference and Its Application in Advanced Manufacturing

Privacy-Preserving Decision Making: The framework enables manufacturers to understand cause-and-effect relationships in their processes without sharing sensitive local data. By using federated Bayesian learning, each facility can keep its data private while still contributing to causal analysis across the manufacturing network.

Practical Application Beyond Theory: While causal inference has been widely adopted in biology and healthcare, this work demonstrates its first significant application to manufacturing systems. The real-world validation using Electrohydrodynamic (EHD) printing data shows the method works in actual industrial settings, not just simulations.

Scalability for Distributed Systems: The "xFBCI" framework is designed to work across multiple distributed manufacturing sites simultaneously. This addresses a key challenge in responsible AI: enabling collaborative learning and decision-making across organizations without centralizing potentially proprietary data.

Explainability Built In: By combining explainable AI (xAI) with causal inference, the system doesn't just make recommendations—it can explain why certain manufacturing conditions cause specific outcomes. This transparency is essential for manufacturers to trust and validate AI-driven process improvements.

Superior Performance: The approach outperformed both standard Bayesian causal inference methods and state-of-the-art federated learning benchmarks, suggesting that responsible AI methods (privacy-preserving + explainable) don't require sacrificing accuracy or effectiveness.