DESIGNING EXPLAINABLE AI FRAMEWORKS FOR REAL-TIME DECISION-MAKING IN DISTRIBUTED CLOUD ARCHITECTURES: A GENERALISED ARCHITECTURE WITH HEALTHCARE AS A USE CASEEXPLAINABLE AI ON HEART DISEASE DATASET
Keywords:
Explainable AI, Real-Time Decision Systems, Distributed Architecture, SHAP, LIME, Cloud-Native AI, XAI-as-a-Service, NIST RMF, FHIRAbstract
This study proposes a generalised, cloud-native Explainable AI (XAI) framework for real-time decision-making in distributed environments. By integrating SHAP and LIME within a modular microservices architecture, the framework enhances interpretability, latency, scalability, and compliance across diverse industry applications, with healthcare as a demonstration case. The framework explicitly aligns with NIST AI RMF 1.0, ISO/IEC 23894, HIPAA, and SOC 2 requirements. Evaluated on the UCI Heart Disease dataset, it demonstrates how XAI can bridge the trust gap in mission-critical domains through consistent, interpretable, and high-performance predictions.
