Change at scale requires a
phased framework
Phase III: Anomaly detection
Phase IV: Supervised ML to train classifiers on historical fraud cases
Phase II: Graph Extraction and Feature Engineering
Phase I: Data Collection and Integration
Phase V: Apply methods like GraphSAGE to analyze relationships in identity networks
Phase VI: Explainable AI (XAI)
Phase VII: Monitoring for Data and Concept Drift
Phase VIII: Event Driven Architecture
Dialog with the agency is essential throughout the project and after delivery
Questions?
Arch Systems
By Daina Bouquin
