Regulatory Graphs and GenAI for Real-Time Transaction Monitoring and Compliance Explanation in Banking

2025-06-01

AI Summary

To address the challenges of real-time transaction monitoring and automated regulatory explanation in banking, this paper proposes a unified framework integrating regulatory graph modeling and generative explanation. First, it constructs a dynamic regulatory transaction graph that fuses heterogeneous financial entities and behavioral relationships from multiple sources. Second, it designs a graph neural network (GNN)-based anomaly detection module enabling millisecond-scale identification of suspicious transactions. Third, it introduces a regulation-aligned retrieval-augmented generation (RAG) mechanism, incorporating narrative field embedding and clause-aware decoding to produce auditable, compliance-conformant natural language explanations. This work is the first to deeply integrate dynamic graph modeling, GNN-driven detection, and regulation-guided generation. Evaluated on synthetic financial flow data, the framework achieves an F1-score of 98.2% and is validated by domain experts for explanation accuracy, logical coherence, and audit readiness.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

This paper presents a real-time transaction monitoring framework that integrates graph-based modeling, narrative field embedding, and generative explanation to support automated financial compliance. The system constructs dynamic transaction graphs, extracts structural and contextual features, and classifies suspicious behavior using a graph neural network. A retrieval-augmented generation module generates natural language explanations aligned with regulatory clauses for each flagged transaction. Experiments conducted on a simulated stream of financial data show that the proposed method achieves superior results, with 98.2% F1-score, 97.8% precision, and 97.0% recall. Expert evaluation further confirms the quality and interpretability of generated justifications. The findings demonstrate the potential of combining graph intelligence and generative models to support explainable, audit-ready compliance in high-risk financial environments.

Problem

• Real-time monitoring of financial transactions for compliance
• Detecting suspicious behavior using graph neural networks
• Generating regulatory-aligned explanations for flagged transactions

Innovation

• Graph neural network classifies suspicious transactions
• Retrieval-augmented generation creates regulatory explanations
• Dynamic transaction graphs model real-time financial data