Graph Convolutional Network for Financial Crime Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Financial institutions face challenges in efficiently detecting unauthorized or suspicious financial activities due to the large amounts of data that need to be processed quickly, while also ensuring high reliability and compliance with evolving regulations, which existing technologies struggle to address effectively.
Innovation Solution
The use of Graph Convolutional Neural Networks (GCNs) to represent financial data as graphs, where nodes represent entities and edges represent transactions or relationships, with aggregated data to enhance processing speed and reliability, allowing for the calculation of risk scores for entities involved in suspicious activities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If complex ANN algorithms and detailed data processing are used to improve detection reliability, then the reliability of suspicious activity detection is improved, but the processing speed decreases
Solution Approach 1:
The patent segments the financial data processing into distinct graph-based representations where entities and transactions are separated into nodes and edges. This segmentation allows the GCN to process structured relationships efficiently while maintaining high detection reliability through preserved entity-transaction connections.
Solution Approach 2:
The patent transforms traditional tabular financial data into a graph-based dimensional structure, adding the dimension of relational connectivity between entities and transactions. This dimensional change enables the GCN to capture complex patterns more efficiently, improving both speed and reliability simultaneously.
2Measurement precision
If large amounts of detailed financial data are processed to improve detection accuracy, then the detection accuracy is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary action by pre-structuring financial data into graph representations with entities as nodes and transactions as edges before analysis. This pre-structuring organizes large volumes of data into an efficient format that the GCN can process quickly while maintaining high detection accuracy through preserved relational structures.
Solution Approach 2:
The patent changes the structural parameters of data representation from traditional formats to graph-based formats, transforming how financial data is organized and processed. This parameter change enables efficient handling of large datasets while maintaining detection accuracy through the preservation of entity-transaction relationships.
3Reliability
If traditional transaction monitoring methods are used to comply with regulations, then regulatory compliance is achieved, but the ability to detect evolving suspicious patterns decreases
Solution Approach 1:
The patent introduces dynamics by using graph-based representations that can adapt to evolving suspicious patterns while maintaining regulatory compliance. The GCN dynamically processes relational structures, allowing the system to detect new types of money laundering and terrorist financing schemes without sacrificing compliance with existing regulations.
Solution Approach 2:
The patent achieves universality by creating a detection system that simultaneously handles multiple regulatory requirements and detects various types of suspicious activities. The graph-based approach with entities and transactions as nodes and edges provides a unified framework that can adapt to different regulatory frameworks and suspicious pattern types.
Data Source
AI summary
In a method for detecting unauthorized or suspicious financial activity, a graph convolutional network for financial crime prevention, a separate node is created for each entity: each account, each person, each address (e.g. email address), etc. Separate attributes are provided to aggregate transactions in which the node acts as a sender; transactions in which the node acts as a receiver; transactions using a specific channel (e.g. ATM); and transactions of a specific type (e.g. online money transfer). In some embodiments, the attributes exclude data on individual transactions to reduce the amount of data and hence provide more effective computer utilization. The approach is suitable for many applications, including anti-money laundering. Other features are also provided, as well as systems for such detection.