Multimodal Entity Resolution for Financial Transaction Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Financial institutions face challenges in obtaining a cohesive view of connected entities involved in transactions due to limited available information, making it difficult to manage risk and provide enhanced services, and this gap is exploited by fraudsters.
Innovation Solution
The system performs multimodal entity resolution using deterministic and graph traversal methods to analyze financial transaction data, combining relational databases and graph databases to identify entities, and provides human-in-the-loop review for accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional single-mode entity resolution is used, then system complexity is reduced, but measurement precision and reliability of entity identification deteriorate
Solution Approach 1:
The patent merges multiple entity resolution modes (deterministic mode using explicit identifiers and graph traversal mode using relationship paths) into a unified multimodal system. The system combines relational database queries with graph database traversals to achieve higher identification accuracy by leveraging multiple data representations simultaneously.
Solution Approach 2:
The entity resolution system is designed with multi-functionality to handle different identification scenarios through different modes. The same system can perform deterministic identification when identifiers are available and graph traversal identification when relationships are needed, making the system universally applicable to various financial entity resolution tasks.
2Reliability
If more data sources are integrated, then reliability of entity view improves, but loss of information and processing time increase
Solution Approach 1:
The system dynamically selects between different entity resolution modes based on the availability and quality of identifiers in the transaction data. When deterministic identifiers are present, the system uses fast deterministic mode; when only relationships are available, it switches to graph traversal mode. This dynamic adaptation optimizes processing time while maintaining reliable entity identification.
Solution Approach 2:
The entity resolution process is segmented into two distinct modes that can be executed independently or in combination. The deterministic mode handles straightforward identifier matching, while the graph traversal mode handles complex relationship-based identification. This segmentation allows the system to process different types of data efficiently without overwhelming the entire system.
3Productivity
If automated entity resolution is increased, then productivity improves, but reliability of complex entity relationships deteriorates
Solution Approach 1:
The graph database serves as an intermediary layer between the transaction data and the final entity resolution results. It stores and processes complex relationship paths, allowing the system to automatically traverse multiple relationships to identify entities while maintaining accuracy. The graph database mediates between fast automated processing and reliable complex relationship analysis.
Data Source
AI summary
Technologies for performing multimodal entity resolution include a compute device. The compute device includes circuitry configured to obtain financial transaction data indicative of financial transactions associated with a financial institution. The circuitry may be further configured to perform one or more deterministic operations on the obtained financial transaction data in a relational database format to resolve identities of entities associated with the financial transactions. Additionally, the circuitry may be configured to perform one or more graph traversal operations on the financial transaction data to resolve additional identities of the entities associated with the financial transactions.