Generative Feature Extraction for Watchlist Identity Resolution

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional watchlist identification methods rely heavily on manual review of structured data sources, leading to errors due to name misspellings, inconsistent formatting, and lack of consideration for aliases, while external data sources like social media and news articles are underutilized due to difficulty in normalizing and assessing their relevance.

Innovation Solution

A system and method utilizing natural language processing and machine learning to process and integrate external data sources, such as social media and news articles, to construct holistic identity profiles for transaction applicants, enhancing watchlist matching accuracy through unsupervised semantic identity modeling and graph-based clustering.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual review of traditional watchlist databases is used, then structured data can be processed, but errors occur due to name misspellings, inconsistent formatting, and lack of consideration for aliases

Engineering Contradiction:

Improvewatchlist identification accuracyVSAvoidmanual review process complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces manual mechanical review processes with automated natural language processing (NLP) and machine learning systems. The NLP models automatically parse unstructured text from external sources, extract entity information, and perform matching without human intervention, thereby eliminating errors associated with manual review while maintaining high reliability.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the parameters of data processing by transforming unstructured external data into structured formats suitable for watchlist matching. Through entity extraction, normalization, and feature engineering, the system adapts diverse data sources (social media, news articles) to match the structured requirements of traditional watchlist databases, enabling accurate automated comparison.

Inventive Principle:

Principle #35Parameter changes

2Loss of information

If external data sources like social media and news articles are used, then more comprehensive identity information can be obtained, but difficulty arises in normalizing and assessing relevance of the unstructured data

Engineering Contradiction:

Improveidentity information completenessVSAvoiddata normalization and relevance assessment difficulty

Core Design Contradiction:

Loss of informationVSDifficulty of detecting and measuring

Solution Approach 1:

The patent introduces NLP models and entity extraction systems as intermediary components between external unstructured data sources and the watchlist matching process. These intermediaries automatically parse text, extract relevant entities (names, dates, locations), normalize formats, and assess relevance through semantic analysis, thereby bridging the gap between diverse external sources and structured matching requirements.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary processing of external data sources before they enter the main matching pipeline. Through pre-processing steps including text cleaning, entity recognition, normalization, and feature extraction, the system prepares unstructured data in advance, making it ready for efficient and accurate watchlist comparison without requiring complex processing during the main matching operation.

Inventive Principle:

Principle #10Preliminary action

3Stability of the object's composition

If traditional structured data sources are used, then data consistency is maintained, but valuable information from external sources remains underutilized

Engineering Contradiction:

Improvedata consistencyVSAvoiddata source diversity

Core Design Contradiction:

Stability of the object's compositionVSAdaptability or versatility

Solution Approach 1:

The patent creates a universal processing framework that can handle multiple data types and sources through a single NLP-based pipeline. The system is designed to process both traditional structured watchlist data and unstructured external data (social media posts, news articles, reports) using the same entity extraction and normalization mechanisms, enabling versatile data source integration while maintaining consistent processing quality.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20250322162A1Generative Systems and Methods of Feature Extraction for Enhancing Entity Resolution for Watchlist Screening

Publication Date: 2025.10.16 SOCURE INC

AI summary

Provided are a method and system for identity correlation between a transaction applicant (TA) and a watchlist entity (WE). Preexisting watchlist data and other aggregated identity data (AID) are processed to provide for comparison to a collective identity of at least the TA. Using various categorizations for the AID and the collective identity, watchlist tags are generated that can then be matched to the collective identity. As a result of the matching, a watchlist candidacy demonstrating a probability that the identity of the TA does or does not correspond to that of the WE can be generated. Natural language processing may be used in connection with sourced textual information such as from news articles, social media posts and the like to supplement the AID available to the system and to further enhance matching functionality.