Entity Record Resolution with Instance-Dependent Cluster Optimization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing entity record resolution systems are resource-intensive, require significant human intervention, and struggle to resolve large volumes of noisy data efficiently in real-time, especially when handling tens of millions of records.

Innovation Solution

A system utilizing independently trained machine learning models and an enhanced elastic search environment for instance-dependent optimization, which automatically optimizes processing based on entity record characteristics, reducing system load and memory requirements while increasing processing speed and precision.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional entity record resolution systems process large volumes of data records, then data resolution capability is improved, but system resource consumption and processing time increase significantly

Engineering Contradiction:

Improvedata resolution capabilityVSAvoidprocessing time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system segments the entity record resolution process into distinct phases: candidate generation using machine learning models, candidate filtering using confidence bands, and final resolution. This segmentation allows each phase to be optimized independently, reducing overall processing time while maintaining resolution capability for large data volumes

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary actions by pre-computing confidence bands for different data record attributes and pre-training machine learning models with historical data. This preliminary preparation enables faster real-time resolution of entity records without requiring full system re-processing, thus reducing processing time for large datasets

Inventive Principle:

Principle #10Preliminary action

2Quantity of substance

If traditional systems handle tens of millions of entity records, then data coverage is improved, but memory requirements and system load increase

Engineering Contradiction:

Improvedata volume handledVSAvoidsystem resource consumption

Core Design Contradiction:

Quantity of substanceVSUse of energy by moving object

Solution Approach 1:

The system extracts only the most relevant features from entity records using machine learning models, and extracts high-confidence candidates using confidence bands. By taking out only the essential information needed for resolution rather than processing all data attributes equally, the system can handle tens of millions of records with reduced memory requirements and system load

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system dynamically adjusts processing parameters such as confidence band thresholds and model complexity based on data characteristics and resource availability. This allows the system to optimize the balance between handling large data volumes and controlling resource consumption, enabling efficient processing of tens of millions of entity records

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If manual intervention is used for entity record resolution, then resolution accuracy is improved, but automation level decreases and processing speed reduces

Engineering Contradiction:

Improveresolution accuracyVSAvoidautomation level

Core Design Contradiction:

Measurement precisionVSExtent of automation

Solution Approach 1:

The system implements feedback mechanisms where resolution outcomes are continuously monitored and used to adjust confidence bands and retrain machine learning models. This automated feedback loop maintains high resolution accuracy while increasing automation level, eliminating the need for manual intervention in the resolution process itself while preserving precision through continuous optimization

Inventive Principle:

Principle #23Feedback

4Measurement precision

If complex machine learning models are used for entity record resolution, then resolution precision is improved, but processing speed and system complexity increase

Engineering Contradiction:

Improveresolution precisionVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSSpeed

Solution Approach 1:

The system applies partial action by using confidence bands to filter and process only the most promising candidate records in detail, while applying quicker heuristic methods to less certain cases. This selective approach maintains high resolution precision for critical records while improving overall processing speed across the entire dataset

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20260087034A1Computer-based systems configured for instance dependent optimization of information processing systems and methods of use thereof

Publication Date: 2026.03.26 CAPITAL ONE SERVICES LLC

AI summary

In order to facilitate entity resolution, systems and methods include a processor receiving a plurality of records associated with one or more entity records. The processor utilizes a first natural language processing model to determine a set of clusters. The processor then utilizes instance inputs to determine adjustments to the natural language processing model to determine from the groups of clusters a second map of clusters from the entity feature, then determines a merge of the entity records, and displays the merged entity records.