Hybrid Knowledge Graph Update System for Real-Time Entity Linking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing knowledge graph update methods face challenges in maintaining real-time accuracy and availability due to issues with offline batch processing's timeliness and online real-time processing's incompleteness, leading to reduced service processing effectiveness.

Innovation Solution

A hybrid approach combining offline incremental entity linking and online real-time entity linking to update knowledge graphs, where full entity linking initializes the graph, followed by cyclic updates using real-time and incremental updates based on predetermined conditions, ensuring both real-time performance and data accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If offline batch processing is used to update knowledge graph, then data completeness is improved, but timeliness deteriorates

Engineering Contradiction:

Improvedata completenessVSAvoidtimeliness

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

The patent segments the knowledge graph update process into two distinct modes: offline batch processing for comprehensive data updates and online real-time processing for timely incremental updates. This segmentation allows each mode to operate independently with optimized parameters, resolving the contradiction between completeness and timeliness by applying the appropriate mode based on update requirements

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system implements periodic switching between offline and online update modes based on predetermined conditions (such as data volume thresholds or time intervals). This periodic action ensures that the knowledge graph receives both comprehensive periodic full updates and continuous real-time incremental updates, balancing completeness and timeliness

Inventive Principle:

Principle #19Periodic action

2Loss of time

If online real-time processing is used to update knowledge graph, then timeliness is improved, but data completeness deteriorates

Engineering Contradiction:

ImprovetimelinessVSAvoiddata completeness

Core Design Contradiction:

Loss of timeVSQuantity of substance

Solution Approach 1:

The system performs preliminary offline batch processing to establish a complete baseline knowledge graph before transitioning to online real-time processing. This preliminary action ensures that the foundation is comprehensive, while subsequent real-time updates maintain timeliness without sacrificing overall completeness

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent merges the results of offline batch processing and online real-time processing into a unified knowledge graph. The offline component provides comprehensive data coverage, while the online component adds timely incremental updates, and their combination resolves the contradiction by integrating both completeness and timeliness

Inventive Principle:

Principle #5Merging (Combining)

3Measurement precision

If full entity linking is performed on all service data, then entity linking accuracy is improved, but processing time increases

Engineering Contradiction:

Improveentity linking accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

Instead of performing full entity linking on all service data continuously, the system applies partial entity linking only when necessary (during offline batch processing or when confidence thresholds are not met). This partial action maintains accuracy for critical operations while reducing unnecessary processing time for routine updates

Inventive Principle:

Principle #16Partial or excessive action

4Measurement precision

If incremental update operations are performed frequently, then real-time accuracy is improved, but system resource consumption increases

Engineering Contradiction:

Improvereal-time accuracyVSAvoidsystem resource consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system performs incremental update operations periodically based on predetermined conditions such as data volume thresholds, time intervals, or confidence scores. This periodic execution maintains real-time accuracy by updating when necessary while avoiding excessive resource consumption from overly frequent updates

Inventive Principle:

Principle #19Periodic action

Data Source

PatentUS20250209350A1Method and apparatus for updating knowledge graph

Publication Date: 2025.06.26 ALIPAY (HANGZHOU) INFORMATION TECH CO LTD

AI summary

The disclosure provides a method and an apparatus for updating a knowledge graph. In a process of providing knowledge graph-based data support for a current service, a knowledge graph is updated by combining online and offline manners. The knowledge graph is constructed offline by using full service data, and full entity linking and entity normalization are performed to initialize the knowledge graph. An incremental update condition is set to perform a plurality of rounds of incremental update. During one round of incremental update, real-time linking is performed based on service data generated in real time, to provide online knowledge graph update; and when the preset incremental update condition is met, incremental linking is performed based on service data newly added in a current incremental update period, to provide offline knowledge graph update and use an updated knowledge graph as an initial knowledge graph in a next round of incremental update.