Entity Linking Strategy Using Multi-Factor Decision Logic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing entity linking technologies suffer from poor accuracy due to incorrect linking of entities with the same name in knowledge bases, as they rely on similarity scores that may not accurately reflect the relevance of entities in text statements.
Innovation Solution
An entity linking method that acquires a target entity most relevant to a to-be-linked entity in a specified statement and decides whether to link based on a preset linking decision strategy, incorporating correlation, type, prior probabilities, and coherence features, rather than solely relying on similarity scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If entity linking relies on similarity scores between entity mentions and knowledge base entities, then the linking process is simple and fast, but the accuracy of entity linking deteriorates due to incorrect linking of entities with the same name
Solution Approach 1:
The patent segments the entity linking process into multiple independent decision dimensions: entity mention identification, candidate entity retrieval, and linking decision. Each dimension is handled separately with specialized strategies, allowing the system to consider multiple factors (co-reference, coreference, entity type, background knowledge) rather than relying on a single similarity score, thus improving accuracy while maintaining efficiency
Solution Approach 2:
The patent introduces an intermediary linking decision strategy that acts as a mediator between the entity mention and the knowledge base entity. This intermediary layer evaluates multiple factors including co-reference relations, coreference relations, entity type compatibility, and background knowledge consistency before making the final linking decision, thereby resolving the contradiction between simple similarity-based linking and accurate multi-factor-based linking
2Measurement precision
If the system considers multiple factors (correlation, type, prior probabilities, coherence) for entity linking decisions, then the accuracy of entity linking improves, but the complexity of the linking process increases
Solution Approach 1:
The patent divides the complex multi-factor evaluation into segmented, modular components: co-reference relation detection, coreference relation detection, entity type matching, and background knowledge verification. Each component operates independently and contributes to the overall linking decision, making the complex process more manageable and interpretable while maintaining high accuracy
Solution Approach 2:
The patent creates a universal linking decision framework that can handle multiple types of entity relations (co-reference, coreference, entity type matching, background knowledge) through a single integrated strategy. This multi-functional approach consolidates various evaluation factors into one cohesive process, reducing overall system complexity while improving linking accuracy
Data Source
AI summary
There is provided an entity linking method, an electronic device, and a storage medium, which relates to the technical field of artificial intelligence such as machine learning, natural language processing, and intelligent search. A specific implementation solution involves: acquiring a target entity in a knowledge base and most relevant to a to-be-linked entity in a specified statement; and deciding, based on a linking decision strategy, whether to link the to-be-linked entity to the target entity in the knowledge base.