Entity Linking via Embedding Vector Splicing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing entity linking technologies rely heavily on labor-intensive feature engineering and training sample construction, making them inefficient for complex scenarios and reducing accuracy in natural language processing and information retrieval tasks.
Innovation Solution
A method and apparatus for entity linking that acquire and process target texts, determine entity mentions, and use a preset knowledge base to find candidate entities, along with additional feature information, to improve the accuracy of entity linking by incorporating external knowledge.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If supervised learning or deep learning based entity linking is used, then entity linking capability is achieved, but labor consumption increases due to feature engineering and training sample construction
Solution Approach 1:
The system automatically retrieves reference texts from knowledge bases and extracts feature information without manual intervention. The embedding model automatically processes texts and generates vectors, eliminating the need for manual feature engineering and training sample construction while maintaining entity linking capability
Solution Approach 2:
The system pre-retrieves reference texts and feature information from knowledge bases before the actual entity linking process. This preliminary preparation of external knowledge allows the model to focus on comparison and matching rather than data collection and feature extraction
2Reliability
If traditional entity linking methods are used, then processing speed is maintained, but accuracy decreases in complex scenarios
Solution Approach 1:
The system transforms text data into embedding vectors that capture semantic meaning in a multi-dimensional space. By comparing vectors rather than raw text, the system achieves higher accuracy in disambiguation while the vector comparison process remains computationally efficient
Solution Approach 2:
The system introduces embedding vectors as an intermediary representation between raw text and entity matching. These vectors serve as a bridge that captures semantic information, allowing for more accurate comparison while the underlying computation remains streamlined through vector operations
Data Source
AI summary
A method, apparatus, device, and storage medium for entity linking is disclosed. The method includes: acquiring a target text; determining at least one entity mention included in the target text; determining a candidate entity corresponding to each of the entity mention based on a preset knowledge base; determining a reference text of each of the candidate entity and determining additional feature information of each of the candidate entity; and determining an entity linking result based on the target text, each of the reference text, and each piece of the additional feature information, wherein determining the entity linking result includes determining a probability of linking each of the candidate entity to the entity mention based on a splicing of a first embedding vector and a second embedding vector of the target text and a splicing of a first embedding vector and a second embedding vector of each respective reference text.