Query Processing Using Word and Entity Vector Representations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engine and question-answer systems rely on literal sorting methods that fail to understand semantic meanings and have limited generalization ability, requiring large training datasets and lacking universality in information retrieval.
Innovation Solution
A query processing method and apparatus that determines word and entity vector representations for query sequences and paragraphs using knowledge graphs, calculating similarity based on these representations to improve retrieval accuracy and generalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If literal sorting methods are used in search engines and question-answer systems, then the system structure remains simple and easy to implement, but the system fails to understand semantic meanings and has limited generalization ability
Solution Approach 1:
The patent introduces vector representations as an intermediary between literal text matching and semantic understanding. By converting words, entities, and paragraphs into vector forms, the system enables semantic comparison without requiring complex neural network structures, thus improving adaptability while controlling complexity
Solution Approach 2:
The patent transforms the representation parameters of text from literal strings to numerical vectors. This parameter change allows the system to capture semantic meanings through vector operations and similarity calculations, enhancing the system's semantic understanding capability without significantly increasing structural complexity
2Measurement precision
If traditional sorting methods based on literal matching are used, then the system is easy to operate and implement, but the retrieval accuracy and generalization ability are insufficient
Solution Approach 1:
The patent replaces the mechanical literal string matching system with a vector-based semantic comparison system. By substituting exact match operations with vector similarity calculations, the system achieves higher retrieval accuracy while maintaining operational simplicity through standardized vector operations
3Adaptability or versatility
If knowledge graphs and dual vector representations are introduced to improve semantic understanding, then retrieval accuracy and generalization improve, but the processing complexity and computational requirements increase
Solution Approach 1:
The patent segments the semantic representation task into two distinct vector components: word vectors for lexical meaning and entity vectors for conceptual meaning. This segmentation allows the system to capture multiple aspects of semantics independently, improving generalization ability while keeping each processing module relatively simple and manageable
Data Source
AI summary
Embodiments of the present disclosure provide a query processing method and an apparatus, a server and a storage medium. The method includes: determining a word vector representation of a query sequence and an entity vector representation of the query sequence respectively based on respective words and respective entities included in the query sequence; determining a word vector representation of a paragraph and an entity vector representation of the paragraph respectively based on respective words and respective entities included in the paragraph; and determining a similarity between the query sequence and the paragraph according to the word vector representation of the query sequence, the entity vector representation of the query sequence, the word vector representation of the paragraph, and the entity vector representation of the paragraph.