Knowledge Graph Corpus Generation for Language Model Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current language models require high-quality training corpora to improve their predictive capabilities, but existing methods struggle to generate corpora with strong logical structure and higher quality.
Innovation Solution
The method involves using a knowledge graph to generate a text corpus by reading graph data and ontology information, generating sentences based on pre-constructed sentence templates, and determining a text corpus for training language models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional corpus generation methods are used, then the process is simple, but the quality and logical structure of the generated corpus are insufficient
Solution Approach 1:
The patent introduces a knowledge graph as an intermediary structure between traditional corpus generation methods and the final training corpus. The knowledge graph organizes entities, relationships, and attributes in a structured manner, enabling systematic sentence generation that ensures both high quality and strong logical structure in the generated corpus
Solution Approach 2:
The patent segments the corpus generation process into distinct components: extracting entities and relationships from knowledge graphs, selecting appropriate sentence templates, filling template variables with extracted information, and generating final sentences. This segmentation allows each component to be optimized independently while ensuring overall corpus quality
2Adaptability or versatility
If more diverse sentence templates are used, then the variety of generated sentences increases, but the complexity of template construction and management increases
Solution Approach 1:
The patent designs sentence templates with universal structures that can handle multiple types of knowledge graph relationships. The templates use variable placeholders that can be dynamically filled with different entity types and relationship types, allowing a single template framework to generate diverse sentences across different domains and relationship types without requiring separate templates for each case
Data Source
AI summary
Embodiments of this specification provide a method and an apparatus for generating a text corpus by using a knowledge graph. Graph elements of the knowledge graph include a node representing an entity and a connecting edge representing a relationship between nodes. Privacy data can be organized into structured data by using the knowledge graph. The graph data includes several triplets including graph elements in the subgraph, and the ontology information includes at least a type of each graph element in the subgraph. Several sentences are generated based on several pre-constructed sentence templates, the graph data, and the ontology information. The several sentences are classified into a generated sentence set, and at least one of the several sentence templates is constructed based on the ontology information. A text corpus corresponding to the subgraph is determined based on the generated sentence set. The text corpus is used to train a language model.