Knowledge-Driven Recommendation System Using Ontology Binding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large Language Models (LLMs) like GPT-4 are limited in recommending new or related topics during conversational contextual querying due to the lack of a comprehensive, hierarchically organized knowledge base, restricting their ability to extend recommendations beyond the provided context.
Innovation Solution
A knowledge-driven recommendation system that utilizes a custom neural network to navigate a knowledge graph through taxonomy and ontology pairs, enabling the extraction of relevant topics and related documents by leveraging a custom neural network trained on high-quality data, and employing techniques such as semantic search and link hopping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If LLMs are used for conversational contextual querying, then natural language interaction capability is improved, but topic recommendation capability deteriorates due to lack of comprehensive knowledge base
Solution Approach 1:
The patent introduces a recommendation system as an intermediary component that sits between the user and the LLM. This recommendation system maintains a comprehensive knowledge base and taxonomy structure, and uses a separate neural network model to generate topic recommendations. The LLM focuses on natural language understanding and generation, while the recommendation system handles topic suggestion, allowing each component to excel at its specific function without compromising the other.
2Measurement precision
If entire knowledge base is shared with LLM, then recommendation accuracy is improved, but system complexity and resource consumption worsen
Solution Approach 1:
The patent divides the knowledge base system into two separate segments: (1) a comprehensive knowledge base maintained by the recommendation system with full taxonomy and ontology structures, and (2) a streamlined interface where only relevant information is shared with the LLM. The recommendation system processes queries against the full knowledge base independently, then shares only the necessary topic suggestions with the LLM, avoiding the need to load the entire knowledge base into the LLM's context.
3Measurement precision
If comprehensive knowledge base is maintained, then recommendation quality is improved, but information retrieval time worsens
Solution Approach 1:
The patent implements preliminary action by pre-processing and organizing the knowledge base into structured taxonomy and ontology formats before queries arrive. The recommendation system pre-computes relationships between topics and maintains ready-to-query data structures. When a user query arrives, the system can quickly traverse the pre-organized knowledge base using the separate neural network model, rather than searching through unstructured data, significantly reducing retrieval time while maintaining comprehensive coverage.
Data Source
AI summary
A knowledge-driven recommendation system comprises a processor and a memory with computer code instructions. The executed code instructions cause the system to receive a user query, extract a topic from the query, and submit the topic and query to a neural network. The instructions may further cause the system to return, from the neural network, a collection of taxonomy and ontology pairs, and use the pairs to select information that expands on the query and topic. The taxonomy and ontology pairs are the closest matched pairs from a knowledge graph. The closest matched pairs are retrieved when the input taxonomy topic semantically matches closest to a taxonomy topic from the custom neural network, the input ontology semantically matches closest to an ontology from the custom neural network, and the taxonomy topic from the custom neural network matches closest to one of the entities in the ontology of the neural network.