LLM Query Conversion for Automatic Search Filter Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional search systems face challenges in optimally selecting filters, leading to irrelevant or overinclusive search results, resource wastage, outdated filter definitions, limitations due to evolving vocabulary, and subjective user interpretations of filter types and values, which affect search efficiency and user experience.
Innovation Solution
A natural language processing (NLP) engine uses a large language model to convert natural language queries into structured search queries, automatically identifying filter types and values without user explicit selection, leveraging generative models and retrieval augmented generation to maintain up-to-date mappings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional search systems use traditional filter selection methods, then users can manually select filters, but search results become irrelevant or overinclusive and user interface becomes cluttered
Solution Approach 1:
The system automatically selects filters using large language models and retrieval augmented generation, eliminating the need for manual filter selection by users. The NLP engine autonomously interprets natural language queries and determines appropriate filters, allowing the system to serve itself rather than requiring user intervention for filter configuration.
Solution Approach 2:
The patent replaces the mechanical interaction of manual filter selection with an automated NLP-based system. Large language models process natural language queries and automatically determine filter settings, substituting the traditional mechanical approach of users clicking and selecting filters with an intelligent automated system that interprets and executes filter selection based on query understanding.
2Stability of the object's composition
If conventional search systems use static filter definitions, then filter types and values remain fixed, but they become outdated and cannot adapt to evolving vocabulary
Solution Approach 1:
The system dynamically adapts filter definitions by using large language models to interpret evolving natural language queries. Instead of static filter definitions, the NLP engine continuously learns from and adapts to changing vocabulary and query patterns, allowing filter interpretations to evolve dynamically while maintaining core functionality.
Solution Approach 2:
The retrieval augmented generation component provides feedback mechanisms that allow the system to learn from query-performance patterns. The system uses feedback from query interpretations and search results to continuously refine and update filter definitions, ensuring they remain current with evolving vocabulary and user intent while maintaining stability in their fundamental structure.
3Productivity
If conventional search systems require explicit user filter selection, then users have control over filters, but search efficiency decreases and resources are wasted
Solution Approach 1:
The system performs preliminary filter selection automatically based on natural language query interpretation before the search execution. Large language models analyze the query and pre-determine appropriate filters and their values, so that when the search is executed, the filtering is already optimized, eliminating the need for users to manually adjust filters and reducing iterative search attempts that waste resources.
Solution Approach 2:
The NLP engine autonomously performs filter selection and optimization without requiring user intervention. The system serves itself by automatically interpreting queries, determining relevant filters, and configuring search parameters, which eliminates the time and computational resources that would be wasted on manual filter adjustment and multiple trial searches.
4Ease of operation
If conventional search systems use manual filter interpretation, then users can control filter application, but subjective interpretations lead to inconsistent results
Solution Approach 1:
The patent replaces subjective human interpretation of filters with an automated large language model-based system. The NLP engine objectively analyzes natural language queries and consistently applies filter logic based on learned patterns, eliminating the variability and subjectivity inherent in manual filter interpretation while maintaining ease of query input for users.
Data Source
AI summary
Embodiments of the disclosed technologies are capable of generating natural language queries. The embodiments describe generating a training natural language query of a training structured search query using a first LLM and a first prompt. The embodiments further describe fine-tuning a second LLM using the training natural language query of the training structured search query and the training structured search query. The fine-tuned second LLM generates a structured version of a natural language query. The embodiments further describe generating the structured version of a received natural language query using the fine-tuned second LLM and a second prompt.