LLM Query Supplementation for Intuitive Personalized Search Results
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online search approaches lack the ability to effectively organize and explain personalized search results, resulting in a less intuitive and personalized experience.
Innovation Solution
Utilize a large language model (LLM) to generate supplemental search results by analyzing user engagement data and personas, generating groups of related queries, and presenting them alongside base search results to enhance personalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If existing online search approaches are used to provide personalized search results, then basic personalization can be achieved through analyzing past user interactions, but the ability to effectively organize and explain the results is insufficient
Solution Approach 1:
The patent introduces large language models as an intermediary component between the search query processing and result presentation. The LLM analyzes user engagement data, generates related queries, and organizes results into coherent groups with explanations, thereby bridging the gap between basic personalization and intuitive result presentation without requiring complete system redesign
Solution Approach 2:
The search results are segmented into multiple groups based on different dimensions (e.g., product categories, user intents, engagement patterns). Each group is independently organized and explained, making the overall result set more manageable and easier for users to navigate while maintaining personalized relevance
2Adaptability or versatility
If supplemental search results are generated using large language models by analyzing user engagement data, then personalization and organization of results are improved, but system complexity increases
Solution Approach 1:
The large language model serves multiple functions within the search system: it analyzes user engagement data, generates related queries, organizes results into groups, and provides explanations. This multi-functionality reduces the need for separate specialized components, thereby managing complexity while enhancing personalization quality
Solution Approach 2:
User engagement data is pre-processed and stored in structured formats before being fed to the LLM. The LLM itself is pre-trained on extensive datasets, enabling it to perform complex analysis and generation tasks without requiring real-time computational resources for training, thus reducing operational complexity
Data Source
AI summary
An online system retrieves engagement data associated with a base query made by a user for an item, the engagement data describing in part subsequent queries for other items following the base query in a single search session. The system generates a prompt that is provided to a machine learned model. The prompt instructs the machine learned model to generate one or more groups of related queries using the subsequent queries. The system selects a group of related queries from the one or more groups of related queries. The system queries an online catalog using at least some of the related queries from the selected group to determine supplemental search results. The system provides, to a user client device associated with the user, the supplemental search results.