Dynamic Content Instance Creation for Information Providers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies lack the capability to dynamically create instances for content of information providers based on user prompts and large language model results, limiting the customization and relevance of provided information.
Innovation Solution
A method and system that dynamically create an instance for content of an information provider by verifying large language model results, combining pre-registered assets with keywords and user information, and modifying the content's expression, format, and tone to enhance relevance and user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If large language model results are used directly to respond to user prompts, then the system can provide quick responses, but the content lacks customization and relevance to specific information providers
Solution Approach 1:
The system segments the content generation process into multiple stages: LLM generates base results, then a dynamic content creation system processes these results through verification, asset combination, and instance creation. This segmentation allows both quick LLM responses and customized information provider content to coexist without compromising overall productivity.
Solution Approach 2:
Information providers pre-register their assets (images, texts, URLs, etc.) and content templates in advance. When a user query is processed, the system can quickly match these pre-prepared assets with LLM results to create customized content instances, avoiding the need to generate everything from scratch and maintaining response speed while improving customization.
2Adaptability or versatility
If content is dynamically created by combining multiple assets and modifying expressions, then content relevance and user engagement improve, but system complexity increases
Solution Approach 1:
The dynamic content creation system serves multiple functions within a single integrated platform: it verifies LLM results, combines assets from multiple information providers, creates customized content instances, and displays them contextually. This multi-functionality manages complexity by consolidating diverse operations into one unified system rather than requiring separate systems for each function.
Solution Approach 2:
The patent introduces an intermediary dynamic content creation system that sits between the LLM and the user interface. This intermediary handles the complex tasks of verifying LLM output, matching information provider assets, and generating customized content instances, thereby shielding the rest of the system from complexity while enabling high content relevance.
3Adaptability or versatility
If multiple information providers' assets are combined and modified, then content richness and customization improve, but processing time and computational resources increase
Solution Approach 1:
Information providers submit and register their assets (images, texts, URLs, contact information) in advance before they are needed. The system stores these pre-registered assets and their associated metadata, enabling rapid retrieval and combination during content creation without requiring time-consuming processing of raw materials at query time.
Solution Approach 2:
The system creates content instances with varying degrees of customization based on query requirements. For some queries, it may use only LLM results with minimal asset combination, while for others it extensively combines multiple information provider assets. This partial action approach balances content richness with processing efficiency by applying full customization only when necessary.
Data Source
AI summary
A method for dynamically creating an instance for content of an information provider includes verifying LLM results created based on a large language model (LLM) for a prompt of a user, creating an instance for content of an information provider using the LLM results and a pre-registered asset of the information provider; and providing the created instance such that the created instance is displayed in relation to the LLM results.