LLM-Based Intelligent Assistant Query Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional intelligent assistants face challenges in providing comprehensive and personalized responses due to the high manpower and technical costs associated with pre-building FAQ databases or graph question answering systems, which struggle to cover a full range of questions and offer insufficient intelligence in user experiences.
Innovation Solution
An information processing method based on a large language model that obtains user query information, determines relevant memory information, selects appropriate tools for processing, invokes these tools to gather auxiliary information, and generates responses, thereby enhancing the intelligent assistant's capabilities and user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional intelligent assistants use pre-built FAQ databases or graph question answering systems, then they can provide structured responses, but they require high manpower and technical costs for maintenance and cannot cover a full range of questions
Solution Approach 1:
The patent replaces the mechanical system of pre-built FAQ databases and graph question answering systems with a large language model-based system. The LLM processes user queries through understanding, planning, tool selection, and response generation, eliminating the need for manual database construction and maintenance while achieving broader question coverage and more natural interactions.
Solution Approach 2:
The system enables self-service by allowing the large language model to autonomously process queries without requiring manual curation of FAQ databases. The model independently performs understanding, planning, tool selection, and response generation, reducing manpower requirements for system maintenance and expansion.
2Ease of operation
If conventional intelligent assistants use pre-built FAQ databases, then they can provide structured responses, but they offer insufficient intelligence in user experiences
Solution Approach 1:
The patent changes the fundamental parameter of how questions are processed by transitioning from static FAQ database matching to dynamic large language model generation. The LLM adapts its response based on the specific query context, user history, and selected tools, providing more comprehensive and intelligent responses that go beyond pre-defined structures.
Solution Approach 2:
The system incorporates feedback mechanisms where the large language model considers user history and query context to generate personalized responses. The model iteratively selects and invokes appropriate tools based on the query, refining its approach to provide more comprehensive and intelligent answers over time.
3Measurement precision
If the system uses large language models to process queries, then it can provide accurate and personalized responses, but it requires determining and selecting appropriate tools from multiple options
Solution Approach 1:
The patent introduces an intermediary planning mechanism that acts as a mediator between the user query and the tool execution. The large language model first plans the appropriate tool selection and invocation sequence based on the query, then executes these plans. This planning intermediary simplifies the overall process by breaking down complex queries into manageable tool invocation steps.
Solution Approach 2:
The system segments the query processing into distinct stages: understanding the query, planning the response strategy, selecting appropriate tools, invoking tools to gather information, and generating the final response. This segmentation allows the complex task of providing accurate and personalized responses to be managed through a series of simpler, organized steps.
Data Source
AI summary
A computer-implemented method for information processing based on a large language model is provided. The method includes obtaining query information provided by a user. The method further includes determining memory information related to the query information. The method further includes determining, based on the query information and the memory information, a tool for processing the query information. The method further includes invoking the tool to obtain auxiliary information. The method further includes generating, based on the query information and the auxiliary information, a result of processing the query information.