Interactive Intent Refinement for More Accurate LLM Responses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large language models often fail to accurately meet user needs, resulting in poor user experience due to mismatched instruction execution results.
Innovation Solution
An interactive processing method that identifies user intent through intent identification models, adjusts it based on target reference information such as user profiles, context, and feedback to obtain a more accurate second user intent, followed by a response operation tailored to this intent.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a large language model directly processes user input data to generate instructions, then the processing speed is fast, but the instruction execution results may not meet user needs, resulting in poor user experience
Solution Approach 1:
The patent segments the intent processing into multiple stages: first user intent identification, then second user intent refinement based on reference information. This segmentation allows each stage to focus on specific tasks, improving overall reliability without requiring complete system redesign.
Solution Approach 2:
The patent introduces target reference information as an intermediary element between user input and instruction generation. This intermediary provides additional context and constraints that guide the large language model to generate more accurate instructions, improving user experience quality.
2Measurement precision
If target reference information is introduced to refine user intent, then the accuracy of instruction execution results improves, but the processing time and system complexity increase
Solution Approach 1:
The patent performs preliminary actions by identifying the first user intent before introducing target reference information. This preliminary intent identification provides a foundation that guides subsequent refinement, reducing the time required for second intent analysis compared to starting from scratch.
Solution Approach 2:
The patent applies partial action by selectively using target reference information based on the specific input data and context. Not all inputs require full reference information processing, allowing the system to balance accuracy improvement with time efficiency on a case-by-case basis.
Data Source
AI summary
An interactive processing method includes identifying an obtained target input data to obtain a first user intent, processing the first user intent based on target reference information to obtain a second user intent and performing a response operation matching the second user intent, the target reference information corresponding to different target input data being different or the same.