Task Scene Intention Refinement Using Confidence-Based Supplementary Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing user intention recognition methods based on large models lack human-like reasoning and planning capabilities, leading to unreliable recognition results in scene tasks such as situational awareness and intention understanding, limiting their application range.
Innovation Solution
A task processing method that determines a task scene, user intention, and confidence, obtains supplementary data to adjust the user intention, and generates a processing result based on enhanced user intention, using an electronic device with processors and memory to execute these steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If large models are used for user intention recognition in scene tasks, then the processing capability is improved, but the reliability of recognition results deteriorates due to lack of human-like reasoning and planning capabilities
Solution Approach 1:
The patent segments the user intention recognition process into multiple stages: initial recognition by large model, confidence assessment, supplementary information acquisition, and iterative refinement. This segmentation allows the system to maintain high processing capability while improving reliability through structured multi-step verification and refinement cycles.
2Speed
If large models directly generate processing results, then the response speed is improved, but the accuracy of user intention recognition deteriorates
Solution Approach 1:
The patent implements preliminary action by having the large model perform initial user intention recognition and confidence assessment before final result generation. Supplementary task data is acquired in advance when confidence is insufficient, allowing the system to maintain fast response for high-confidence cases while ensuring accuracy for uncertain cases through pre-computed refinements.
3Measurement precision
If supplementary task data is obtained to adjust user intention, then the recognition accuracy is improved, but the processing time increases
Solution Approach 1:
The patent applies partial action by acquiring supplementary task data only when the initial recognition confidence falls below a threshold. This selective approach ensures that full processing time and data acquisition resources are consumed only when necessary, maintaining fast processing for clear cases while achieving high accuracy for ambiguous cases through targeted supplementary data collection.
Data Source
AI summary
A task processing method includes obtaining a to-be-processed task, determining current task data of the to-be-processed task, determining a task scene to which the to-be-processed task belongs based on the current task data, determining a first user intention corresponding to the to-be-processed task and intention confidence based on the task scene, obtaining supplementary task data for the to-be-processed task based on the intention confidence, adjusting the first user intention based on the supplementary task data to obtain a second user intention, and generating a processing result for the to-be-processed task based on the second user intention.