User Intent Recognition Using Utterance Frequency Threshold Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional voice recognition-based devices struggle to accurately interpret user utterances, leading to misinterpretation of user intent, such as distinguishing between local commands (e.g., searching for a song) and global commands (e.g., searching for a location) due to unclear intent recognition.
Innovation Solution
A method and apparatus that utilize utterance frequency data to determine the proportion of local commands and adjust the confidence score threshold based on this proportion, enabling accurate determination of user intent by distinguishing between local and global commands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional voice recognition systems use fixed confidence score thresholds, then the system structure remains simple, but the accuracy of user intent recognition deteriorates
Solution Approach 1:
The patent implements dynamic adjustment of the confidence score threshold based on the user's utterance frequency data. The threshold is no longer fixed but adapts according to the determined proportion of local commands from historical data, allowing the system to maintain high accuracy while managing complexity through data-driven parameter adjustment
Solution Approach 2:
The patent changes the confidence score threshold parameter dynamically based on utterance frequency analysis. By calculating the proportion of local commands from historical utterance data and using this to adjust the threshold, the system optimizes recognition accuracy without requiring complex structural changes
2Reliability
If the system uses a single threshold for all commands, then the operation is simple, but the reliability of command classification deteriorates
Solution Approach 1:
The patent performs preliminary analysis of utterance frequency data to determine the proportion of local commands before processing new utterances. This pre-calculated proportion is then used to set an optimized confidence score threshold, improving command classification reliability without adding complexity during real-time operation
Solution Approach 2:
The system automatically adjusts the confidence score threshold using its own historical utterance data without requiring external intervention or manual calibration. The utterance frequency data from the system's normal operation is leveraged to self-optimize the threshold parameter
Data Source
AI summary
A method and apparatus for understanding user intent by using user's utterance frequency data. An aspect of the present disclosure provides a method for understanding an utterance intent using utterance frequency data of a user, the method comprising: checking a screen ID of a previous screen before a user utterance is input; obtaining a number of utterances per screen ID from the utterance frequency data; obtaining a ratio of user utterances intended as a local command from the utterance frequency data when the number of utterances per screen ID is larger than or equal to a predetermined number for all screen ID; and adjusting a threshold of a confidence score based on the ratio.