AI Speech Recognition Confidence Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current devices that perform control operations using image or sound recognition lack the ability to determine if input data is sufficiently recognized, leading to inaccurate control operations and inadequate feedback when recognition fails, resulting in a low confidence level.
Innovation Solution
An AI apparatus that measures the confidence level of speech recognition and performs control operations only when the confidence level meets a predetermined value, while providing feedback on failure and analyzing the cause of low confidence for improved recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If speech recognition is performed without confidence level measurement, then the device complexity is reduced, but the reliability of control operation is worsened due to inaccurate recognition
Solution Approach 1:
The patent applies preliminary action by measuring the confidence level of speech recognition results before executing control operations. The processor determines whether to perform a control operation based on the measured confidence level, ensuring that only sufficiently reliable recognition results trigger actions. This prevents inaccurate control operations while maintaining system simplicity.
2Productivity
If only generic feedback is provided when recognition fails, then the ease of operation is improved, but the productivity is worsened due to repeated recognition failures
Solution Approach 1:
The patent implements feedback by providing specific information about recognition failure causes to the user. When speech recognition fails, the processor identifies and communicates the specific reason (e.g., low confidence level, noise interference), enabling users to adjust their input accordingly. This targeted feedback improves recognition rates by addressing root causes rather than merely indicating failure.
3Manufacturing precision
If control operations are performed without verifying sufficient recognition, then the speed of operation is improved, but the manufacturing precision is worsened due to erroneous speech recognition
Solution Approach 1:
The patent applies preliminary action by verifying the confidence level of speech recognition before executing control operations. The processor measures the confidence level and determines whether to proceed with the control operation based on this verification. This ensures high precision in control operations by preventing erroneous executions, while the automated confidence level check maintains operational speed.
Data Source
AI summary
An embodiment of the present invention provides an artificial intelligence (AI) apparatus for recognizing a speech of a user, the artificial intelligence apparatus includes a memory to store a speech recognition model and a processor to obtain a speech signal for a user speech, to convert the speech signal into a text using the speech recognition model, to measure a confidence level for the conversion, to perform a control operation corresponding to the converted text if the measured confidence level is greater than or equal to a reference value, and to provide feedback for the conversion if the measured confidence level is less than the reference value.