Automated Dictation Correction Using Similarity-Based Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech-to-text conversion technologies often make mistakes when users speak with accents, in noisy environments, or use uncommon acronyms, requiring manual user corrections that are inconvenient.
Innovation Solution
An electronic device analyzes user corrections to similar text portions, automatically correcting or recommending corrections for similar mistakes, enhancing the speech-to-text process by learning from user inputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If speech-to-text conversion is performed without automated correction, then the conversion process is simple and quick, but the accuracy of the converted text deteriorates due to mistakes from accents, noisy environments, or uncommon acronyms
Solution Approach 1:
The system analyzes user corrections to identified errors and uses this feedback to automatically correct similar errors elsewhere in the text. When a user corrects an error, the system learns from this correction and applies it to similar instances, creating a feedback loop that continuously improves accuracy without requiring full manual review
Solution Approach 2:
The system performs automated correction of similar errors based on user corrections, enabling the text conversion process to self-correct without requiring continuous user intervention. The system serves itself by automatically identifying and correcting patterns of errors based on learned corrections
2Measurement precision
If manual user corrections are required for all errors, then the text conversion accuracy can be improved, but the time and effort required for correction increases significantly
Solution Approach 1:
Instead of requiring users to manually correct all errors in the converted text, the system applies corrections only to similar instances based on identified errors. This partial action approach corrects multiple similar errors with a single user correction, significantly reducing the time and effort required while maintaining high accuracy
Solution Approach 2:
The system automatically identifies and corrects similar errors without requiring continuous user intervention. Once a user corrects one instance of an error, the system self-corrects similar instances throughout the text, reducing correction time while maintaining accuracy
3Measurement precision
If the system analyzes all text portions for corrections, then the accuracy of error identification improves, but the processing time and computational resources increase
Solution Approach 1:
The system focuses its analysis on specific portions of text that are similar to identified errors rather than analyzing the entire text uniformly. By applying local quality analysis to similar instances, the system achieves high error identification accuracy while reducing overall processing time through targeted analysis
Data Source
AI summary
Automated dictation correction technology, in which voice input of one or more sentences spoken by a user is captured by a microphone and converted to text data representing the one or more sentences. The text data is displayed on a user interface and a text correction of a first portion of the text data is entered by the user. Based on receipt of the text correction of the first portion of the text data, the text data is analyzed to assess similarity of the first portion of the text data to other portions of the text data. Based on the assessment, the system determines that an additional correction of a second portion of the text data is recommended. Based on the determination that the additional correction is recommended, the system performs an operation related to automated correction of the second portion of the text data.