Dynamic Voice Recognition Database Scoring for Undefined Instructions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional voice recognition systems fail to understand user intentions when instructions are not predefined in the database, leading to recognition failures and inability to execute desired operations.
Innovation Solution
An electronic apparatus and method that uses machine learning to increase the score of categories in a database based on user utterances, allowing for the registration of new instructions and creation of new categories, enabling the system to recognize and execute user intentions even if instructions are not previously defined.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a database with predefined instructions is used for voice recognition, then recognition accuracy for predefined commands is improved, but the system fails to recognize and execute undefined user instructions
Solution Approach 1:
The patent implements a dynamic voice recognition system where the database is continuously updated with new instructions learned from user utterances. The system transitions from a static predefined database to a dynamic one that adapts to user needs, allowing recognition of both predefined and undefined instructions while maintaining accuracy through machine learning-based category scoring.
Solution Approach 2:
The system performs self-learning by automatically analyzing user utterances, determining whether they represent new instructions, and updating the database accordingly. This self-service mechanism allows the system to expand its instruction recognition capabilities without requiring manual database updates, thereby improving both accuracy and adaptability.
2Adaptability or versatility
If the database is updated with new instructions from user utterances, then the system's ability to recognize undefined instructions is improved, but the complexity of database management increases
Solution Approach 1:
The system implements a feedback mechanism where user utterances are analyzed to determine if they represent new instructions. The analysis results feed back into the database update process, automatically incorporating learned instructions. This feedback loop manages database complexity by providing structured updates based on systematic analysis rather than arbitrary modifications.
Solution Approach 2:
The patent replaces manual database management mechanisms with automated machine learning-based systems. The machine learning model automatically analyzes utterances, determines new instructions, updates category scores, and manages database structure, substituting complex manual processes with intelligent automation that handles the complexity internally.
3Reliability
If machine learning is used to learn new instructions from user utterances, then the success probability of voice recognition is improved, but the processing time and computational resources increase
Solution Approach 1:
The system performs preliminary analysis of user utterances to determine whether they represent new instructions before full processing. By pre-evaluating utterances and only initiating comprehensive machine learning analysis when necessary, the system reduces overall processing time while maintaining high recognition success rates through targeted application of computational resources.
Data Source
AI summary
Disclosed is an electronic apparatus capable of controlling voice recognition. The electronic apparatus increases a score of a category corresponding to a word included in user's utterance in a database when the instruction included in the user's utterance is present in the database. The electronic apparatus checks whether the score of the category corresponding to the word is equal to or greater than a preset value when the instruction is not present in the database. The electronic apparatus registers the instruction in the database so that the instruction is included in the category corresponding to the word when the score is equal to or greater than the preset value as the check result.