Voice Input Keyword Matching for Application Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing information processing device struggles to accurately distinguish between applications with similar keywords, leading to inadequate selection accuracy when using voice input for control targets.
Innovation Solution
An information processing device that identifies search keywords from input data through voice input and selects control targets based on calculated numeric values from predefined expressions, enhancing the accuracy of selection by classifying keywords into required and preferred words.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If weights are assigned to first voice and second voice respectively, then voice priority determination is achieved, but application selection accuracy cannot be sufficiently improved
Solution Approach 1:
The patent segments the search keyword processing into multiple stages: extracting main words from voice input, identifying search keywords from predefined lists, and calculating numeric values through expressions. This segmentation allows the system to handle complex search requirements systematically, improving application selection accuracy by breaking down the keyword matching process into manageable steps that can be optimized independently.
Solution Approach 2:
The patent transforms search keywords into numeric values through mathematical expressions, changing the parameter representation from textual keywords to numerical scores. This parameter transformation enables more precise comparison and selection among applications, as numeric values provide a standardized metric for evaluating match quality between voice input and application keywords, thereby improving selection accuracy without proportionally increasing complexity.
2Measurement precision
If search keywords are matched with main words from voice input, then control target identification is achieved, but distinction between applications with similar keywords remains difficult
Solution Approach 1:
The patent introduces numeric values as an intermediary representation between search keywords and application selections. By converting keywords into numeric values through expressions, the system creates a intermediate layer that captures the semantic relationships and distinctions between keywords more effectively than direct textual matching. This intermediary transformation preserves distinction information between similar keywords, enabling accurate control target identification even when applications have similar keywords.
3Measurement precision
If multiple search keywords are processed through expressions, then selection accuracy is improved, but processing time increases
Solution Approach 1:
The patent prepares predefined search keywords and expressions in advance, before the actual search operation. By pre-organizing the keyword lists and preparing the expression templates, the system eliminates the need to create these structures during real-time processing. This preliminary preparation significantly reduces processing time during actual search operations, as the system only needs to match voice input against pre-defined keywords and evaluate pre-prepared expressions, thereby improving accuracy without proportionally increasing processing time.
Data Source
AI summary
A control section includes: an identifying section configured to, by referring to one or more search keywords set for one or more control targets, identify at least one search keyword from among the one or more search keywords, the at least one search keyword matching any of one or more main words contained in input data acquired through voice input; and a selecting section configured to select at least one control target from among the one or more control targets based on one or more numeric values obtained through calculation of one or more expressions each of which batch-converts, into numerical form, one or more of the one or more search keywords set for the one or more control targets.