Phonetic Search False Positive Detection via ASR Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Phonetically similar phrases often cause false positives in search systems, leading to incorrect identification of documents, as systems struggle to differentiate between similar sounding phrases.
Innovation Solution
The use of automatic speech recognition (ASR) to identify and differentiate between searched phrases and phonetically similar phrases, with a system that generates a list of potential causes for false positives, including confidence scores and thresholds to filter out confusable phrases, and generates a query operator to exclude these phrases from search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If phonetic search is used to search for documents, then the search coverage is improved, but false positives increase due to phonetically similar phrases
Solution Approach 1:
The patent segments the phonetic search results into two categories: exact phrase matches and potential false positives. By dividing the search results and applying different verification methods to each segment, the system maintains broad search coverage while filtering out phonetically similar but semantically different phrases through ASR verification.
Solution Approach 2:
The patent introduces automatic speech recognition (ASR) as an intermediary verification layer between phonetic search and final result delivery. The ASR system acts as a mediator that verifies whether phonetically similar phrases are actual matches or false positives, thereby resolving the contradiction between comprehensive search and accurate results.
2Productivity
If phonetic search is used to find documents, then more relevant documents are found, but measurement precision deteriorates due to incorrect phrase identification
Solution Approach 1:
The patent performs preliminary phonetic search to retrieve potential matches efficiently, then applies ASR verification as a subsequent filtering step. This two-stage approach enables high productivity in the first stage while ensuring measurement precision in the second stage through confidence score-based filtering.
Solution Approach 2:
The patent changes the verification parameter from exact string matching to phonetic similarity matching with confidence scores. By adjusting the confidence threshold parameter, the system can balance between retrieving more documents (high productivity) and ensuring accurate phrase identification (high precision).
Data Source
AI summary
A system and method for detecting phonetically similar imposter phrases may include using automatic speech recognition (ASR) to search for a first phrase in a set of objects; producing a list of references by searching for the first phrase in the set of objects using phonetic search; using output produced by the ASR to determine whether or not a reference in the list points to a phrase that is the same as the first phrase; and if it is determined that the reference points to a second phrase that is different from the first phrase then marking the second phrase as a potential cause for a phrase search false positive.