Speech Recognition Search Term Scoring via Phonetic-Linguistic Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Automated speech recognition systems face challenges in accurately detecting specific words or phrases due to the need for carefully tuned search terms, which is a time-intensive and manual process, often resulting in false positives or false negatives, and requires repeated iterations for improvement.

Innovation Solution

A system and method for statistically analyzing search terms using a textual-linguistic analysis module, phonetic conversion module, and phonemic-linguistic analysis module to generate scores indicating the likelihood of accurate detection by a speech recognition engine, incorporating textual and phonemic features and their weighting factors to normalize search term scores.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual iterative tuning of search terms is performed by professionals, then speech recognition accuracy is improved, but time consumption and operational complexity increase significantly

Engineering Contradiction:
Improvespeech recognition accuracyVSAvoidtuning time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs self-tuning by automatically analyzing phonemic characteristics of search terms and identifying optimal terms without requiring manual intervention from linguistics professionals. The automated analysis includes computing phoneme frequencies, detecting phonemic patterns, and ranking search terms based on their expected recognition accuracy.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The manual mechanical process of professional tuning is replaced with an automated computational system that uses phonemic analysis algorithms, statistical modeling, and computer-based evaluation to determine optimal search terms, eliminating the need for human experts to manually iterate through tuning cycles.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If manually tuned search terms are used, then detection accuracy is improved, but device complexity and operational difficulty increase

Engineering Contradiction:
Improvedetection accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system introduces an intermediary automated analysis layer between the search term database and the speech recognition engine. This intermediary performs phonemic analysis, computes statistical characteristics, and filters/ranks search terms automatically, simplifying the overall system operation while maintaining high detection accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system changes the operational parameters from manual expert judgment to automated phonemic feature extraction and statistical analysis. By transforming the tuning process into computable parameters such as phoneme frequency, phonemic complexity, and pattern matching scores, the system reduces operational complexity while improving consistency and accuracy.

Inventive Principle:
Principle #35Parameter changes

3Ease of operation

If short search terms are used, then system simplicity is maintained, but detection accuracy decreases due to phonemic characteristics

Engineering Contradiction:
Improvesystem simplicityVSAvoiddetection accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The system performs preliminary phonemic analysis and evaluation of search terms before they are used in speech recognition. By pre-computing phonemic characteristics, frequencies, and detection probabilities, the system can automatically select or weight search terms to optimize accuracy without requiring complex manual configuration or lengthy search terms.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2301013B1Systems and methods of improving automated speech recognition accuracy using statistical analysis of search terms
Publication Date: 2017.05.03 CALABRIO INC
  • EP2301013B1 patent drawingFigure 1
  • EP2301013B1 patent drawingFigure 2
  • EP2301013B1 patent drawingFigure 3

AI summary

Systems and methods of improving speech recognition accuracy using statistical analysis of word or phrase-based search terms are disclosed. An illustrative system for statistically analyzing search terms includes an interface adapted to receive a text-based search term, a textual-linguistic analysis module that detects textual features within the search term and generates a first score, a phonetic conversion module that converts the search term into a phoneme string, a phonetic-linguistic analysis module that detects phonemic features within the phoneme string and generates a second score, and a score normalization module that normalizes the first and second scores and outputs a search term score to a user or process.