Automated Speech Recognition Testing via Recursive Path Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current testing methods for automatic speech recognition applications are labor-intensive, incomplete, and prone to errors, as they rely heavily on human testers and text-based simulations that fail to comprehensively test all functions of a voice application.

Innovation Solution

A system and method that generates automated test applications using a recursive algorithm to identify all possible paths through a voice application, analyzing 'speak' and 'listen' elements, and deploying a test application to execute and report on voice application functionality, including resource fetching, grammar accuracy, and response time testing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If human testers manually perform speech recognition application testing, then testing can be performed with simple tools, but the testing process becomes labor-intensive and incomplete

Engineering Contradiction:
Improvetesting operation simplicityVSAvoidtesting efficiency
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The system enables automated self-testing of speech recognition applications by generating test applications that automatically execute test cases, collect results, and produce test reports without requiring human testers to manually perform each testing step

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical human testing process with an automated system that uses speech recognition technology to generate and execute test cases, substituting manual human operations with automated computational processes

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

2Ease of operation

If human testers manually test speech recognition applications, then testing can be performed interactively, but errors and omissions occur when testers enter results into reports

Engineering Contradiction:
Improvetesting interactivityVSAvoidtesting accuracy
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system replaces manual data entry and result recording with automated electronic capture of test results directly from the speech recognition application, eliminating human error in transcription and reporting

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

Solution Approach 2:

The system implements automated feedback loops where test results are automatically captured, processed, and reported, ensuring accurate and complete recording of testing outcomes without manual intervention

Inventive Principle:
Principle #23Feedback

3Extent of automation

If text-based testing is used for voice applications, then testing can be automated to some extent, but it does not test the entire voice application comprehensively

Engineering Contradiction:
Improvetesting automation levelVSAvoidtesting coverage
Core Design Contradiction:
Extent of automationVSAdaptability or versatility

Solution Approach 1:

The system creates a universal testing framework that can test all functions of a voice application including speech recognition, text-to-speech, grammar validation, and resource fetching through a single automated test application generation process

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system dynamically generates test applications based on the specific voice application being tested, adapting the test cases to cover all relevant functions and paths of the target application rather than using static predefined test scripts

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP1936607B1Automated speech recognition application testing
Publication Date: 2012.07.25 SAP SE
  • EP1936607B1 patent drawingFigure 1
  • EP1936607B1 patent drawingFigure 2
  • EP1936607B1 patent drawingFigure 3

AI summary

The present application relates to speech recognition programs and, more particularly, automated speech recognition application testing. Various embodiments described herein provide systems, methods, and software that analyze voice applications and automatically generate test applications to test the voice applications.