Machine Learning Test Script Recommendation for Web UI Automation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The process of generating and maintaining test scripts for web application user interfaces (UIs) is time-consuming and inefficient, as it involves numerous scripts and requires manual effort to validate interactions and operations, making it difficult to effectively automate UI testing.

Innovation Solution

A system and method that uses machine learning to recommend existing test scripts to customers based on page object characteristics and test script requirements, grouping similar scripts and using a neural network to generate generic page object representations, which can be compiled into multiple programming languages for use in UI test automation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual test script generation is used, then test scripts can be created with full customization, but the process is time-consuming and inefficient

Engineering Contradiction:
Improvetest script generation speedVSAvoidtime for generating and maintaining test scripts
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The system enables self-service automation where the machine learning model automatically generates test scripts by analyzing page object characteristics and matching them with existing test script patterns, eliminating the need for manual script creation while maintaining high customization through intelligent recommendation

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system creates copies of existing test scripts by matching page object characteristics with historical test script patterns, generating new test scripts that replicate successful patterns from existing scripts while adapting to specific page object requirements

Inventive Principle:
Principle #26Copying

2Reliability

If numerous test scripts are created to cover all UI interactions, then testing coverage is improved, but the complexity of managing and maintaining these scripts increases

Engineering Contradiction:
ImproveUI testing coverageVSAvoidcomplexity of test script management
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system creates universal test scripts that can be applied across multiple similar page objects by identifying common patterns and characteristics, allowing a single test script template to serve multiple functions and reduce the total number of scripts needed while maintaining comprehensive coverage

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

Solution Approach 2:

Instead of creating unique scripts for each UI element, the system copies and adapts existing test script patterns to match new page objects, reducing management complexity by reusing proven scripts rather than creating new ones from scratch

Inventive Principle:
Principle #26Copying

3Reliability

If manual maintenance of test scripts is performed, then scripts can be kept up-to-date with application changes, but the effort and time required for maintenance increases significantly

Engineering Contradiction:
Improvetest script accuracyVSAvoidtime for maintaining test scripts
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system implements feedback loops where test execution results and page object characteristic changes automatically trigger script updates, continuously refining test scripts based on actual usage patterns and application state changes without requiring manual intervention

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The maintenance process is automated through self-service mechanisms where the machine learning model continuously learns from page object characteristics and test results, automatically updating and optimizing test scripts without human effort while maintaining high accuracy

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS11636025B2Intelligent generation of automated user interface testing methods
Publication Date: 2023.04.25 SALESFORCE INC
  • US11636025B2 patent drawing
  • US11636025B2 patent drawing
  • US11636025B2 patent drawing

AI summary

Systems and methods for recommending test scripts may include identifying a plurality of test scripts associated with a web application, each of the test scripts associated with a set of page objects from a plurality of page objects of the web application, wherein a set of page objects to be tested by a test script is associated with a set of page object characteristics; generating a plurality of groupings of test scripts from the plurality of test scripts based on page objects with similar sets of page object characteristics; receiving a request for a first test script to test a first set of page objects, the request including a first set of page object characteristics associated with the first test script; and identifying the first test script from the plurality of groupings of test scripts based on the first set of page objects and the first set of page object characteristics.