Generative Test Script Creation for Digital Therapeutics V&V
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing V&V processes for digital therapeutic applications suffer from discrepancies between the designed and implemented software, leading to delays, resource wastage, and compromised therapeutic efficacy due to mismatches in functionalities, user interface design, performance, compliance, and security, which are often identified late in the development cycle.
Innovation Solution
A generative machine learning model is used to automate the generation of test configurations, including test documentation and scripts, to ensure precise verification and validation, reducing discrepancies and accelerating the V&V process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual creation of test scripts and documentation is used, then quality and thoroughness of V&V can be maintained, but time consumption increases significantly (hundreds of hours and several weeks)
Solution Approach 1:
The patent uses generative AI models to create test scripts and documentation by copying patterns from existing high-quality V&V materials. The AI learns from training data comprising previously approved test cases, documentation templates, and regulatory guidelines, then generates new test scripts that replicate the quality and structure of manual creations, thereby reducing time while maintaining reliability
Solution Approach 2:
The patent replaces the mechanical manual writing process with an automated AI-based system. Instead of quality assurance teams manually crafting each test script and documentation piece, the system uses generative AI models to automatically produce these artifacts based on input requirements, significantly reducing the time investment while maintaining the required quality standards through AI-generated content
2Reliability
If extensive V&V documentation and script writing is performed, then compliance with regulatory requirements is improved, but development cycles are extended
Solution Approach 1:
The patent implements preliminary action by generating complete V&V documentation and test scripts before the actual testing begins. The AI model creates all necessary documentation artifacts in advance based on the software requirements and regulatory guidelines, allowing the development team to proceed to execution phases without delays, thus maintaining compliance while accelerating the overall development cycle
Solution Approach 2:
The system copies proven compliance structures and documentation formats from training data that includes previously approved regulatory submissions. By replicating the successful templates and structures from existing compliant documentation, the system ensures regulatory compliance is met while reducing the time needed to create these documents from scratch
3Measurement precision
If test plans are executed late in the development cycle, then discrepancies between designed and implemented software can be identified, but significant adjustments are required causing further delays
Solution Approach 1:
The patent applies preliminary action by generating test scripts and documentation in advance, before the software development is complete. This allows discrepancies between designed and implemented functionality to be detected early in the cycle, enabling timely adjustments to be made while the software architecture is still flexible, rather than discovering issues late when significant rework would be required
4Reliability
If manual V&V processes are used, then thorough testing can be performed, but computing resources and network bandwidth are wasted due to incongruent test plans
Solution Approach 1:
The AI model copies accurate patterns from training data to generate congruent and coherent test plans that align with the actual software implementation. By learning from previously successful test cases and documentation, the AI ensures that generated test plans are internally consistent and match the software requirements, preventing resource waste from executing incongruent or contradictory test scenarios
Data Source
AI summary
Aspects of the present disclosure are directed to systems, methods, and computer readable media for automated generation of test scripts and documentation for verifying and validating digital therapeutics applications. A service may receive a test configuration identifying a plurality of test cases to check an application executable on a user device for addressing an indication of a user. The service may provide a model input generated using the test configuration to a generative model. The generative model may be trained a set of corpora identifying test cases and test packages. The service may generate, based on providing the model input to the generative model, a test package defining execution of the plurality of test cases to check the application. The service may store an association between the application and the test package on a database.