Natural-Language Test Script Mapping Using LSTM Code Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing testing methods for devices and systems lack efficient and accurate automated generation of test scripts, particularly in mapping test steps to code functions, which hinders comprehensive defect detection and product improvement.
Innovation Solution
An apparatus utilizing a machine learning system, specifically a Long Short-Term Memory (LSTM) neural network, maps test step information in natural language to code functions, determining parameter values in multiple phases to generate automated test scripts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated test script generation is implemented using traditional methods, then testing coverage can be improved, but the accuracy of mapping test steps to code functions deteriorates
Solution Approach 1:
The patent replaces traditional mechanical/manual test script generation methods with an LSTM-based neural network system. The LSTM model automatically learns the mapping between natural language test steps and code functions from training data, eliminating the need for manual mapping while achieving high accuracy through pattern recognition and sequence modeling.
Solution Approach 2:
The patent transforms the parameters of test step description from structured formal languages to natural language inputs. The LSTM model processes variable-length sequences of natural language tokens and maps them to corresponding code functions, allowing flexible input while maintaining accurate mapping through learned parameters from training.
2Measurement precision
If manual test script generation is used, then mapping accuracy can be maintained, but testing efficiency deteriorates
Solution Approach 1:
The patent implements a self-service system where the LSTM model automatically generates test scripts from natural language descriptions without requiring manual intervention. The model serves itself by learning from training data and autonomously performing the mapping task, thereby maintaining accuracy while dramatically improving efficiency.
Solution Approach 2:
The patent performs preliminary training of the LSTM model on annotated test case data before actual test script generation. This preliminary action enables the model to learn effective mappings in advance, so that during actual use, accurate and efficient script generation can be achieved without manual effort.
3Ease of manufacture
If traditional test script generation methods are used, then implementation simplicity can be maintained, but automation level deteriorates
Solution Approach 1:
The patent introduces an intermediary LSTM-based natural language processing system between the test requirements and the code generation process. This intermediary automatically translates natural language test steps into code functions, simplifying the overall implementation by eliminating complex manual mapping procedures while achieving high automation.
Data Source
AI summary
An apparatus in an illustrative embodiment comprises at least one processing device that includes at least a processor and a memory coupled to the processor. The at least one processing device is configured to obtain test step information in natural language, to apply the test step information to a machine learning system configured to map the test step information to one or more code functions, to determine values for one or more parameters in the one or more code functions, and to execute the one or more code functions in a test script execution environment utilizing the determined values for the one or more parameters. The machine learning system in some embodiments comprises a long short-term memory (LSTM) neural network configured to receive a sequence of text tokens of the test step information and to map the sequence of text tokens to a particular code function.