Automated Software Testing Using Machine Learning: A Systematic Mapping Study

In today's digital world, software quality assurance is a crucial part of the Software Development Life Cycle (SDLC), and automated testing is essential to this effort. This study investigates how machine learning (ML) can augment automated testing by analyzing research from 2006 onward to determine its possible uses, prevalent techniques, and related advantages and disadvantages. Our findings reveal a growing interest in leveraging ML for testing, particularly in tasks like test case generation and user interface validation. Common ML techniques such as symbolic AI, evolutionary algorithms, and deep learning are emerging as the primary methods. ML holds promise for accelerating testing processes, enhancing accuracy, and improving adaptability. However, challenges such as sourcing high-quality training data, understanding complex ML models, and integrating ML with existing tools persist. This study illuminates the transformative potential of ML in testing and provides valuable insights for guiding future research in this dynamic field.