The integration of machine learning into automated test generation: A systematic mapping study

Machine learning (ML) may enable effective automated test generation. We characterize emerging research, examining testing practices, researcher goals, ML techniques applied, evaluation, and challenges in this intersection by performing. perform a systematic mapping study on sample of 124 publications. generates input for system, GUI, unit, performance, combinatorial or improves the performance existing generation methods. is also used to generate verdicts, property-based, expected output oracles. Supervised learning—often based neural networks—and reinforcement Q-learning—are common, some publications employ unsupervised semi-supervised learning. (Semi-/Un-)Supervised approaches are evaluated using both traditional metrics ML-related (e.g., accuracy), while often tied reward function. The work-to-date shows great promise, but there open regarding training data, retraining, scalability, evaluation complexity, algorithms employed—and how they applied—benchmarks, replicability. Our findings can serve as roadmap inspiration researchers field.