Machine Learning-assisted Fault Injection

Fault Injection (FI) is a method for system validation and verification in which the tester evaluates the system behavior resulting from the introduction of faults into the system under test. This paper proposes a model-based approach to improve the efficiency of the FI process by utilizing Machine Learning (ML) and formalized domain knowledge. This ML algorithm uses a probabilistic automaton to reduce the manual effort required in the testing procedure as the algorithm can automatically make decisions and predictions about catastrophic fault parameters. This assists the tester in dealing with complicated and broad-scale systems by enabling higher fault coverage with fewer simulations.