Prophet: Automatic patch generation via learning from successful human patches

We present Prophet, a novel patch generation system that learns a probabilistic model over candidate patches from a large code database that contains many past successful hu-man patches. It defines the probabilistic model as the com-bination of a distribution over program points based on error localization algorithms and a parameterized log-linear distri-bution over modification operations. It then learns the model parameters via maximum log-likelihood, which identifies important characteristics of the successful human patches. For a new defect, Prophet generates a search space that con-tains many candidate patches, applies the learned model to prioritize those potentially correct patches that are consis-tent with the identified successful patch characteristics, and then validates the candidate patches with a user supplied test suite. 1