Gradient-Based Program Repair: Fixing Bugs in Continuous Program Spaces

Created by W.Langdon from gp-bibliography.bib Revision:1.8414

@Misc{silva2025gradientbasedprogramrepairfixing,

• author = "Andre Silva and Gustav Thoren and Martin Monperrus",
• title = "Gradient-Based Program Repair: Fixing Bugs in Continuous Program Spaces",
• howpublished = "ArXiv 2505.17703",
• year = "2025",
• month = "23 " # may,
• keywords = "genetic algorithms, genetic programming, genetic improvement, APR, bugfixing, SBSE, ANN, RASP, Tracr, mutation testing, Dyck2",
• primaryclass = "cs.PL",
• URL = "https://arxiv.org/abs/2505.17703",
• size = "14 pages",
• abstract = "... We introduce Gradient-Based Program Repair (GBPR), a new paradigm that reframes program repair as continuous optimization in a differentiable numerical program space. Our core insight is to compile symbolic programs into differentiable numerical representations, enabling search in the numerical program space directly guided by program behaviour. To evaluate GBPR, we present RaspBugs, a new benchmark of 1466 buggy symbolic RASP programs and their respective numerical representations. Our experiments demonstrate that GBPR can effectively repair buggy symbolic programs by gradient-based optimization in the numerical program space, with convincing repair trajectories...",

}

Genetic Programming entries for Andre Silva Gustav Thoren Martin Monperrus

Citations