Wiki Home
Automated program repair
Definition
Automated program repair (APR) takes a program that fails some correctness criterion and produces a modified program that passes it, with no human writing the patch. The criterion is an oracle: usually a regression test suite, sometimes a formal property or a crashing input run under a sanitizer. Classical generate-and-validate systems search a space of candidate edits and keep any that pass the tests; GenProg, which mutates the program with genetic operators, is the canonical example (Le Goues et al., 2012). A recurring pitfall is that a patch can be plausible, passing the supplied tests, without being correct, so weak test suites overstate repair rates (Qi et al., 2015). Learning-based repair instead trains neural sequence-to-sequence models on mined bug-fix pairs to predict a developer-like patch (Tufano et al., 2019; Chen et al., 2021). Repairing a security vulnerability is a specialization in which the oracle must also witness that the exploited weakness is gone.
Papers that use this concept
- Examining Zero-Shot Vulnerability Repair with Large Language Models — asks whether prompting a pretrained code model can stand in for a task-specific repair model, evaluated inside a generate-and-validate framework with separate security and functional oracles.
See also
References
- Le Goues, C., Nguyen, T., Forrest, S., and Weimer, W. “GenProg: A Generic Method for Automatic Software Repair.” IEEE Transactions on Software Engineering, 38(1):54-72, 2012.
- Qi, Z., Long, F., Achour, S., and Rinard, M. “An Analysis of Patch Plausibility and Correctness for Generate-and-Validate Patch Generation Systems.” International Symposium on Software Testing and Analysis (ISSTA), 2015.
- Tufano, M., Watson, C., Bavota, G., Di Penta, M., White, M., and Poshyvanyk, D. “An Empirical Study on Learning Bug-Fixing Patches in the Wild via Neural Machine Translation.” ACM Transactions on Software Engineering and Methodology, 28(4):19:1-19:29, 2019.
- Chen, Z., Kommrusch, S., Tufano, M., Pouchet, L.-N., Poshyvanyk, D., and Monperrus, M. “SequenceR: Sequence-to-Sequence Learning for End-to-End Program Repair.” IEEE Transactions on Software Engineering, 47(9):1943-1959, 2021.