- author = "Linsey Kitt and Myra B. Cohen",
- title = "Partial Specifications for Program Repair",
- booktitle = "GI @ ICSE 2021",
- year = "2021",
- month = "30 " # may,
- editor = "Justyna Petke and Bobby R. Bruce and Yu Huang and Aymeric Blot and Westley Weimer and W. B. Langdon",
- publisher = "IEEE",
- address = "internet",
- pages = "19--20",
- keywords = "genetic algorithms, genetic programming, genetic improvement, APR, program repair, overfitting, test suites, PyGGI 2.0, triangle program",
- isbn13 = "978-1-6654-4466-8/21",
-
URL = "
https://geneticimprovementofsoftware.com/paper_pdfs/gi2021icse/kitt_gi-icse_2021.pdf",
-
video_url = "
https://www.youtube.com/watch?v=NrmaeHMR_Lw&list=PLI8fiFpB7BoKDaxvS7SQp0iA7fN7rrvDD&index=21",
-
video_url = "https://www.youtube.com/watch?v=GMFU4DpnO3g&list=PLI8fiFpB7BoKDaxvS7SQp0iA7fN7rrvDD&index=22",
-
video_url = "https://www.youtube.com/watch?v=oJzYRJTYgyk&list=PLXTjhGKkSnI-se7uQneCX-pEDiDrQ7TIS&index=10",
-
DOI = "
doi:10.1109/GI52543.2021.00012",
- size = "2 pages",
- abstract = "we argue for using many partial test suites instead of one full test suite during program repair. This may provide a pool of simpler, yet correct patches, addressing both the overfitting and poor repair quality problem. To support this idea, we present some insight obtained running APR partial test suites on the well studied triangle program.",
-
notes = "Triangle Program. Partial specification. Removing
tests from test suite => many more simple repairs
Video oJzYRJTYgyk Linsey Kitt
2:23 Discussion Chair:Aymeric Blot. A: Linsey Kitt(Blue) and Myra B. Cohen (black+white)
2:23 Q: Aymeric Blot fewer test cases leads to simpler repairs, question why will these smaller repair generalise? A: Linsey Kitt code may be modular so fix may also be modular
3:47 Q: over fitting A: Linsey Kitt sometimes we do see overfitting, but post evolution can use full test suite to may be clean up.
4:35 Q: Erik Fredericks, Lexicase selection. A: Linsey Kitt
5:08 Q: Justyna Petke, specification A: Linsey Kitt remove set of related tests (eg all those to do with scalene triangles). More studies.
6:01 Alexandre Bergel stop early? A: Linsey Kitt
6:58 Q: Westley Weimer. A: Linsey Kitt. A: Myra B. Cohen \cite{Guizzo:2021:ICSE} Relaxing the constraints, post evolution clean up. A: Linsey Kitt. Here we are choosing specific tests, ie not random subsets of tests.
part of \cite{Petke:2021:ICSEworkshop} http://geneticimprovementofsoftware.com/events/icse2021.html",
