Automated Program Repair and Test Overfitting: Measurements and Approaches using Formal Methods

Amirfarhad Nilizadeh

doi:10.1109/icst53961.2022.00061

Automated Program Repair and Test Overfitting: Measurements and Approaches using Formal Methods

Amirfarhad Nilizadeh

https://doi.org/10.1109/icst53961.2022.00061

Copy DOI

Publication Date: Apr 1, 2022

Citations: 8

Affiliation: University of Central Florida

#Automated Program Repair #Automated Program Repair Tools + Show 7 more

Abstract
Full-Text PDF
Similar Papers

Abstract

Automated program repair (APR) techniques that use test suites for bug localization and assessing candidate patches have promising results. Studies have shown that generated patches by APR tools that pass its test suite do not always correctly implement the program's expected behavior, known as test overfitting. My research aims to provide fresh insight into the test overfitting problem by using formal methods.

Full Text