2021/02/11 by Yigit Kucuk, Kucuk, Yigit, Tim A. D. Henderson +3 · 1 citation
Computer Science · Mathematics · #Artificial intelligence #Causal inference #Causation #Computer science #Copula (linguistics) #Data mining #Dependence analysis #Econometrics #FOS: Computer and information sciences #Inference #Machine learning #Mathematics #Natural language processing #Predicate (mathematical logic) #Programming language #Software Engineering (cs.SE) #Software Engineering Research #Software Reliability and Analysis Research #Software Testing and Debugging Techniques #Statistical inference #Statistics #Variable (mathematics) #cs.SE
paper · pdf · doi:10.48550/arxiv.2102.06292
published in arXiv (Cornell University) (Cornell University)
arxiv created 2021/02/11 · openalex publication_date 2021/02/11 · arxiv updated 2021/02/15 · openalex created_date 2022/07/25 · openalex updated_date 2026/08/06
Statistical fault localization (SFL) techniques use execution profiles and\nsuccess/failure information from software executions, in conjunction with\nstatistical inference, to automatically score program elements based on how\nlikely they are to be faulty. SFL techniques typically employ one type of\nprofile data: either coverage data, predicate outcomes, or variable values.\nMost SFL techniques actually measure correlation, not causation, between\nprofile values and success/failure, and so they are subject to confounding bias\nthat distorts the scores they produce. This paper presents a new SFL technique,\nnamed \UniVal, that uses causal inference techniques and machine learning\nto integrate information about both predicate outcomes and variable values to\nmore accurately estimate the true failure-causing effect of program statements.\n\UniVal was empirically compared to several coverage-based,\npredicate-based, and value-based SFL techniques on 800 program versions with\nreal faults.\n