Improving Fault Localization by Integrating Value and Predicate Based Causal Inference Techniques

TL;DR

This paper introduces UniVal, a novel statistical fault localization method that combines predicate and variable data using causal inference to improve fault detection accuracy.

Contribution

UniVal is the first SFL technique to integrate predicate outcomes and variable values with causal inference, reducing confounding bias in fault localization.

Findings

Outperforms existing SFL techniques on 800 program versions with real faults.

More accurately identifies faulty program elements.

Reduces false positives in fault localization.

Abstract

Statistical fault localization (SFL) techniques use execution profiles and success/failure information from software executions, in conjunction with statistical inference, to automatically score program elements based on how likely they are to be faulty. SFL techniques typically employ one type of profile data: either coverage data, predicate outcomes, or variable values. Most SFL techniques actually measure correlation, not causation, between profile values and success/failure, and so they are subject to confounding bias that distorts the scores they produce. This paper presents a new SFL technique, named \emph{UniVal}, that uses causal inference techniques and machine learning to integrate information about both predicate outcomes and variable values to more accurately estimate the true failure-causing effect of program statements. \emph{UniVal} was empirically compared to several coverage-based, predicate-based, and value-based SFL techniques on 800 program versions with real faults.