Scalable statistical bug isolation

By Ben Liblit, Mayur Naik, Alice X. Zheng, A. Aiken, Michael I. Jordan. 2005

912 citations recorded by the source corpus.

Subfields: Information Systems, Software, Optimization & Training

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

In Software

In Optimization & Training

Citations in the corpus

A Survey of Machine Learning for Big Code and Naturalness (2017, background): “[117] are indicative of the possibilities within this area.”
Prochlo: Strong Privacy for Analytics in the Crowd (2017, background): “, for understanding software behavior or bugs, software performance or resource use, or for other software improvements [13, 39, 46, 59, 61, 77].”
Where should the bugs be fixed? More accurate information retrieval-based bug localization based on bug reports (2012, methodology): “Unlike spectrum-based fault localization techniques [1, 18, 19, 22, 23], information retrieval (IR) based bug localization does not require program execution information (such as passing and failing traces).”
The ManyBugs and IntroClass Benchmarks for Automated Repair of C Programs (2015, background): “Since there are not enough developer resources to repair all of these defects before deployment, it is well known that programs ship with both known and unknown bugs [27].”
Fault Analysis and Debugging of Microservice Systems: Industrial Survey, Benchmark System, and Empirical Study (2018, background): “Similarly related fault localization techniques are designed mainly for sequential programs, such as statistic-based fault localization [39], and machine-learning-based ones [40], [41].”
Statistical debugging: simultaneous identification of multiple bugs (2006, methodology): “In earlier work (Liblit et al., 2005), we proposed a predicate ranking algorithm based on the harmonic mean of predicate sensitivity and specificity.”
BugFix: A learning-based tool to assist developers in fixing bugs (2009, methodology): “Other approaches [17, 21, 24, 25] use runtime information to rank program entities in the order of their likelihood of being faulty.”
A Survey on Software Fault Localization (2016, methodology): “A statistical debugging technique (Liblit05) that can isolate bugs in programs with instrumented predicates at particular points is presented in [222].”
Debugging Machine Learning Tasks (2016, methodology): “The CBI project and its variants [22, 23, 24] use information collected from program runs together with a statistical analysis in order to compute predicates that are highly correlated with failures.”
A model for spectra-based software diagnosis (2011, methodology): “Some bugs always result in failure when they are executed (called deterministic bugs in [Liblit et al. 2005]) whereas some return the correct result for nearly all tests in which they are executed.”
TBar: revisiting template-based automated program repair (2019, methodology): “…defect locations in a program, several automated FL techniques have been proposed [67]: slice-based [41, 65], spectrum-based [3, 52], statistics-based [28, 29], etc. Spectrum-based FL (SBFL) techniques are widely adopted in APR pipelines since they identify bug position at the statement level.”
PROCHLO: Strong Privacy for Analytics in the Crowd (2017, background): “, for understanding software behavior or bugs, software performance or resource use, or for other software improvements [13, 39, 46, 59, 61, 77].”
What Makes a Good Bug Report? (2008, neutral): “introduced statistical debugging [21].”
Carat: collaborative energy diagnosis for mobile devices (2013, methodology): “Such methods have been used to identify code paths correlated with failure [16, 17], concurrency bugs [14], shared influence (surprising behavior that is correlated in time) [26, 28], invariant violation [13], and configuration errors [41].”
HDD: hierarchical delta debugging (2006, background): “They also propose to use statistical learning techniques to infer the failures from many sampled runs [9,10].”
Statistical Debugging: A Hypothesis Testing-Based Approach (2006, methodology): “However, by exploring detailed statistics about predicate evaluation, our method can detect more and subtler faults than the stateof-the-art statistical debugging approach proposed by Liblit et al. [12].”
SOBER: statistical model-based bug localization (2005, methodology): “According to [15], neither P nor P ′ is ranked top since they are either negative or too small.”
Empirical evaluation of the tarantula automatic fault-localization technique (2005, methodology): “In later work [8], they utilize equations similar to those used by the Tarantula tool to prune and rank the predicates in the program for inspection by the programmer.”