Towards Quality Assurance of Software Product Lines with Adversarial Configurations

Temple, Paul; Acher, Mathieu; Perrouin, Gilles; Biggio, Battista; Jézéquel, Jean-Marc; Roli, Fabio

doi:10.1145/3336294.3336309

Cited by 6 publications

(13 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This new generator produces adversarial configurations up to 50% faster than our initial algorithm [85]. Overall our empirical assessment suggests that generating adversarial configurations is effective to investigate the quality of ML-based SPLs.…”

Section: Introductionmentioning

confidence: 87%

“…In some cases, configurations can lead to undesirable variants despite being logically valid within the variability model. For instance, when considering the MOTIV SPL [39,85], some video variants may contain too much noise or not enough contrast. Variants can still be generated but these videos are not exploitable for any object recognition task, since the non-functional property (here: the visual quality of a video) does meet expectations.…”

Section: Spl Frameworkmentioning

confidence: 99%

“…Algorithm 1 A dedicated algorithm [85] conducting the gradient-descent evasion attack inspired by [16] Input: x 0 , the initial configuration; t, the step size; nb_disp, the number of displacements; g, the discriminant function Output: x * , the final attack point…”

Section: A Dedicated Evasion Algorithmmentioning

confidence: 99%

“…(5) x m = x m−1 − t∇F (x m−1 ); end while (6) return x * = x m ; Algorithm 1 presents an adaptation of Biggio et al's evasion attack [16], initially presented at the SPLC 2019 conference [85]. First, we select an initial configuration to be moved (x 0 ); multiple strategies can be used to select x 0 , we will simply use a random strategy selecting one configuration labeled with the class from which the attack starts.…”

Section: A Dedicated Evasion Algorithmmentioning

confidence: 99%

“…This article is an extension of "Towards quality assurance of software product lines with adversarial configurations" published at SPLC 2019 (full paper, research track) [85]. In this extension, we describe the implementation of a new adversarial generator for SPLs with a new preprocessing technique on top of SecML, a Python library that implements state-of-the-art adversarial algorithms.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Empirical assessment of generating adversarial configurations for software product lines

et al. 2021

Self Cite

View full text Add to dashboard Cite

Software product line (SPL) engineering allows the derivation of products tailored to stakeholders' needs through the setting of a large number of configuration options.Unfortunately, options and their interactions create a huge configuration space which is either intractable or too costly to explore exhaustively. Instead of covering all products, machine learning (ML) approximates the set of acceptable products (e.g., successful builds, passing tests) out of a training set (a sample of configurations). However, ML techniques can make prediction errors yielding non-acceptable products wasting time, energy and other resources.We apply adversarial machine learning techniques to the world of SPLs and craft new configurations faking to be acceptable configurations but that are not and vice-versa. It allows to diagnose prediction errors and take appropriate actions. We develop two adversarial configuration generators on top of state-

show abstract

Section: Introductionmentioning

confidence: 87%

Section: Spl Frameworkmentioning

confidence: 99%

Section: A Dedicated Evasion Algorithmmentioning

confidence: 99%

Section: A Dedicated Evasion Algorithmmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Empirical assessment of generating adversarial configurations for software product lines

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

Empirical analysis of the tool support for software product lines

2022

View full text Add to dashboard Cite

For the last ten years, software product line (SPL) tool developers have been facing the implementation of different variability requirements and the support of SPL engineering activities demanded by emergent domains. Despite systematic literature reviews identifying the main characteristics of existing tools and the SPL activities they support, these reviews do not always help to understand if such tools provide what complex variability projects demand. This paper presents an empirical research in which we evaluate the degree of maturity of existing SPL tools focusing on their support of variability modeling characteristics and SPL engineering activities required by current application domains. We first identify the characteristics and activities that are essential for the development of SPLs by analyzing a selected sample of case studies chosen from application domains with high variability. Second, we conduct an exploratory study to analyze whether the existing tools support those characteristics and activities. We conclude that, with the current tool support, it is possible to develop a basic SPL approach. But we have also found out that these tools present several limitations when dealing with complex variability requirements demanded by emergent application domains, such as non-Boolean features or large configuration spaces. Additionally, we identify the necessity for an integrated approach with appropriate tool support to completely cover all the activities and phases of SPL engineering. To mitigate this problem, we propose different road map using the existing tools to partially or entirely support SPL engineering activities, from variability modeling to product derivation.

show abstract

Feature-oriented defect prediction

Strüder

Mukelabai

Strüber

et al. 2020

Proceedings of the 24th ACM Conference on Systems and Software Product Line: Volume a - Volume A

View full text Add to dashboard Cite

Software errors are a major nuisance in software development and can lead not only to reputation damages, but also to considerable financial losses for companies. Therefore, numerous techniques for predicting software defects, largely based on machine learning methods, have been developed over the past decades. These techniques usually rely on code and process metrics in order to predict defects at the granularity of typical software assets, such as subsystems, components, and files. In this paper, we present the first systematic investigation of feature-oriented defect prediction: the prediction of defects at the granularity of features-domainoriented entities abstractly representing (and often cross-cutting) typical software assets. Feature-oriented prediction can be beneficial, since: (i) particular features might be more error-prone than others, (ii) characteristics of features known as defective might be useful to predict other error-prone features, (iii) feature-specific code might be especially prone to faults arising from feature interactions. We present a dataset derived from 12 software projects and introduce two metric sets for feature-oriented defect prediction. We evaluated seven machine learning classifiers with three different attribute sets each, using our two new metric sets as well as an existing metric set from the literature. We observe precision and recall values of around 85% and better robustness when more diverse metrics sets with richer feature information are used. CCS CONCEPTS• Software and its engineering → Software product lines; Software defect analysis; • Computing methodologies → Supervised learning by classification.

show abstract

Towards Quality Assurance of Software Product Lines with Adversarial Configurations

Cited by 6 publications

References 48 publications

Empirical assessment of generating adversarial configurations for software product lines

Empirical assessment of generating adversarial configurations for software product lines

Empirical analysis of the tool support for software product lines

Feature-oriented defect prediction

Contact Info

Product

Resources

About