Importance-driven deep learning system testing

Gerasimou, Simos; Enişer, Hasan Ferit; Şen, Alper; Cakan, Alper

doi:10.1145/3377811.3380391

Cited by 51 publications

(70 citation statements)

References 38 publications

(15 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, we selected the coverage metrics that we could apply and replicate on our models and datasets. The first two factors yielded four coverage metrics: Likelihood-based Surprise Coverage, Distance-based Surprise Coverage, Importance-Driven Coverage (IDC) [11] and Sign-Sign Coverage (SSC) [14]. However, we could not apply IDC and SSC on our datasets and models.…”

Section: Rq3 How Does Coverage Relate To Fault Detection?mentioning

confidence: 96%

“…In fact, state-of-the-art coverage metrics highly rely on artificial inputs generated based on adversarial methods [12], [13], [14], [10], [11]. However, their positive correlation with the presence of adversarial inputs does not necessarily mean that they are efficient to reveal the fault detection capability of natural test input sets.…”

Section: • Rq3 How Does Coverage Relate To Fault Detection?mentioning

confidence: 99%

“…Nevertheless, the obtained results indicate that GD can be used as guidance to effectively test DNNs by devising input sets with maximum diversity so as to increase their fault revealing capabilities. Let us recall that GD has also the practical advantage of being black-box, as opposed stateof-the-art DNN coverage metrics [12], [13], [10], [11] that require access to the internals of DNN models or their training sets.…”

Section: Rq2 How Does Diversity Relate To Fault Detection?mentioning

confidence: 99%

“…While full coverage does not ensure functional correctness, high coverage increases stakeholders' confidence in the testing results since it triggers more code execution paths. Inspired by code coverage, several coverage criteria have been introduced to measure the adequacy of test data in the context of DNNs [10], [11], [12], [13], [14]. Neuron coverage measures the extent to which neurons in a DNN are activated based on certain input data.…”

Section: Introductionmentioning

confidence: 99%

“…However these assumptions break down in DNN testing as (1) as opposed to code coverage, neuron coverage does not necessarily fully exercise the implicit logic embedded in DNNs, (2) the homogeneity assumption is broken with adversarial inputs, and (3) increasing the diversity of inputs does not necessarily increase DNNs coverage [15]. Further, most coverage studies rely on adversarial inputs to validate their proposed criteria [12], [13], [14], [10], [11]. These inputs are, however, mostly unrealistic and are used to study the robustness of the DNN model instead of its accuracy.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Black-Box Testing of Deep Neural Networks Through Test Case Diversity

Aghababaeyan¹,

Abdellatif²,

Briand³

et al. 2021

Preprint

View full text Add to dashboard Cite

Deep Neural Networks (DNNs) have been extensively used in many areas including image processing, medical diagnostics, and autonomous driving. However, DNNs can exhibit erroneous behaviours that may lead to critical errors, especially when used in safety-critical systems. Inspired by testing techniques for traditional software systems, researchers have proposed neuron coverage criteria, as an analogy to source code coverage, to guide the testing of DNN models. Despite very active research on DNN coverage, several recent studies have questioned the usefulness of such criteria in guiding DNN testing. Further, from a practical standpoint, these criteria are white-box as they require access to the internals or training data of DNN models, which is in many contexts not feasible or convenient. In this paper, we investigate black-box input diversity metrics as an alternative to white-box coverage criteria. To this end, we first select and adapt three diversity metrics and study, in a controlled manner, their capacity to measure actual diversity in input sets. We then analyse their statistical association with fault detection using two datasets and three DNN models. We further compare diversity with state-of-the-art white-box coverage criteria. Our experiments show that relying on the diversity of image features embedded in test input sets is a more reliable indicator than coverage criteria to effectively guide the testing of DNNs. Indeed, we found that one of our selected black-box diversity metrics far outperforms existing coverage criteria in terms of fault-revealing capability and computational time. Results also confirm the suspicions that state-of-the-art coverage metrics are not adequate to guide the construction of test input sets to detect as many faults as possible with natural inputs.

show abstract

Section: Rq3 How Does Coverage Relate To Fault Detection?mentioning

confidence: 96%

Section: • Rq3 How Does Coverage Relate To Fault Detection?mentioning

confidence: 99%

Section: Rq2 How Does Diversity Relate To Fault Detection?mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Black-Box Testing of Deep Neural Networks Through Test Case Diversity

Aghababaeyan¹,

Abdellatif²,

Briand³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

Revisiting deep neural network test coverage from the test effectiveness perspective

Yan

Chen

Cao

et al. 2023

J Software Evolu Process

View full text Add to dashboard Cite

Many test coverage metrics have been proposed to measure the deep neural network (DNN) testing effectiveness, including structural coverage and nonstructural coverage. These test coverage metrics are proposed based on the fundamental assumption: They are correlated with test effectiveness. However, the fundamental assumption is still not validated sufficiently and reasonably, which brings question on the usefulness of DNN test coverage. This paper conducted a revisiting study on the existing DNN test coverage from the test effectiveness perspective, to effectively validate the fundamental assumption. Here, we carefully considered the diversity of subjects, three test effectiveness criteria, and both typical and state‐of‐the‐art test coverage metrics. Different from all the existing studies that deliver negative conclusions on the usefulness of existing DNN test coverage, we identified some positive conclusions on their usefulness from the test effectiveness perspective. In particular, we found the complementary relationship between structural and nonstructural coverage and identified the practical usage scenarios and promising research directions for these existing test coverage metrics.

show abstract

Validity Matters: Uncertainty‐Guided Testing of Deep Neural Networks

Jiang,

Li,

Wang

et al. 2024

Software Testing Verif & Rel

View full text Add to dashboard Cite

Despite numerous applications of deep learning technologies on critical tasks in various domains, advanced deep neural networks (DNNs) face persistent safety and security challenges, such as the overconfidence in predicting out‐of‐distribution samples and susceptibility to adversarial examples. Thorough testing by exploring the input space serves as a key strategy to ensure their robustness and trustworthiness of these networks. However, existing testing methods focus on disclosing more erroneous model behaviours, overlooking the validity of the generated test inputs. To mitigate this issue, we investigate devising valid test input generation method for DNNs from a predictive uncertainty perspective. Through a large‐scale empirical study across 11 predictive uncertainty metrics for DNNs, we explore the correlation between validity and uncertainty of test inputs. Our findings reveal that the predictive entropy‐based and ensemble‐based uncertainty metrics effectively characterize the input validity demonstration. Building on these insights, we introduce UCTest, an uncertainty‐guided deep learning testing approach, to efficiently generate valid and authentic test inputs. We formulate a joint optimization objective: to uncover the model's misbehaviours by maximizing the loss function and concurrently generate valid test input by minimizing uncertainty. Extensive experiments demonstrate that our approach outperforms the current testing methods in generating valid test inputs. Furthermore, incorporating natural variation through data augmentation techniques into UCTest effectively boosts the diversity of generated test inputs.

show abstract

Importance-driven deep learning system testing

Cited by 51 publications

References 38 publications

Black-Box Testing of Deep Neural Networks Through Test Case Diversity

Black-Box Testing of Deep Neural Networks Through Test Case Diversity

Revisiting deep neural network test coverage from the test effectiveness perspective

Validity Matters: Uncertainty‐Guided Testing of Deep Neural Networks

Contact Info

Product

Resources

About