English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep learning has revolutionized various real-world applications, but thequality of Deep Neural Networks (DNNs) remains a concern. DNNs are complex andhave millions of parameters, making it difficult to determine theircontributions to fulfilling a task. Moreover, the behavior of a DNN is highlyinfluenced by the data used during training, making it challenging to collectenough data to exercise all potential DNN behavior under all possiblescenarios. This paper proposes NP SBFL method to locate faulty neural pathways(NP) using spectrum-based fault localization (SBFL). Our method identifiescritical neurons using the layer-wise relevance propagation (LRP) technique anddetermines which critical neurons are faulty. Moreover, we propose amulti-stage gradient ascent (MGA), an extension of gradient ascent (GA), toeffectively activate a sequence of neurons one at a time while maintaining theactivation of previous neurons, so we are able to test the reported faultypathways. We evaluated the effectiveness of our method, i.e. NP-SBFL-MGA, ontwo commonly used datasets, MNIST and CIFAR-10, two baselines DeepFault andNP-SBFL-GA, and three suspicious neuron measures, Tarantula, Ochiai, andBarinel. The empirical results showed that NP-SBFL-MGA is statistically moreeffective than the baselines at identifying suspicious paths and synthesizingadversarial inputs. Particularly, Tarantula on NP-SBFL-MGA had the highestfault detection rate at 96.75%, surpassing DeepFault on Ochiai (89.90%) andNP-SBFL-GA on Ochiai (60.61%). Our approach also yielded comparable results tothe baselines in synthesizing naturalness inputs, and we found a positivecorrelation between the coverage of critical paths and the number of failedtests in DNN fault localization.

Path Analysis for Effective Fault Localization in Deep Neural Networks