Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 1, 2026Journal of Software Evolution and Process

Adaptive Multi‐Metric Test Case Selection for Deep Neural Networks Based on Genetic Algorithm

View Full Paper
Ask AI
Bookmark
Share

Authors

WWWeiwei WangQCQingshuai ChenZZZh. Zhao

Discussion

Loading...

Member takes

Overview

Randomized trial demonstrates improved fault detection in deep neural networks, indicating enhanced testing efficiency.

Key Points

  • The aim is to optimize test case selection for deep neural networks to improve fault detection.
  • Implemented a framework utilizing adaptive multi-metric optimization based on genetic algorithms.
  • Constructed multiple slightly mutated models to assess test cases' behavioral discrepancies.
  • Established prediction uncertainty and divergence as effectiveness metrics for test case sensitivity and diversity.
  • Achieved an average improvement of 3.32% in fault detection rate for the Top-5% candidate subset.
  • Achieved an average improvement of 9.54% in fault detection rate for the Top-10% candidate subset.

Cite This Study

Wang et al. (2026) studied this question.

synapsesocial.com/papers/6a57a3d0d7502ad8fc0fbfe6https://doi.org/10.1002/smr.70150
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Neuron Sensitivity Guided Test Case Selection2024 · 1 citations
  2. 2A Mutation‐Score–Guided Test Case Generation Framework for Deep Neural Networks2026
  3. 3LeF‐MTP: Prioritizing GNN Test Cases by Fusing Model Uncertainty and Feature‐Space Confusability2026 · 12 citations
  4. 4Evaluating the effectiveness of neuron coverage metrics: a metamorphic-testing approach2024
  5. 5An empirical study of defect clustering in deep neural networks and its implications for testing2026