Test Input

Test input prioritization aims to optimize the order in which test cases are executed, maximizing the early detection of software bugs with minimal resources. Current research focuses on developing and comparing machine learning-based prioritization techniques, often leveraging uncertainty quantification methods or feature selection to improve the effectiveness of prioritization, particularly for deep neural networks. These advancements are significant for improving software testing efficiency and reliability across various applications, especially in resource-constrained continuous integration environments. The field is actively exploring the robustness and generalizability of different approaches, seeking methods that perform well across diverse datasets, model architectures, and testing scenarios.

Papers

September 13, 2024

FAST: Boosting Uncertainty-based Test Prioritization Methods for Neural Networks via Feature Selection
Jialuo Chen, Jingyi Wang, Xiyue Zhang, Youcheng Sun, Marta Kwiatkowska, Jiming Chen, Peng Cheng
Neural Network Deep Neural Network Feature Selection Test Case High Confidence Prediction Various Error Test Input

July 18, 2023

CertPri: Certifiable Prioritization for Deep Neural Networks via Movement Cost in Feature Space
Haibin Zheng, Jinyin Chen, Haibo Jin
Deep Neural Network Feature Space Priority Based Action Cost Test Input

May 2, 2022

Simple Techniques Work Surprisingly Well for Neural Network Test Prioritization and Active Learning (Replicability Study)
Michael Weiss, Paolo Tonella
Deep Neural Network Active Learning Research Reproducibility Simple Method Test Datasets Test Input

April 22, 2022

Comparative Study of Machine Learning Test Case Prioritization for Continuous Integration Testing
Dusica Marijan
Comparative Study Test Case Fault Prediction Continuous Integration Test Input

Test Input

Papers

FAST: Boosting Uncertainty-based Test Prioritization Methods for Neural Networks via Feature Selection

CertPri: Certifiable Prioritization for Deep Neural Networks via Movement Cost in Feature Space

Simple Techniques Work Surprisingly Well for Neural Network Test Prioritization and Active Learning (Replicability Study)

Comparative Study of Machine Learning Test Case Prioritization for Continuous Integration Testing