摘要
针对智能计算系统的自动化测试,设计和实现了IntelTest,用于智能计算系统的自动化白盒测试系统。介绍了IntelTest的总体框架、组成模块及神经元覆盖率的定义,并将测试问题建模为联合优化问题,采用梯度上升算法来求解该联合优化问题,最后采用ImageNet数据集和两个深度神经网络(DNN)模型来评估IntelTest的性能。
For the automatic test of intelligent computing system,IntelTest is designed and implemented,which is used in the automatic white box test system of intelligent computing system.This paper introduces the general framework of IntelTest,its constituent modules and the definition of neuron coverage.The test problem is modeled as a joint optimization problem.The joint optimization problem is solved by a gradient rise algorithm.Finally,the ImageNet dataset and two deep neural networks(DNN)models are used to evaluate the performance of IntelTest.
作者
苏警
SU Jing(College of Software,Anhui Vocational College of Electronics&Information Technology,Bengbu,Anhui 233000,China)
出处
《大庆师范学院学报》
2019年第6期76-81,共6页
Journal of Daqing Normal University
基金
2018年安徽电子信息职业技术学院自然科学研究项目“基于selenium的web自动化测试框架的设计与研究”(ADZX1811)
