• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Annual Pacific northwest software quality conference >Al for Software Testing
【24h】

Al for Software Testing

机译：Al用于软件测试

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Software testing is fundamentally an exercise in applying sample inputs to a system, and measuring the outputs to determine the correctness of the application's behavior. This testing input and output function is very similar to the basic operations of training and executing modern Artificial Intelligence (Al) and Machine Learning (ML) systems. The similarity implies that the field of Software Testing is ripe to benefit from recent advances in Al and ML. Software Testing and approaches to measure Software Quality have been relatively stagnant over the past decade or so, with only slight technical improvements in test frameworks and methodologies; meanwhile product development, DevOps, and deployment have benefited greatly from increasing availability of compute, network, and storage. Software testing has remained an often linear activity with respect to test coverage-adding tests one at a time or in small batches while products add complexity at a much higher rate, meaning almost no product is tested efficiently. The complexity of products has been growing exponentially, where testing has remained more or less a linear activity with respect to test coverage. Applying the advancements in Al and ML will help testing 'catch up' with development advances, increase the efficiency, speed and efficacy of software testing, and free testers from many mundane testing tasks. Al and ML, when used to abstract the test input, test execution, and test evaluation problems enable testing at scale, standardization of quality metrics, benchmarking, and a global set of reusable test cases.
机译：从根本上说，软件测试是将示例输入应用于系统并测量输出以确定应用程序行为的正确性的一种练习。该测试输入和输出功能与培训和执行现代人工智能（Al）和机器学习（ML）系统的基本操作非常相似。相似之处表明，可以从Al和ML的最新进展中受益于软件测试领域。在过去的十年左右的时间里，软件测试和衡量软件质量的方法一直处于停滞状态，测试框架和方法仅进行了少许技术改进;同时，产品开发，DevOps和部署从计算，网络和存储可用性的提高中受益匪浅。相对于一次或小批量增加测试覆盖率，软件测试一直是一项线性活动，而产品以更高的速度增加了复杂性，这意味着几乎没有产品能得到有效测试。产品的复杂性呈指数增长，其中测试相对于测试覆盖率或多或少保持线性活动。应用Al和ML的进步将有助于测试“赶上”开发的进展，提高软件测试的效率，速度和功效，并使测试人员摆脱许多平凡的测试任务。 A1和ML在用于抽象测试输入，测试执行和测试评估问题时，可以进行大规模测试，质量指标的标准化，基准测试以及一组可重复使用的测试用例。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号