| Title | An AI Software Test Method Based on Scene Deductive Approach |
| Publication Type | Conference Paper |
| Year of Publication | 2018 |
| Authors | Zhao, Xinghan, Gao, Xiangfei |
| Conference Name | 2018 IEEE International Conference on Software Quality, Reliability and Security Companion (QRS-C) |
| Date Published | July 2018 |
| Publisher | IEEE |
| ISBN Number | 978-1-5386-7839-8 |
| Keywords | AI software testing method, algorithm complexity, artificial intelligence, artificial intelligence software, automobiles, i-o systems security, i/o systems security, intelligent tracking car, machine learning algorithms, Mathematical model, position control, program testing, pubcrawl, road vehicles, Scalability, scene deductive approach, Software, Software algorithms, Software Testing |
| Abstract | Artificial intelligence (AI) software has high algorithm complexity, and the scale and dimension of the input and output parameters are high, and the test oracle isn't explicit. These features make a lot of difficulties for the design of test cases. This paper proposes an AI software testing method based on scene deductive approach. It models the input, output parameters and the environment, uses the random algorithm to generate the inputs of the test cases, then use the algorithm of deductive approach to make the software testing automatically, and use the test assertions to verify the results of the test. After description of the theory, this paper uses intelligent tracking car as an example to illustrate the application of this method and the problems needing attention. In the end, the paper describes the shortcoming of this method and the future research directions. |
| URL | https://ieeexplore.ieee.org/document/8431945 |
| DOI | 10.1109/QRS-C.2018.00017 |
| Citation Key | zhao_ai_2018 |
An AI Software Test Method Based on Scene Deductive Approach