Properties of Machine Learning Applications for Use in Metamorphic Testing

Christian Murphy; Gail E. Kaiser; Lifeng Hu

On Saturday, May 7, Academic Commons will undergo planned system upgrades from 11:30 am until 9:30 pm North American Eastern Daylight Time. During this maintenance some downloads may be unavailable.
Properties of Machine Learning Applications for Use in Metamorphic Testing
Murphy, Christian
Kaiser, Gail E.
Hu, Lifeng
Technical reports
Computer Science
Permanent URL:
Columbia University Computer Science Technical Reports
Part Number:
Department of Computer Science, Columbia University
Publisher Location:
New York
It is challenging to test machine learning (ML) applications, which are intended to learn properties of data sets where the correct answers are not already known. In the absence of a test oracle, one approach to testing these applications is to use metamorphic testing, in which properties of the application are exploited to define transformation functions on the input, such that the new output will be unchanged or can easily be predicted based on the original output; if the output is not as expected, then a defect must exist in the application. Here, we seek to enumerate and classify the metamorphic properties of some machine learning algorithms, and demonstrate how these can be applied to reveal defects in the applications of interest. In addition to the results of our testing, we present a set of properties that can be used to define these metamorphic relationships so that metamorphic testing can be used as a general approach to testing machine learning applications.
Computer science
Item views:
text | xml
Suggested Citation:
Christian Murphy, Gail E. Kaiser, Lifeng Hu, 2008, Properties of Machine Learning Applications for Use in Metamorphic Testing, Columbia University Academic Commons, http://hdl.handle.net/10022/AC:P:29550.

In Partnership with the Center for Digital Research and Scholarship at Columbia University Libraries/Information Services | Terms of Use | Copyright