Can Offline Testing of Deep Neural Networks Replace Their Online Testing?

Abstract

We distinguish two general modes of testing for Deep Neural Networks (DNNs): Offline testing where DNNs are tested as individual units based on test datasets obtained independently from the DNNs under test, and online testing where DNNs are embedded into a specific application environment and tested in a closed-loop mode in interaction with the application environment. Typically, DNNs are subjected to both types of testing during their development life cycle where offline testing is applied immediately after DNN training and online testing follows after offline testing and once a DNN is deployed within a specific application environment. In this paper, we study the relationship between offline and online testing. Our goal is to determine how offline testing and online testing differ or complement one another and if we can use offline testing results to run fewer tests during online testing to reduce the testing cost. Though these questions are generally relevant to all autonomous systems, we study them in the context of automated driving systems where, as study subjects, we use DNNs automating end-to-end controls of steering functions of self-driving vehicles. Our results show that offline testing is more optimistic than online testing as many safety violations identified by online testing could not be identified by offline testing, while large prediction This work has received funding from the European Research Council under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 694277), Luxembourg’s National Research Fund (FNR) under grant BRIDGES2020/IS/14711346/FUNTASY, IEE S.A. Luxembourg, and the Canada Research Chair programme. Donghwan Shin was partially supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2019R1A6A3A03033444). Fitash Ul Haq · Donghwan Shin · Lionel Briand SnT, University of Luxembourg E-mail: [email protected], [email protected], [email protected] Shiva Nejati · Lionel Briand University of Ottawa E-mail: [email protected], [email protected] ar X iv :2 10 1. 11 11 8v 1 [ cs .S E ] 2 6 Ja n 20 21 2 Fitash Ul Haq et al. errors generated by offline testing always led to severe safety violations detectable by online testing. Further, we cannot use offline testing results to run fewer tests during online testing in practice since we are not able to identify specific situations where offline testing could be as accurate as online testing in identifying safety requirement violations.