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 without involving 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 offline testing results can be used to help reduce the cost of online testing? 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 less effective than online testing as many safety violations identified by online testing could not be identified by offline testing, while large prediction errors generated by offline testing always

Ergo, SMIRK is Safe: A Safety Case for a Machine Learning Component in a Pedestrian Automatic Emergency Brake System

“…If any of the objects within the red bounding boxes were on a collision course with the ego car, commencing PAEB would indeed be the right action for SMIRK and thus not violate SYS-SAF-REQ1. This observation corroborates the position by (Haq et al, 2021), i.e., system level testing that goes beyond model testing on single frames is critically needed. All results from running ML model testing, i.e., ML Verification Results [Z], are documented in the Protocols folder.…”

“…Second, SMIRK could be used as a realistic test benchmark for automotive ML testing. The testing community has largely worked on offline testing of single frames, but we know that this is insufficient (Haq et al, 2021). Third, we recommend the community to port SMIRK to other simulators beyond ESI Pro-SiVIC.…”

“…In contrast to most previous work that stop at pedestrian detection, we present an ADAS that subsequently commences emergency braking in a simulated environment. This addition responds to calls for researchers to go from offline to online testing (Haq et al, 2021), as many safety violations identified by online testing could not be identified by offline testing. We hope that SMIRK can contribute to a shift in the testing community away from standalone image data sets.…”

SMIRK: A machine learning-based pedestrian automatic emergency braking system with a complete safety case

“…If False, the Rule Engine performs a sanity check based on laws of physics. (9) If UM remains confident that collision with a pedestrian is imminent, the signal to perform PAEB propagates to ego car.…”

“…SMIRK allows researchers to explore data testing, ML model testing, integration testing, and system testing since data sets, ML model architectures, and the source code are publicly available. For example, offline model testing can be compared to online system testing, as Haq et al recently proved important [9]. Concrete test techniques that could be evaluated using SMIRK include search-based software testing, metamorphic testing, fuzz testing, neural network test adequacy assessments, and testing for explainable AI.…”