Continuous Safety Verification of Neural Networks

TL;DR

This paper addresses the challenge of maintaining neural network safety verification over time as DNNs are iteratively improved or adapted in autonomous driving, proposing methods for efficient transfer of verification results.

Contribution

It introduces approaches to transfer safety verification results to modified DNNs using abstractions and Lipschitz constants, reducing the need for full re-verification.

Findings

Effective transfer conditions for safety verification in DNNs.

Validation on a scaled vehicle platform with DNN controller.

Reduction in verification effort for iterative DNN improvements.

Abstract

Deploying deep neural networks (DNNs) as core functions in autonomous driving creates unique verification and validation challenges. In particular, the continuous engineering paradigm of gradually perfecting a DNN-based perception can make the previously established result of safety verification no longer valid. This can occur either due to the newly encountered examples (i.e., input domain enlargement) inside the Operational Design Domain or due to the subsequent parameter fine-tuning activities of a DNN. This paper considers approaches to transfer results established in the previous DNN safety verification problem to the modified problem setting. By considering the reuse of state abstractions, network abstractions, and Lipschitz constants, we develop several sufficient conditions that only require formally analyzing a small part of the DNN in the new problem. The overall concept is evaluated in a $1/10$-scaled vehicle that equips a DNN controller to determine the visual waypoint from the perceived image.