Verification for Machine Learning, Autonomy, and Neural Networks Survey

TL;DR

This survey reviews recent verification techniques for AI-driven autonomous systems, especially neural networks, highlighting advances in formal methods for ensuring safety and correctness in learning-enabled components.

Contribution

It provides a comprehensive overview of the latest formal verification methods applied to neural networks and autonomous cyber-physical systems.

Findings

Recent formal methods enable behavior characterization of neural networks

Verification techniques are advancing for safety-critical autonomous systems

Survey covers a wide range of approaches and tools in the field

Abstract

This survey presents an overview of verification techniques for autonomous systems, with a focus on safety-critical autonomous cyber-physical systems (CPS) and subcomponents thereof. Autonomy in CPS is enabling by recent advances in artificial intelligence (AI) and machine learning (ML) through approaches such as deep neural networks (DNNs), embedded in so-called learning enabled components (LECs) that accomplish tasks from classification to control. Recently, the formal methods and formal verification community has developed methods to characterize behaviors in these LECs with eventual goals of formally verifying specifications for LECs, and this article presents a survey of many of these recent approaches.