Abstract:While neural networks (NNs) have potential as autonomous controllers for Cyber-Physical Systems, verifying the safety of NN based control systems (NNCSs) poses significant challenges for the practical use of NNs, especially when safety is needed for unbounded time horizons. One reason is the intractability of analyzing NNs, ODEs and hybrid systems. To this end, we introduce VerSAILLE (Verifiably Safe AI via Logically Linked Envelopes): The first general approach that allows reusing control theory results for NNCS verification. By joining forces, we exploit the efficiency of NN verification tools while retaining the rigor of differential dynamic logic (dL). Based on provably safe control envelopes in dL, we derive specifications for the NN which is proven via NN verification. We show that a proof of the NN adhering to the specification is mirrored by a dL proof on the infinite-time safety of the NNCS. The NN verification properties resulting from hybrid systems typically contain nonlinear arithmetic and arbitrary logical structures while efficient NN verification merely supports linear constraints. To overcome this divide, we present Mosaic: An efficient, sound and complete verification approach for polynomial real arithmetic properties on piece-wise linear NNs. Mosaic partitions complex verification queries into simple queries and lifts off-the-shelf linear constraint tools to the nonlinear setting in a completeness-preserving manner by combining approximation with exact reasoning for counterexample regions. Our evaluation demonstrates the versatility of VerSAILLE and Mosaic: We prove infinite-time safety on the classical Vertical Airborne Collision Avoidance NNCS verification benchmark for two scenarios while (exhaustively) enumerating counterexample regions in unsafe scenarios. We also show that our approach significantly outperforms State-of-the-Art tools in closed-loop NNV.

Parallel Composition and Modular Verification of Computer Controlled Systems in Differential Dynamic Logic

Complete Dynamic Logic of Communicating Hybrid Programs

Collaborative Verification-Driven Engineering of Hybrid Systems

Co-Verification Approach to Control Software Program for CPS

Formal Verification of Hybrid Systems and Its Application on Chemical Process Control

Embedding CCSL into Dynamic Logic - A Logical Approach for the Verification of CCSL Specifications.

Towards Compositional Verification for Modular Robotic Systems

Contract Composition for Dynamical Control Systems: Definition and Verification using Linear Programming

Temporal Logic of Composable Distributed Components

A Clock-Based Dynamic Logic for the Verification of CCSL Specifications in Synchronous Systems

Unified Graphical Co-modelling, Analysis and Verification of Cyber-physical Systems by Combining AADL and Simulink/Stateflow

Logic Verification Of Collision Avoidance System In Train Control Systems

Modular System Verification by Inference, Testing and Reachability Analysis

Provably Safe Neural Network Controllers via Differential Dynamic Logic

Compositional Model Checking of Real Time Systems

TLC: temporal logic of distributed components

Formal Modelling and Verification of Hybrid Systems by Hybrid Relational Approach

Study on modeling and verification methods of digital reactor control system based on differential dynamic logic

A Unified Modeling and Verifying Framework for Cyber Physical Systems

Compositional Model-Checking for Real-Time Systems