Reachable Polyhedral Marching (RPM): An Exact Analysis Tool for Deep-Learned Control Systems
share
We present a tool for computing exact forward and backward reachable sets of deep neural networks with rectified linear unit (ReLU) activation. We then develop algorithms using this tool to compute invariant sets and regions of attraction (ROAs) for control systems with neural networks in the feedback loop. Our algorithm is unique in that it builds the reachable sets by incrementally enumerating polyhedral regions in the input space, rather than iterating layer-by-layer through the network as in other methods. When performing safety verification, if an unsafe region is found, our algorithm can return this result without completing the full reachability computation, thus giving an anytime property that accelerates safety verification. Furthermore, we introduce a method to accelerate the computation of ROAs in the case that deep learned components are homeomorphisms, which we find is surprisingly common in practice. We demonstrate our tool in several test cases. We compute a ROA for a learned van der Pol oscillator model. We find a control invariant set for a learned torque-controlled pendulum model. We also verify specific safety properties for multiple deep networks related to the ACAS Xu aircraft collision advisory system. Finally, we apply our algorithm to find ROAs for an image-based aircraft runway taxi problem. Algorithm source code: https://github.com/StanfordMSL/Neural-Network-Reach .
READ FULL TEXT