Parallel and Multi-objective Falsification with Scenic and VerifAI

Falsification has emerged as an important tool for simulation-based verification of autonomous systems. In this paper, we present extensions to the Scenic scenario specification language and VerifAI toolkit that improve scalability sampling-based falsification methods by using parallelism extend multi-objective specifications. We first a parallelized framework is interfaced with both simulation sampling capabilities VerifAI, reducing execution time bottleneck inherently in testing. then extension VerifAI's algorithms support optimization during sampling, concept rulebooks specify preference ordering over multiple metrics can be used guide counterexample search process. Lastly, evaluate benefits these comprehensive set benchmarks written language.