E-grāmata: Automated Technology for Verification and Analysis: 20th International Symposium, ATVA 2022, Virtual Event, October 25-28, 2022, Proceedings

Invited Paper.- Learning Monitorable Operational Design Domains for Assured Autonomy.- Reinforcement Learning.- Dynamic Shielding for Reinforcement Learning in Black-Box Environments.- An Impossibility Result in Automata-Theoretic Reinforcement Learning.- Reusable Contracts for Safe Integration of Reinforcement Learning in Hybrid Systems.- Program Analysis and Verification.- SISL: Concolic Testing of Structured Binary Input Formats via Partial Specification.- Fence Synthesis under the C11 Memory Model.- Checking Scheduling-induced Violations of Control Safety Properties.- Symbolic Runtime Verification for Monitoring under Uncertainties and Assumptions.- SMT and Verification.- Handling Polynomial and Transcendental Functions in SMT via Unconstrained Optimisation and Topological Degree Test.- Verification of SMT Systems with Quantifiers.- Projected Model Counting: Beyond Independent Support.- Automata and Applications.- Minimization of Automata for Liveness Languages.- Temporal Causality in Reactive Systems.- PDAAAL: A Library for Reachability Analysis of Weighted Pushdown Systems.- Active Learning.- Learning Deterministic One-Clock Timed Automata via Mutation Testing.- Active Learning of One-Clock Timed Automata using Constraint Solving.- Learning and Characterizing Fully-Ordered Lattice Automata.- Probabilistic and Stochastic Systems.- Optimistic and Topological Value Iteration for Simple Stochastic Games.- Alternating Good-for-MDPs Automata.- PET - A Partial Exploration Tool for Probabilistic Verification.- STOMPC: Stochastic Model-Predictive Control with Uppaal Stratego.- Synthesis and Repair.- Synthesis of Parametric Hybrid Automata from Time Series.- Optimal Repair For Omega-regular Properties.- Repairing Real-Time Requirements.- Verification of Neural Networks.- An Abstraction-Refinement Approach to Verifying Convolutional Neural Networks.- Prioritizing Corners in OoD Detectors via Symbolic String Manipulation.- POLAR: A Polynomial Arithmetic Framework for Verifying Neural-Network Controlled Systems.