Learning Neural Hybrid Surrogates for Gradient-Based Falsification
For engineers verifying safety of hybrid systems, this work extends surrogate-based falsification to handle mode-dependent dynamics and discrete transitions, addressing a known bottleneck in the field.
The paper proposes a neural hybrid surrogate model for falsification of hybrid dynamical systems, enabling gradient-based optimization to find safety violations. The method achieves competitive or improved sample efficiency on most benchmarks compared to existing tools.
Falsification of hybrid dynamical systems remains challenging due to mode-dependent dynamics and discrete transitions. In this work, we propose a surrogate-based falsification approach that enables hybrid systems by learning a differentiable hybrid automaton model from data. This extends previous surrogate-based falsification methods, which were limited to purely continuous dynamics. Specifically, we employ neural hybrid automata to learn both a latent mode encoder and the corresponding mode-conditioned vector fields. Once the surrogate has paired each mode with an associated vector field, the transition guards are inferred using existing trajectory data. The learned surrogate is subsequently subjected to a gradient-based optimal control formulation, which minimizes a smooth approximation of the safety specification to find safety violations. In the last step, an experiment with the optimal control solution is carried out on the original system to ensure soundness. The proposed method consistently uncovers counterexamples on a majority of evaluated benchmark specifications; on these cases, it achieves competitive or improved sample efficiency than other tools while using a reduced simulation budget.