Falsification of Learning-Based Controllers through Multi-Fidelity Bayesian Optimization
This work addresses the challenge of efficient safety validation for learning-based systems, which is crucial for real-world deployment, though it is incremental as it builds on existing falsification and Bayesian optimization techniques.
The paper tackles the problem of computationally expensive safety testing for learning-based controllers by proposing a multi-fidelity Bayesian optimization framework that uses simulators with varying accuracy and cost. The result is falsification performance comparable to single-fidelity methods but at much lower computational cost, as demonstrated in simulation experiments.
Simulation-based falsification is a practical testing method to increase confidence that the system will meet safety requirements. Because full-fidelity simulations can be computationally demanding, we investigate the use of simulators with different levels of fidelity. As a first step, we express the overall safety specification in terms of environmental parameters and structure this safety specification as an optimization problem. We propose a multi-fidelity falsification framework using Bayesian optimization, which is able to determine at which level of fidelity we should conduct a safety evaluation in addition to finding possible instances from the environment that cause the system to fail. This method allows us to automatically switch between inexpensive, inaccurate information from a low-fidelity simulator and expensive, accurate information from a high-fidelity simulator in a cost-effective way. Our experiments on various environments in simulation demonstrate that multi-fidelity Bayesian optimization has falsification performance comparable to single-fidelity Bayesian optimization but with much lower cost.