Verifying the Safety of Autonomous Systems with Neural Network Controllers
Verifying the Safety of Autonomous Systems with Neural Network Controllers
复制标题
DOI：
10.1145/3419742
复制
发表时间：
2021-01-01
影响因子：
2
通讯作者：
Lee, Insup
中科院分区：
文献类型：
--
作者：
Ivanov, Radoslav;Carpenter, Taylor J.;Lee, Insup
本文解决了使用神经网络 (NN) 控制器验证自主系统安全性的问题。我们关注具有 sigmoid/tanh 激活的神经网络，并利用 sigmoid/tanh 是二次微分方程的解这一事实。这使得我们能够将神经网络转换为等效的混合系统，并将问题转化为混合系统验证问题，可以通过现有工具来解决。此外，我们通过使用具有最坏情况误差范围的泰勒级数逼近 sigmoid 来提高所提出方法的可扩展性。最后，我们提供了对四个基准的评估，包括与基于混合整数线性规划以及星集的替代方法的比较。
This article addresses the problem of verifying the safety of autonomous systems with neural network (NN) controllers. We focus on NNs with sigmoid/tanh activations and use the fact that the sigmoid/tanh is the solution to a quadratic differential equation. This allows us to convert the NN into an equivalent hybrid system and cast the problem as a hybrid system verification problem, which can be solved by existing tools. Furthermore, we improve the scalability of the proposed method by approximating the sigmoid with a Taylor series with worst-case error bounds. Finally, we provide an evaluation over four benchmarks, including comparisons with alternative approaches based on mixed integer linear programming as well as on star sets.