一句话：提出保语义归约，将多类 DNN 安全性质统一转化为可用对抗攻击求解的等价正确性问题。
深度神经网络正被越来越多地部署于自动驾驶、医疗设备等安全关键领域，此时仅靠高测试精度不足以提供行为保证，需要更强的正确性论证手段。本文主张并证明：现有对抗攻击技术可作为一种可扩展的性质证伪算法库。为使其广泛适用，作者提出一种保语义的归约方法，将多种安全性质类型（其涵盖先前工作）转化为一组等价的正确性问题，从而可直接交由对抗攻击进行证伪。作者在一系列 DNN 正确性问题上评估该归约对证伪的使能作用，结果表明该方法具有较低代价与良好可扩展性。
Deep Neural Networks (DNN) are increasingly being deployed in safety-critical domains, from autonomous vehicles to medical devices, where the consequences of errors demand techniques that can provide stronger guarantees about behavior than just high test accuracy. This paper explores broadening the application of existing adversarial attack techniques for the falsification of DNN safety properties. We contend and later show that such attacks provide a powerful repertoire of scalable algorithms for property falsification. To enable the broad application of falsification, we introduce a semantics-preserving reduction of multiple safety property types, which subsume prior work, into a set of equivalid correctness problems amenable to adversarial attacks. We evaluate our reduction approach as an enabler of falsification on a range of DNN correctness problems and show its cost-effectiveness and scalability.