一句话：用可证明保证的验证与测试技术，为深度神经网络在有输入扰动时的安全性与鲁棒性提供输出可达集的上下界逼近。
随着计算系统日益复杂并大量依赖深度学习组件，自动驾驶、医学诊断等安全关键应用亟需涵盖机器学习部件的严格软件开发方法学。然而深度神经网络对对抗扰动不稳定，其决策缺乏可靠性保证。该报告介绍面向深度神经网络决策安全性与鲁棒性的自动化验证与测试技术进展：这些技术利用网络的 Lipschitz 连续性，针对给定输入集合，以 anytime 方式逼近网络可达输出集的上下界，并给出可证明的保证。作者提出基于特征引导搜索、博弈、全局优化与贝叶斯方法的新算法，并在当前最先进的网络上进行实验评估。报告最后概述该领域仍面临的挑战。
Computing systems are becoming ever more complex, with decisions increasingly often based on deep learning components. A wide variety of applications are being developed, many of them safety-critical, such as self-driving cars and medical diagnosis. Since deep learning is unstable with respect to adversarial perturbations, there is a need for rigorous software development methodologies that encompass machine learning components. This lecture will describe progress with developing automated verification and testing techniques for deep neural networks to ensure safety and robustness of their decisions with respect to input perturbations. The techniques exploit Lipschitz continuity of the networks and aim to approximate, for a given set of inputs, the reachable set of network outputs in terms of lower and upper bounds, in anytime manner, with provable guarantees. We develop novel algorithms based on feature-guided search, games, global optimisation and Bayesian methods, and evaluate them on state-of-the-art networks. The lecture will conclude with an overview of the challenges in this field.
DOI 原文 ·
@inproceedings{paperbot5168,
title = {Safety and robustness for deep learning with provable guarantees (keynote)},
author = {Marta Kwiatkowska},
booktitle = {FSE},
year = {2019},
doi = {10.1145/3338906.3342812}
}