arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1807.03571cs.LGcs.AIstat.ML
收藏

一种基于博弈的可证明保证的深度神经网络近似验证

A Game-Based Approximate Verification of Deep Neural Networks with Provable Guarantees

University of Oxford(牛津大学);University of Georgia(佐治亚大学);University of Liverpool(利物浦大学)

Min Wu, Matthew Wicker, Wenjie Ruan, Xiaowei Huang, Marta Kwiatkowska

摘要:

尽管深度神经网络的准确性有所提高,但对抗性例子的发现引发了严重的安全问题。在本文中,我们研究了两个变种的逐点鲁棒性,最大安全半径问题,这对于一个给定的输入样本计算的最小距离的对抗性的例子,和功能的鲁棒性问题,其目的是量化的鲁棒性个别功能对抗性扰动。我们证明,在Lipschitz连续性的假设下,这两个问题都可以通过离散输入空间使用有限优化来近似,并且近似具有可证明的保证,即,误差是有界的。然后,我们表明,由此产生的优化问题可以减少到两个玩家回合制游戏的解决方案,其中第一个玩家选择功能和第二扰动的图像内的功能。虽然第二个玩家的目标是最小化与敌对示例的距离,但根据优化目标,第一个玩家可以是合作的或竞争的。我们采用了一种随时随地的方法来解决游戏,在这个意义上的近似值的游戏,通过单调提高其上限和下限。应用蒙特卡罗树搜索算法计算两种博弈的上界,并分别使用容许A* 和Alpha-Beta剪枝算法计算最大安全半径和特征鲁棒性博弈的下界。当处理最大安全半径问题的上限时,我们的工具与现有的对抗性示例制作算法相比表现出了竞争力。此外,我们展示了如何部署我们的框架来评估神经网络在安全关键应用中的逐点鲁棒性,例如自动驾驶汽车中的交通标志识别。

英文摘要:

Despite the improved accuracy of deep neural networks, the discovery of adversarial examples has raised serious safety concerns. In this paper, we study two variants of pointwise robustness, the maximum safe radius problem, which for a given input sample computes the minimum distance to an adversarial example, and the feature robustness problem, which aims to quantify the robustness of individual features to adversarial perturbations. We demonstrate that, under the assumption of Lipschitz continuity, both problems can be approximated using finite optimisation by discretising the input space, and the approximation has provable guarantees, i.e., the error is bounded. We then show that the resulting optimisation problems can be reduced to the solution of two-player turn-based games, where the first player selects features and the second perturbs the image within the feature. While the second player aims to minimise the distance to an adversarial example, depending on the optimisation objective the first player can be cooperative or competitive. We employ an anytime approach to solve the games, in the sense of approximating the value of a game by monotonically improving its upper and lower bounds. The Monte Carlo tree search algorithm is applied to compute upper bounds for both games, and the Admissible A* and the Alpha-Beta Pruning algorithms are, respectively, used to compute lower bounds for the maximum safety radius and feature robustness games. When working on the upper bound of the maximum safe radius problem, our tool demonstrates competitive performance against existing adversarial example crafting algorithms. Furthermore, we show how our framework can be deployed to evaluate pointwise robustness of neural networks in safety-critical applications such as traffic sign recognition in self-driving cars.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑