AI/ML for Network Security: The Emperor has no Clothes
AI/ML for Network Security: The Emperor has no Clothes
复制标题
DOI：
10.1145/3548606.3560609
复制
发表时间：
2022-11
期刊：
影响因子：
--
通讯作者：
A. Jacobs;Roman Beltiukov;W. Willinger;R. Ferreira;Arpit Gupta;L. Granville
中科院分区：
文献类型：
--
作者：
A. Jacobs;Roman Beltiukov;W. Willinger;R. Ferreira;Arpit Gupta;L. Granville
最近的一些研究工作提出了基于机器学习（ML）的解决方案，可以检测网络流量中的复杂模式，以解决各种网络安全问题。然而，如果不了解这些黑盒模型是如何做出决策的，网络运营商就不愿意信任它们并将其部署到生产环境中。这种不情愿的一个关键原因是，这些模型容易出现规格不足的问题，这里定义为未能详细说明模型。这个问题并不是网络安全领域所独有的，它表现在ML模型中，当部署在现实世界的环境中时，ML模型表现出意想不到的糟糕行为，并且促使人们对开发可解释的ML解决方案越来越感兴趣（例如，决策树），用于向人类“解释”给定的黑盒模型如何做出决策。然而，合成这样的可解释的模型，其以高保真度捕获给定的黑盒模型的决策，同时也是实用的（即，小到足以让人类理解）是一个挑战。在本文中，我们专注于合成高保真度和低复杂度的决策树，以帮助网络运营商确定他们的ML模型是否存在规格不足的问题。为此，我们提出了Trustee，这是一个框架，它以现有的ML模型和训练数据集作为输入，并生成高保真，易于解释的决策树和相关的信任报告作为输出。使用已发布的完全可复制的ML模型，我们展示了从业者如何使用Trustee来识别模型欠规范的三个常见实例;即，捷径学习的证据，虚假相关性的存在，以及对分布外样本的脆弱性。
Several recent research efforts have proposed Machine Learning (ML)-based solutions that can detect complex patterns in network traffic for a wide range of network security problems. However, without understanding how these black-box models are making their decisions, network operators are reluctant to trust and deploy them in their production settings. One key reason for this reluctance is that these models are prone to the problem of underspecification, defined here as the failure to specify a model in adequate detail. Not unique to the network security domain, this problem manifests itself in ML models that exhibit unexpectedly poor behavior when deployed in real-world settings and has prompted growing interest in developing interpretable ML solutions (e.g., decision trees) for "explaining'' to humans how a given black-box model makes its decisions. However, synthesizing such explainable models that capture a given black-box model's decisions with high fidelity while also being practical (i.e., small enough in size for humans to comprehend) is challenging. In this paper, we focus on synthesizing high-fidelity and low-complexity decision trees to help network operators determine if their ML models suffer from the problem of underspecification. To this end, we present Trustee, a framework that takes an existing ML model and training dataset as input and generates a high-fidelity, easy-to-interpret decision tree and associated trust report as output. Using published ML models that are fully reproducible, we show how practitioners can use Trustee to identify three common instances of model underspecification; i.e., evidence of shortcut learning, presence of spurious correlations, and vulnerability to out-of-distribution samples.