召唤恶魔：在机器学习中寻找可利用的bug
Summoning Demons: The Pursuit of Exploitable Bugs in Machine Learning
摘要：
政府和企业越来越依赖数据分析和机器学习（ML）来提高其在消费者满意度、威胁情报、决策制定和产品效率等领域的竞争优势。然而，通过巧妙地破坏用作目标机器学习算法输入的数据子集，对手可以扰乱结果并损害机器学习技术的有效性。虽然对抗性机器学习领域的先前工作已经研究了输入操作对正确的ML算法的影响，但我们考虑了ML实现中的漏洞的利用。在本文中，我们描述了ML程序的攻击面，我们表明，恶意输入利用实现错误，使严格更强大的攻击比经典的对抗性机器学习技术。我们提出了一种半自动化的技术，称为转向模糊，探索这个攻击面，并发现可利用的错误，在机器学习程序，以证明这种威胁的严重性。作为我们工作的结果，我们负责任地披露了五个漏洞，建立了三个新的CVE-ID，并阐明了许多机器学习系统中常见的不安全做法。最后，我们概述了几个研究方向，以进一步了解和减轻这种威胁。
英文摘要：
Governments and businesses increasingly rely on data analytics and machine learning (ML) for improving their competitive edge in areas such as consumer satisfaction, threat intelligence, decision making, and product efficiency. However, by cleverly corrupting a subset of data used as input to a target's ML algorithms, an adversary can perturb outcomes and compromise the effectiveness of ML technology. While prior work in the field of adversarial machine learning has studied the impact of input manipulation on correct ML algorithms, we consider the exploitation of bugs in ML implementations. In this paper, we characterize the attack surface of ML programs, and we show that malicious inputs exploiting implementation bugs enable strictly more powerful attacks than the classic adversarial machine learning techniques. We propose a semi-automated technique, called steered fuzzing, for exploring this attack surface and for discovering exploitable bugs in machine learning programs, in order to demonstrate the magnitude of this threat. As a result of our work, we responsibly disclosed five vulnerabilities, established three new CVE-IDs, and illuminated a common insecure practice across many machine learning systems. Finally, we outline several research directions for further understanding and mitigating this threat.
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。
