arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2006.06841cs.LGcs.CRstat.ML
收藏

Source Code神经模型中的后门

Backdoors in Neural Models of Source Code

University of Wisconsin–Madison(威斯康星大学麦迪逊分校)

Goutham Ramakrishnan, Aws Albarghouthi

摘要:

深度神经网络容易受到一系列对手的攻击。一类特别有害的漏洞是后门,在输入中存在微妙触发的情况下,模型预测会出现分歧。攻击者可以通过毒化训练数据来植入后门,以在触发的输入上产生期望的目标预测。我们在源代码深度学习的背景下研究后门。(1)我们为源代码任务定义了一系列后门类,并展示了如何毒化数据集以安装此类后门。(2)我们从强大的统计数据中调整和改进了最近的算法,表明后门在源代码的学习表示中留下了频谱签名,从而能够检测中毒数据。(3)我们对不同的架构和语言进行了全面的评估,显示了注入后门的容易性和我们消除它们的能力。

英文摘要:

Deep neural networks are vulnerable to a range of adversaries. A particularly pernicious class of vulnerabilities are backdoors, where model predictions diverge in the presence of subtle triggers in inputs. An attacker can implant a backdoor by poisoning the training data to yield a desired target prediction on triggered inputs. We study backdoors in the context of deep-learning for source code. (1) We define a range of backdoor classes for source-code tasks and show how to poison a dataset to install such backdoors. (2) We adapt and improve recent algorithms from robust statistics for our setting, showing that backdoors leave a spectral signature in the learned representation of source code, thus enabling detection of poisoned data. (3) We conduct a thorough evaluation on different architectures and languages, showing the ease of injecting backdoors and our ability to eliminate them.

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。

↑