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 2312.10467cs.LG
收藏

TrojFSP:少样本提示调优中的木马插入

TrojFSP: Trojan Insertion in Few-shot Prompt Tuning

University of Central Florida(中佛罗里达大学);Samsung Research America(三星美国研究院);University of Pittsburgh(匹兹堡大学);Indiana University Bloomington(印第安纳大学伯明顿分校)

Mengxin Zheng, Jiaqi Xue, Xun Chen, YanShan Wang, Qian Lou, Lei Jiang

AI总结:

针对少样本提示调优中的木马攻击,提出TrojFSP方法,通过目标类收缩和选择性令牌投毒解决投毒不平衡与过拟合问题,实现高攻击成功率且几乎不损失干净数据准确率。

摘要:

提示调优是使固定的预训练语言模型(PLM)适应各种下游任务的最有效解决方案之一,尤其是在仅有少量输入样本的情况下。然而,在少量数据样本上进行提示调优的安全问题(例如木马攻击)尚未得到充分研究。将已有的数据投毒攻击直接迁移到少样本提示调优中会带来多重挑战。一个显著的问题是“投毒不平衡问题”,即非目标类样本被添加到目标类中,导致目标类样本数量多于非目标类样本。虽然这个问题在常规调优中并不关键,但它会严重阻碍少样本提示调优,使得难以同时实现高攻击成功率(ASR)并保持干净数据准确率(CDA)。此外,少样本提示在ASR和CDA方面都容易出现过拟合。在本文中,我们引入了TrojFSP,一种旨在解决这些挑战的方法。为了解决投毒不平衡问题,我们开发了一种“目标类收缩(TC-Shrink)”技术,旨在均衡投毒样本的数量。为了对抗过拟合,我们采用了一种“选择性令牌投毒”技术来提升攻击性能。此外,我们引入了一个“木马触发器注意力”目标函数,以增强被投毒的木马提示对触发器的注意力。实验表明,我们的TrojFSP在各种PLM和数据集上实现了超过99%的ASR,同时将CDA的下降保持在可忽略的水平。

英文摘要:

Prompt tuning is one of the most effective solutions to adapting a fixed pre-trained language model (PLM) for various downstream tasks, especially with only a few input samples. However, the security issues, e.g., Trojan attacks, of prompt tuning on a few data samples are not well-studied. Transferring established data poisoning attacks directly to few-shot prompt tuning presents multiple challenges. One significant issue is the \textit{poisoned imbalance issue}, where non-target class samples are added to the target class, resulting in a greater number of target-class samples compared to non-target class. While this issue is not critical in regular tuning, it significantly hampers the few-shot prompt tuning, making it difficult to simultaneously achieve a high attack success rate (ASR) and maintain clean data accuracy (CDA). Additionally, few-shot prompting is prone to overfitting in terms of both ASR and CDA. In this paper, we introduce \textit{TrojFSP}, a method designed to address the challenges. To solve the poisoned imbalance issue, we develop a \textit{Target-Class Shrink (TC-Shrink)} technique, which aims to equalize the number of poisoning samples. To combat overfitting, we employ a \textit{Selective Token Poisoning} technique to boost attack performance. Furthermore, we introduce a \textit{Trojan-Trigger Attention} objective function to amplify the attention of the poisoned trojan prompt on triggers. Experiments show that our TrojFSP achieves an ASR of over 99\% while maintaining negligible decreases in CDA across various PLMs and datasets.

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。

↑