arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.10013cs.CL
收藏

模型提取和对抗性可移植性,您的BERT易受攻击!

Model Extraction and Adversarial Transferability, Your BERT is Vulnerable!

Monash University(莫纳什大学);Ant Group(蚂蚁集团);The Australian National University(澳大利亚国立大学);Data61 CSIRO;Lehigh University(里海大学)

Xuanli He, Lingjuan Lyu, Qiongkai Xu, Lichao Sun

摘要:

自然语言处理(NLP)任务,从文本分类到文本生成,已经被预先训练的语言模型(如BERT)彻底改变。这允许公司通过为下游任务封装微调的BERT模型来轻松构建强大的API。然而,当一个经过微调的BERT模型被部署为服务时,它可能会遭受恶意用户发起的不同攻击。在这项工作中,我们首先介绍了对手如何在有限的先验知识和查询的情况下,在多个基准数据集上窃取基于BERT的API服务(受害者/目标模型)。我们进一步表明,提取的模型可以导致对受害者模型的高度可转移的对抗攻击。我们的研究表明,即使受害者模型和攻击模型之间存在架构不匹配,基于BERT的API服务的潜在漏洞仍然存在。最后,我们研究了两种保护受害者模型的防御策略,发现除非牺牲受害者模型的性能,否则模型抽取和对抗转移都可以有效地危害目标模型

英文摘要:

Natural language processing (NLP) tasks, ranging from text classification to text generation, have been revolutionised by the pre-trained language models, such as BERT. This allows corporations to easily build powerful APIs by encapsulating fine-tuned BERT models for downstream tasks. However, when a fine-tuned BERT model is deployed as a service, it may suffer from different attacks launched by malicious users. In this work, we first present how an adversary can steal a BERT-based API service (the victim/target model) on multiple benchmark datasets with limited prior knowledge and queries. We further show that the extracted model can lead to highly transferable adversarial attacks against the victim model. Our studies indicate that the potential vulnerabilities of BERT-based API services still hold, even when there is an architectural mismatch between the victim model and the attack model. Finally, we investigate two defence strategies to protect the victim model and find that unless the performance of the victim model is sacrificed, both model ex-traction and adversarial transferability can effectively compromise the target models

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。

↑