arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2108.05490cs.CLcs.LG
收藏

基于文本嵌入的排名算法攻击--以招聘算法为例

Attacks against Ranking Algorithms with Text Embeddings: a Case Study on Recruitment Algorithms

University of Texas at Arlington（德克萨斯大学阿灵顿分校）

Anahita Samadi, Debapriya Banerjee, Shirin Nilizadeh

阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

摘要：

最近的一些研究表明，文本分类任务容易受到中毒和逃避攻击。然而，很少有工作调查攻击决策算法，使用文本嵌入，其输出是一个排名。在本文中，我们专注于招聘过程中的排名算法，采用文本嵌入排名申请人的简历相比，职位描述。我们展示了白盒和黑盒攻击，识别文本项，根据其位置在嵌入空间中，有显着的贡献，在增加简历和职位描述之间的相似性得分。然后，对手使用这些文本项来提高其简历的排名。我们测试了使用从通用句子编码器（USE）和词频逆文档频率（TF IDF）向量获得的相似性得分的招聘算法。我们的研究结果表明，在这两种对抗性设置中，平均而言，攻击者是成功的。我们还发现，与USE相比，对TF IDF的攻击更成功。

英文摘要：

Recently, some studies have shown that text classification tasks are vulnerable to poisoning and evasion attacks. However, little work has investigated attacks against decision making algorithms that use text embeddings, and their output is a ranking. In this paper, we focus on ranking algorithms for recruitment process, that employ text embeddings for ranking applicants resumes when compared to a job description. We demonstrate both white box and black box attacks that identify text items, that based on their location in embedding space, have significant contribution in increasing the similarity score between a resume and a job description. The adversary then uses these text items to improve the ranking of their resume among others. We tested recruitment algorithms that use the similarity scores obtained from Universal Sentence Encoder (USE) and Term Frequency Inverse Document Frequency (TF IDF) vectors. Our results show that in both adversarial settings, on average the attacker is successful. We also found that attacks against TF IDF is more successful compared to USE.

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。

↑