arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2205.04321cs.LG
收藏

评估差异私有合成数据的公平性影响

Evaluating the Fairness Impact of Differentially Private Synthetic Data

Harvard University(哈佛大学);Microsoft(微软)

Blake Bullwinkel, Kristen Grabarz, Lily Ke, Scarlett Gong, Chris Tanner, Joshua Allen

摘要:

差分私有合成数据是一种很有前途的方法,可以最大限度地利用包含敏感信息的数据。然而,由于压制代表性不足的阶层往往需要实现隐私,这可能与公平相冲突。我们评估了四个DP合成器,并提出了实证结果表明,这些模型中的三个经常降低下游二进制分类任务的公平性结果。我们在公平性和生成的合成数据中存在的少数群体的比例之间建立了联系,并发现通过多标签欠采样方法预处理的数据训练合成器可以在不降低准确性的情况下促进更公平的结果。

英文摘要:

Differentially private (DP) synthetic data is a promising approach to maximizing the utility of data containing sensitive information. Due to the suppression of underrepresented classes that is often required to achieve privacy, however, it may be in conflict with fairness. We evaluate four DP synthesizers and present empirical results indicating that three of these models frequently degrade fairness outcomes on downstream binary classification tasks. We draw a connection between fairness and the proportion of minority groups present in the generated synthetic data, and find that training synthesizers on data that are pre-processed via a multi-label undersampling method can promote more fair outcomes without degrading accuracy.

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。

↑