论文标题
在元安装中的性别偏见
Gender Bias in Meta-Embeddings
论文作者
论文摘要
已经提出了不同的方法来从给定的一组源嵌入中开发元嵌入。但是，源嵌入可能包含不公平的性别相关偏见，并且尚未研究这些元素的影响。我们研究在三种不同的情况下创建的元安装中的性别偏见：（1）元安装多个来源而不执行任何证据（多源无偏差），（2）通过单个方法（多source单debiasing）和单个方法（3）单一的方法（3）单一的方法（3）单个方法（3）元来的多个来源（3）元来启用多个来源（3）元用户（3）元来（3）元来（3）元来（3）元来（3）meta-emping（3）多态度）。我们的实验结果表明，与输入源嵌入相比，荟萃提示会放大性别偏见。我们发现，不仅需要来源，而且还需要其元装置来减轻这些偏见。此外，我们提出了一种基于元嵌入学习的新型伪造方法，在该方法中，我们在单个源嵌入中使用多种偏见方法，然后创建一个单一的无偏元装置。
Different methods have been proposed to develop meta-embeddings from a given set of source embeddings. However, the source embeddings can contain unfair gender-related biases, and how these influence the meta-embeddings has not been studied yet. We study the gender bias in meta-embeddings created under three different settings: (1) meta-embedding multiple sources without performing any debiasing (Multi-Source No-Debiasing), (2) meta-embedding multiple sources debiased by a single method (Multi-Source Single-Debiasing), and (3) meta-embedding a single source debiased by different methods (Single-Source Multi-Debiasing). Our experimental results show that meta-embedding amplifies the gender biases compared to input source embeddings. We find that debiasing not only the sources but also their meta-embedding is needed to mitigate those biases. Moreover, we propose a novel debiasing method based on meta-embedding learning where we use multiple debiasing methods on a single source embedding and then create a single unbiased meta-embedding.
