BriefGPT.xyz
Aug, 2022
使用非线性几何修正词向量偏置
Debiasing Word Embeddings with Nonlinear Geometry
HTML
PDF
Lu Cheng, Nayoung Kim, Huan Liu
TL;DR
本文研究了多个社会类别上的偏差问题，并构建了交叉子空间来去除这些交叉偏差，实验证明这种方法的有效性。
Abstract
debiasing
word embeddings
has been largely limited to individual and independent
social categories
. However, real-world corpora typically
→