Gender Bias Hidden Behind Chinese Word Embeddings: The Case of Chinese Adjectives

TL;DR

This paper investigates gender bias in Chinese adjective word embeddings, revealing how such biases differ from human attitudes and highlighting the importance of language-specific analysis.

Contribution

It introduces a novel focus on Chinese adjectives in static word embeddings and compares computational bias with human attitudes to understand cultural differences.

Findings

Gender bias exists in Chinese adjective embeddings.

Bias in embeddings differs from human attitudes.

Analysis highlights language-specific bias characteristics.

Abstract

Gender bias in word embeddings gradually becomes a vivid research field in recent years. Most studies in this field aim at measurement and debiasing methods with English as the target language. This paper investigates gender bias in static word embeddings from a unique perspective, Chinese adjectives. By training word representations with different models, the gender bias behind the vectors of adjectives is assessed. Through a comparison between the produced results and a human-scored data set, we demonstrate how gender bias encoded in word embeddings differentiates from people's attitudes.