Gender bias in (non)-contextual clinical word embeddings for stereotypical medical categories. (arXiv:2208.01341v2 [cs.CL] UPDATED)

Clinical word embeddings are extensively used in various Bio-NLP problems as
a state-of-the-art feature vector representation. Although they are quite
successful at the semantic representation of words, due to the dataset - which
potentially carries statistical and societal bias - on which they are trained,
they might exhibit gender stereotypes. This study analyses gender bias of
clinical embeddings on three medical categories: mental disorders, sexually
transmitted diseases, and personality traits. To this extent, we analyze two
different pre-trained embeddings namely …

arxiv bias gender gender bias medical word embeddings