On Measuring and Mitigating Biased Inferences of Word Embeddings

08/25/2019

∙

by   Sunipa Dev, et al.

∙

0

∙

share

Word embeddings carry stereotypical connotations from the text they are trained on, which can lead to invalid inferences. We use this observation to design a mechanism for measuring stereotypes using the task of natural language inference. We demonstrate a reduction in invalid inferences via bias mitigation strategies on static word embeddings (GloVe), and explore adapting them to contextual embeddings (ELMo).