On Measuring and Mitigating Biased Inferences of Word Embeddings

Authors

  • Sunipa Dev University of Utah
  • Tao Li University of Utah
  • Jeff M. Phillips University of Utah
  • Vivek Srikumar University of Utah

DOI:

https://doi.org/10.1609/aaai.v34i05.6267

Abstract

Word embeddings carry stereotypical connotations from the text they are trained on, which can lead to invalid inferences in downstream models that rely on them. 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). Further, we show that for gender bias, these techniques extend to contextualized embeddings when applied selectively only to the static components of contextualized embeddings (ELMo, BERT).

Downloads

Published

2020-04-03

How to Cite

Dev, S., Li, T., Phillips, J. M., & Srikumar, V. (2020). On Measuring and Mitigating Biased Inferences of Word Embeddings. Proceedings of the AAAI Conference on Artificial Intelligence, 34(05), 7659-7666. https://doi.org/10.1609/aaai.v34i05.6267

Issue

Section

AAAI Technical Track: Natural Language Processing