-
MNLI: Multi-Genre Natural Language Inference
Propose a method for evaluating gender bias in contextualised word embeddings. -
SEAT: Sentence Encoder Association Test
Propose a method for evaluating gender bias in contextualised word embeddings. -
MNLI and FEVER datasets
The MNLI and FEVER datasets are used to evaluate the proposed MoCaD framework. -
Debias Results of LSDM on Neutral Occupational Pronouns
The authors used the Professions Dataset to generate biased sentences and evaluate the effectiveness of the proposed debiasing method. -
DeBayes: a Bayesian Method for Debiasing Network Embeddings
DeBayes: a Bayesian method for debiasing network embeddings -
Compressing and Debiasing Vision-Language Pre-Trained Models for Visual Quest...
This paper investigates whether a VLP can be compressed and debiased simultaneously by searching sparse and robust subnetworks.
