Mitigating bias in training on biased datasets is an important open prob...
Real-world datasets often encode stereotypes and societal biases. Such b...
This paper presents fairlib, an open-source framework for assessing and
...
Trained classification models can unintentionally lead to biased
represe...
While pretrained language models ("LM") have driven impressive gains ove...
The quality of a document is affected by various factors, including
gram...