Data-driven statistical natural language processing (NLP) techniques leverage
large amounts of language data to build models that can understand language.
However, most language data reflect the public discourse at the time the data
was produced, and hence NLP models are susceptible to