TaCo: Targeted Concept Removal in Output Embeddings for NLP via Information Theory and Explainability

arXiv:2312.06499v3 Announce Type: replace-cross
Abstract: The fairness of Natural Language Processing (NLP) models has emerged as a crucial concern. Information theory indicates that to achieve fairness, a model should not be able to predict sensitive variables, such as gender, ethnicity, and age. However, information related to these variables often appears implicitly in language, posing a challenge in identifying and mitigating biases effectively. To tackle this issue, we present a novel approach that operates at the embedding level of an NLP …

arxiv concept cs.cl embeddings explainability information nlp stat.ml theory type via