Double-Hard Debias: Tailoring Word Embeddings for Gender Bias Mitigation

Tianlu Wang · Xi Lin · Nazneen Fatema Rajani · Vicente Ordóñez · Caiming Xiong

44 citations0/0 files storedAnnual Meeting of the Association for Computational Linguistics (ACL)

No readable copy stored yet — this record is metadata only.

No open-access location is known for this record yet.

Literature note Cite in a project Citation map

Ethics and Social Impacts of AI Topic Modeling Hate Speech and Cyberbullying Detection

Abstract

Word embeddings derived from humangenerated corpora inherit strong gender bias which can be further amplified by downstream models.Some commonly adopted debiasing approaches, including the seminal Hard Debias algorithm (Bolukbasi et al., 2016), apply post-processing procedures that project pre-trained word embeddings into a subspace orthogonal to an inferred gender subspace.We discover that semantic-agnostic corpus regularities such as word frequency captured by the word embeddings negatively impact the performance of these algorithms.We propose a simple but effective technique, Double-Hard Debias, which purifies the word embeddings against such corpus regularities prior to inferring and removing the gender subspace.Experiments on three bias mitigation benchmarks show that our approach preserves the distributional semantics of the pre-trained word embeddings while reducing gender bias to a significantly larger degree than prior approaches.

Paper brief

Loading…

Cite this work

Citation unavailable for this work.

BibTeX RIS CSL-JSON

No public copy is stored for this record, so only the metadata is cited.

Citations in this library

Citation map →

Stored only when both papers are in the corpus — 44 citations exist upstream.

References

2017Semantics derived automatically from language corpora contain human-like biases 1969Category norms of verbal items in 56 categories A replication and extension of the Connecticut category norms.2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings 2019Measuring Bias in Contextualized Word Representations 2019Gender Bias in Contextualized Word Embeddings 2019Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them 2019Evaluating the Underlying Gender Bias in Contextualized Word Embeddings 2019Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them

Cited by

2022Word Embeddings via Causal Inference: Gender Bias Reducing and Semantic Information Preserving 2020Monolingual and Multilingual Reduction of Gender Bias in Contextualized Representations 2021WordBias: An Interactive Visual Tool for Discovering Intersectional Biases Encoded in Word Embeddings 2021Dynamically Disentangling Social Bias from Task-Oriented Representations with Adversarial Attack

Explore the map →

Papers that cite the same references — the relational fallback while embeddings are still backfilling.

Imported from OpenAlex (CC0) — nothing here is generated.