# Understanding Undesirable Word Embedding Associations

arXiv: https://arxiv.org/abs/1908.06361 · HTML: https://mlsanity.com/paper/1908.06361/

## Key facts

- Cited by 59 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Aug 2019.

## Built on by (most cited first, 30 of 59)

- [RedditBias: A Real-World Resource for Bias Evaluation and Debiasing of Conversational Language Models](https://mlsanity.com/paper/2106.03521/) (2021)
- [A Survey on Gender Bias in Natural Language Processing](https://mlsanity.com/paper/2112.14168/) (2021)
- [Theories of "Gender" in NLP Bias Research](https://mlsanity.com/paper/2205.02526/) (2022)
- [Unmasking the Mask -- Evaluating Social Biases in Masked Language Models](https://mlsanity.com/paper/2104.07496/) (2021)
- [Problems with Cosine as a Measure of Embedding Similarity for High Frequency Words](https://mlsanity.com/paper/2205.05092/) (2022)
- [Probing Classifiers are Unreliable for Concept Removal and Detection](https://mlsanity.com/paper/2207.04153/) (2022)
- [Low Frequency Names Exhibit Bias and Overfitting in Contextualizing Language Models](https://mlsanity.com/paper/2110.00672/) (2021)
- [Gold Doesn't Always Glitter: Spectral Removal of Linear and Nonlinear Guarded Attribute Information](https://mlsanity.com/paper/2203.07893/) (2022)
- [MABEL: Attenuating Gender Bias using Textual Entailment Data](https://mlsanity.com/paper/2210.14975/) (2022)
- [Choose Your Lenses: Flaws in Gender Bias Evaluation](https://mlsanity.com/paper/2210.11471/) (2022)
- [Sociodemographic Bias in Language Models: A Survey and Forward Path](https://mlsanity.com/paper/2306.08158/) (2023)
- [Dictionary-based Debiasing of Pre-trained Word Embeddings](https://mlsanity.com/paper/2101.09525/) (2021)
- [Frequency-based Distortions in Contextualized Word Embeddings](https://mlsanity.com/paper/2104.08465/) (2021)
- [Fairness and Bias in Multimodal AI: A Survey](https://mlsanity.com/paper/2406.19097/) (2024)
- [Evaluating Gender Bias in Hindi-English Machine Translation](https://mlsanity.com/paper/2106.08680/) (2021)
- [Idiosyncratic but not Arbitrary: Learning Idiolects in Online Registers Reveals Distinctive yet Consistent Individual Styles](https://mlsanity.com/paper/2109.03158/) (2021)
- [Sense Embeddings are also Biased--Evaluating Social Biases in Static and Contextualised Sense Embeddings](https://mlsanity.com/paper/2203.07523/) (2022)
- [Better Hit the Nail on the Head than Beat around the Bush: Removing Protected Attributes with a Single Projection](https://mlsanity.com/paper/2212.04273/) (2022)
- [The Birth of Bias: A case study on the evolution of gender bias in an English language model](https://mlsanity.com/paper/2207.10245/) (2022)
- [Fairness in Language Models Beyond English: Gaps and Challenges](https://mlsanity.com/paper/2302.12578/) (2023)
- [Assessing the Reliability of Word Embedding Gender Bias Measures](https://mlsanity.com/paper/2109.04732/) (2021)
- [Diachronic Analysis of German Parliamentary Proceedings: Ideological Shifts through the Lens of Political Biases](https://mlsanity.com/paper/2108.06295/) (2021)
- [How Gender Debiasing Affects Internal Model Representations, and Why It Matters](https://mlsanity.com/paper/2204.06827/) (2022)
- [Learning to Generate Equitable Text in Dialogue from Biased Training Data](https://mlsanity.com/paper/2307.04303/) (2023)
- [Residual2Vec: Debiasing graph embedding with random graphs](https://mlsanity.com/paper/2110.07654/) (2021)
- [Comparing Intrinsic Gender Bias Evaluation Measures without using Human Annotated Examples](https://mlsanity.com/paper/2301.12074/) (2023)
- [Evaluating Metrics for Bias in Word Embeddings](https://mlsanity.com/paper/2111.07864/) (2021)
- [Richer Countries and Richer Representations](https://mlsanity.com/paper/2205.05093/) (2022)
- [Trustworthy Social Bias Measurement](https://mlsanity.com/paper/2212.11672/) (2022)
- [Mitigating Gender Stereotypes in Hindi and Marathi](https://mlsanity.com/paper/2205.05901/) (2022)

## Abstract (arXiv)

Word embeddings are often criticized for capturing undesirable word associations such as gender stereotypes. However, methods for measuring and removing such biases remain poorly understood. We show that for any embedding model that implicitly does matrix factorization, debiasing vectors post hoc using subspace projection (Bolukbasi et al., 2016) is, under certain conditions, equivalent to training on an unbiased corpus. We also prove that WEAT, the most common association test for word embeddings, systematically overestimates bias. Given that the subspace projection method is provably effective, we use it to derive a new measure of association called the textitrelational inner product association (RIPA). Experiments with RIPA reveal that, on average, skipgram with negative sampling (SGNS) does not make most words any more gendered than they are in the training corpus. However, for gender-stereotyped words, SGNS actually amplifies the gender association in the corpus.