Attenuating Bias in Word Vectors

Abstract

Word vector representations are well developed tools for various NLP and Machine Learning tasks and are known to retain significant semantic and syntactic structure of languages. But they are prone to carrying and amplifying bias which can perpetrate discrimination in various applications. In this work, we explore new simple ways to detect the most stereotypically gendered words in an embedding and remove the bias from them. We verify how names are masked carriers of gender bias and then use that as a tool to attenuate bias in embeddings. Further, we extend this property of names to show how names can be used to detect other types of bias in the embeddings such as bias based on race, ethnicity, and age.

To appear in AIStats 2019

Cited by in corpus (19)

• Two Simple Ways to Learn Individual Fairness Metrics from Data
• Robustness and Reliability of Gender Bias Assessment in Word Embeddings: The Role of Base Pairs
• USCORE: An Effective Approach to Fully Unsupervised Evaluation Metrics for Machine Translation
• Debiasing Pre-trained Contextualised Embeddings
• Neutralizing Gender Bias in Word Embedding with Latent Disentanglement and Counterfactual Generation
• Examining Gender Bias in Languages with Grammatical Gender
• AraWEAT: Multidimensional Analysis of Biases in Arabic Word Embeddings
• Harms of Gender Exclusivity and Challenges in Non-Binary Representation in Language Technologies
• On the Limitations of Cross-lingual Encoders as Exposed by Reference-Free Machine Translation Evaluation
• Argument from Old Man's View: Assessing Social Bias in Argumentation
• User Acceptance of Gender Stereotypes in Automated Career Recommendations
• Intentonomy: a Dataset and Study towards Human Intent Understanding
• MDR Cluster-Debias: A Nonlinear WordEmbedding Debiasing Pipeline
• VERB: Visualizing and Interpreting Bias Mitigation Techniques for Word Representations
• The Geometry of Distributed Representations for Better Alignment, Attenuated Bias, and Improved Interpretability
• Obstructing Classification via Projection
• OSCaR: Orthogonal Subspace Correction and Rectification of Biases in Word Embeddings
• Constrained Non-Affine Alignment of Embeddings
• Gender Stereotype Reinforcement: Measuring the Gender Bias Conveyed by Ranking Algorithms