Biased Embeddings from Wild Data: Measuring, Understanding and Removing

TL;DR

This paper introduces methods to measure, understand, and mitigate biases in NLP embeddings derived from real-world data, highlighting the connection between embedding bias and societal biases, and proposing a simple bias reduction technique.

Contribution

It provides a rigorous bias measurement approach, analyzes the reflection of societal biases in embeddings, and demonstrates an effective bias removal method.

Findings

Bias measurement correlates with social psychology word lists.

Gender bias in embeddings mirrors real-world occupational gender bias.

Simple projection reduces embedding bias significantly.

Abstract

Many modern Artificial Intelligence (AI) systems make use of data embeddings, particularly in the domain of Natural Language Processing (NLP). These embeddings are learnt from data that has been gathered "from the wild" and have been found to contain unwanted biases. In this paper we make three contributions towards measuring, understanding and removing this problem. We present a rigorous way to measure some of these biases, based on the use of word lists created for social psychology applications; we observe how gender bias in occupations reflects actual gender bias in the same occupations in the real world; and finally we demonstrate how a simple projection can significantly reduce the effects of embedding bias. All this is part of an ongoing effort to understand how trust can be built into AI systems.