On Measures of Biases and Harms in NLP
It addresses the need for cohesive understanding and comparison of bias measures in NLP, which is crucial for practitioners to mitigate societal biases, though it is incremental in building on existing work.
This paper tackles the problem of measuring biases and harms in NLP by proposing a practical framework and documentation questions to guide bias measure development, validated through case studies aligning existing measures with different harms.
Recent studies show that Natural Language Processing (NLP) technologies propagate societal biases about demographic groups associated with attributes such as gender, race, and nationality. To create interventions and mitigate these biases and associated harms, it is vital to be able to detect and measure such biases. While existing works propose bias evaluation and mitigation methods for various tasks, there remains a need to cohesively understand the biases and the specific harms they measure, and how different measures compare with each other. To address this gap, this work presents a practical framework of harms and a series of questions that practitioners can answer to guide the development of bias measures. As a validation of our framework and documentation questions, we also present several case studies of how existing bias measures in NLP -- both intrinsic measures of bias in representations and extrinsic measures of bias of downstream applications -- can be aligned with different harms and how our proposed documentation questions facilitates more holistic understanding of what bias measures are measuring.