Multi-Group Proportional Representation in Retrieval
This work addresses representational harms in retrieval systems for marginalized intersectional groups, offering a novel approach beyond incremental improvements.
The paper tackles the problem of harmful stereotypes and social disparities in image search and retrieval by addressing the oversight of intersectional groups defined by combinations of attributes like gender and race. It introduces Multi-Group Proportional Representation (MPR), a metric and optimization method that yields more proportional representation across intersectional groups with minimal compromise in retrieval accuracy.
Image search and retrieval tasks can perpetuate harmful stereotypes, erase cultural identities, and amplify social disparities. Current approaches to mitigate these representational harms balance the number of retrieved items across population groups defined by a small number of (often binary) attributes. However, most existing methods overlook intersectional groups determined by combinations of group attributes, such as gender, race, and ethnicity. We introduce Multi-Group Proportional Representation (MPR), a novel metric that measures representation across intersectional groups. We develop practical methods for estimating MPR, provide theoretical guarantees, and propose optimization algorithms to ensure MPR in retrieval. We demonstrate that existing methods optimizing for equal and proportional representation metrics may fail to promote MPR. Crucially, our work shows that optimizing MPR yields more proportional representation across multiple intersectional groups specified by a rich function class, often with minimal compromise in retrieval accuracy.