CAIRe: Cultural Attribution of Images by Retrieval-Augmented Evaluation

TL;DR

CAIRe is a novel evaluation metric that measures the cultural relevance of images across diverse contexts, addressing biases in text-to-image models with strong alignment to human judgment.

Contribution

We introduce CAIRe, a new metric for evaluating cultural relevance in images, grounded in knowledge bases and factual judgments, improving bias measurement in text-to-image models.

Findings

CAIRe outperforms baselines by 22% F1 on a curated dataset.

Achieves Pearson's correlation of 0.56 and 0.66 with human ratings.

Demonstrates strong alignment with human cultural relevance judgments.

Abstract

As text-to-image models become increasingly prevalent, ensuring their equitable performance across diverse cultural contexts is critical. Efforts to mitigate cross-cultural biases have been hampered by trade-offs, including a loss in performance, factual inaccuracies, or offensive outputs. Despite widespread recognition of these challenges, an inability to reliably measure these biases has stalled progress. To address this gap, we introduce CAIRe, an evaluation metric that assesses the degree of cultural relevance of an image, given a user-defined set of labels. Our framework grounds entities and concepts in the image to a knowledge base and uses factual information to give independent graded judgments for each culture label. On a manually curated dataset of culturally salient but rare items built using language models, CAIRe surpasses all baselines by 22% F1 points. Additionally, we construct two datasets for culturally universal concepts, one comprising T2I-generated outputs and another retrieved from naturally occurring data. CAIRe achieves Pearson's correlations of 0.56 and 0.66 with human ratings on these sets, based on a 5-point Likert scale of cultural relevance. This demonstrates its strong alignment with human judgment across diverse image sources.