BiasConnect: Investigating Bias Interactions in Text-to-Image Models

2025-03-12

AI Summary

Social biases (e.g., gender, race, age) in text-to-image (TTI) models are not independent; debiasing along one dimension may inadvertently amplify or mitigate biases along others—an interaction effect poorly quantified by existing methods. Method: Departing from the conventional assumption of independent biases, we propose BiasConnect, the first framework grounded in counterfactual reasoning to model and quantify cross-dimensional bias interactions via a causal graph of bias interdependence. It integrates counterfactual image generation, pairwise causal modeling, bias distribution shift estimation, and cross-sensitivity analysis. Contribution/Results: Experiments demonstrate reliable interaction estimation (correlation +0.69), enabling optimal debiasing path selection, cross-model comparison of bias dependencies, and mechanistic analysis of intersectional bias amplification. BiasConnect provides an interpretable, quantifiable foundation for fair generative modeling—advancing both theory and practical tooling for bias mitigation in TTI systems.

Technology Category

• Computer Vision: Bias, Fairness & Privacy
• Reasoning under Uncertainty: Causality
• Philosophy and Ethics of AI: Bias, Fairness & Equity

Application Category

• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

The biases exhibited by Text-to-Image (TTI) models are often treated as if they are independent, but in reality, they may be deeply interrelated. Addressing bias along one dimension, such as ethnicity or age, can inadvertently influence another dimension, like gender, either mitigating or exacerbating existing disparities. Understanding these interdependencies is crucial for designing fairer generative models, yet measuring such effects quantitatively remains a challenge. In this paper, we aim to address these questions by introducing BiasConnect, a novel tool designed to analyze and quantify bias interactions in TTI models. Our approach leverages a counterfactual-based framework to generate pairwise causal graphs that reveals the underlying structure of bias interactions for the given text prompt. Additionally, our method provides empirical estimates that indicate how other bias dimensions shift toward or away from an ideal distribution when a given bias is modified. Our estimates have a strong correlation (+0.69) with the interdependency observations post bias mitigation. We demonstrate the utility of BiasConnect for selecting optimal bias mitigation axes, comparing different TTI models on the dependencies they learn, and understanding the amplification of intersectional societal biases in TTI models.

Problem

• Analyzing bias interactions in Text-to-Image models.
• Quantifying how bias mitigation affects other dimensions.
• Understanding intersectional societal biases in generative models.

Innovation

• BiasConnect tool analyzes bias interactions
• Counterfactual framework generates causal graphs
• Empirical estimates quantify bias shifts