Is What You Ask For What You Get? Investigating Concept Associations in Text-to-Image Models

2024-10-06

arXiv.org

AI Summary

Auditing implicit biases in text-to-image (T2I) models lacks human-interpretable explanatory mechanisms. To address this, we propose Concept2Concept—a novel framework that establishes the first explainable, vision-language concept–based conditional distribution representation paradigm, mapping prompt–image associations onto a semantically grounded concept space. Our method enables dual-path auditing: user-defined concepts and real-world distribution alignment, integrating concept-space modeling, quantitative conditional distribution analysis, and vision-language alignment evaluation. We release an open-source, interactive web tool built with D3.js and Gradio. Empirical evaluation across multiple state-of-the-art T2I models reveals significant conceptual biases—e.g., strong gender–occupation and race–scene associations. The tool has been widely adopted by the research community, with over 5,000 demo accesses to date.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Text-to-image (T2I) models are increasingly used in impactful real-life applications. As such, there is a growing need to audit these models to ensure that they generate desirable, task-appropriate images. However, systematically inspecting the associations between prompts and generated content in a human-understandable way remains challenging. To address this, we propose Concept2Concept, a framework where we characterize conditional distributions of vision language models using interpretable concepts and metrics that can be defined in terms of these concepts. This characterization allows us to use our framework to audit models and prompt-datasets. To demonstrate, we investigate several case studies of conditional distributions of prompts, such as user-defined distributions or empirical, real-world distributions. Lastly, we implement Concept2Concept as an open-source interactive visualization tool to facilitate use by non-technical end-users. A demo is available at https://tinyurl.com/Concept2ConceptDemo.

Problem

• Auditing text-to-image models
• Understanding prompt-content associations
• Developing interpretable visualization tools

Innovation

• Characterizes vision language models
• Audits models using interpretable concepts
• Implements open-source visualization tool