# KG-FairDiff: Knowledge Graph-Guided Prompt Refinement for Demographically Fair Text-to-Image Generation

arXiv: https://arxiv.org/abs/2606.01282 · HTML: https://mlsanity.com/paper/2606.01282/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 39 references that are in the map.
- First posted on arXiv 31 May 2026.

## Builds on (most cited first, 28 of 28)

- [Bias and Fairness in Large Language Models: A Survey](https://mlsanity.com/paper/2309.00770/) (2023)
- [Stable Bias: Analyzing Societal Representations in Diffusion Models](https://mlsanity.com/paper/2303.11408/) (2023)
- [Holistic Evaluation of Text-To-Image Models](https://mlsanity.com/paper/2311.04287/) (2023)
- [Having Beer after Prayer? Measuring Cultural Bias in Large Language Models](https://mlsanity.com/paper/2305.14456/) (2023)
- [CultureBank: An Online Community-Driven Knowledge Base Towards Culturally Aware Language Technologies](https://mlsanity.com/paper/2404.15238/) (2024)
- [Measuring Political Bias in Large Language Models: What Is Said and How It Is Said](https://mlsanity.com/paper/2403.18932/) (2024)
- [Survey of Bias In Text-to-Image Generation: Definition, Evaluation, and Mitigation](https://mlsanity.com/paper/2404.01030/) (2024)
- [ITI-GEN: Inclusive Text-to-Image Generation](https://mlsanity.com/paper/2309.05569/) (2023)
- [OpenBias: Open-set Bias Detection in Text-to-Image Generative Models](https://mlsanity.com/paper/2404.07990/) (2024)
- [Investigating Bias in LLM-Based Bias Detection: Disparities between LLMs and Human Perception](https://mlsanity.com/paper/2403.14896/) (2024)
- [CultureVLM: Characterizing and Improving Cultural Understanding of Vision-Language Models for over 100 Countries](https://mlsanity.com/paper/2501.01282/) (2025)
- [T2IAT: Measuring Valence and Stereotypical Biases in Text-to-Image Generation](https://mlsanity.com/paper/2306.00905/) (2023)
- [Fair Text-to-Image Diffusion via Fair Mapping](https://mlsanity.com/paper/2311.17695/) (2023)
- [CDEval: A Benchmark for Measuring the Cultural Dimensions of Large Language Models](https://mlsanity.com/paper/2311.16421/) (2023)
- [Rethinking Training for De-biasing Text-to-Image Generation: Unlocking the Potential of Stable Diffusion](https://mlsanity.com/paper/2408.12692/) (2024)
- [Mitigating stereotypical biases in text to image generative systems](https://mlsanity.com/paper/2310.06904/) (2023)
- [SCoFT: Self-Contrastive Fine-Tuning for Equitable Image Generation](https://mlsanity.com/paper/2401.08053/) (2024)
- [CIVICS: Building a Dataset for Examining Culturally-Informed Values in Large Language Models](https://mlsanity.com/paper/2405.13974/) (2024)
- [Metrics for Dataset Demographic Bias: A Case Study on Facial Expression Recognition](https://mlsanity.com/paper/2303.15889/) (2023)
- [Images Speak Louder than Words: Understanding and Mitigating Bias in Vision-Language Model from a Causal Mediation Perspective](https://mlsanity.com/paper/2407.02814/) (2024)
- [Analyzing Quality, Bias, and Performance in Text-to-Image Generative Models](https://mlsanity.com/paper/2407.00138/) (2024)
- [Exploring Bias in over 100 Text-to-Image Generative Models](https://mlsanity.com/paper/2503.08012/) (2025)
- [Diverse Diffusion: Enhancing Image Diversity in Text-to-Image Generation](https://mlsanity.com/paper/2310.12583/) (2023)
- [Culture-TRIP: Culturally-Aware Text-to-Image Generation with Iterative Prompt Refinement](https://mlsanity.com/paper/2502.16902/) (2025)
- [FairImagen: Post-Processing for Bias Mitigation in Text-to-Image Models](https://mlsanity.com/paper/2510.21363/) (2025)
- [A Survey of Automatic Prompt Optimization with Instruction-focused Heuristic-based Search Algorithm](https://mlsanity.com/paper/2502.18746/) (2025)
- [BiasGuard: A Reasoning-enhanced Bias Detection Tool For Large Language Models](https://mlsanity.com/paper/2504.21299/) (2025)
- [DebiasPI: Inference-time Debiasing by Prompt Iteration of a Text-to-Image Generative Model](https://mlsanity.com/paper/2501.18642/) (2025)

## Abstract (arXiv)

Text-to-Image (TTI) systems are now everyday infrastructure for journalism, education, advertising, and public communication, and the demographic and cultural stereotypes they inherit from training data (rendering women, people of colour, older adults, and non-Western cultures as under-represented or caricatured) become a population-level harm at deployment scale. Existing mitigations either require costly retraining, infeasible for the closed-source backbones that dominate consumer products, or rely on fixed demographic templates that ignore cultural context. We present KG-FairDiff, a model-agnostic, inference-time framework that formalises fairness-aware prompt refinement as a constrained optimisation problem and operationalises it as a closed-loop pipeline: a knowledge graph of ~1,200 culture- and bias-related triples retrieves structured context, an LLM rewriter proposes refinements, and a validator accepts only prompts that reduce a divergence-based fairness loss while preserving semantic fidelity to the user's original intent. We prove a finite-termination bound for the refinement loop, contribute a mathematically consistent evaluation suite linking Bias-P/Bias-W to divergence from target distributions and ENS to KL divergence, and audit eight widely-deployed backbone generators. KG-FairDiff substantially reduces gender, race, age, and intersectional disparities while preserving prompt semantics, offering a practical, deployment-ready route to more equitable generative AI.