# Inference Time Debiasing Concepts in Diffusion Models

arXiv: https://arxiv.org/abs/2508.14933 · HTML: https://mlsanity.com/paper/2508.14933/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models.
- Builds on 18 of its 25 references that are in the map.
- First posted on arXiv 19 Aug 2025.

## Placed in

- Diffusion models / [Image generation, text-to-image and editing](https://mlsanity.com/diffusion/image-generation-text-to-image-and-editing/)

## Builds on (most cited first, 18 of 18)

- [Classifier-Free Diffusion Guidance](https://mlsanity.com/paper/2207.12598/) (2022)
- [SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis](https://mlsanity.com/paper/2307.01952/) (2023)
- [Deep Learning Face Attributes in the Wild](https://mlsanity.com/paper/1411.7766/) (2014)
- [CLIPScore: A Reference-free Evaluation Metric for Image Captioning](https://mlsanity.com/paper/2104.08718/) (2021)
- [A Survey on Bias and Fairness in Machine Learning](https://mlsanity.com/paper/1908.09635/) (2019)
- [The Woman Worked as a Babysitter: On Biases in Language Generation](https://mlsanity.com/paper/1909.01326/) (2019)
- [Tackling the Generative Learning Trilemma with Denoising Diffusion GANs](https://mlsanity.com/paper/2112.07804/) (2021)
- [Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Models](https://mlsanity.com/paper/2211.05105/) (2022)
- [DALL-Eval: Probing the Reasoning Skills and Social Biases of Text-to-Image Generation Models](https://mlsanity.com/paper/2202.04053/) (2022)
- [Mitigating Bias in Algorithmic Hiring: Evaluating Claims and Practices](https://mlsanity.com/paper/1906.09208/) (2019)
- [Fair Generative Modeling via Weak Supervision](https://mlsanity.com/paper/1910.12008/) (2019)
- [The Bias Amplification Paradox in Text-to-Image Generation](https://mlsanity.com/paper/2308.00755/) (2023)
- [T2ISafety: Benchmark for Assessing Fairness, Toxicity, and Privacy in Image Generation](https://mlsanity.com/paper/2501.12612/) (2025)
- [The Dark Side of Dataset Scaling: Evaluating Racial Classification in Multimodal Models](https://mlsanity.com/paper/2405.04623/) (2024)
- [Debiasing Methods for Fairer Neural Models in Vision and Language Research: A Survey](https://mlsanity.com/paper/2211.05617/) (2022)
- [The Compute Divide in Machine Learning: A Threat to Academic Contribution and Scrutiny?](https://mlsanity.com/paper/2401.02452/) (2024)
- [Fairness and Bias in Multimodal AI: A Survey](https://mlsanity.com/paper/2406.19097/) (2024)
- [MIST: Mitigating Intersectional Bias with Disentangled Cross-Attention Editing in Text-to-Image Diffusion Models](https://mlsanity.com/paper/2403.19738/) (2024)

## Built on by (most cited first, 1 of 1)

- [A Framework for Benchmarking Fairness-Utility Trade-offs in Text-to-Image Models via Pareto Frontiers](https://mlsanity.com/paper/2508.16752/) (2025)

## Abstract (arXiv)

We propose DeCoDi, a debiasing procedure for text-to-image diffusion-based models that changes the inference procedure, does not significantly change image quality, has negligible compute overhead, and can be applied in any diffusion-based image generation model. DeCoDi changes the diffusion process to avoid latent dimension regions of biased concepts. While most deep learning debiasing methods require complex or compute-intensive interventions, our method is designed to change only the inference procedure. Therefore, it is more accessible to a wide range of practitioners. We show the effectiveness of the method by debiasing for gender, ethnicity, and age for the concepts of nurse, firefighter, and CEO. Two distinct human evaluators manually inspect 1,200 generated images. Their evaluation results provide evidence that our method is effective in mitigating biases based on gender, ethnicity, and age. We also show that an automatic bias evaluation performed by the GPT4o is not significantly statistically distinct from a human evaluation. Our evaluation shows promising results, with reliable levels of agreement between evaluators and more coverage of protected attributes. Our method has the potential to significantly improve the diversity of images it generates by diffusion-based text-to-image generative models.