How I Met Your Bias: Investigating Bias Amplification in Diffusion Models

TL;DR

This paper investigates how sampling algorithms and hyperparameters in diffusion models influence the amplification or reduction of dataset biases during image generation, revealing controllable bias effects.

Contribution

It is the first to analyze the impact of sampling hyperparameters on bias amplification in diffusion models, demonstrating their significant influence.

Findings

Sampling hyperparameters can both reduce and amplify bias.

Bias effects are controllable through sampling choices.

Bias amplification varies across datasets and models.

Abstract

Diffusion-based generative models demonstrate state-of-the-art performance across various image synthesis tasks, yet their tendency to replicate and amplify dataset biases remains poorly understood. Although previous research has viewed bias amplification as an inherent characteristic of diffusion models, this work provides the first analysis of how sampling algorithms and their hyperparameters influence bias amplification. We empirically demonstrate that samplers for diffusion models -- commonly optimized for sample quality and speed -- have a significant and measurable effect on bias amplification. Through controlled studies with models trained on Biased MNIST, Multi-Color MNIST and BFFHQ, and with Stable Diffusion, we show that sampling hyperparameters can induce both bias reduction and amplification, even when the trained model is fixed. Source code is available at https://github.com/How-I-met-your-bias/how_i_met_your_bias.