# A Note on Data Biases in Generative Models

arXiv: https://arxiv.org/abs/2012.02516 · HTML: https://mlsanity.com/paper/2012.02516/

## Key facts

- Cited by 13 papers in the map (data through 13 Oct 2026).
- First posted on arXiv 4 Dec 2020.

## Built on by (most cited first, 13 of 13)

- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [VBench: Comprehensive Benchmark Suite for Video Generative Models](https://mlsanity.com/paper/2311.17982/) (2023)
- [Vector-quantized Image Modeling with Improved VQGAN](https://mlsanity.com/paper/2110.04627/) (2021)
- [DiffusionCLIP: Text-Guided Diffusion Models for Robust Image Manipulation](https://mlsanity.com/paper/2110.02711/) (2021)
- [Score-Based Generative Models Detect Manifolds](https://mlsanity.com/paper/2206.01018/) (2022)
- [Collaborative Diffusion for Multi-Modal Face Generation and Editing](https://mlsanity.com/paper/2304.10530/) (2023)
- [ReVersion: Diffusion-Based Relation Inversion from Images](https://mlsanity.com/paper/2303.13495/) (2023)
- [Semi-Parametric Neural Image Synthesis](https://mlsanity.com/paper/2204.11824/) (2022)
- [Toward Spatially Unbiased Generative Models](https://mlsanity.com/paper/2108.01285/) (2021)
- [DDMI: Domain-Agnostic Latent Diffusion Models for Synthesizing High-Quality Implicit Neural Representations](https://mlsanity.com/paper/2401.12517/) (2024)
- [Robust 3D Shape Reconstruction in Zero-Shot from a Single Image in the Wild](https://mlsanity.com/paper/2403.14539/) (2024)
- [M-VADER: A Model for Diffusion with Multimodal Context](https://mlsanity.com/paper/2212.02936/) (2022)
- [InsertDiffusion: Identity Preserving Visualization of Objects through a Training-Free Diffusion Architecture](https://mlsanity.com/paper/2407.10592/) (2024)

## Abstract (arXiv)

It is tempting to think that machines are less prone to unfairness and prejudice. However, machine learning approaches compute their outputs based on data. While biases can enter at any stage of the development pipeline, models are particularly receptive to mirror biases of the datasets they are trained on and therefore do not necessarily reflect truths about the world but, primarily, truths about the data. To raise awareness about the relationship between modern algorithms and the data that shape them, we use a conditional invertible neural network to disentangle the dataset-specific information from the information which is shared across different datasets. In this way, we can project the same image onto different datasets, thereby revealing their inherent biases. We use this methodology to (i) investigate the impact of dataset quality on the performance of generative models, (ii) show how societal biases of datasets are replicated by generative models, and (iii) present creative applications through unpaired transfer between diverse datasets such as photographs, oil portraits, and animes. Our code and an interactive demonstration are available at https://github.com/CompVis/net2net .