# Contrastive Examples for Addressing the Tyranny of the Majority

arXiv: https://arxiv.org/abs/2004.06524 · HTML: https://mlsanity.com/paper/2004.06524/

## Key facts

- Cited by 21 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Apr 2020.

## Built on by (most cited first, 21 of 21)

- [Diversify Your Vision Datasets with Automatic Diffusion-Based Augmentation](https://mlsanity.com/paper/2305.16289/) (2023)
- [Contrastive Explanations for Model Interpretability](https://mlsanity.com/paper/2103.01378/) (2021)
- [Towards Intersectionality in Machine Learning: Including More Identities, Handling Underrepresentation, and Performing Evaluation](https://mlsanity.com/paper/2205.04610/) (2022)
- [Gender Artifacts in Visual Datasets](https://mlsanity.com/paper/2206.09191/) (2022)
- [Using Language to Extend to Unseen Domains](https://mlsanity.com/paper/2210.09520/) (2022)
- [Men Also Do Laundry: Multi-Attribute Bias Amplification](https://mlsanity.com/paper/2210.11924/) (2022)
- [LARGE: Latent-Based Regression through GAN Semantics](https://mlsanity.com/paper/2107.11186/) (2021)
- [From Fake to Real: Pretraining on Balanced Synthetic Images to Prevent Spurious Correlations in Image Recognition](https://mlsanity.com/paper/2308.04553/) (2023)
- [Fair Bilevel Neural Network (FairBiNN): On Balancing fairness and accuracy via Stackelberg Equilibrium](https://mlsanity.com/paper/2410.16432/) (2024)
- [A Multidimensional Analysis of Social Biases in Vision Transformers](https://mlsanity.com/paper/2308.01948/) (2023)
- [GeNIe: Generative Hard Negative Images Through Diffusion](https://mlsanity.com/paper/2312.02548/) (2023)
- [InvGAN: Invertible GANs](https://mlsanity.com/paper/2112.04598/) (2021)
- [Resampled Datasets Are Not Enough: Mitigating Societal Bias Beyond Single Attributes](https://mlsanity.com/paper/2407.03623/) (2024)
- [CAT: Controllable Attribute Translation for Fair Facial Attribute Classification](https://mlsanity.com/paper/2209.06850/) (2022)
- [Through a fair looking-glass: mitigating bias in image datasets](https://mlsanity.com/paper/2209.08648/) (2022)
- [Relevant Irrelevance: Generating Alterfactual Explanations for Image Classifiers](https://mlsanity.com/paper/2405.05295/) (2024)
- [Distraction is All You Need for Fairness](https://mlsanity.com/paper/2203.07593/) (2022)
- [RealPatch: A Statistical Matching Framework for Model Patching with Real Samples](https://mlsanity.com/paper/2208.02192/) (2022)
- [Harnessing Diffusion-Generated Synthetic Images for Fair Image Classification](https://mlsanity.com/paper/2511.08711/) (2025)
- [Targeted Image Data Augmentation Increases Basic Skills Captioning Robustness](https://mlsanity.com/paper/2309.15991/) (2023)
- [SimpleStrat: Diversifying Language Model Generation with Stratification](https://mlsanity.com/paper/2410.09038/) (2024)

## Abstract (arXiv)

Computer vision algorithms, e.g. for face recognition, favour groups of individuals that are better represented in the training data. This happens because of the generalization that classifiers have to make. It is simpler to fit the majority groups as this fit is more important to overall error. We propose to create a balanced training dataset, consisting of the original dataset plus new data points in which the group memberships are intervened, minorities become majorities and vice versa. We show that current generative adversarial networks are a powerful tool for learning these data points, called contrastive examples. We experiment with the equalized odds bias measure on tabular data as well as image data (CelebA and Diversity in Faces datasets). Contrastive examples allow us to expose correlations between group membership and other seemingly neutral features. Whenever a causal graph is available, we can put those contrastive examples in the perspective of counterfactuals.