# Feature and Label Embedding Spaces Matter in Addressing Image Classifier Bias

arXiv: https://arxiv.org/abs/2110.14336 · HTML: https://mlsanity.com/paper/2110.14336/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 38 of its 56 references that are in the map.
- First posted on arXiv 27 Oct 2021.

## Builds on (most cited first, 30 of 38)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [Deep Learning Face Attributes in the Wild](https://mlsanity.com/paper/1411.7766/) (2014)
- [Domain-Adversarial Training of Neural Networks](https://mlsanity.com/paper/1505.07818/) (2015)
- [Deep contextualized word representations](https://mlsanity.com/paper/1802.05365/) (2018)
- [Equality of Opportunity in Supervised Learning](https://mlsanity.com/paper/1610.02413/) (2016)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Datasheets for Datasets](https://mlsanity.com/paper/1803.09010/) (2018)
- [Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings](https://mlsanity.com/paper/1607.06520/) (2016)
- [Semantics derived automatically from language corpora contain human-like biases](https://mlsanity.com/paper/1608.07187/) (2016)
- [InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets](https://mlsanity.com/paper/1606.03657/) (2016)
- [Maximum Classifier Discrepancy for Unsupervised Domain Adaptation](https://mlsanity.com/paper/1712.02560/) (2017)
- [Mitigating Unwanted Biases with Adversarial Learning](https://mlsanity.com/paper/1801.07593/) (2018)
- [MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition](https://mlsanity.com/paper/1607.08221/) (2016)
- [Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints](https://mlsanity.com/paper/1707.09457/) (2017)
- [AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias](https://mlsanity.com/paper/1810.01943/) (2018)
- [Taking A Closer Look at Domain Shift: Category-level Adversaries for Semantics Consistent Domain Adaptation](https://mlsanity.com/paper/1809.09478/) (2018)
- [Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them](https://mlsanity.com/paper/1903.03862/) (2019)
- [Learning Not to Learn: Training Deep Neural Networks with Biased Data](https://mlsanity.com/paper/1812.10352/) (2018)
- [Word Embeddings Quantify 100 Years of Gender and Ethnic Stereotypes](https://mlsanity.com/paper/1711.08412/) (2017)
- [Towards Fairness in Visual Recognition: Effective Strategies for Bias Mitigation](https://mlsanity.com/paper/1911.11834/) (2019)
- [Large image datasets: A pyrrhic win for computer vision?](https://mlsanity.com/paper/2006.16923/) (2020)
- [Women also Snowboard: Overcoming Bias in Captioning Models](https://mlsanity.com/paper/1803.09797/) (2018)
- [Simultaneous Deep Transfer Across Domains and Tasks](https://mlsanity.com/paper/1510.02192/) (2015)
- [Gender Bias in Contextualized Word Embeddings](https://mlsanity.com/paper/1904.03310/) (2019)
- [Learning Gender-Neutral Word Embeddings](https://mlsanity.com/paper/1809.01496/) (2018)
- [Data Decisions and Theoretical Implications when Adversarially Learning Fair Representations](https://mlsanity.com/paper/1707.00075/) (2017)
- [Towards Fairer Datasets: Filtering and Balancing the Distribution of the People Subtree in the ImageNet Hierarchy](https://mlsanity.com/paper/1912.07726/) (2019)

## Built on by (most cited first, 11 of 11)

- [Gender and Racial Bias in Visual Question Answering Datasets](https://mlsanity.com/paper/2205.08148/) (2022)
- [Quantifying Societal Bias Amplification in Image Captioning](https://mlsanity.com/paper/2203.15395/) (2022)
- [Model-Agnostic Gender Debiased Image Captioning](https://mlsanity.com/paper/2304.03693/) (2023)
- [Beyond Skin Tone: A Multidimensional Measure of Apparent Skin Color](https://mlsanity.com/paper/2309.05148/) (2023)
- [Epistemic Uncertainty-Weighted Loss for Visual Bias Mitigation](https://mlsanity.com/paper/2204.09389/) (2022)
- [ImageCaptioner2: Image Captioner for Image Captioning Bias Amplification Assessment](https://mlsanity.com/paper/2304.04874/) (2023)
- [Beard Segmentation and Recognition Bias](https://mlsanity.com/paper/2308.15740/) (2023)
- [Information Removal at the bottleneck in Deep Neural Networks](https://mlsanity.com/paper/2210.00891/) (2022)
- [Variation of Gender Biases in Visual Recognition Models Before and After Finetuning](https://mlsanity.com/paper/2303.07615/) (2023)
- [Evaluating Fairness and Mitigating Bias in Machine Learning: A Novel Technique using Tensor Data and Bayesian Regression](https://mlsanity.com/paper/2506.11627/) (2025)
- [Implicit Visual Bias Mitigation by Posterior Estimate Sharpening of a Bayesian Neural Network](https://mlsanity.com/paper/2303.16564/) (2023)

## Abstract (arXiv)

This paper strives to address image classifier bias, with a focus on both feature and label embedding spaces. Previous works have shown that spurious correlations from protected attributes, such as age, gender, or skin tone, can cause adverse decisions. To balance potential harms, there is a growing need to identify and mitigate image classifier bias. First, we identify in the feature space a bias direction. We compute class prototypes of each protected attribute value for every class, and reveal an existing subspace that captures the maximum variance of the bias. Second, we mitigate biases by mapping image inputs to label embedding spaces. Each value of the protected attribute has its projection head where classes are embedded through a latent vector representation rather than a common one-hot encoding. Once trained, we further reduce in the feature space the bias effect by removing its direction. Evaluation on biased image datasets, for multi-class, multi-label and binary classifications, shows the effectiveness of tackling both feature and label embedding spaces in improving the fairness of the classifier predictions, while preserving classification performance.