# Biased Binary Attribute Classifiers Ignore the Majority Classes

arXiv: https://arxiv.org/abs/2403.14435 · HTML: https://mlsanity.com/paper/2403.14435/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 32 references that are in the map.
- First posted on arXiv 21 Mar 2024.

## Builds on (most cited first, 21 of 21)

- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Deep Learning Face Attributes in the Wild](https://mlsanity.com/paper/1411.7766/) (2014)
- [Learning Deep Features for Discriminative Localization](https://mlsanity.com/paper/1512.04150/) (2015)
- [Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks](https://mlsanity.com/paper/1710.11063/) (2017)
- [Long-tail learning via logit adjustment](https://mlsanity.com/paper/2007.07314/) (2020)
- [Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networks](https://mlsanity.com/paper/1910.01279/) (2019)
- [Full-Gradient Representation for Neural Network Visualization](https://mlsanity.com/paper/1905.00780/) (2019)
- [Axiom-based Grad-CAM: Towards Accurate Visualization and Explanation of CNNs](https://mlsanity.com/paper/2008.02312/) (2020)
- [Use HiResCAM instead of Grad-CAM for faithful explanations of convolutional neural networks](https://mlsanity.com/paper/2011.08891/) (2020)
- [MOON: A Mixed Objective Optimization Network for the Recognition of Facial Attributes](https://mlsanity.com/paper/1603.07027/) (2016)
- [Escaping Saddle Points for Effective Generalization on Class-Imbalanced Data](https://mlsanity.com/paper/2212.13827/) (2022)
- [Deep Multi-task Multi-label CNN for Effective Facial Attribute Classification](https://mlsanity.com/paper/2002.03683/) (2020)
- [Improving Facial Attribute Prediction using Semantic Segmentation](https://mlsanity.com/paper/1704.08740/) (2017)
- [Consistency and Accuracy of CelebA Attribute Values](https://mlsanity.com/paper/2210.07356/) (2022)
- [A Survey of Stealth Malware: Attacks, Mitigation Measures, and Steps Toward Autonomous Open World Solutions](https://mlsanity.com/paper/1603.06028/) (2016)
- [Facial Attributes: Accuracy and Adversarial Robustness](https://mlsanity.com/paper/1801.02480/) (2018)
- [Human Gender Classification: A Review](https://mlsanity.com/paper/1507.05122/) (2015)
- [AFFACT - Alignment-Free Facial Attribute Classification Technique](https://mlsanity.com/paper/1611.06158/) (2016)
- [Multi-label Learning Based Deep Transfer Neural Network for Facial Attribute Classification](https://mlsanity.com/paper/1805.01282/) (2018)
- [Are Facial Attributes Adversarially Robust?](https://mlsanity.com/paper/1605.05411/) (2016)
- [Leveraging Mid-Level Deep Representations For Predicting Face Attributes in the Wild](https://mlsanity.com/paper/1602.01827/) (2016)

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

- [Refining Tuberculosis Detection in CXR Imaging: Addressing Bias in Deep Neural Networks via Interpretability](https://mlsanity.com/paper/2407.14064/) (2024)

## Abstract (arXiv)

To visualize the regions of interest that classifiers base their decisions on, different Class Activation Mapping (CAM) methods have been developed. However, all of these techniques target categorical classifiers only, though most real-world tasks are binary classification. In this paper, we extend gradient-based CAM techniques to work with binary classifiers and visualize the active regions for binary facial attribute classifiers. When training an unbalanced binary classifier on an imbalanced dataset, it is well-known that the majority class, i.e. the class with many training samples, is mostly predicted much better than minority class with few training instances. In our experiments on the CelebA dataset, we verify these results, when training an unbalanced classifier to extract 40 facial attributes simultaneously. One would expect that the biased classifier has learned to extract features mainly for the majority classes and that the proportional energy of the activations mainly reside in certain specific regions of the image where the attribute is located. However, we find very little regular activation for samples of majority classes, while the active regions for minority classes seem mostly reasonable and overlap with our expectations. These results suggest that biased classifiers mainly rely on bias activation for majority classes. When training a balanced classifier on the imbalanced data by employing attribute-specific class weights, majority and minority classes are classified similarly well and show expected activations for almost all attributes