# Compressed Models Decompress Race Biases: What Quantized Models Forget for Fair Face Recognition

arXiv: https://arxiv.org/abs/2308.11840 · HTML: https://mlsanity.com/paper/2308.11840/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 22 references that are in the map.
- First posted on arXiv 23 Aug 2023.

## Builds on (most cited first, 15 of 15)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Training Generative Adversarial Networks with Limited Data](https://mlsanity.com/paper/2006.06676/) (2020)
- [MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition](https://mlsanity.com/paper/1607.08221/) (2016)
- [Adversarial Domain Adaptation with Domain Mixup](https://mlsanity.com/paper/1912.01805/) (2019)
- [MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices](https://mlsanity.com/paper/1804.07573/) (2018)
- [What Do Compressed Deep Neural Networks Forget?](https://mlsanity.com/paper/1911.05248/) (2019)
- [Characterising Bias in Compressed Models](https://mlsanity.com/paper/2010.03058/) (2020)
- [ElasticFace: Elastic Margin Loss for Deep Face Recognition](https://mlsanity.com/paper/2109.09416/) (2021)
- [SynFace: Face Recognition with Synthetic Data](https://mlsanity.com/paper/2108.07960/) (2021)
- [Meta Balanced Network for Fair Face Recognition](https://mlsanity.com/paper/2205.06548/) (2022)
- [QuantFace: Towards Lightweight Face Recognition by Synthetic Data Low-bit Quantization](https://mlsanity.com/paper/2206.10526/) (2022)
- [StRDAN: Synthetic-to-Real Domain Adaptation Network for Vehicle Re-Identification](https://mlsanity.com/paper/2004.12032/) (2020)
- [The Effect of Model Compression on Fairness in Facial Expression Recognition](https://mlsanity.com/paper/2201.01709/) (2022)
- [FocusFace: Multi-task Contrastive Learning for Masked Face Recognition](https://mlsanity.com/paper/2110.14940/) (2021)
- [Causality-Inspired Taxonomy for Explainable Artificial Intelligence](https://mlsanity.com/paper/2208.09500/) (2022)

## Built on by (most cited first, 9 of 9)

- [Second FRCSyn-onGoing: Winning Solutions and Post-Challenge Analysis to Improve Face Recognition with Synthetic Data](https://mlsanity.com/paper/2412.01383/) (2024)
- [Model Compression Techniques in Biometrics Applications: A Survey](https://mlsanity.com/paper/2401.10139/) (2024)
- [Can the accuracy bias by facial hairstyle be reduced through balancing the training data?](https://mlsanity.com/paper/2405.20062/) (2024)
- [Fairness Under Cover: Evaluating the Impact of Occlusions on Demographic Bias in Facial Recognition](https://mlsanity.com/paper/2408.10175/) (2024)
- [Massively Annotated Datasets for Assessment of Synthetic and Real Data in Face Recognition](https://mlsanity.com/paper/2404.15234/) (2024)
- [MST-KD: Multiple Specialized Teachers Knowledge Distillation for Fair Face Recognition](https://mlsanity.com/paper/2408.16563/) (2024)
- [Fairer Analysis and Demographically Balanced Face Generation for Fairer Face Verification](https://mlsanity.com/paper/2412.03349/) (2024)
- [How Knowledge Distillation Mitigates the Synthetic Gap in Fair Face Recognition](https://mlsanity.com/paper/2408.17399/) (2024)
- [Balancing Beyond Discrete Categories: Continuous Demographic Labels for Fair Face Recognition](https://mlsanity.com/paper/2506.01532/) (2025)

## Abstract (arXiv)

With the ever-growing complexity of deep learning models for face recognition, it becomes hard to deploy these systems in real life. Researchers have two options: 1) use smaller models; 2) compress their current models. Since the usage of smaller models might lead to concerning biases, compression gains relevance. However, compressing might be also responsible for an increase in the bias of the final model. We investigate the overall performance, the performance on each ethnicity subgroup and the racial bias of a State-of-the-Art quantization approach when used with synthetic and real data. This analysis provides a few more details on potential benefits of performing quantization with synthetic data, for instance, the reduction of biases on the majority of test scenarios. We tested five distinct architectures and three different training datasets. The models were evaluated on a fourth dataset which was collected to infer and compare the performance of face recognition models on different ethnicity.