Fairness meets Cross-Domain Learning: a new perspective on Models and Metrics
Not on the map. Why?
Abstract (arXiv)
Deep learning-based recognition systems are deployed at scale for several real-world applications that inevitably involve our social life. Although being of great support when making complex decisions, they might capture spurious data correlations and leverage sensitive attributes (e.g. age, gender, ethnicity). How to factor out this information while keeping a high prediction performance is a task with still several open questions, many of which are shared with those of the domain adaptation and generalization literature which focuses on avoiding visual domain biases. In this work, we propose an in-depth study of the relationship between cross-domain learning (CD) and model fairness by introducing a benchmark on face and medical images spanning several demographic groups as well as classification and localization tasks. After having highlighted the limits of the current evaluation metrics, we introduce a new Harmonic Fairness (HF) score to assess jointly how fair and accurate every model is with respect to a reference baseline. Our study covers 14 CD approaches alongside three state-of-the-art fairness algorithms and shows how the former can outperform the latter. Overall, our work paves the way for a more systematic analysis of fairness problems in computer vision. Code available at: https://github.com/iurada/fairness_crossdomain
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 59 of its 92 references that are in the map. First posted on arXiv 25 Mar 2023.
Builds on
59 of 92 references.
References · 59
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2015Rethinking the Inception Architecture for Computer Visioncited by 5,475
- 2016Image-to-Image Translation with Conditional Adversarial Networkscited by 4,527
- 2014Deep Learning Face Attributes in the Wildcited by 3,693
- 2015Domain-Adversarial Training of Neural Networkscited by 3,044
- 2016Equality of Opportunity in Supervised Learningcited by 2,007
- 2020Shortcut Learning in Deep Neural Networkscited by 1,969
- 2015Learning Transferable Features with Deep Adaptation Networkscited by 1,329
- 2019Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalizationcited by 1,247
- 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,209
- 2017Learning to Generalize: Meta-Learning for Domain Generalizationcited by 763
- 2017Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networkscited by 707
38 more
- 2018Simple Baselines for Human Pose Estimation and Trackingcited by 502
- 2022ViTPose: Simple Vision Transformer Baselines for Human Pose Estimationcited by 456
- 2019Domain Generalization by Solving Jigsaw Puzzlescited by 450
- 2021SWAD: Domain Generalization by Seeking Flat Minimacited by 434
- 2017Age Progression/Regression by Conditional Adversarial Autoencodercited by 429
- 2020Self-Challenging Improves Cross-Domain Generalizationcited by 426
- 2019Semi-supervised Domain Adaptation via Minimax Entropycited by 335
- 2017Fairness in Criminal Justice Risk Assessments: The State of the Artcited by 315
- 2021Gradient Matching for Domain Generalizationcited by 258
- 2019Towards Fairness in Visual Recognition: Effective Strategies for Bias Mitigationcited by 252
- 2019Adversarial Domain Adaptation with Domain Mixupcited by 250
- 2019Episodic Training for Domain Generalizationcited by 245
- 2018Multiaccuracy: Black-Box Post-Processing for Fairness in Classificationcited by 222
- 2021SelfReg: Self-supervised Contrastive Regularization for Domain Generalizationcited by 189
- 2021Evaluating Deep Neural Networks Trained on Clinical Images in Dermatology with the Fitzpatrick 17k Datasetcited by 185
- 2021Learning to Diversify for Single Domain Generalizationcited by 170
- 2020Minimax Pareto Fairness: A Multi Objective Perspectivecited by 153
- 2019Mitigate Bias in Face Recognition using Skewness-Aware Reinforcement Learningcited by 145
- 2018Look at Boundary: A Boundary-Aware Face Alignment Algorithmcited by 138
- 2020Fair Attribute Classification through Latent Space De-biasingcited by 125
- 2021Learning Debiased Representation via Disentangled Feature Augmentationcited by 116
- 2017Wing Loss for Robust Facial Landmark Localisation with Convolutional Neural Networkscited by 109
- 2021EnD: Entangling and Disentangling deep representations for bias correctioncited by 105
- 2020COVID-19 Image Data Collection: Prospective Predictions Are the Futurecited by 100
- 2021HyperInverter: Improving StyleGAN Inversion via Hypernetworkcited by 85
- 2020Fairness by Learning Orthogonal Disentangled Representationscited by 76
- 2020On the Effectiveness of Image Rotation for Open Set Domain Adaptationcited by 72
- 2022Fair Contrastive Learning for Facial Attribute Classificationcited by 71
- 2020Attract, Perturb, and Explore: Learning a Feature Alignment Network for Semi-supervised Domain Adaptationcited by 69
- 2020Learning Invariant Representations and Risks for Semi-supervised Domain Adaptationcited by 69
- 2022MEDFAIR: Benchmarking Fairness for Medical Imagingcited by 68
- 2021Fair Feature Distillation for Visual Recognitioncited by 59
- 2022Leveling Down in Computer Vision: Pareto Inefficiencies in Fair Deep Classifierscited by 53
- 2019Discrimination in the Age of Algorithmscited by 51
- 2022Fairness-aware Adversarial Perturbation Towards Bias Mitigation for Deployed Deep Modelscited by 47
- 2020Self-Supervised Learning Across Domainscited by 44
- 2020Deep Co-Training with Task Decomposition for Semi-Supervised Domain Adaptationcited by 41
- 2021PASS: Protected Attribute Suppression System for Mitigating Bias in Face Recognitioncited by 38
50 most cited of 59 All 59 (JSON)
Built on by
Not cited yet.