mlsanityData to 7 Oct 2026

SF-PATE: Scalable, Fair, and Private Aggregation of Teacher Ensembles

arXiv 2204.05157 ↗First posted 11 Apr 2022cs.LG cs.AIcited by 8

Not on the map. Why?

Abstract (arXiv)

A critical concern in data-driven processes is to build models whose outcomes do not discriminate against some demographic groups, including gender, ethnicity, or age. To ensure non-discrimination in learning tasks, knowledge of the group attributes is essential. However, in practice, these attributes may not be available due to legal and ethical requirements. To address this challenge, this paper studies a model that protects the privacy of the individuals' sensitive information while also allowing it to learn non-discriminatory predictors. A key characteristic of the proposed model is to enable the adoption of off-the-selves and non-private fair models to create a privacy-preserving and fair model. The paper analyzes the relation between accuracy, privacy, and fairness, and the experimental evaluation illustrates the benefits of the proposed models on several prediction tasks. In particular, this proposal is the first to allow both scalable and accurate training of private and fair models for very large neural networks.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 23 of its 42 references that are in the map. First posted on arXiv 11 Apr 2022.

Builds on

23 of 42 references.

References · 23

  1. 2016Deep Learning with Differential Privacycited by 2,682
  2. 2016Equality of Opportunity in Supervised Learningcited by 2,000
  3. 2017Renyi Differential Privacycited by 643
  4. 2018A Reductions Approach to Fair Classificationcited by 585
  5. 2019On the Efficacy of Knowledge Distillationcited by 377
  6. 2021Opacus: User-Friendly Differential Privacy Library in PyTorchcited by 375
  7. 2016Fairness Beyond Disparate Treatment & Disparate Impact: Learning Classification without Disparate Mistreatmentcited by 368
  8. 2016Semi-supervised Knowledge Transfer for Deep Learning from Private Training Datacited by 347
  9. 2019Differential Privacy Has Disparate Impact on Model Accuracycited by 316
  10. 2018Scalable Private Learning with PATEcited by 266
  11. 2018Differentially Private Fair Learningcited by 76
  12. 2020Lagrangian Duality for Constrained Deep Learningcited by 72
11 more
  1. 2021Fairness via Representation Neutralizationcited by 63
  2. 2019Fair Decision Making using Privacy-Protected Datacited by 58
  3. 2020Enabling Fast Differentially Private SGD via Just-in-Time Compilation and Vectorizationcited by 58
  4. 2020Differentially Private and Fair Deep Learning: A Lagrangian Dual Approachcited by 53
  5. 2022Differential Privacy and Fairness in Decisions and Learning Tasks: A Surveycited by 47
  6. 2020Fair Learning with Private Demographic Datacited by 46
  7. 2021Decision Making with Differential Privacy under a Fairness Lenscited by 32
  8. 2021DP-SGD vs PATE: Which Has Less Disparate Impact on Model Accuracy?cited by 32
  9. 2020FairFaceGAN: Fairness-aware Facial Image-to-Image Translationcited by 21
  10. 2022Learning fair representation with a parametric integral probability metriccited by 20
  11. 2021A Fairness Analysis on Private Aggregation of Teacher Ensemblescited by 11

Built on by

Cited by year

Citing papers per year: 2022: 1; 2023: 2; 2024: 4; 2025: 1.

8 papers.

  1. 2022Stochastic Differentially Private and Fair Learningcited by 16
  2. 2023Survey on AI Ethics: A Socio-technical Perspectivecited by 4
  3. 2023FairDP: Certified Fairness with Differential Privacycited by 3
  4. 2024Differentially Private Fair Binary Classificationscited by 3
  5. 2024PFGuard: A Generative Framework with Privacy and Fairness Safeguardscited by 3
  6. 2024Differentially Private Data Release on Graphs: Inefficiencies and Unfairnesscited by 0
  7. 2024A Stochastic Optimization Framework for Private and Fair Learning From Decentralized Datacited by 0
  8. 2025Private Rate-Constrained Optimization with Applications to Fair Learningcited by 0