mlsanityData to 14 Oct 2026

Impact of Dataset Properties on Membership Inference Vulnerability of Deep Transfer Learning

arXiv 2402.06674 ↗First posted 7 Feb 2024cs.CR cs.LGcited by 7

On the map: LLM safety and security → Membership inference attacks

Abstract (arXiv)

Membership inference attacks (MIAs) are used to test practical privacy of machine learning models. MIAs complement formal guarantees from differential privacy (DP) under a more realistic adversary model. We analyse MIA vulnerability of fine-tuned neural networks both empirically and theoretically, the latter using a simplified model of fine-tuning. We show that the vulnerability of non-DP models when measured as the attacker advantage at a fixed false positive rate reduces according to a simple power law as the number of examples per class increases. A similar power-law applies even for the most vulnerable points, but the dataset size needed for adequate protection of the most vulnerable points is very large.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Placed in Membership inference attacks, one of 12 directions in LLM safety and security. Builds on 28 of its 47 references that are in the map. First posted on arXiv 7 Feb 2024.

Builds on

28 of 47 references.

Closest ancestors · 3

Same direction, newest first

  1. 2023Low-Cost High-Power Membership Inference Attacks6 Dec
  2. 2021Membership Inference Attacks From First Principles7 Dec
  3. 2021Membership Inference Attacks on Machine Learning: A Survey14 Mar

Other references · 25

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  2. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  3. 2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,501
  4. 2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,333
  5. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,281
  6. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,496
  7. 2017Remote Sensing Image Scene Classification: Benchmark and State of the Artcited by 797
  8. 2019Big Transfer (BiT): General Visual Representation Learningcited by 769
  9. 2020What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimationcited by 497
  10. 2023Scalable Extraction of Training Data from (Production) Language Modelscited by 434
  11. 2019A Large-scale Study of Representation Learning with the Visual Task Adaptation Benchmarkcited by 425
  12. 2018Rotation Equivariant CNNs for Digital Pathologycited by 422
13 more
  1. 2019Differential Privacy Has Disparate Impact on Model Accuracycited by 316
  2. 2013The Composition Theorem for Differential Privacycited by 297
  3. 2020Systematic Evaluation of Privacy Risks of Machine Learning Modelscited by 292
  4. 2022Reconstructing Training Data with Informed Adversariescited by 148
  5. 2020On the Privacy Risks of Algorithmic Fairnesscited by 82
  6. 2020Chasing Your Long Tails: Differentially Private Prediction in Health Care Settingscited by 37
  7. 2019Disparate Vulnerability to Membership Inference Attackscited by 35
  8. 2023Achilles' Heels: Vulnerable Record Identification in Synthetic Data Publishingcited by 28
  9. 2022FiT: Parameter Efficient Few-shot Transfer Learning for Personalized and Federated Image Classificationcited by 26
  10. 2022Individual Privacy Accounting for Differentially Private Stochastic Gradient Descentcited by 20
  11. 2023On the Efficacy of Differentially Private Few-shot Image Classificationcited by 12
  12. 2020Data and Model Dependencies of Membership Inference Attackcited by 11
  13. 2023Privacy and Accuracy Implications of Model Complexity and Integration in Heterogeneous Federated Learningcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 4; 2026: 3.

7 papers.

  1. 2025Membership Inference Attacks on Large-Scale Models: A Surveycited by 18
  2. 2025Estimating Model-Level Membership Inference Vulnerability Without Reference Modelscited by 1
  3. 2026On Reliability of Membership Inference Vulnerability Evaluationcited by 1
  4. 2025Mitigating Membership Inference Vulnerability in Personalized Federated Learningcited by 0
  5. 2025Empirical Comparison of Membership Inference Attacks in Deep Transfer Learningcited by 0
  6. 2026Quantifying the Privacy of Counterfactuals by Leveraging Membership Inference Attacks Against Synthetic Datacited by 0
  7. 2026SoK: Rethinking Membership Inference in Tabular Machine Learningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2021On the Privacy Risks of Deploying Recurrent Neural Networks in Machine Learning Models
  2. 2025Membership Inference Attacks Beyond Overfitting
  3. 2026A Critical Review on the Effectiveness and Privacy Threats of Membership Inference Attacks
  4. 2023Fundamental Limits of Membership Inference Attacks on Machine Learning Models
  5. 2025Hyperparameters in Score-Based Membership Inference Attacks
  6. 2025Membership Inference Attacks as Privacy Tools: Reliability, Disparity and Ensemble
  7. 2022On the Vulnerability of Data Points under Multiple Membership Inference Attacks and Target Models
  8. 2025Exposing and Defending Membership Leakage in Vulnerability Prediction Models