# Discriminative Adversarial Privacy: Balancing Accuracy and Membership Privacy in Neural Networks

arXiv: https://arxiv.org/abs/2306.03054 · HTML: https://mlsanity.com/paper/2306.03054/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 34 references that are in the map.
- First posted on arXiv 5 Jun 2023.

## Builds on (most cited first, 20 of 20)

- [Generative Adversarial Networks](https://mlsanity.com/paper/2203.00667/) (2022)
- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Deep Learning with Differential Privacy](https://mlsanity.com/paper/1607.00133/) (2016)
- [Membership Inference Attacks against Machine Learning Models](https://mlsanity.com/paper/1610.05820/) (2016)
- [EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification](https://mlsanity.com/paper/1709.00029/) (2017)
- [Membership Inference Attacks on Machine Learning: A Survey](https://mlsanity.com/paper/2103.07853/) (2021)
- [Differential Privacy Has Disparate Impact on Model Accuracy](https://mlsanity.com/paper/1905.12101/) (2019)
- [Are Diffusion Models Vulnerable to Membership Inference Attacks?](https://mlsanity.com/paper/2302.01316/) (2023)
- [Membership Inference Attacks Against Text-to-image Generation Models](https://mlsanity.com/paper/2210.00968/) (2022)
- [Membership Inference of Diffusion Models](https://mlsanity.com/paper/2301.09956/) (2023)
- [RelaxLoss: Defending Membership Inference Attacks without Losing Utility](https://mlsanity.com/paper/2207.05801/) (2022)
- [On Privacy and Personalization in Cross-Silo Federated Learning](https://mlsanity.com/paper/2206.07902/) (2022)
- [Amplifying Membership Exposure via Data Poisoning](https://mlsanity.com/paper/2211.00463/) (2022)
- [This Person (Probably) Exists. Identity Membership Attacks Against GAN Generated Faces](https://mlsanity.com/paper/2107.06018/) (2021)
- [Interaction-level Membership Inference Attack Against Federated Recommender Systems](https://mlsanity.com/paper/2301.10964/) (2023)
- [SGDE: Secure Generative Data Exchange for Cross-Silo Federated Learning](https://mlsanity.com/paper/2109.12062/) (2021)
- [Debiasing Learning for Membership Inference Attacks Against Recommender Systems](https://mlsanity.com/paper/2206.12401/) (2022)
- [Semi-Leak: Membership Inference Attacks Against Semi-supervised Learning](https://mlsanity.com/paper/2207.12535/) (2022)
- [On the utility and protection of optimization with differential privacy and classic regularization techniques](https://mlsanity.com/paper/2209.03175/) (2022)
- [Membership Inference Attacks Against Semantic Segmentation Models](https://mlsanity.com/paper/2212.01082/) (2022)

## Built on by (most cited first, 4 of 4)

- [Federated Knowledge Recycling: Privacy-Preserving Synthetic Data Sharing](https://mlsanity.com/paper/2407.20830/) (2024)
- [Synthetic Image Learning: Preserving Performance and Preventing Membership Inference Attacks](https://mlsanity.com/paper/2407.15526/) (2024)
- [Your Image Generator Is Your New Private Dataset](https://mlsanity.com/paper/2504.04582/) (2025)
- [Harnessing the Computing Continuum across Personalized Healthcare, Maintenance and Inspection, and Farming 4.0](https://mlsanity.com/paper/2403.14650/) (2024)

## Abstract (arXiv)

The remarkable proliferation of deep learning across various industries has underscored the importance of data privacy and security in AI pipelines. As the evolution of sophisticated Membership Inference Attacks (MIAs) threatens the secrecy of individual-specific information used for training deep learning models, Differential Privacy (DP) raises as one of the most utilized techniques to protect models against malicious attacks. However, despite its proven theoretical properties, DP can significantly hamper model performance and increase training time, turning its use impractical in real-world scenarios. Tackling this issue, we present Discriminative Adversarial Privacy (DAP), a novel learning technique designed to address the limitations of DP by achieving a balance between model performance, speed, and privacy. DAP relies on adversarial training based on a novel loss function able to minimise the prediction error while maximising the MIA's error. In addition, we introduce a novel metric named Accuracy Over Privacy (AOP) to capture the performance-privacy trade-off. Finally, to validate our claims, we compare DAP with diverse DP scenarios, providing an analysis of the results from performance, time, and privacy preservation perspectives.