Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

PATE-TripleGAN: Privacy-Preserving Image Synthesis with Gaussian Differential Privacy (2404.12730v1)

Published 19 Apr 2024 in cs.CV, cs.CR, and cs.LG

Abstract: Conditional Generative Adversarial Networks (CGANs) exhibit significant potential in supervised learning model training by virtue of their ability to generate realistic labeled images. However, numerous studies have indicated the privacy leakage risk in CGANs models. The solution DPCGAN, incorporating the differential privacy framework, faces challenges such as heavy reliance on labeled data for model training and potential disruptions to original gradient information due to excessive gradient clipping, making it difficult to ensure model accuracy. To address these challenges, we present a privacy-preserving training framework called PATE-TripleGAN. This framework incorporates a classifier to pre-classify unlabeled data, establishing a three-party min-max game to reduce dependence on labeled data. Furthermore, we present a hybrid gradient desensitization algorithm based on the Private Aggregation of Teacher Ensembles (PATE) framework and Differential Private Stochastic Gradient Descent (DPSGD) method. This algorithm allows the model to retain gradient information more effectively while ensuring privacy protection, thereby enhancing the model's utility. Privacy analysis and extensive experiments affirm that the PATE-TripleGAN model can generate a higher quality labeled image dataset while ensuring the privacy of the training data.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (33)
  1. Deep learning with differential privacy, in: Proceedings of the 2016 ACM SIGSAC conference on computer and communications security, pp. 308–318.
  2. Individualized pate: Differentially private machine learning with individual privacy guarantees. arXiv preprint arXiv:2202.10517 .
  3. Deep learning with gaussian differential privacy. Harvard data science review 2020, 10–1162.
  4. Gs-wgan: A gradient-sanitized approach for learning differentially private generators. Advances in Neural Information Processing Systems 33, 12673–12684.
  5. Gaussian differential privacy. Journal of the Royal Statistical Society Series B: Statistical Methodology 84, 3–37.
  6. Differential privacy, in: International colloquium on automata, languages, and programming, Springer. pp. 1–12.
  7. The algorithmic foundations of differential privacy. Foundations and Trends® in Theoretical Computer Science 9, 211–407.
  8. Model inversion attacks that exploit confidence information and basic countermeasures, in: Proceedings of the 22nd ACM SIGSAC conference on computer and communications security, pp. 1322–1333.
  9. Dp-sgd vs pate: Which has less disparate impact on gans? arXiv preprint arXiv:2111.13617 .
  10. Generative adversarial nets. Advances in neural information processing systems 27.
  11. Deep residual learning for image recognition, in: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 770–778.
  12. Crd-cgan: Category-consistent and relativistic constraints for diverse text-to-image generation. Frontiers of Computer Science 18, 181304.
  13. Densely connected convolutional networks, in: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 4700–4708.
  14. Pate-gan: Generating synthetic data with differential privacy guarantees, in: International conference on learning representations.
  15. The composition theorem for differential privacy, in: International conference on machine learning, PMLR. pp. 1376–1385.
  16. Learning multiple layers of features from tiny images .
  17. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems 25.
  18. The mnist database of handwritten digits. http://yann. lecun. com/exdb/mnist/ .
  19. Triple generative adversarial nets. Advances in neural information processing systems 30.
  20. Sccgan: style and characters inpainting based on cgan. Mobile networks and applications 26, 3–12.
  21. Privacy and security issues in deep learning: A survey. IEEE Access 9, 4566–4593.
  22. Machine learning for synthetic data generation: a review. arXiv preprint arXiv:2302.04062 .
  23. Ml-cgan: conditional generative adversarial network with a meta-learner structure for high-quality image generation with few training data. Cognitive Computation 13, 418–430.
  24. Rényi differential privacy, in: 2017 IEEE 30th computer security foundations symposium (CSF), IEEE. pp. 263–275.
  25. Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784 .
  26. Semi-supervised knowledge transfer for deep learning from private training data. arXiv preprint arXiv:1610.05755 .
  27. Scalable private learning with pate. arXiv preprint arXiv:1802.08908 .
  28. Style transfer in conditional gans for cross-modality synthesis of brain magnetic resonance images. Computers in Biology and Medicine 148, 105928.
  29. Membership inference attacks against machine learning models, in: 2017 IEEE symposium on security and privacy (SP), IEEE. pp. 3–18.
  30. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556 .
  31. Dp-cgan: Differentially private synthetic data and label generation, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pp. 0–0.
  32. Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms. arXiv preprint arXiv:1708.07747 .
  33. Differentially private generative adversarial network. arXiv preprint arXiv:1802.06739 .
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Zepeng Jiang (1 paper)
  2. Weiwei Ni (2 papers)
  3. Yifan Zhang (245 papers)

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now