Deep Learning with Differential Privacy
Martı́n Abadi, Andy Chu, Ian Goodfellow et al. · ACM SIGSAC Conference on Computer and Communications Security (CCS) · 2016
Machine learning techniques based on neural networks are achieving remarkable results in a wide variety of domains. Often, the training of models requires large, representative datasets, which may be crowdsourced and contain sensitive information. The models should not expose private information in these datasets. Addressing this goal, we develop new algorithmic techniques for learning and a refined analysis of privacy costs within the framework of differential privacy. Our implementation and experiments demonstrate that we can train deep neural networks with non-convex objectives, under a modest privacy budget, and at a manageable cost in software complexity, training efficiency, and model quality.
Cited by 6,180 works. doi:10.1145/2976749.2978318
Authors
References
- ImageNet classification with deep convolutional neural networks
- Gradient-based learning applied to document recognition
- Going deeper with convolutions
- Learning representations by back-propagating errors
- Learning Multiple Layers of Features from Tiny Images
- Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
- Efficient Estimation of Word Representations in Vector Space
- Glove: Global Vectors for Word Representation
- What is the best multi-stage architecture for object recognition?
- Adaptive Subgradient Methods for Online Learning and Stochastic Optimization