arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1607.00133stat.MLcs.CRcs.LG
收藏

深度学习与差分隐私

Deep Learning with Differential Privacy

Google(谷歌);OpenAI(开放人工智能研究中心)

Martín Abadi, Andy Chu, Ian Goodfellow, H. Brendan McMahan, Ilya Mironov, Kunal Talwar, Li Zhang

摘要:

基于神经网络的机器学习技术在广泛的领域中取得了显著的成果。通常,模型的训练需要大型的代表性数据集,这些数据集可能是众包的,并且包含敏感信息。模型不应该暴露这些数据集中的私有信息。为了实现这一目标,我们开发了新的学习算法技术,并在差分隐私的框架内对隐私成本进行了精细的分析。我们的实现和实验表明,我们可以训练深度神经网络与非凸目标,在适度的隐私预算,并在软件复杂性,训练效率和模型质量的可控成本。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑