基于迁移学习的多源数据隐私保护方法研究

隐私保护的多源数据分析是大数据分析的研究热点,在多方隐私数据中学习分类器具有重要应用。提出两阶段的隐私保护分析器模型,首先在本地使用具有隐私保护性的PATE-T模型对隐私数据训练分类器;然后集合多方分类器,使用迁移学习将集合知识迁移到全局分类器,建立一个准确的、具有差分隐私的全局分类器。该全局分类器无需访问任何一方隐私数据。实验结果表明,全局分类器不仅能够很好地诠释各个本地分类器,而且还可以保护各方隐私训练数据的细节。

Multi-source data privacy protection based on transfer learning

Multi source data analysis with privacy protection is a research hotspot in big data analysis. Learning classifiers from multi party privacy data has important applications. We propose a two stage privacy protection analyzer model. Firstly, we use the PATE T model with privacy protection to train the classifier for private data. Then we gather the multi-party classifier, and use transfer learning to transfer the set knowledge to the global classifier to establish an accurate global classifier with differential privacy. The global classifier does not need to access any party’s private data. Experimental results show that the global classifier can not only interpret each local classifier well, but also protect the details of the privacy training data of all parties.