报告题目：R&D on Privacy-Preserving Data Mining

主 讲 人：Lihua Wang 副教授

报告时间：3月26日（周二）下午16:00--17:00

报告地点：数理楼221

报告摘要：

Recently, computational outsourcing using cloud services is getting popular for data analysis, and many cloud sourcing providers provide machine learning platforms. On the other hand, there still remains a big hurdle to analyze personal data on cloud services because the leakage of personal information is a critical issue. In order to address the issue, the homomorphic encryption (HE) which allows computation on encrypted data is considered as a useful techneque because it can delegate the complicated calculation operation to the semi-reliable cloud server without information leakage. Especially, inner product of vectors is one of the most basic and useful operations in statistical calculations and machine learning. In this talk, we first introduce our recernt study on homomorphic encryption techneque of constructing generic yet efficient secure inner product, then present our recent works on privacy-preserving machine learning (ex., privacy-preserving logistic regression, three-layer feedforward neural networks, and so on).