Outsourcing of privacy preserving data mining for large-scale non-structured information

2016 Fiscal Year Final Research Report

• Project/Area Number: 24680015
• Research Category: Grant-in-Aid for Young Scientists (A)
• Allocation Type: Partial Multi-year Fund
• Research Field: Intelligent informatics
• Research Institution: University of Tsukuba
• Principal Investigator: SAKUMA Jun 筑波大学, システム情報系, 教授 (90376963)
• Project Period (FY): 2012-04-01 – 2017-03-31
• Keywords: 準同型暗号 / 差分プライバシー / プライバシ保護データマイニング / アウトソーシング

Outline of Final Research Achievements

For outsourcing of privacy-preserving data mining with large-scale data, we consider two tasks: secure computation and statistical privacy. When the data contains private information and is distributed over multiple locations, the former provides a methodology that computes a specified function with the distributed data sources with keeping the secrecy of data in the process of computation. The latter aims to prevent inference of input from the computed results. In this research project, we developed studies on secure computation and statistical privacy (including differential privacy) for various statistics and data mining tasks. More specifically, we developed secure computation of vector matrix multiplication, set intersection cardinality, statistical testing, and so on. For statistical privacy, we studied differential privacy of outlier detection and privacy protection by interval release.

Free Research Field

プライバシー, 機械学習