Privacy Preserving Schema Matching

In the context of information integration between organizations, schema matching task is required to establish mappings between attributes of two given database schemas. In some scenarios, organizations do not want to reveal their schema structure due to privacy concerns. However, high protection of attributes’ information (name, type, structure, etc.) might lead to poor matching quality due to the lack of matching contexts. There is a need to overcome this trade-off.

The goal of this project is to leverage schema covering techniques (with concepts) for privacy preserving schema matching. At first, by interpreting a schema in terms of concepts, we maintain the structural correlation (context) between attributes. Then, the tasks of matching between concepts in the concept repository can be outsourced to different mediators (e.g., paid services).

The project involves the following steps:

• Design and implement the schema covering solution for privacy preserving schema matching.
• Formulate the integration scenario and privacy concerns (privacy policy, access control, object to be protected, etc.)
• Define the metrics to measure the privacy loss and matching performance.
• Simulate various kinds of attack to show the robustness of the proposed framework.

Prerequisites

• Information systems: structured and semi-structured databases
• Good programming skills (Java, Distributed Services)
• Privacy background (k-protection, obfuscation, and encryption techniques)