Policy Compliance of Queries for Private Information Retrieval

Abstract

The use of Private Information Retrieval (PIR) techniques enable clients to retrieve items from cooperating databases without revealing either the queries or the information being retrieved. In order to prevent clients from accessing information that they are not authorized to access, it must be possible to prove that the queries being posed are compliant with a set of privacy policies previously agreed upon by the clients and database owners. Efforts to address privacy in these situations have been dominated by techniques that assume that most clients are malicious and focus on helping database owners restrict access to data. With the current push towards need-to-share, we suggest alternative approaches such as the application of accountability mechanisms. These mechanisms include the use of formalisms that can express realistic data-use policies, automated reasoning engines that can interpret those policies, automatically determining whether particular queries are policy-compliant, and justifications to enable users to understand the compliance decision and the policies.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 2010
Accession Number
ADA533742

Entities

People

  • Lalana Kagal

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Biomedical
  • Human Systems

DTIC Thesaurus Topics

  • Accountability
  • Air Force
  • Air Force Research Laboratories
  • Artificial Intelligence
  • Artificial Intelligence Computing
  • Computer Access Control
  • Computer Science
  • Department Of Homeland Security
  • Governments
  • Information Exchange
  • Information Retrieval
  • Information Systems
  • Law
  • New England
  • Semantic Models
  • Test And Evaluation
  • User Interface

Fields of Study

  • Computer science

Readers

  • Cybersecurity.
  • Geospatial Intelligence and Artificial Intelligence Analytics
  • Theoretical Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Information Retrieval