Security and Privacy of the Data Science Knowledge Graph – SecureKG

Co-PIs:

  • Prof. Jean-Pierre Hubaux (EPFL)
  • Prof. Bryan Ford (EPFL)

Project presentation

April 20th 2018

Problem:

  • Large-scale multidisciplinary distributed scenarios
  • Security, privacy and accountability challenges on the diverse and heterogeneous data sets whose lineage is managed by the Renga platform.
  • Strict legal landscapes (e.g., LPD, EU GDPR) also call for stronger protection for personal data.

Solution:

Blockchain-based tech for federated self-sovereign identities, decentralized logging and access control:

  • Consistency, integrity and accountability: blockchain-based immutable and traceable logging
  • Multi-level access control with federated identity: distributed access rights and consensus rules
  • Confidentiality and privacy protection: collective homomorphic encryption, and query auditing for inference-resistance
  • Configurable and modular architecture, adaptable to the domain requirements

Impact:

Key enabler for effective, secure, privacy-conscious and data protection-compliant provenance for distributed data science.