PET and data spaces internship

MAster assignment

PET and data spaces internship

This is the description of an internship project at TNO for a graduating master's student.

Type: Master CS (either from cyber security, or semantics)

Period: Start date: as soon as possible

Student: Unassigned

If you are interested please contact:

prof.dr.ir. P.J.M. Veugen (Thijs) Honorary Professor

+31534893639

 p.j.m.veugen@utwente.nl

Building: Zilverling 2058

Personal page

Topic:

The project is about connecting two domains that are currently gaining a lot of traction:

1. International Data Spaces [1] define a European architecture for data sharing over various domains. They specify a future-proof way of sharing data that is flexible and robust.
2. Privacy-enhancing technologies (PETs) are a set of technologies that can be used to protect the privacy (or security) of sensitive data. They are typically used in distributed data-sharing situations where (privacy-) sensitive data is used. Examples are secure multi-party computation (MPC) [2] and federated learning [3].

Eventually, we would like to integrate these two domains. This would yield a privacy-friendly and flexible way of sharing data over various domains that is easily accessible. However, the current data space architecture focuses on bilateral data-sharing agreements, whereas sensitive data-sharing problems usually involve multiple parties that would like to link their data together to enrich it. Furthermore, the (cryptographic) protocols that lie underneath have complicated communication patterns: many parties are involved to send data to each other within many different communication rounds, with synchronization and security issues.

Goal:

The goal is to extend the data space architecture such that distributed PETs could eventually be used within (sensitive) data-sharing problems with multiple parties.

Method:

1. Literature research on data spaces and PETs.
2. Describe the communication patterns and specific processing actions that will be needed for generic MPCs and how they could be incorporated into the data space architecture.
3. Implement a simple PET algorithm that follows the generic MPC patterns on top of a standard data space as a data-sharing mechanism.
4. Analyze whether and how these extensions could be incorporated into data spaces.

Further information:

This assignment is part of an internship project at TNO where you will be working within the Data Ecosystems department [4]. Throughout your internship, office presence is expected on team days.

Literature:

1. European data spaces and the role of data.europe.eu, European Commission, 2023. https://op.europa.eu/en/publication-detail/-/publication/70d01867-8ce1-11ee-8aa6-01aa75ed71a1/language-en
2. https://www.tno.nl/en/technology-science/technologies/secure-multi-party-computation/
3. https://www.tno.nl/en/technology-science/technologies/federated-learning/
4. https://www.tno.nl/en/about-tno/organisation/units/information-communication-technology/data-ecosystems/