Evaluation of privacy-preserving machine learning solutions
Organisatsiooni nimi
Cybernetica
Kokkuvõte
We are looking for someone to evaluate existing open source privacy-preserving machine learning (PPML) platforms like EzPC or CrypTen. The goal is to set up and run some PPML frameworks, write about their secure computation technology, sequcity levels, which algorithms (e.g. learning vs inference and types of models) and data types they support. The goal is to find a common set of benchmarks and deployments that can be used to compare these frameworks and report on the results of the comparisons. This involves learning about secure multiparty computation, machine learning and working with setting up systems developed in various programming languages.
Lõputöö kaitsmise aasta
2024-2025
Juhendaja
Pille Pullonen-Raudvere
Suhtlemiskeel(ed)
eesti keel, inglise keel
Nõuded kandideerijale
Tase
Magister
Märksõnad
Kandideerimise kontakt
Nimi
Pille Pullonen-Raudvere
Tel