Efficient Multi-Party Private Function Evaluation for Depth-optimized Circuits

Master Thesis


When a loanee aims to borrow money from a loaner, the loaner needs to check the credit worthiness of the loanee. Privacy-preserving credit checking [1] is a prominent application for the cryptographic primitive Private Function Evaluation (PFE), which allows two parties to securely evaluate the loaner’s private function on the loanee’ secret inputs without revealing anything but the output. Today, however, it is important to additionally include the General Credit Protection Agency into the computation as they are able to estimate the credit worthiness of the loanee more precisely. Unfortunately, state-of-the-art PFE frameworks [2,3] are designed for only n=2 parties. Mohassel and Sadeghian [4] propose a multi-party PFE protocol for n>2 parties. However, this protocol is only applicable to Boolean circuits, but efficient privacy-preserving credit checking becomes more efficient via arithmetic circuits.


The goal of this thesis is to implement an interactive credit checking application between the loaner, the loanee, and the General Credit Protection Agency. For this, it is necessary to design, implement, and evaluate a multi-party PFE protocol that can evaluate Boolean and arithmetic circuits. The resulting protocol shall be additionally benchmarked with multiple Boolean and arithmetic circuits, and compared to state-of-the-art PFE frameworks [2,3].


  • Good programming skills in C++
  • At least basic knowledge of cryptography
  • High motivation + ability to work independently
  • Knowledge of the English language, Git, and LaTeX