Conceptual illustration of gradient-based cryptanalysis (AI-generated).

Cryptanalysis through Differentiable Cyphers

Conceptual illustration of gradient-based cryptanalysis (AI-generated).

Cryptanalysis through Differentiable Cyphers

This project aims to develop state-of-the-art cryptanalytic attacks on cyphers and encryption functions by implementing soft differentiable versions and optimizing for the unknown key using gradient descent.

The approach replaces discrete operations with continuous relaxations so that gradients can guide the search over candidate keys. We will investigate how the choice of relaxation and optimization method affects key recovery, and evaluate candidate solutions using the original discrete encryption functions.

Avatar
Razvan Marinescu
Assistant Professor

My research interests are in Machine Learning, and it’s applications in Healthcare and Molecular Biology. I am doing research in generative models, bayesian modelling, causal ML, compositional ML and multimodal modelling.