gabzai / Conditional-Variational-Autoencoder-based-Stochastic-Attacks

Conditional Variational AutoEncoder-based on Stochastic Attacks

The current repository is associated with the article "Conditional Variational AutoEncoder based on Stochastic Attacks" available on IACR Transactions on Cryptographic Hardware and Embedded Systems (TCHES) and the eprints.

This project has been developed in Python 3.6.9.

This project is composed of the following scripts and repositories:

• training_phase.py: provides the training process interface (ie. projection to the orthonormal monomial basis, construction of the cVAE-SA, hyperparameters' setting, application of the training process);
• cvaest_architecture.py: implements the cVAE-SA architecture (ie. the encoder and the decoder) as well as the ELBO, the reconstruction and the KL-divergence loss functions;
• key_recovery_phase.py: conducts the key recovery phase which follows the modus operandi detailed in Sec.3.4 in the paper;
• log.py: captures and saves the training process log;
• plot_history.py: plots the evolution of the loss functions (ie. the ELBO loss, the reconstruction loss and the KL-divergence loss);
• "training_history": contains information related to the ELBO, the reconstruction and the KL-divergence loss functions,
• "trained_models": containts the model trained with the "training_phase.py" script.
• "models_in_paper": containts the models used in the article.

All the simulated traces used in this paper are fully available in the "Simulated_traces.zip" zip file. In addition, an additional script, namely "simulation_trace_exemple.py" is proposed to conduct additional simulations with other setup configurations. The use of this script uses the Lascar library (Ledger-Donjon) for computing the Signal-to-Noise Ratio (SNR).

Raw data files hashes

The zip file SHA-256 hash value is:

---

Simulated_traces.zip: 5f9aceeabaaca2d4d9bb356f81fa47c795d25172bef5c9e10882fc93698c8869

---

Citation

If you use our code, models or wish to refer to our results, please use the following BibTex entry:

@article{Zaid_Bossuet_Carbone_Habrard_Venelli_2023, 
title={Conditional Variational AutoEncoder based on Stochastic Attacks}, 
volume={2023}, 
url={https://tches.iacr.org/index.php/TCHES/article/view/10286}, 
DOI={10.46586/tches.v2023.i2.310-357}, 
number={2}, 
journal={IACR Transactions on Cryptographic Hardware and Embedded Systems}, 
author={Zaid, Gabriel and Bossuet, Lilian and Carbone, Mathieu and Habrard, Amaury and Venelli, Alexandre}, 
year={2023}, 
month={Mar.}, 
pages={310–357} 
}