Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/kuleuven-cosic/master-attack

https://github.com/kuleuven-cosic/master-attack

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (8.5%) to scientific vocabulary
Last synced: 4 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: KULeuven-COSIC
  • License: mit
  • Language: Python
  • Default Branch: main
  • Size: 9.73 MB
Statistics
  • Stars: 0
  • Watchers: 2
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created 10 months ago · Last pushed 10 months ago
Metadata Files
Readme License

README.md

MaSTer Attack

Implementation of the techniques presented in paper "Exploring Adversarial Attacks on the MaSTer Truncation Protocol"

Overview

Exploring adversarial influence on NN evaluation through MaSTer truncation

Installation

Dependencies

Ensure you have Python installed along with the necessary libraries: bash pip install tensorflow numpy matplotlib seaborn h5py pandas scikit-learn argparse cleverhans tikzplotlib

Usage

Training Models

To train all models: bash python3 main.py --train

Running Attacks

You can choose between three attacks: 1. Adversarial Example Attack (AE) 2. Inference (destination) Attack (DEST) 3. Optimisation Attack (OPT)

Example:

To run an AE attack: bash python3 main.py --attack AE --optimised --realistic --budget

To run a DEST attack: bash python3 main.py --attack DEST --optimised --realistic --budget

To run an OPT attack: bash python3 main.py --attack OPT --optimised --budget The script runs an attack on all specified models and fixed-point precisions as specified in main.py.

Project Structure

  • train.py - Trains all models.
  • main.py - Main entry point.
  • AE_attack.py - Runs adversarial example attacks.
  • dest_attack.py - Runs inference attacks.
  • optimisation_attack.py - Runs optimisation attacks.
  • network.py - Defines the neural network.
  • layers.py - Implements layers like Dense and Conv2D.
  • data_loader.py - Loads datasets.
  • model_init.py - Initializes models.
  • visualiser.py - Generates plots for analysis.

Expected Output

Training will save models in models/. Attack results will be stored in model_plots/{attack_type}.

Dataset Credits

This project uses publicly available datasets for training and evaluation. We gratefully acknowledge the following sources:

Owner

  • Name: KU Leuven - COSIC
  • Login: KULeuven-COSIC
  • Kind: organization

GitHub Events

Total
  • Member event: 1
  • Push event: 2
  • Create event: 2
Last Year
  • Member event: 1
  • Push event: 2
  • Create event: 2