Skip to main content

AIA and MIA attack versus synthetic and real data.

Project description

Impact of using synthetic data on MIA and AIA

API documentation

The documentation is generated using sphinx. To view it, clone the repo and use your web browser to open docs/build/html/index.html. To regenerate the API documentation:

sphinx-apidoc -f -F -o docs src;
sphinx-build -M html docs docs/build/

Tests

Unit tests are setup in the test directory. They use the python unittest packages. To run all tests:

python -m unittest discover -v -p "*.py" -s test

Installation

Instal with pip in a venv. In the directory containing pyproject.toml :

 pip install --editable .

Usage

Running experiments through the command line interface

python -m synthetic_aia_mia.experiments.adult <generator> <k>;
  • Generator is either identity or ctgan.
  • k folding step in 0,1,2,3,4.

Key features

#Load data
from synthetic_aia_mia.fetch_data import adult
from synthetic_aia_mia.fetch_data import utk
data_adult = adult.load()
data_utk = utk.load()

#Train a neural network an adult 
form synthetic.predictor.adult import AdultNN
adultNN = AdultNN()
adultNN.fit(data_adult["train"])

#Evalute trained neural network
adultNN.predict(data_adult["test"])

#Generate synthetic data
from synthetic_aia_mia.generator.adult import ctgan 
from synthetic_aia_mia.generator import identity 
synthetic_identity = identity(data_adult(["train"]))
synthetic_ctgan = ctgan(data_adult(["train"]))

Datasets

Adult

We are using folktables adult.

UTKFaces

From Kaggle: jangedoo/utkfaces-new. For loading and parsing files we use aia_fairness.dataset_processing

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

synthetic_aia_mia-0.0.45.tar.gz (2.0 MB view details)

Uploaded Source

Built Distribution

synthetic_aia_mia-0.0.45-py3-none-any.whl (46.1 kB view details)

Uploaded Python 3

File details

Details for the file synthetic_aia_mia-0.0.45.tar.gz.

File metadata

  • Download URL: synthetic_aia_mia-0.0.45.tar.gz
  • Upload date:
  • Size: 2.0 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.11.8

File hashes

Hashes for synthetic_aia_mia-0.0.45.tar.gz
Algorithm Hash digest
SHA256 92ea7479f9f9372ddafd57d3acf63f94da9b7465c0b4aee476b083e9108457b0
MD5 46bdeff8d573ed4a58f58ff1978e0572
BLAKE2b-256 bee13b4b57312817c4c0330a92ce981232c2dd629605b50e6723324ed68f34d1

See more details on using hashes here.

File details

Details for the file synthetic_aia_mia-0.0.45-py3-none-any.whl.

File metadata

File hashes

Hashes for synthetic_aia_mia-0.0.45-py3-none-any.whl
Algorithm Hash digest
SHA256 1d83481656301a972193d821c8a092bc89b86ebc8b038fe647819a4ba4d469e7
MD5 831cfc3c19531d4173855da0033d5b39
BLAKE2b-256 8e8529a92e3af2b42b7db3351ddd02397f1f627e2513e6c737de1560cd7601fa

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page