README

   1    ___       _   _
   2   / _ \_   _| |_| |__   ___  _ __
   3  / /_)/ | | | __| '_ \ / _ \| '_ \
   4 / ___/| |_| | |_| | | | (_) | | | |
   5 \/     \__, |\__|_| |_|\___/|_| |_|
   6        |___/  ___ _  __  __                     _   _       _
   7              /   (_)/ _|/ _| ___ _ __ ___ _ __ | |_(_) __ _| |
   8             / /\ / | |_| |_ / _ \ '__/ _ \ '_ \| __| |/ _` | |
   9            / /_//| |  _|  _|  __/ | |  __/ | | | |_| | (_| | |
  10           /___,' |_|_| |_|  \___|_|  \___|_| |_|\__|_|\__,_|_|
  11                        ___      _
  12                       / _ \_ __(_)_   ____ _  ___ _   _
  13                      / /_)/ '__| \ \ / / _` |/ __| | | |
  14                     / ___/| |  | |\ V / (_| | (__| |_| |
  15                     \/    |_|  |_| \_/ \__,_|\___|\__, |
  16                                                   |___/
  17 --------------------------------------------------------------------------------
  18
  19 This repository shows some example implementations of differential privacy[1]
  20 in Python via Jupyter Notebooks. Each notebook contains further references and
  21 information. This work may contain blunders, and it's accuracy has not been
  22 verified, however it should still be able to give you a good starting point
  23 for work with differential privacy.
  24
  25 You can use nbviewer[2] or GitHub[3] to view the notebooks online.
  26
  27 This work was completed under the guidance of Praneeth Vepakomma [4]
  28
  29 1: https://en.wikipedia.org/wiki/Differential_privacy
  30 2: https://nbviewer.org
  31 3: https://github.com/acheam0/python_dp
  32 4: https://praneeth.mit.edu/
  33
  34 --------------------------------------------------------------------------------
  35
  36 +-------+
  37 | INDEX |
  38 +-------+
  39
  40 - Approximate differential privacy (epsilon-delta):
  41     - gaussian_mechanism.ipynb
  42
  43 - Bayes rule:
  44     - bayesian_attacker.ipynb
  45
  46 - Choice query:
  47     - exponential_mechanism.ipynb
  48
  49 - Clipping:
  50     - laplace_example_class_height.ipynb
  51
  52 - Compositions:
  53     - gaussian_mechanism.ipynb
  54
  55 - Exponential mechanism:
  56     - exponential_mechanism.ipynb
  57
  58 - Histogram query:
  59     - laplace_mechanism.ipynb
  60
  61 - Helper functions:
  62     - common.py
  63
  64 - Privacy loss distributions:
  65     - gaussian_mechanism.ipynb
  66
  67 - Privacy loss random variable:
  68     - gaussian_mechanism.ipynb
  69
  70 - Pure differential privacy (epsilon):
  71     - laplace_example_class_height.ipynb
  72     - laplace_mechanism.ipynb
  73
  74 - Sum, count, mean queries:
  75     - laplace_example_class_height.ipynb
  76     - laplace_mechanism.ipynb
  77
  78 In addition to these, there are some incomplete notebooks in the "Incomplete/"
  79 subdirectory.
  80
  81 --------------------------------------------------------------------------------
  82 Copyright (c) 2023 Armaan Bhojwani <me@armaanb.net>, MIT License