{ "cells": [ { "cell_type": "markdown", "id": "e2c530b2", "metadata": {}, "source": [ "# Example implementation of pure differential privacy\n", "\n", "By [Armaan Bhojwani](https://armaanb.net) under [Praneeth Vepakomma](https://praneeth.mit.edu/)\n", "\n", "Privatizing the height data of a class using the laplace mechanism.\n", "\n", "In this notebook we do the following:\n", "1. Define parameters with which to operate\n", "1. Generate a dataset to work on\n", "1. Implement functions that return differentially private versions of sum, count, and mean\n", "1. Clip for sensitivity\n", "1. Analyze the accuracy of the private mean across various epsilons\n", "\n", "### Dependencies\n", "- matplotlib\n", "- scipy\n", "- tqdm\n", "\n", "### Status\n", "- Complete\n", "- Could make the clipping process epsilon-delta DP" ] }, { "cell_type": "markdown", "id": "2413eb87", "metadata": {}, "source": [ "## Parameters\n", "\n", "### Data parameters\n", "\n", "For our data, we will draw `num_records` number of samples from a gaussian distribution with parameters `height_loc` and `height_std` for mean and standard deviation respectively. This should create a distribution of heights fairly accurate to what you might find in the real world. The default values of 175 and 7 for the height distribution parameters are accurate for the US in centimeters." ] }, { "cell_type": "code", "execution_count": 1, "id": "77642ebc", "metadata": {}, "outputs": [], "source": [ "num_records = 8000 # Number of records to create\n", "height_loc = 175 # Mean of height distribution (175 is USA mean)\n", "height_std = 7 # Standard deviation of height distributlon (7 is USA std)" ] }, { "cell_type": "markdown", "id": "d1ebce1f", "metadata": {}, "source": [ "### Analysis parameters\n", "The `max_epsilon` and `epsilon_step` parameters are used to create a list of epsilons which we will use to show how the efficacy of DP changes with varying epsilon.\n", "\n", "`num_samples` is the number of times to run these tests. Higher is better, but will take longer." ] }, { "cell_type": "code", "execution_count": 2, "id": "80c7682b", "metadata": {}, "outputs": [], "source": [ "max_epsilon = 10 # Largest epsilon value to test\n", "epsilon_step = 0.25 # Step size between epsilons\n", "# With these parameters, we test epsilons [0.25, 0.5... 9.75, 10]\n", "\n", "num_samples = 20000 # Number of times to run lim x->inf functions" ] }, { "cell_type": "markdown", "id": "5cdd14c9", "metadata": {}, "source": [ "### Clipping parameters\n", "\n", "Reccomended reading: https://programming-dp.com/ch5.html#clipping\n", "\n", "These parameters configure the process for clipping for the bounds of the `sum` function sensitivity. The `max_height` parameter is the upper bound for calculating the clipping of the sensitivity of the sum query. It should be higher than any expected height in the distribution, but not so high as to waste computational power. `clipping_epsilon` is the epsilon value used when adding noise during the clipping process. With small datasets, epsilon should be high so that the scale of the noise does not overcome the scale of the data. For larger datasets, the `clipping_epsilon` can be more reasonable. `smoothing_width` and `slope_width` determine how much smoothing to do during the smoothing and slope steps respectively. Lower epsilons can allow for lower `smoothing_width` and `slope_width`. `growth_bound` is the value at which growth in the cumulative sum can be considered negligible." ] }, { "cell_type": "code", "execution_count": 3, "id": "e5315ef9", "metadata": {}, "outputs": [], "source": [ "max_height = 250 # Clipping bound greater than any height in the dataset\n", "clipping_epsilon = 3 # Epsilon to use when calculating clipping bound\n", "smoothing_width = 15 # Width of window to use for initial smoothing\n", "slope_width = 10 # Width of window to use when calculating rolling slope\n", "growth_bound = 0.25 # Slope at which to consider change negligible" ] }, { "cell_type": "markdown", "id": "358975fd", "metadata": {}, "source": [ "## Building the data\n" ] }, { "cell_type": "code", "execution_count": 4, "id": "89abe686", "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import numpy as np\n", "import matplotlib.pyplot as plt\n", "\n", "# Create height distribution and draw from it to create our height data\n", "rng = np.random.default_rng()\n", "heights = rng.normal(loc=height_loc, scale=height_std, size=num_records)\n", "heights = np.asarray([round(height, 2) for height in heights])\n", "\n", "# Plot\n", "plt.title(\"Distribution of Heights\")\n", "plt.xlabel(\"Height (cm)\")\n", "plt.ylabel(\"Count\")\n", "plt.hist(heights, bins=15)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "c7c5b3d4", "metadata": {}, "source": [ "## Differential Privacy\n", "### Laplace noise functions\n", "\n", "Implementations of adding laplace noise to given functions." ] }, { "cell_type": "code", "execution_count": 5, "id": "ca8ec0a9", "metadata": {}, "outputs": [], "source": [ "def laplace_mech(epsilon, sensitivity):\n", " \"\"\" Sample adequate laplace noise\n", " Inputs:\n", " epsilon: epsilon value to use\n", " sensitivity: sensitivity to use\n", " \n", " Output:\n", " Laplace noise given the parameters\n", " \"\"\"\n", " return rng.laplace(scale=sensitivity / epsilon)\n", "\n", "\n", "def dp_sum(data, epsilon, sensitivity):\n", " \"\"\" Differentially private sum\n", " Inputs:\n", " data: array of numbers to sum\n", " epsilon: epsilon value to use\n", " sensitivity: sensitivity to use.\n", "\n", " Output:\n", " Non private sum + appropriate laplace noise\n", " \"\"\"\n", " return np.sum(data) + laplace_mech(epsilon, sensitivity)\n", "\n", "\n", "def dp_count(data, epsilon):\n", " \"\"\" Differentially private count\n", " Inputs:\n", " data: array of numbers to sum\n", " epsilon: epsilon value to use\n", "\n", " Output:\n", " Non private count + appropriate laplace noise\n", " \"\"\"\n", " sensitivity = 1\n", " return np.size(data) + laplace_mech(epsilon, sensitivity)\n", "\n", "\n", "def dp_mean(data, epsilon, sum_sensitivity=0):\n", " \"\"\" Differentially private mean utilizing dp_sum and dp_count functions\n", " Inputs:\n", " data: array of numbers to sum\n", " epsilon: epsilon value to use\n", " sum_sensitivity: sensitivity to use for the sum query\n", " Defaults to the maximum value in the dataset\n", "\n", " Output:\n", " Differentially private mean\n", " \"\"\"\n", " return dp_sum(data, epsilon, sum_sensitivity) / dp_count(data, epsilon)" ] }, { "cell_type": "markdown", "id": "98a732fa", "metadata": {}, "source": [ "### Clipping\n", "See https://programming-dp.com/ch5.html#clipping and the documentation in the parameters section\n", "\n", "#### Step 1" ] }, { "cell_type": "code", "execution_count": 6, "id": "0b381105", "metadata": {}, "outputs": [], "source": [ "from tqdm.notebook import tqdm\n", "\n", "\n", "def clipping_step1(heights, clipping_epsilon, max_height):\n", " \"\"\" Step 1 of the clipping process\n", " Calculates privacy cumulative sums of the data\n", " Inputs:\n", " heights: height data\n", " clipping_epsilon: epsilon to use in clipping\n", " max_height: value to stop testing at\n", "\n", " Output:\n", " Array of cumulative sums with laplace noise\n", " \"\"\"\n", " epsilon_i = clipping_epsilon / max_height\n", " data1 = [\n", " heights.clip(min=0, max=i).sum() + laplace_mech(epsilon_i, i)\n", " for i in range(max_height)\n", " ]\n", "\n", " return data1\n", "\n", "\n", "data1 = clipping_step1(heights, clipping_epsilon, max_height)" ] }, { "cell_type": "markdown", "id": "9faebe55", "metadata": {}, "source": [ "#### Step 2" ] }, { "cell_type": "code", "execution_count": 7, "id": "e217b0d9", "metadata": {}, "outputs": [], "source": [ "def clipping_step2(data1, smoothing_width):\n", " \"\"\" Step 2 of the clipping process\n", " Smooths the data gathered in step 1\n", " Inputs:\n", " data1: data collected in step 1\n", " smoothing_width: width of the window to smooth over\n", "\n", " Output:\n", " Smoothed version of data from step1\n", " \"\"\"\n", " cumsum_vec = np.cumsum(np.insert(data1, 0, 0))\n", " data2 = (cumsum_vec[smoothing_width:] -\n", " cumsum_vec[:-smoothing_width]) / smoothing_width\n", "\n", " return data2\n", "\n", "\n", "data2 = clipping_step2(data1, smoothing_width)" ] }, { "cell_type": "markdown", "id": "e76c572b", "metadata": {}, "source": [ "#### Step 3" ] }, { "cell_type": "code", "execution_count": 8, "id": "476de948", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "10449.964151012013\n" ] } ], "source": [ "from scipy.stats import linregress\n", "\n", "\n", "def clipping_step3(data2, slope_width, growth_bound):\n", " \"\"\" Step 3 of the clipping process\n", " Calculates the rate of change over a rolling window of data\n", " gathered in step 2\n", " Inputs:\n", " data2: data collected in step 2\n", " slope_width: width of the window over which to calculate the slope\n", " growth_bound: slope at which to consider growth negligible\n", "\n", " Output:\n", " Derivative of step 2, differentially private sensitivity to use\n", " \"\"\"\n", " # Calculate maximum slope between heights\n", " deviations = np.diff(data2)\n", " print(max(deviations))\n", "\n", " data3 = []\n", " for i in range(slope_width, len(data2) - slope_width + 1):\n", " area = range(i - slope_width, i + slope_width)\n", " data3.append(linregress(area, [data2[i] for i in area])[0])\n", "\n", " data3 = np.asarray(data3)\n", " upper_bound = np.where(data3 <= growth_bound)[0][0]\n", "\n", " return data3, upper_bound\n", "\n", "\n", "data3, sum_sensitivity = clipping_step3(data2, slope_width, growth_bound)" ] }, { "cell_type": "markdown", "id": "2a3b1433", "metadata": {}, "source": [ "#### Plotting" ] }, { "cell_type": "code", "execution_count": 9, "id": "d13dc1d8", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax1 = plt.subplots()\n", "\n", "colors = plt.rcParams['axes.prop_cycle'].by_key()['color']\n", "plt.title(\"Summary of the clipping process\")\n", "plt.xlabel(\"Height (cm)\")\n", "\n", "# Axis 1\n", "plt.ylabel(\"Sums\")\n", "plt.plot(data1, label=\"Step 1 (Cumulative sum)\")\n", "plt.plot(data2, label=\"Step 2 (Smoothed cumulative sum)\")\n", "\n", "# Axis 2\n", "ax2 = ax1.twinx()\n", "plt.ylabel(\"Slopes\")\n", "\n", "plt.plot(data3,\n", " label=\"Step 3 (Smoothed derivative of step 2)\",\n", " color=colors[2])\n", "plt.axhline(growth_bound,\n", " linestyle=\"dashed\",\n", " label=\"Growth bound (negligible growth cutoff)\",\n", " color=colors[3])\n", "\n", "# Intersect\n", "plt.axvline(sum_sensitivity,\n", " linestyle=\"dashed\",\n", " label=\"Growth bound intersect with step 3\",\n", " color=colors[4])\n", "plt.xticks(np.append(ax2.get_xticks(), sum_sensitivity).clip(min=0))\n", "\n", "# Legend\n", "handles, labels = [], []\n", "for ax in fig.axes:\n", " for h, l in zip(*ax.get_legend_handles_labels()):\n", " handles.append(h)\n", " labels.append(l)\n", "\n", "plt.legend(handles, labels, loc='center left', bbox_to_anchor=(1.12, 0.5))\n", "\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "971f3107", "metadata": {}, "source": [ "### Analysis of mean\n", "\n", "Find the mean height at various epsilons given the parameters above and record the data." ] }, { "cell_type": "code", "execution_count": 10, "id": "f2c9ec81", "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "17f9e67eac1a489987a6a6ea91da1df1", "version_major": 2, "version_minor": 0 }, "text/plain": [ " 0%| | 0/39 [00:00" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "dp_height_mean_maxes = [np.max(i) for i in data]\n", "dp_height_mean_mins = [np.min(i) for i in data]\n", "\n", "fig, ax = plt.subplots()\n", "ax.fill_between(epsilons, dp_height_mean_maxes, dp_height_mean_mins)\n", "\n", "heights_mean = np.mean(heights)\n", "plt.axhline(heights_mean,\n", " color=\"orange\",\n", " label=f\"True mean ({round(heights_mean, 2)})\")\n", "\n", "plt.xticks(np.arange(max(epsilons)))\n", "plt.title(\"Possible Private Responses vs Epsilon\\n(Mean Query)\")\n", "plt.xlabel(\"Epsilon\")\n", "plt.ylabel(\"Possible Private Responses (cm)\")\n", "ax.legend()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "715f7cef", "metadata": {}, "source": [ "From this data, we can see that you start to get diminishing returns around an epsilon of 5, thus for good privacy you should select an epsilon value lower than that." ] }, { "cell_type": "markdown", "id": "846912ac", "metadata": {}, "source": [ "### Plot range data\n", "\n", "Plot the range of the DP results found, and their difference from the true (non-private) mean." ] }, { "cell_type": "code", "execution_count": 12, "id": "98cc7a55", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.plot(epsilons, range_data, label=\"Range (max - min) of DP result\")\n", "plt.plot(epsilons, [i / 2 for i in range_data],\n", " label=\"Difference of DP result from true mean\")\n", "plt.title(\"Accuracy of Private Responses\\n(Mean Query)\")\n", "plt.xlabel(\"Epsilon\")\n", "plt.ylabel(\"Centimeters\")\n", "plt.legend()\n", "plt.show()" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.9.7" } }, "nbformat": 4, "nbformat_minor": 5 }