diff --git a/examples/demo_reweighing_preproc.ipynb b/examples/demo_reweighing_preproc.ipynb index 25b3a9af..c0bc99a1 100644 --- a/examples/demo_reweighing_preproc.ipynb +++ b/examples/demo_reweighing_preproc.ipynb @@ -1,956 +1,1340 @@ { - "cells": [ - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### This notebook demonstrates the use of a reweighing pre-processing algorithm for bias mitigation\n" - ] - }, - { - "cell_type": "code", - "execution_count": 1, - "metadata": {}, - "outputs": [], - "source": [ - "%matplotlib inline\n", - "# Load all necessary packages\n", - "import sys\n", - "sys.path.append(\"../\")\n", - "import numpy as np\n", - "from tqdm import tqdm\n", - "\n", - "from aif360.datasets import BinaryLabelDataset\n", - "from aif360.datasets import AdultDataset, GermanDataset, CompasDataset\n", - "from aif360.metrics import BinaryLabelDatasetMetric\n", - "from aif360.metrics import ClassificationMetric\n", - "from aif360.algorithms.preprocessing.reweighing import Reweighing\n", - "from aif360.algorithms.preprocessing.optim_preproc_helpers.data_preproc_functions\\\n", - " import load_preproc_data_adult, load_preproc_data_german, load_preproc_data_compas\n", - "from sklearn.linear_model import LogisticRegression\n", - "from sklearn.preprocessing import StandardScaler\n", - "from sklearn.metrics import accuracy_score\n", - "\n", - "from IPython.display import Markdown, display\n", - "import matplotlib.pyplot as plt\n", - "\n", - "from common_utils import compute_metrics" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Load dataset and set options" - ] - }, - { - "cell_type": "code", - "execution_count": 2, - "metadata": {}, - "outputs": [], - "source": [ - "## import dataset\n", - "dataset_used = \"adult\" # \"adult\", \"german\", \"compas\"\n", - "protected_attribute_used = 1 # 1, 2\n", - "\n", - "\n", - "if dataset_used == \"adult\":\n", - "# dataset_orig = AdultDataset()\n", - " if protected_attribute_used == 1:\n", - " privileged_groups = [{'sex': 1}]\n", - " unprivileged_groups = [{'sex': 0}]\n", - " dataset_orig = load_preproc_data_adult(['sex'])\n", - " else:\n", - " privileged_groups = [{'race': 1}]\n", - " unprivileged_groups = [{'race': 0}]\n", - " dataset_orig = load_preproc_data_adult(['race'])\n", - " \n", - "elif dataset_used == \"german\":\n", - "# dataset_orig = GermanDataset()\n", - " if protected_attribute_used == 1:\n", - " privileged_groups = [{'sex': 1}]\n", - " unprivileged_groups = [{'sex': 0}]\n", - " dataset_orig = load_preproc_data_german(['sex'])\n", - " else:\n", - " privileged_groups = [{'age': 1}]\n", - " unprivileged_groups = [{'age': 0}]\n", - " dataset_orig = load_preproc_data_german(['age'])\n", - " \n", - "elif dataset_used == \"compas\":\n", - "# dataset_orig = CompasDataset()\n", - " if protected_attribute_used == 1:\n", - " privileged_groups = [{'sex': 1}]\n", - " unprivileged_groups = [{'sex': 0}]\n", - " dataset_orig = load_preproc_data_compas(['sex'])\n", - " else:\n", - " privileged_groups = [{'race': 1}]\n", - " unprivileged_groups = [{'race': 0}]\n", - " dataset_orig = load_preproc_data_compas(['race'])\n", - "\n", - "all_metrics = [\"Statistical parity difference\",\n", - " \"Average odds difference\",\n", - " \"Equal opportunity difference\"]\n", - "\n", - "#random seed for calibrated equal odds prediction\n", - "np.random.seed(1)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Split into train, and test" - ] - }, - { - "cell_type": "code", - "execution_count": 3, - "metadata": {}, - "outputs": [], - "source": [ - "# Get the dataset and split into train and test\n", - "dataset_orig_train, dataset_orig_vt = dataset_orig.split([0.7], shuffle=True)\n", - "dataset_orig_valid, dataset_orig_test = dataset_orig_vt.split([0.5], shuffle=True)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Clean up training data" - ] - }, - { - "cell_type": "code", - "execution_count": 4, - "metadata": {}, - "outputs": [ - { - "data": { - "text/markdown": [ - "#### Training Dataset shape" + "cells": [ + { + "cell_type": "markdown", + "metadata": { + "id": "oqoGd0kN6QV1" + }, + "source": [ + "#### This notebook demonstrates the use of a reweighing pre-processing algorithm for bias mitigation\n" + ] + }, + { + "cell_type": "code", + "source": [ + "!pip install 'aif360[all]'" ], - "text/plain": [ - "" + "metadata": { + "id": "21RrRdmW6RET", + "outputId": "2bec7003-4c0c-4059-d933-99b829eeb8a2", + "colab": { + "base_uri": "https://localhost:8080/" + } + }, + "execution_count": 1, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n", + "Requirement already satisfied: aif360[all] in /usr/local/lib/python3.7/dist-packages (0.5.0)\n", + "Requirement already satisfied: scipy>=1.2.0 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.7.3)\n", + "Requirement already satisfied: numpy>=1.16 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.21.6)\n", + "Requirement already satisfied: scikit-learn>=1.0 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.0.2)\n", + "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (3.5.3)\n", + "Requirement already satisfied: pandas>=0.24.0 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.3.5)\n", + "Requirement already satisfied: lime in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (0.2.0.1)\n", + "Requirement already satisfied: sphinx<2 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.8.6)\n", + "Requirement already satisfied: rpy2 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (3.4.5)\n", + "Requirement already satisfied: seaborn in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (0.11.2)\n", + "Requirement already satisfied: adversarial-robustness-toolbox>=1.0.0 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.11.1)\n", + "Requirement already satisfied: torch in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.12.1+cu113)\n", + "Requirement already satisfied: jupyter in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.0.0)\n", + "Requirement already satisfied: cvxpy>=1.0 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.2.1)\n", + "Requirement already satisfied: jinja2<3.1.0 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (2.11.3)\n", + "Requirement already satisfied: pytest>=3.5 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (3.6.4)\n", + "Requirement already satisfied: ipympl in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (0.9.2)\n", + "Requirement already satisfied: tempeh in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (0.1.12)\n", + "Requirement already satisfied: tqdm in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (4.64.1)\n", + "Requirement already satisfied: sphinx-rtd-theme in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (1.0.0)\n", + "Requirement already satisfied: BlackBoxAuditing in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (0.1.54)\n", + "Requirement already satisfied: tensorflow>=1.13.1 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (2.8.2+zzzcolab20220719082949)\n", + "Requirement already satisfied: fairlearn~=0.7 in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (0.7.0)\n", + "Requirement already satisfied: lightgbm in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (2.2.3)\n", + "Requirement already satisfied: igraph[plotting] in /usr/local/lib/python3.7/dist-packages (from aif360[all]) (0.10.1)\n", + "Requirement already satisfied: setuptools in /usr/local/lib/python3.7/dist-packages (from adversarial-robustness-toolbox>=1.0.0->aif360[all]) (57.4.0)\n", + "Requirement already satisfied: six in /usr/local/lib/python3.7/dist-packages (from adversarial-robustness-toolbox>=1.0.0->aif360[all]) (1.15.0)\n", + "Requirement already satisfied: numba>=0.53.1 in /usr/local/lib/python3.7/dist-packages (from adversarial-robustness-toolbox>=1.0.0->aif360[all]) (0.56.2)\n", + "Requirement already satisfied: scs>=1.1.6 in /usr/local/lib/python3.7/dist-packages (from cvxpy>=1.0->aif360[all]) (3.2.0)\n", + "Requirement already satisfied: osqp>=0.4.1 in /usr/local/lib/python3.7/dist-packages (from cvxpy>=1.0->aif360[all]) (0.6.2.post0)\n", + "Requirement already satisfied: ecos>=2 in /usr/local/lib/python3.7/dist-packages (from cvxpy>=1.0->aif360[all]) (2.0.10)\n", + "Requirement already satisfied: MarkupSafe>=0.23 in /usr/local/lib/python3.7/dist-packages (from jinja2<3.1.0->aif360[all]) (2.0.1)\n", + "Requirement already satisfied: importlib-metadata in /usr/local/lib/python3.7/dist-packages (from numba>=0.53.1->adversarial-robustness-toolbox>=1.0.0->aif360[all]) (4.12.0)\n", + "Requirement already satisfied: llvmlite<0.40,>=0.39.0dev0 in /usr/local/lib/python3.7/dist-packages (from numba>=0.53.1->adversarial-robustness-toolbox>=1.0.0->aif360[all]) (0.39.1)\n", + "Requirement already satisfied: qdldl in /usr/local/lib/python3.7/dist-packages (from osqp>=0.4.1->cvxpy>=1.0->aif360[all]) (0.1.5.post2)\n", + "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.24.0->aif360[all]) (2.8.2)\n", + "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas>=0.24.0->aif360[all]) (2022.2.1)\n", + "Requirement already satisfied: more-itertools>=4.0.0 in /usr/local/lib/python3.7/dist-packages (from pytest>=3.5->aif360[all]) (8.14.0)\n", + "Requirement already satisfied: py>=1.5.0 in /usr/local/lib/python3.7/dist-packages (from pytest>=3.5->aif360[all]) (1.11.0)\n", + "Requirement already satisfied: atomicwrites>=1.0 in /usr/local/lib/python3.7/dist-packages (from pytest>=3.5->aif360[all]) (1.4.1)\n", + "Requirement already satisfied: attrs>=17.4.0 in /usr/local/lib/python3.7/dist-packages (from pytest>=3.5->aif360[all]) (22.1.0)\n", + "Requirement already satisfied: pluggy<0.8,>=0.5 in /usr/local/lib/python3.7/dist-packages (from pytest>=3.5->aif360[all]) (0.7.1)\n", + "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from scikit-learn>=1.0->aif360[all]) (3.1.0)\n", + "Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.7/dist-packages (from scikit-learn>=1.0->aif360[all]) (1.1.0)\n", + "Requirement already satisfied: imagesize in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (1.4.1)\n", + "Requirement already satisfied: docutils<0.18,>=0.11 in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (0.17.1)\n", + "Requirement already satisfied: babel!=2.0,>=1.3 in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (2.10.3)\n", + "Requirement already satisfied: sphinxcontrib-websupport in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (1.2.4)\n", + "Requirement already satisfied: snowballstemmer>=1.1 in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (2.2.0)\n", + "Requirement already satisfied: Pygments>=2.0 in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (2.6.1)\n", + "Requirement already satisfied: requests>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (2.23.0)\n", + "Requirement already satisfied: alabaster<0.8,>=0.7 in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (0.7.12)\n", + "Requirement already satisfied: packaging in /usr/local/lib/python3.7/dist-packages (from sphinx<2->aif360[all]) (21.3)\n", + "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests>=2.0.0->sphinx<2->aif360[all]) (3.0.4)\n", + "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests>=2.0.0->sphinx<2->aif360[all]) (2.10)\n", + "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests>=2.0.0->sphinx<2->aif360[all]) (1.24.3)\n", + "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests>=2.0.0->sphinx<2->aif360[all]) (2022.6.15)\n", + "Requirement already satisfied: tensorboard<2.9,>=2.8 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (2.8.0)\n", + "Requirement already satisfied: gast>=0.2.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (0.5.3)\n", + "Requirement already satisfied: protobuf<3.20,>=3.9.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (3.17.3)\n", + "Requirement already satisfied: typing-extensions>=3.6.6 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (4.1.1)\n", + "Requirement already satisfied: astunparse>=1.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (1.6.3)\n", + "Requirement already satisfied: libclang>=9.0.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (14.0.6)\n", + "Requirement already satisfied: keras-preprocessing>=1.1.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (1.1.2)\n", + "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (1.1.0)\n", + "Requirement already satisfied: wrapt>=1.11.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (1.14.1)\n", + "Requirement already satisfied: google-pasta>=0.1.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (0.2.0)\n", + "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (3.3.0)\n", + "Requirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (0.26.0)\n", + "Requirement already satisfied: h5py>=2.9.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (3.1.0)\n", + "Requirement already satisfied: absl-py>=0.4.0 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (1.2.0)\n", + "Requirement already satisfied: flatbuffers>=1.12 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (2.0.7)\n", + "Requirement already satisfied: grpcio<2.0,>=1.24.3 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (1.48.1)\n", + "Requirement already satisfied: tensorflow-estimator<2.9,>=2.8 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (2.8.0)\n", + "Requirement already satisfied: keras<2.9,>=2.8.0rc0 in /usr/local/lib/python3.7/dist-packages (from tensorflow>=1.13.1->aif360[all]) (2.8.0)\n", + "Requirement already satisfied: wheel<1.0,>=0.23.0 in /usr/local/lib/python3.7/dist-packages (from astunparse>=1.6.0->tensorflow>=1.13.1->aif360[all]) (0.37.1)\n", + "Requirement already satisfied: cached-property in /usr/local/lib/python3.7/dist-packages (from h5py>=2.9.0->tensorflow>=1.13.1->aif360[all]) (1.5.2)\n", + "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (1.8.1)\n", + "Requirement already satisfied: google-auth<3,>=1.6.3 in /usr/local/lib/python3.7/dist-packages (from tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (1.35.0)\n", + "Requirement already satisfied: tensorboard-data-server<0.7.0,>=0.6.0 in /usr/local/lib/python3.7/dist-packages (from tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (0.6.1)\n", + "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.7/dist-packages (from tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (0.4.6)\n", + "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.7/dist-packages (from tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (3.4.1)\n", + "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.7/dist-packages (from tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (1.0.1)\n", + "Requirement already satisfied: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (0.2.8)\n", + "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (4.2.4)\n", + "Requirement already satisfied: rsa<5,>=3.1.4 in /usr/local/lib/python3.7/dist-packages (from google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (4.9)\n", + "Requirement already satisfied: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.7/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (1.3.1)\n", + "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata->numba>=0.53.1->adversarial-robustness-toolbox>=1.0.0->aif360[all]) (3.8.1)\n", + "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.7/dist-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (0.4.8)\n", + "Requirement already satisfied: oauthlib>=3.0.0 in /usr/local/lib/python3.7/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.9,>=2.8->tensorflow>=1.13.1->aif360[all]) (3.2.0)\n", + "Requirement already satisfied: networkx in /usr/local/lib/python3.7/dist-packages (from BlackBoxAuditing->aif360[all]) (2.6.3)\n", + "Requirement already satisfied: texttable>=1.6.2 in /usr/local/lib/python3.7/dist-packages (from igraph[plotting]->aif360[all]) (1.6.4)\n", + "Requirement already satisfied: cairocffi>=1.2.0 in /usr/local/lib/python3.7/dist-packages (from igraph[plotting]->aif360[all]) (1.3.0)\n", + "Requirement already satisfied: cffi>=1.1.0 in /usr/local/lib/python3.7/dist-packages (from cairocffi>=1.2.0->igraph[plotting]->aif360[all]) (1.15.1)\n", + "Requirement already satisfied: pycparser in /usr/local/lib/python3.7/dist-packages (from cffi>=1.1.0->cairocffi>=1.2.0->igraph[plotting]->aif360[all]) (2.21)\n", + "Requirement already satisfied: pillow in /usr/local/lib/python3.7/dist-packages (from ipympl->aif360[all]) (7.1.2)\n", + "Requirement already satisfied: ipywidgets<9,>=7.6.0 in /usr/local/lib/python3.7/dist-packages (from ipympl->aif360[all]) (7.7.1)\n", + "Requirement already satisfied: ipython<9 in /usr/local/lib/python3.7/dist-packages (from ipympl->aif360[all]) (7.9.0)\n", + "Requirement already satisfied: ipython-genutils in /usr/local/lib/python3.7/dist-packages (from ipympl->aif360[all]) (0.2.0)\n", + "Requirement already satisfied: traitlets<6 in /usr/local/lib/python3.7/dist-packages (from ipympl->aif360[all]) (5.1.1)\n", + "Requirement already satisfied: prompt-toolkit<2.1.0,>=2.0.0 in /usr/local/lib/python3.7/dist-packages (from ipython<9->ipympl->aif360[all]) (2.0.10)\n", + "Requirement already satisfied: jedi>=0.10 in /usr/local/lib/python3.7/dist-packages (from ipython<9->ipympl->aif360[all]) (0.18.1)\n", + "Requirement already satisfied: backcall in /usr/local/lib/python3.7/dist-packages (from ipython<9->ipympl->aif360[all]) (0.2.0)\n", + "Requirement already satisfied: pickleshare in /usr/local/lib/python3.7/dist-packages (from ipython<9->ipympl->aif360[all]) (0.7.5)\n", + "Requirement already satisfied: pexpect in /usr/local/lib/python3.7/dist-packages (from ipython<9->ipympl->aif360[all]) (4.8.0)\n", + "Requirement already satisfied: decorator in /usr/local/lib/python3.7/dist-packages (from ipython<9->ipympl->aif360[all]) (4.4.2)\n", + "Requirement already satisfied: jupyterlab-widgets>=1.0.0 in /usr/local/lib/python3.7/dist-packages (from ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (3.0.3)\n", + "Requirement already satisfied: ipykernel>=4.5.1 in /usr/local/lib/python3.7/dist-packages (from ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (5.3.4)\n", + "Requirement already satisfied: widgetsnbextension~=3.6.0 in /usr/local/lib/python3.7/dist-packages (from ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (3.6.1)\n", + "Requirement already satisfied: jupyter-client in /usr/local/lib/python3.7/dist-packages (from ipykernel>=4.5.1->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (6.1.12)\n", + "Requirement already satisfied: tornado>=4.2 in /usr/local/lib/python3.7/dist-packages (from ipykernel>=4.5.1->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (5.1.1)\n", + "Requirement already satisfied: parso<0.9.0,>=0.8.0 in /usr/local/lib/python3.7/dist-packages (from jedi>=0.10->ipython<9->ipympl->aif360[all]) (0.8.3)\n", + "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->aif360[all]) (0.11.0)\n", + "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.7/dist-packages (from matplotlib->aif360[all]) (4.37.2)\n", + "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->aif360[all]) (1.4.4)\n", + "Requirement already satisfied: pyparsing>=2.2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->aif360[all]) (3.0.9)\n", + "Requirement already satisfied: wcwidth in /usr/local/lib/python3.7/dist-packages (from prompt-toolkit<2.1.0,>=2.0.0->ipython<9->ipympl->aif360[all]) (0.2.5)\n", + "Requirement already satisfied: notebook>=4.4.1 in /usr/local/lib/python3.7/dist-packages (from widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (5.3.1)\n", + "Requirement already satisfied: nbconvert in /usr/local/lib/python3.7/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (5.6.1)\n", + "Requirement already satisfied: jupyter-core>=4.4.0 in /usr/local/lib/python3.7/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (4.11.1)\n", + "Requirement already satisfied: Send2Trash in /usr/local/lib/python3.7/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (1.8.0)\n", + "Requirement already satisfied: terminado>=0.8.1 in /usr/local/lib/python3.7/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.13.3)\n", + "Requirement already satisfied: nbformat in /usr/local/lib/python3.7/dist-packages (from notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (5.4.0)\n", + "Requirement already satisfied: pyzmq>=13 in /usr/local/lib/python3.7/dist-packages (from jupyter-client->ipykernel>=4.5.1->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (23.2.1)\n", + "Requirement already satisfied: ptyprocess in /usr/local/lib/python3.7/dist-packages (from terminado>=0.8.1->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.7.0)\n", + "Requirement already satisfied: jupyter-console in /usr/local/lib/python3.7/dist-packages (from jupyter->aif360[all]) (6.1.0)\n", + "Requirement already satisfied: qtconsole in /usr/local/lib/python3.7/dist-packages (from jupyter->aif360[all]) (5.3.2)\n", + "Requirement already satisfied: scikit-image>=0.12 in /usr/local/lib/python3.7/dist-packages (from lime->aif360[all]) (0.18.3)\n", + "Requirement already satisfied: tifffile>=2019.7.26 in /usr/local/lib/python3.7/dist-packages (from scikit-image>=0.12->lime->aif360[all]) (2021.11.2)\n", + "Requirement already satisfied: PyWavelets>=1.1.1 in /usr/local/lib/python3.7/dist-packages (from scikit-image>=0.12->lime->aif360[all]) (1.3.0)\n", + "Requirement already satisfied: imageio>=2.3.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image>=0.12->lime->aif360[all]) (2.9.0)\n", + "Requirement already satisfied: mistune<2,>=0.8.1 in /usr/local/lib/python3.7/dist-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.8.4)\n", + "Requirement already satisfied: entrypoints>=0.2.2 in /usr/local/lib/python3.7/dist-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.4)\n", + "Requirement already satisfied: bleach in /usr/local/lib/python3.7/dist-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (5.0.1)\n", + "Requirement already satisfied: defusedxml in /usr/local/lib/python3.7/dist-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.7.1)\n", + "Requirement already satisfied: pandocfilters>=1.4.1 in /usr/local/lib/python3.7/dist-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (1.5.0)\n", + "Requirement already satisfied: testpath in /usr/local/lib/python3.7/dist-packages (from nbconvert->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.6.0)\n", + "Requirement already satisfied: fastjsonschema in /usr/local/lib/python3.7/dist-packages (from nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (2.16.1)\n", + "Requirement already satisfied: jsonschema>=2.6 in /usr/local/lib/python3.7/dist-packages (from nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (4.3.3)\n", + "Requirement already satisfied: importlib-resources>=1.4.0 in /usr/local/lib/python3.7/dist-packages (from jsonschema>=2.6->nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (5.9.0)\n", + "Requirement already satisfied: pyrsistent!=0.17.0,!=0.17.1,!=0.17.2,>=0.14.0 in /usr/local/lib/python3.7/dist-packages (from jsonschema>=2.6->nbformat->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.18.1)\n", + "Requirement already satisfied: webencodings in /usr/local/lib/python3.7/dist-packages (from bleach->nbconvert->notebook>=4.4.1->widgetsnbextension~=3.6.0->ipywidgets<9,>=7.6.0->ipympl->aif360[all]) (0.5.1)\n", + "Requirement already satisfied: qtpy>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from qtconsole->jupyter->aif360[all]) (2.2.0)\n", + "Requirement already satisfied: tzlocal in /usr/local/lib/python3.7/dist-packages (from rpy2->aif360[all]) (1.5.1)\n", + "Requirement already satisfied: sphinxcontrib-serializinghtml in /usr/local/lib/python3.7/dist-packages (from sphinxcontrib-websupport->sphinx<2->aif360[all]) (1.1.5)\n", + "Requirement already satisfied: shap in /usr/local/lib/python3.7/dist-packages (from tempeh->aif360[all]) (0.41.0)\n", + "Requirement already satisfied: memory-profiler in /usr/local/lib/python3.7/dist-packages (from tempeh->aif360[all]) (0.60.0)\n", + "Requirement already satisfied: psutil in /usr/local/lib/python3.7/dist-packages (from memory-profiler->tempeh->aif360[all]) (5.4.8)\n", + "Requirement already satisfied: cloudpickle in /usr/local/lib/python3.7/dist-packages (from shap->tempeh->aif360[all]) (1.5.0)\n", + "Requirement already satisfied: slicer==0.0.7 in /usr/local/lib/python3.7/dist-packages (from shap->tempeh->aif360[all]) (0.0.7)\n" + ] + } ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "(34189, 18)\n" - ] + "cell_type": "code", + "execution_count": 2, + "metadata": { + "id": "1qtya2sO6QV3" + }, + "outputs": [], + "source": [ + "%matplotlib inline\n", + "# Load all necessary packages\n", + "import sys\n", + "sys.path.append(\"../\")\n", + "import numpy as np\n", + "from tqdm import tqdm\n", + "\n", + "from aif360.datasets import BinaryLabelDataset\n", + "from aif360.datasets import AdultDataset, GermanDataset, CompasDataset\n", + "from aif360.metrics import BinaryLabelDatasetMetric\n", + "from aif360.metrics import ClassificationMetric\n", + "from aif360.algorithms.preprocessing.reweighing import Reweighing\n", + "from aif360.algorithms.preprocessing.optim_preproc_helpers.data_preproc_functions\\\n", + " import load_preproc_data_adult, load_preproc_data_german, load_preproc_data_compas\n", + "from sklearn.linear_model import LogisticRegression\n", + "from sklearn.preprocessing import StandardScaler\n", + "from sklearn.metrics import accuracy_score\n", + "\n", + "from IPython.display import Markdown, display\n", + "import matplotlib.pyplot as plt" + ] }, { - "data": { - "text/markdown": [ - "#### Favorable and unfavorable labels" + "cell_type": "code", + "source": [ + "# Metrics function\n", + "from collections import OrderedDict\n", + "from aif360.metrics import ClassificationMetric\n", + "\n", + "def compute_metrics(dataset_true, dataset_pred, \n", + " unprivileged_groups, privileged_groups,\n", + " disp = True):\n", + " \"\"\" Compute the key metrics \"\"\"\n", + " classified_metric_pred = ClassificationMetric(dataset_true,\n", + " dataset_pred, \n", + " unprivileged_groups=unprivileged_groups,\n", + " privileged_groups=privileged_groups)\n", + " metrics = OrderedDict()\n", + " metrics[\"Balanced accuracy\"] = 0.5*(classified_metric_pred.true_positive_rate()+\n", + " classified_metric_pred.true_negative_rate())\n", + " metrics[\"Statistical parity difference\"] = classified_metric_pred.statistical_parity_difference()\n", + " metrics[\"Disparate impact\"] = classified_metric_pred.disparate_impact()\n", + " metrics[\"Average odds difference\"] = classified_metric_pred.average_odds_difference()\n", + " metrics[\"Equal opportunity difference\"] = classified_metric_pred.equal_opportunity_difference()\n", + " metrics[\"Theil index\"] = classified_metric_pred.theil_index()\n", + " \n", + " if disp:\n", + " for k in metrics:\n", + " print(\"%s = %.4f\" % (k, metrics[k]))\n", + " \n", + " return metrics" ], - "text/plain": [ - "" + "metadata": { + "id": "tGTwc4lE64gj" + }, + "execution_count": 3, + "outputs": [] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "1w414lZE6QV4" + }, + "source": [ + "#### Load dataset and set options" ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "(1.0, 0.0)\n" - ] + "cell_type": "code", + "source": [ + "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data\n", + "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.test\n", + "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.names" + ], + "metadata": { + "id": "0LKP7PHV6USy", + "outputId": "4c9bf373-a87a-4b5b-99b8-796c58e329e2", + "colab": { + "base_uri": "https://localhost:8080/" + } + }, + "execution_count": 4, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "--2022-09-17 00:49:50-- https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data\n", + "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n", + "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n", + "HTTP request sent, awaiting response... 200 OK\n", + "Length: 3974305 (3.8M) [application/x-httpd-php]\n", + "Saving to: ‘adult.data’\n", + "\n", + "adult.data 100%[===================>] 3.79M 16.8MB/s in 0.2s \n", + "\n", + "2022-09-17 00:49:51 (16.8 MB/s) - ‘adult.data’ saved [3974305/3974305]\n", + "\n", + "--2022-09-17 00:49:51-- https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.test\n", + "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n", + "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n", + "HTTP request sent, awaiting response... 200 OK\n", + "Length: 2003153 (1.9M) [application/x-httpd-php]\n", + "Saving to: ‘adult.test’\n", + "\n", + "adult.test 100%[===================>] 1.91M 10.2MB/s in 0.2s \n", + "\n", + "2022-09-17 00:49:51 (10.2 MB/s) - ‘adult.test’ saved [2003153/2003153]\n", + "\n", + "--2022-09-17 00:49:51-- https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.names\n", + "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n", + "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n", + "HTTP request sent, awaiting response... 200 OK\n", + "Length: 5229 (5.1K) [application/x-httpd-php]\n", + "Saving to: ‘adult.names’\n", + "\n", + "adult.names 100%[===================>] 5.11K --.-KB/s in 0s \n", + "\n", + "2022-09-17 00:49:51 (82.6 MB/s) - ‘adult.names’ saved [5229/5229]\n", + "\n" + ] + } + ] }, { - "data": { - "text/markdown": [ - "#### Protected attribute names" + "cell_type": "code", + "source": [ + "!mv /content/adult* /usr/local/lib/python3.7/dist-packages/aif360/data/raw/adult" ], - "text/plain": [ - "" + "metadata": { + "id": "88cISW5R6Zet" + }, + "execution_count": 5, + "outputs": [] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": { + "id": "l1_gbOzU6QV4" + }, + "outputs": [], + "source": [ + "## import dataset\n", + "dataset_used = \"adult\" # \"adult\", \"german\", \"compas\"\n", + "protected_attribute_used = 1 # 1, 2\n", + "\n", + "\n", + "if dataset_used == \"adult\":\n", + "# dataset_orig = AdultDataset()\n", + " if protected_attribute_used == 1:\n", + " privileged_groups = [{'sex': 1}]\n", + " unprivileged_groups = [{'sex': 0}]\n", + " dataset_orig = load_preproc_data_adult(['sex'])\n", + " else:\n", + " privileged_groups = [{'race': 1}]\n", + " unprivileged_groups = [{'race': 0}]\n", + " dataset_orig = load_preproc_data_adult(['race'])\n", + " \n", + "elif dataset_used == \"german\":\n", + "# dataset_orig = GermanDataset()\n", + " if protected_attribute_used == 1:\n", + " privileged_groups = [{'sex': 1}]\n", + " unprivileged_groups = [{'sex': 0}]\n", + " dataset_orig = load_preproc_data_german(['sex'])\n", + " else:\n", + " privileged_groups = [{'age': 1}]\n", + " unprivileged_groups = [{'age': 0}]\n", + " dataset_orig = load_preproc_data_german(['age'])\n", + " \n", + "elif dataset_used == \"compas\":\n", + "# dataset_orig = CompasDataset()\n", + " if protected_attribute_used == 1:\n", + " privileged_groups = [{'sex': 1}]\n", + " unprivileged_groups = [{'sex': 0}]\n", + " dataset_orig = load_preproc_data_compas(['sex'])\n", + " else:\n", + " privileged_groups = [{'race': 1}]\n", + " unprivileged_groups = [{'race': 0}]\n", + " dataset_orig = load_preproc_data_compas(['race'])\n", + "\n", + "all_metrics = [\"Statistical parity difference\",\n", + " \"Average odds difference\",\n", + " \"Equal opportunity difference\"]\n", + "\n", + "#random seed for calibrated equal odds prediction\n", + "np.random.seed(1)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "bOSDn2pn6QV4" + }, + "source": [ + "#### Split into train, and test" ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "['sex']\n" - ] + "cell_type": "code", + "execution_count": 7, + "metadata": { + "id": "TGCMILsl6QV4" + }, + "outputs": [], + "source": [ + "# Get the dataset and split into train and test\n", + "dataset_orig_train, dataset_orig_vt = dataset_orig.split([0.7], shuffle=True)\n", + "dataset_orig_valid, dataset_orig_test = dataset_orig_vt.split([0.5], shuffle=True)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "Hb41URiP6QV5" + }, + "source": [ + "#### Clean up training data" + ] }, { - "data": { - "text/markdown": [ - "#### Privileged and unprivileged protected attribute values" + "cell_type": "code", + "execution_count": 8, + "metadata": { + "id": "btD6cG2T6QV5", + "outputId": "617160ff-62d7-4bc0-fa70-0b675fd72e28", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 324 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Training Dataset shape" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "(34189, 18)\n" + ] + }, + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Favorable and unfavorable labels" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "1.0 0.0\n" + ] + }, + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Protected attribute names" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "['sex']\n" + ] + }, + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Privileged and unprivileged protected attribute values" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "[array([1.])] [array([0.])]\n" + ] + }, + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Dataset feature names" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "['race', 'sex', 'Age (decade)=10', 'Age (decade)=20', 'Age (decade)=30', 'Age (decade)=40', 'Age (decade)=50', 'Age (decade)=60', 'Age (decade)=>=70', 'Education Years=6', 'Education Years=7', 'Education Years=8', 'Education Years=9', 'Education Years=10', 'Education Years=11', 'Education Years=12', 'Education Years=<6', 'Education Years=>12']\n" + ] + } ], - "text/plain": [ - "" + "source": [ + "# print out some labels, names, etc.\n", + "display(Markdown(\"#### Training Dataset shape\"))\n", + "print(dataset_orig_train.features.shape)\n", + "display(Markdown(\"#### Favorable and unfavorable labels\"))\n", + "print(dataset_orig_train.favorable_label, dataset_orig_train.unfavorable_label)\n", + "display(Markdown(\"#### Protected attribute names\"))\n", + "print(dataset_orig_train.protected_attribute_names)\n", + "display(Markdown(\"#### Privileged and unprivileged protected attribute values\"))\n", + "print(dataset_orig_train.privileged_protected_attributes, \n", + " dataset_orig_train.unprivileged_protected_attributes)\n", + "display(Markdown(\"#### Dataset feature names\"))\n", + "print(dataset_orig_train.feature_names)" ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "([array([1.])], [array([0.])])\n" - ] + "cell_type": "markdown", + "metadata": { + "id": "9Y1vMWO96QV6" + }, + "source": [ + "#### Metric for original training data" + ] }, { - "data": { - "text/markdown": [ - "#### Dataset feature names" + "cell_type": "code", + "execution_count": 9, + "metadata": { + "id": "1oAjG9Dd6QV6", + "outputId": "6fc602cd-7496-46c3-b014-986b9c867a70", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 74 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Original training dataset" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Difference in mean outcomes between unprivileged and privileged groups = -0.190244\n" + ] + } ], - "text/plain": [ - "" + "source": [ + "# Metric for the original dataset\n", + "metric_orig_train = BinaryLabelDatasetMetric(dataset_orig_train, \n", + " unprivileged_groups=unprivileged_groups,\n", + " privileged_groups=privileged_groups)\n", + "display(Markdown(\"#### Original training dataset\"))\n", + "print(\"Difference in mean outcomes between unprivileged and privileged groups = %f\" % metric_orig_train.mean_difference())" ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "['race', 'sex', 'Age (decade)=10', 'Age (decade)=20', 'Age (decade)=30', 'Age (decade)=40', 'Age (decade)=50', 'Age (decade)=60', 'Age (decade)=>=70', 'Education Years=6', 'Education Years=7', 'Education Years=8', 'Education Years=9', 'Education Years=10', 'Education Years=11', 'Education Years=12', 'Education Years=<6', 'Education Years=>12']\n" - ] - } - ], - "source": [ - "# print out some labels, names, etc.\n", - "display(Markdown(\"#### Training Dataset shape\"))\n", - "print(dataset_orig_train.features.shape)\n", - "display(Markdown(\"#### Favorable and unfavorable labels\"))\n", - "print(dataset_orig_train.favorable_label, dataset_orig_train.unfavorable_label)\n", - "display(Markdown(\"#### Protected attribute names\"))\n", - "print(dataset_orig_train.protected_attribute_names)\n", - "display(Markdown(\"#### Privileged and unprivileged protected attribute values\"))\n", - "print(dataset_orig_train.privileged_protected_attributes, \n", - " dataset_orig_train.unprivileged_protected_attributes)\n", - "display(Markdown(\"#### Dataset feature names\"))\n", - "print(dataset_orig_train.feature_names)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Metric for original training data" - ] - }, - { - "cell_type": "code", - "execution_count": 5, - "metadata": {}, - "outputs": [ - { - "data": { - "text/markdown": [ - "#### Original training dataset" + "cell_type": "markdown", + "metadata": { + "id": "Jz1zihmt6QV6" + }, + "source": [ + "#### Train with and transform the original training data" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": { + "id": "q6ihofey6QV6" + }, + "outputs": [], + "source": [ + "RW = Reweighing(unprivileged_groups=unprivileged_groups,\n", + " privileged_groups=privileged_groups)\n", + "RW.fit(dataset_orig_train)\n", + "dataset_transf_train = RW.transform(dataset_orig_train)" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": { + "id": "gjNNCnhM6QV6" + }, + "outputs": [], + "source": [ + "### Testing \n", + "assert np.abs(dataset_transf_train.instance_weights.sum()-dataset_orig_train.instance_weights.sum())<1e-6" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "xO-MQD7D6QV6" + }, + "source": [ + "#### Metric with the transformed training data" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": { + "id": "jCdNm_6w6QV7", + "outputId": "50d24597-a99f-4ae8-ff27-d65a9ae07265", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 74 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Transformed training dataset" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Difference in mean outcomes between unprivileged and privileged groups = 0.000000\n" + ] + } ], - "text/plain": [ - "" + "source": [ + "metric_transf_train = BinaryLabelDatasetMetric(dataset_transf_train, \n", + " unprivileged_groups=unprivileged_groups,\n", + " privileged_groups=privileged_groups)\n", + "display(Markdown(\"#### Transformed training dataset\"))\n", + "print(\"Difference in mean outcomes between unprivileged and privileged groups = %f\" % metric_transf_train.mean_difference())" ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "Difference in mean outcomes between unprivileged and privileged groups = -0.190698\n" - ] - } - ], - "source": [ - "# Metric for the original dataset\n", - "metric_orig_train = BinaryLabelDatasetMetric(dataset_orig_train, \n", - " unprivileged_groups=unprivileged_groups,\n", - " privileged_groups=privileged_groups)\n", - "display(Markdown(\"#### Original training dataset\"))\n", - "print(\"Difference in mean outcomes between unprivileged and privileged groups = %f\" % metric_orig_train.mean_difference())" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Train with and transform the original training data" - ] - }, - { - "cell_type": "code", - "execution_count": 6, - "metadata": {}, - "outputs": [], - "source": [ - "RW = Reweighing(unprivileged_groups=unprivileged_groups,\n", - " privileged_groups=privileged_groups)\n", - "RW.fit(dataset_orig_train)\n", - "dataset_transf_train = RW.transform(dataset_orig_train)" - ] - }, - { - "cell_type": "code", - "execution_count": 7, - "metadata": {}, - "outputs": [], - "source": [ - "### Testing \n", - "assert np.abs(dataset_transf_train.instance_weights.sum()-dataset_orig_train.instance_weights.sum())<1e-6" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Metric with the transformed training data" - ] - }, - { - "cell_type": "code", - "execution_count": 8, - "metadata": {}, - "outputs": [ - { - "data": { - "text/markdown": [ - "#### Transformed training dataset" + "cell_type": "code", + "execution_count": 13, + "metadata": { + "id": "9fJC1Mqj6QV7" + }, + "outputs": [], + "source": [ + "### Testing \n", + "assert np.abs(metric_transf_train.mean_difference()) < 1e-6" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "qfwCwF4N6QV7" + }, + "source": [ + "### Train classifier on original data" + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": { + "id": "OYOyTq2U6QV7" + }, + "outputs": [], + "source": [ + "# Logistic regression classifier and predictions\n", + "scale_orig = StandardScaler()\n", + "X_train = scale_orig.fit_transform(dataset_orig_train.features)\n", + "y_train = dataset_orig_train.labels.ravel()\n", + "w_train = dataset_orig_train.instance_weights.ravel()\n", + "\n", + "lmod = LogisticRegression()\n", + "lmod.fit(X_train, y_train, \n", + " sample_weight=dataset_orig_train.instance_weights)\n", + "y_train_pred = lmod.predict(X_train)\n", + "\n", + "# positive class index\n", + "pos_ind = np.where(lmod.classes_ == dataset_orig_train.favorable_label)[0][0]\n", + "\n", + "dataset_orig_train_pred = dataset_orig_train.copy()\n", + "dataset_orig_train_pred.labels = y_train_pred" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "UbBLhkn16QV7" + }, + "source": [ + "#### Obtain scores for original validation and test sets" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": { + "id": "m5bS2VnM6QV7" + }, + "outputs": [], + "source": [ + "dataset_orig_valid_pred = dataset_orig_valid.copy(deepcopy=True)\n", + "X_valid = scale_orig.transform(dataset_orig_valid_pred.features)\n", + "y_valid = dataset_orig_valid_pred.labels\n", + "dataset_orig_valid_pred.scores = lmod.predict_proba(X_valid)[:,pos_ind].reshape(-1,1)\n", + "\n", + "dataset_orig_test_pred = dataset_orig_test.copy(deepcopy=True)\n", + "X_test = scale_orig.transform(dataset_orig_test_pred.features)\n", + "y_test = dataset_orig_test_pred.labels\n", + "dataset_orig_test_pred.scores = lmod.predict_proba(X_test)[:,pos_ind].reshape(-1,1)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "JTjNV-Xg6QV7" + }, + "source": [ + "### Find the optimal classification threshold from the validation set" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": { + "id": "yryZmeg16QV7", + "outputId": "a65af7f2-78ee-42fb-c1fd-185e5c0b350d", + "colab": { + "base_uri": "https://localhost:8080/" + } + }, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Best balanced accuracy (no reweighing) = 0.7463\n", + "Optimal classification threshold (no reweighing) = 0.2872\n" + ] + } ], - "text/plain": [ - "" + "source": [ + "num_thresh = 100\n", + "ba_arr = np.zeros(num_thresh)\n", + "class_thresh_arr = np.linspace(0.01, 0.99, num_thresh)\n", + "for idx, class_thresh in enumerate(class_thresh_arr):\n", + " \n", + " fav_inds = dataset_orig_valid_pred.scores > class_thresh\n", + " dataset_orig_valid_pred.labels[fav_inds] = dataset_orig_valid_pred.favorable_label\n", + " dataset_orig_valid_pred.labels[~fav_inds] = dataset_orig_valid_pred.unfavorable_label\n", + " \n", + " classified_metric_orig_valid = ClassificationMetric(dataset_orig_valid,\n", + " dataset_orig_valid_pred, \n", + " unprivileged_groups=unprivileged_groups,\n", + " privileged_groups=privileged_groups)\n", + " \n", + " ba_arr[idx] = 0.5*(classified_metric_orig_valid.true_positive_rate()\\\n", + " +classified_metric_orig_valid.true_negative_rate())\n", + "\n", + "best_ind = np.where(ba_arr == np.max(ba_arr))[0][0]\n", + "best_class_thresh = class_thresh_arr[best_ind]\n", + "\n", + "print(\"Best balanced accuracy (no reweighing) = %.4f\" % np.max(ba_arr))\n", + "print(\"Optimal classification threshold (no reweighing) = %.4f\" % best_class_thresh)" ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "Difference in mean outcomes between unprivileged and privileged groups = -0.000000\n" - ] - } - ], - "source": [ - "metric_transf_train = BinaryLabelDatasetMetric(dataset_transf_train, \n", - " unprivileged_groups=unprivileged_groups,\n", - " privileged_groups=privileged_groups)\n", - "display(Markdown(\"#### Transformed training dataset\"))\n", - "print(\"Difference in mean outcomes between unprivileged and privileged groups = %f\" % metric_transf_train.mean_difference())" - ] - }, - { - "cell_type": "code", - "execution_count": 9, - "metadata": {}, - "outputs": [], - "source": [ - "### Testing \n", - "assert np.abs(metric_transf_train.mean_difference()) < 1e-6" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Train classifier on original data" - ] - }, - { - "cell_type": "code", - "execution_count": 10, - "metadata": {}, - "outputs": [], - "source": [ - "# Logistic regression classifier and predictions\n", - "scale_orig = StandardScaler()\n", - "X_train = scale_orig.fit_transform(dataset_orig_train.features)\n", - "y_train = dataset_orig_train.labels.ravel()\n", - "w_train = dataset_orig_train.instance_weights.ravel()\n", - "\n", - "lmod = LogisticRegression()\n", - "lmod.fit(X_train, y_train, \n", - " sample_weight=dataset_orig_train.instance_weights)\n", - "y_train_pred = lmod.predict(X_train)\n", - "\n", - "# positive class index\n", - "pos_ind = np.where(lmod.classes_ == dataset_orig_train.favorable_label)[0][0]\n", - "\n", - "dataset_orig_train_pred = dataset_orig_train.copy()\n", - "dataset_orig_train_pred.labels = y_train_pred" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Obtain scores for original validation and test sets" - ] - }, - { - "cell_type": "code", - "execution_count": 11, - "metadata": {}, - "outputs": [], - "source": [ - "dataset_orig_valid_pred = dataset_orig_valid.copy(deepcopy=True)\n", - "X_valid = scale_orig.transform(dataset_orig_valid_pred.features)\n", - "y_valid = dataset_orig_valid_pred.labels\n", - "dataset_orig_valid_pred.scores = lmod.predict_proba(X_valid)[:,pos_ind].reshape(-1,1)\n", - "\n", - "dataset_orig_test_pred = dataset_orig_test.copy(deepcopy=True)\n", - "X_test = scale_orig.transform(dataset_orig_test_pred.features)\n", - "y_test = dataset_orig_test_pred.labels\n", - "dataset_orig_test_pred.scores = lmod.predict_proba(X_test)[:,pos_ind].reshape(-1,1)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Find the optimal classification threshold from the validation set" - ] - }, - { - "cell_type": "code", - "execution_count": 12, - "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Best balanced accuracy (no reweighing) = 0.7473\n", - "Optimal classification threshold (no reweighing) = 0.2674\n" - ] - } - ], - "source": [ - "num_thresh = 100\n", - "ba_arr = np.zeros(num_thresh)\n", - "class_thresh_arr = np.linspace(0.01, 0.99, num_thresh)\n", - "for idx, class_thresh in enumerate(class_thresh_arr):\n", - " \n", - " fav_inds = dataset_orig_valid_pred.scores > class_thresh\n", - " dataset_orig_valid_pred.labels[fav_inds] = dataset_orig_valid_pred.favorable_label\n", - " dataset_orig_valid_pred.labels[~fav_inds] = dataset_orig_valid_pred.unfavorable_label\n", - " \n", - " classified_metric_orig_valid = ClassificationMetric(dataset_orig_valid,\n", - " dataset_orig_valid_pred, \n", - " unprivileged_groups=unprivileged_groups,\n", - " privileged_groups=privileged_groups)\n", - " \n", - " ba_arr[idx] = 0.5*(classified_metric_orig_valid.true_positive_rate()\\\n", - " +classified_metric_orig_valid.true_negative_rate())\n", - "\n", - "best_ind = np.where(ba_arr == np.max(ba_arr))[0][0]\n", - "best_class_thresh = class_thresh_arr[best_ind]\n", - "\n", - "print(\"Best balanced accuracy (no reweighing) = %.4f\" % np.max(ba_arr))\n", - "print(\"Optimal classification threshold (no reweighing) = %.4f\" % best_class_thresh)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Predictions from the original test set at the optimal classification threshold" - ] - }, - { - "cell_type": "code", - "execution_count": 13, - "metadata": {}, - "outputs": [ - { - "data": { - "text/markdown": [ - "#### Predictions from original testing data" + "cell_type": "markdown", + "metadata": { + "id": "eI9uuQdJ6QV7" + }, + "source": [ + "### Predictions from the original test set at the optimal classification threshold" + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "metadata": { + "id": "heeWB3396QV8", + "outputId": "53f24f57-1e78-4843-d242-57e645748124", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 213 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Predictions from original testing data" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Classification threshold used = 0.2872\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + " 39%|███▉ | 39/100 [00:00<00:00, 70.87it/s]" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Balanced accuracy = 0.7437\n", + "Statistical parity difference = -0.3580\n", + "Disparate impact = 0.2794\n", + "Average odds difference = -0.3181\n", + "Equal opportunity difference = -0.3769\n", + "Theil index = 0.1129\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + " 70%|███████ | 70/100 [00:01<00:00, 57.13it/s]invalid value encountered in double_scalars\n", + "100%|██████████| 100/100 [00:01<00:00, 60.45it/s]\n" + ] + } ], - "text/plain": [ - "" - ] - }, - "metadata": {}, - "output_type": "display_data" - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - " 14%|█▍ | 14/100 [00:00<00:00, 134.43it/s]" - ] - }, - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Classification threshold used = 0.2674\n" - ] - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - " 41%|████ | 41/100 [00:00<00:00, 131.04it/s]" - ] - }, - { - "name": "stdout", - "output_type": "stream", - "text": [ - "Balanced accuracy = 0.7417\n", - "Statistical parity difference = -0.3576\n", - "Disparate impact = 0.2774\n", - "Average odds difference = -0.3281\n", - "Equal opportunity difference = -0.4001\n", - "Theil index = 0.1128\n" - ] - }, - { - "name": "stderr", - "output_type": "stream", - "text": [ - " 67%|██████▋ | 67/100 [00:00<00:00, 127.49it/s]../aif360/metrics/dataset_metric.py:94: RuntimeWarning: invalid value encountered in double_scalars\n", - " return metric_fun(privileged=False) / metric_fun(privileged=True)\n", - "100%|██████████| 100/100 [00:00<00:00, 121.44it/s]\n" - ] - } - ], - "source": [ - "display(Markdown(\"#### Predictions from original testing data\"))\n", - "bal_acc_arr_orig = []\n", - "disp_imp_arr_orig = []\n", - "avg_odds_diff_arr_orig = []\n", - "\n", - "print(\"Classification threshold used = %.4f\" % best_class_thresh)\n", - "for thresh in tqdm(class_thresh_arr):\n", - " \n", - " if thresh == best_class_thresh:\n", - " disp = True\n", - " else:\n", - " disp = False\n", - " \n", - " fav_inds = dataset_orig_test_pred.scores > thresh\n", - " dataset_orig_test_pred.labels[fav_inds] = dataset_orig_test_pred.favorable_label\n", - " dataset_orig_test_pred.labels[~fav_inds] = dataset_orig_test_pred.unfavorable_label\n", - " \n", - " metric_test_bef = compute_metrics(dataset_orig_test, dataset_orig_test_pred, \n", - " unprivileged_groups, privileged_groups,\n", - " disp = disp)\n", - "\n", - " bal_acc_arr_orig.append(metric_test_bef[\"Balanced accuracy\"])\n", - " avg_odds_diff_arr_orig.append(metric_test_bef[\"Average odds difference\"])\n", - " disp_imp_arr_orig.append(metric_test_bef[\"Disparate impact\"])" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Display results for all thresholds" - ] - }, - { - "cell_type": "code", - "execution_count": 14, - "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqUAAAG2CAYAAABGcv+aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4VFX6wPHvSW8QSELvSC9SBBHQDS7iIooFC6yiYgG7IspPReyNXcUuFlbExoKNRV1ARUVFUJGyCNJDTwgESAjp5fz+ODOQDCl3kpm5M5n38zzz3OSWc99Y4M0p71Faa4QQQgghhLBTiN0BCCGEEEIIIUmpEEIIIYSwnSSlQgghhBDCdpKUCiGEEEII20lSKoQQQgghbCdJqRBCCCGEsJ0kpUIIIYQQwnaSlAohhBBCCNtJUiqEEEIIIWwXZncAdgsJCdHR0dF2hyGEEEIIUa3c3Fytta6TnYpBn5RGR0eTk5NjdxgiyOzfvx+Apk2b2hyJEEKIQKKUyrM7Bm8J+qRUCDuMGTMGgKVLl9obiBBCCOEnJCkVwgb333+/3SEIIYQQfkVpre2OwVaxsbFahu+FEEIIEQiUUrla61i74/CGOjlRVgh/t2fPHvbs2WN3GEIIIYTfkOF7IWxw9dVXAzKnVAghhHCyJSlVSt0KTAaaARuAiVrrnyq5dzZwbQWXjndfK6WGAN9XcE9XrfUmT8QshCdNnTrV7hCEEEIIv+LzOaVKqdHAB8CtwDLH8Tqgm9Z6dwX3xwOuhUR/Bn7UWl/nuGcIJintDhwuc99BrXVJVfHInFIhhBBCBIq6PKfUjqT0V2Cd1np8mXNbgU+01g9YeH4wJpkdrLVe7jg3BJOUNtJaZ7gTjySlwg4pKSkAtG/f3uZIhBBCBJK6nJT6dKGTUioCOA342uXS18Agi82MBzY4E1IXvyul0pRS3yqlzq5FqEJ41fXXX8/1119vdxhCCCGE3/D1nNIkIBRIdzmfDpxT3cOOofwrANce1TTgFmAlEAFcDXyrlEquaK6qUmoCMAEgIiLCzR9BiNp77LHH7A5BCCGE8CuBtvp+LKZ39/2yJ7XWm4HNZU6tUEq1xSymOikp1Vq/BbwFZvjeS7EKUank5GS7QxBCCCH8iq/rlGYAJUATl/NNgP0Wnh8PfKq1PlztnfAr0NG98ITwjc2bN7N58+bqbxRCCCHcpJT6i1Lqc6XUPqWUVkqNs/BMT6XUD0qpPMdzDyullA/CPc6nSanWuhBYBQxzuTQMqGiO6HFKqdOBXsBMi6/rjRnWF8Lv3HTTTdx00012hyGEEKJuigPWA3cBedXdrJSqD3yDmU7Z3/HcZGCSF2M8iR3D988D7yulfsOUdroZaA68AaCUeg9Aa32Ny3MTgK1a66WuDSqlJgI7MTVPIzDD/BcDl3rlJxCilp5++mm7QxBCCFFHaa0XAgvheL336lwFxADXaq3zgPVKqS7AJKXU89pHpZp8npRqrecppRKBqZji+euBEVrrXY5bWrs+o5SqB4wBHq+k2QjgWaAl5jeCDcD5jn8pQnhcUUkpmblFNKoXWaPnBw2yWmxCeE1JCWzZYo5C1ETjxuYjROAbCPzkSEidvgKeANoCO3wRhC0LnbTWM4AZlVwbUsG5bExXdGXt/RP4p6fiE6Ks/KISNqRmsX7fUf5MPcqGtCy27D9GYUkpdw7tyKRhndxuc/369QD06NHD0+EKq154ASZPtjsKEcgefxweesjuKITwhKbAXpdz6WWu1d2k1J8kJCTI/uOinJJSTW5hCTmFxeQWlJBbVIJz5KJtiKJLs1Ci24RSWKzIOrCRj7/cQaM493pMJ06cCMCLL77o8fiFNT0/+YSY5s1JmTDB7lBEgMpp1Ypc+ftD+F6YUur3Mt+/5agqFPB8vqOTv5EdnURZn/8vlYlz11CqISxE0bNlPP3bJnBam4b0bBFPs/gonIsRS0o1d81dw5fr0njiou5cPbCt5fesXLkSgP79+3vjxxDVKS2FpCS49FKYaXXtpBBC2M/dHZ2UUseA27XWs6u45z0gUWt9fplz/YHfgPZaa+kpFcKXiktKee6rzXRuWp+HLuhKn1YNiY4IrfT+0BDFC6N7k19UykMLNhAdEcZlp7W09C5JRm22ZQscOQIDB9odiRBC+IMVwD+UUlFa63zHuWFAKmYhuU/4uk6pEH5r0fr97D6cy11DOzLolKQqE1Kn8NAQXr2yD2d1TOL/PvkfX65LtfSutWvXsnbt2tqGLGpqxQpzlKRUCFEHKaXilFK9lVK9Mblea8f3rR3Xn1FKfVvmkTlALjBbKdVDKTUKuB/w2cp7kOF7Gb4XAGitGfnqMnILSlgyKZmQEPfqBecWFnPtrN9YszuTzk3rnXQ9PDSEqPAQosJDiQoL5ctnbkJrGHLPq2TlFZGZW0RWXhHhoSG0ToihbWIMrRNjaZMQw+ntEmiVEOOpH1UATJgAH38Mhw5BiPxuLoQIHFaG75VSQ4DvK7j0rtZ6nKNM1BCtddsyz/QEXgNOB45gSnU+LkmpD0lSKgCWbc1g7Nu/Mm1UT8acflJVMkuy84t4euEmDmbnlzuvNRSWlFJQVEp+cQn5RSVk7NpCWIiiVcduxEeHH/8UFJew61Auuw7lkpaVR6mGepFhfD95CEluLqYSVTj1VGjRAhYtsjsSIYRwi7tzSgOJzCkVAnjzx+00rhfJJX1b1LiNelHhPDOqp8W7k6u9o7C4lHV7Mxn91i+8uGQLT15stW1RpaNHYf16s8hJCCGE35BxKxH01u/L4qetGVx/Zjsiw6qfR+oJK1euPL4CvzIRYSH0a5vAVQNa8+/f9rDtQLZPYqvzfvvNdF/LfFIhhPArkpSKoPfGD9upFxnGlQNqNmxfE5MnT2ayxcLtdw3tSEx4KM8s3OTlqILEihWgFAwYYHckQgghypDhexHUdh/KZeEfaYz/S3vqR4X77L2vvvqq5XsT4yK59ewO/GPxJpZvy2BQhyQvRhYEVqyAbt0gPt7uSIQQQpQhPaUiqM38KYWwkBCuH9zOp+/t0aOHW1uMXje4LS0aRPPUwo2Ulgb34sRaKS2FX36RoXshhPBDkpSKoJVxrICPft/DJX1a0KR+lE/fvXz5cpYvX275/qjwUP5veGc2pB5l/pp9XoysjpOi+UII4bdk+F7UeVm5RSzekMafqUc5klvEkdxCMnOLSMvKp7CklAnJ7X0e05QpUwBY6sa+2SNPbc6sZTt49qvNjOjZzFJxf+FCiuYLIYTfkqRU1Em5hcV882c6X/wvlR+2HKSoRFMvMoyEuAgaxkSQFBdBx8ZxnNkxiVMaxfk8vjfffNPtZ0JCFA+e340r3lzB28tSuP2vHb0QWR23YgU0aACdO9sdiRBCCBdSPF+K5weco/lF/G9PJmt3Z/K/vZlk5xeXu641/LEvi7yiEprWj2Jkr2Zc2KsFPVrURyn3dmryRze9/zvLtmbw+9Rh0lvqrp49oWVLKZovhAhYUjxfCJsVlZTyzMJN/Lj1INsPHkNrU9WnQ6M4EmIjyt2rFFzStwUX9mrO6W0T3N4y1Bd++OEHAJKTqy+i7+rKAW34akM6K3ce5i+dGnk6tLorKws2bIDLL7c7EiGEEBWQpFQEhJe/3cqsn3eQ3KkRF/VqTp/WDTm1VbxPyzh50iOPPAK4N6fUqX/bhkSEhvDztgxJSt0hRfOFEMKvSVIq/N6qXUd47fttXHZaS567vJfd4XjErFmzavxsTEQYfds0YNm2DA9GFASkaL4QQvg1KQkl/FpOQTGTPlpLs/hoHhnZze5wPKZ9+/a0b1/zVf9ndkhiQ+pRDucUejAqP/LFF7B3r2fbXLECuneH+vU9264QQgiPkKRU+LUn/7uR3Ydzef6KXtQL0KH6iixZsoQlS5bU+PnBjl2dlm+vg72lu3fDhReaXZdef90UvK8tKZovhBB+T5JS4be+3ZjOv3/bzYSz2jOgfaLd4XjUk08+yZNPPlnj53u2iKdeVBg/18Uh/NRUc0xKgltvheRk2LSpdm1u3gyZmZKUCiGEH5M5pcIvHTpWwH2frqNL03pMOreT3eF43Pvvv1+r58NCQxjYPrFuzis9cMAc582D9eth0iTo1QsefhhGjTLzQl0/1fnvf81RklIhhPBbkpQKv6O15oHP/uBoXjHv3zCAyLC6V4uzVatWtW7jzI5JfP1nOrsP5dI6McYDUfmJ9HRzbNIE+veH886DO+6AqVPNp6YSE6FT3fsFRwgh6gpJSoVfOZhdwP2fruPbTQeYMqILXZvVzUUpixcvBmD48OE1bsM5r3TZtgyuTGztkbj8grOntHFjc2zaFD7+GJYtM4uftC7/sapbNwiRGUtCCOGvJCkVfuObP9O5/9N1ZBcU89AF3bh+cFu7Q/KaadOmAbVLStsnxdIsPoqft2Vw5YA6lJSmp5sV8lFR5c+feaY98QghhPAJSUqF7XIKinniyz+Zu3IP3ZrV599jetOpST27w/KquXPn1roNpRSDOySxZGM6paXaL3euqpEDB8zQvRBCiKAiSamwVcrBY1w3eyW7D+dyy5BTuPucTkSE1f0h1qZNm3qknTM7JPHJqr38mXaUHi3iPdKm7dLTTwzdCyGECBqSlArbZOUWceO7v5OdX8y8CQM5vV2C3SH5zBdffAHAyJEja9XOoA6mVNaybRl1Jyk9cAC6dLE7CiGEED5W97ukhF8qLinl9n+vZs+RXN68+rSgSkgBpk+fzvTp02vdTuN6UXRuUq9u1StNT5fheyGECELSUyps8eR/N/LT1gz+eemp9G8bXAkpwCeffOKxtgZ3SOLDX3eRX1RCVHiAl88qKoJDh2T4XgghgpD0lAqfm/PrbmYv38mNZ7bjiv61r9cZiJKSkkhKSvJIW2d2TKSguJTVu454pD1bZTh6fKWnVAghgo4kpcKnfkk5xMML1pPcqREPjOhqdzi2+eyzz/jss8880tbp7RIJC1F1Y3cnZ+F86SkVQoigI8P3wmd2H8rllg9W0SYxhleu7ENoXSlhVAMvv/wyAKNGjap1W3GRYfRp3aBuzCt1Fs6XnlIhhAg6kpQKnzhWUMz4936nVMPb1/anflS43SHZasGCBR5tb3CHJF76diuXzPj5pGshShGiTF3TUKUIDwth7IDWnNvdM2WpPEp6SoUQImhJUiq8rrRUc/e8tWw7eIx3rzudtkmxdodku/h4z5ZvGtWnJev3ZVFQXHrSNa2hVGtKtaakVLMvI4cJ769iTP9WPHRBN2Ij/eiPAekpFUKIoOVHfxuJuurFJVv45s90HhnZjTM7emZxT6CbN28eAKNHj/ZIe60TY/jXtf0t3VtYXMoLS7bwxg/bWZFyiBdG96Zv64YeiaPW0tMhIsJsMyqEECKoyEIn4VVfrkvl5e+2cUW/lowb1NbucPzG66+/zuuvv27LuyPCQrhveBfmjj+D4hLN5W+s4IVvtlBccnIvq885txhVwTvfWAghgpXSWtsdg61iY2N1Tk6O3WHUSev3ZXHZG8vp3jyeOeMHEBkW4DU0PSg3NxeAmJgYW+M4ml/Eows28NmafZzeLoHXruxLo3qR9gV03nmmLNTKlfbFIIQQfkwplau1rpPz4KSnVHjFgex8Jrz3Ow1jInhj7GmSkLqIiYmxPSEFqB8VzvOje/Pi6N6s25vJBa/8xOrdNtY7PXBAFjkJIUSQkqRUeNSew7k8/sWfnP3sUg7lFPLW1f3s7XnzUx988AEffPCB3WEcd3GfFnx2y2AiwkIY8+YvzPl1tz2ByBajQggRtGT4XobvPeJ/ezKZ+VMKi9bvRwEXnNqMm4ecQpemsmClIkOGDAFg6dKltsbhKjO3kDvnruXHLQcZ078Vj17Y3Xdbl2oNkZEwaRJMm+abdwohRICpy8P3svpe1NrzX2/m5e+2US8yjBvPbMe1g9rSvEG03WH5tW+++cbuECrUICaCd8b154VvtvDq99tQSvHMqJ6+eXlmJhQVSU+pEEIEKUlKRa19tmYfA9sn8tY1p1EvyIviWxUe7r//nEJDFPf+rTOpWXl8uS6Vxy7sTkSYD2b6SOF8IYQIajKnVNTKgex89h7JY2jXxpKQumH27NnMnj3b7jCqNKJHM7Lzi1mRcsg3L5TC+UIIEdQkKRW1snpXJgB9/KX4eoAIhKT0zI5JxEaEsnh9mm9e6OwplaRUCCGCkiSlolbW7D5CRGgIPVrIgiZ3LF261O8WObmKCg/l7C6N+XpDOiWlPlgQ6ewpleF7IYQISpKUilpZtesI3VvUlzqkddTwHk05lFPI7zsPe/9l6elmJ6ck2YpWCCGCkSSlosYKi0tZty/Lf/ZNDyAzZ85k5syZdodRrbM7NyYiLIRF6/d7/2UHDpiENFR+wRFCiGAkSamosT/TjlJYXCpJaQ3MmzePefPm2R1GtWIjw/hLx0Z8tWE/Xq9pLIXzhRAiqElJKFFjq3eZ7Sj7tmlgcySBZ8mSJXaHYNnwHk1ZsjGddXuz6NXKi/+u09NlPqkQQgQx6SkVNbZ69xGax0fRLF4K5ddl53RtTFiI8v4Q/oED0lMqhBBBTJJSUWNrdmfSp40M3dfEjBkzmDFjht1hWNIgJoKBpySyeH2ad4fwpadUCCGCmiSlokbSj+azLzNP5pPW0BdffMEXX3xhdxiW/a17U3YeymVL+jHvvCA3F44dk55SIYQIYpKUiho5Pp+0tcwnrYlFixaxaNEiu8Ow7NzuTVAKFnmrkL7s5iSEEEFPklJRI6t3HyEiLITuzePtDkX4QON6UfRr05DF3ppXKoXzhRAi6ElSKmpk1a4j9GwRT0SY/CdUEy+99BIvvfSS3WG45W/dm7JpfzY7M3I837hsMSqEEEFPSkIJtxUUl7B+31GuHdTG7lAC1rfffgvAXXfdZXMk1g3v0ZQn/7uRj1ft4dK+Lctdyy0sIS0rn7SsPFIz89mflUdRiaZvm4ac0T6Brk3rExKiKm9cekqFECLoSVIq3LYh9SiFJVI0vzY+//zzmj9cXAxHjkCjRp4LyIKWDWM4tWU8r32/nde+317pfeGhiib1owD47x9mDmp8dDint0ugT+sGJ21JGxGquGjXXuqDJKVCCBHEJCkVbjtRNF+SUlu8+Sbcdx/s2OHzxPTVv/dlzZ4jJ52PDAulWXwUzRpEkRQbebxXNDUzj193HOKX7Yf5ZcchvvkzvcJ2i79dw5ioWP7Yn0f/tlEoVUWvqhBCiDpJklLhtjW7M2nRIPp4b5hw33PPPQfAvffe6/7Da9dCTg589BHcdpuHI6ta68QYWifGWL6/eYNoLunTkkv6mOH+YwXFlJSWr3V66FgBeStmcCAmniveXEHXZvUZN6gNl5/WquohfyGEEHWKLatUlFK3KqV2KKXylVKrlFJnVXHvbKWUruCT43JfsqOtfKVUilLqZu//JMFp9e4j0ktaSytWrGDFihU1ezglxRw//NBzAflIXGQY8dHh5T7tG8XRPTSPlp3bMG1UT7TW3PfpH8z5bbfd4QohhPAhnyelSqnRwEvA00AfYDmwSCnVupJH7gKauXxSgI/KtNkOWOhoqw/wDPCKUupSL/0YQSstK4+0rHypT1pLn376KZ9++mnNHt6+HcLCYMWKEwlqTWgNe/fCp5/CQw/BJ5+Y+ap2OHCA0KZNGXN6axbddRbtkmL5dmPFQ/1CCCHqJjt6SicBs7XWM7XWG7XWdwBpwC0V3ay1ztJa73d+gFOA9sDMMrfdDKRqre9wtDkTeBeowdioqMrqXZkAssjJLoWFsGcPjB1rvne3t1RrMyf1kkugRQto1QouuwyefBIuvxw6dIAXXoCjRz0fe1XS04+Xg1JKkdypEStSDpFfVOLbOIQQQtjGp0mpUioCOA342uXS18Agi82MBzZorZeXOTewgja/AvoppcJrEquo2OrdR4gMC6Frs/p2hxLQpk2bxrRp09x/cNcuKC2FIUMgOdkkpe7sRz9/Ptx8M/zxBwwdCi+/DL/+auaozp8PbdrApEnQsiXcfTfs91Kx/LKKi+HQoXIr75M7NSK/qJSVOw97//1CCCH8gq8XOiUBoYDruFw6cE51Dyul4oErgAdcLjUFllTQZpjjneX2RlRKTQAmAERERFgMXWxMO8ri9fs5taUUza+ttWvX1uxB53B9+/Zw1VUwYQKsWgX9+lX/bHExPPggdO1qktLQ8qWZuPhi81m1yvSWvvqqSVh//hm8uRo+I8Mk1mUK5w9on0BEWAg/bD7IWR19W2FACCGEPQItsxiLifn92jSitX5La91Pa90vLEwKEFSnuKSUGUu3ceGryygoLuHeczvbHVLAmzt3LnPnznX/QWdSesopZtg9IsL6EP5778GmTfDUUycnpGWddhp88AG89pqZt7pwoftxuqOCwvkxEWGc3jaBH7ce9O67hRBC+A1fJ6UZQAngupdgE8DKOOF44FOtteuY3v5K2ix2vFPUUMrBY1z+5gr+uXgzw7o14eu7kxnQPtHusILX9u0QFQVNm0LDhnD++TB3bvULlPLz4ZFH4PTTTW+oFdddZ5LfqVPNlAFvqWSL0eROjdiSfozUzDzvvVsIIYTf8GlSqrUuBFYBw1wuDcOsnK+UUup0oBflFzg5raikzd+11kU1i1Z88MsuRrz8EykHc3hpTG9eu7IvCbEy3cETnnjiCZ544gn3H0xJMUP3IY7/da+6ysz7/O67qp+bMcOstJ82zfpQfHg4PPqoqYv62Wfux2pVJVuMJnc2w/Y/bpHeUiGE8DtKRaLUVSj1PkptQamjKJWNUltRag5KjUOpaHeatGP4/nlgnFLqRqVUV6XUS0Bz4A0ApdR7Sqn3KnhuArBVa720gmtvAC2UUi862rwRGAc855WfIAi89v02pv5nPQPaJfL13X/hot4tZJcdD9q8eTObN292/8Ht201S6nT++RAfX/UQflYWPP00nHsunH22e+/7+9+hWzd4+GEo8dJK+Ep6Sjs2jqNp/SgZwhdCCH+iVCxKPQakAu8BVwIdgDggFlMlaTTwNpCKUk+gVKyVpn0+oVJrPU8plQhMxdQcXQ+M0FrvctxyUr1SpVQ9YAzweCVt7lBKjQBewJSWSgXu1FrXsBBkcHvjh+08+9VmLu7dnOlX9CZUdtXxuA8++MD9h7Q2PaVDhpw4FxVl5pbOmwevvw4xFey2NH26Wd3+9NPuvzM0FB5/3Lzjww/hmmvcb6M66elmbmx8fLnTztJQC9enUVxSSlhooE2BF0KIOmkb0BhwJgcZwDpOTJdMAk51HOOBKcANmA7IKtnyp7zWeobWuq3WOlJrfZrW+scy14ZorYe43J+ttY7TWv+zijZ/0Fr3dbTZTmv9hhd/hDpr5o8pTFu0iZG9mvPc5b0kIfUnBw/CsWNmnmdZV11lzn/++cnPHDgAzz8PV1xhFjDVxCWXQJ8+Zii/yAuzYQ4cMEP3FfTEJ3duRHZ+MWv3ZHr+vUIIIWqiCbAH01HYDa0bo/U5aD3G8TkHrRsD3Rz37OXkdT8Vkq4Hcdzby3bw1MKNnH9qM164opf0THnRww8/zMMPP+zeQ2XLQZWVnGzqilY0hP/UU2aRU03mrzqFhJji+jt2wKxZNW+nMmUK57safEoSIUrmlQohhLvc2dLdcf+VSqm1SqlcpdR+pdQHSqmmFdx6A9ABrR9F602VNqj1JrR+FDOcf6OVmCXrEADM/nkHT3z5J+f1aMqLo3tLQuple/bsYc+ePe49VLYcVFkhIWbu58KFZv5n2c+MGXD99dCpU+0CPu88GDTIJLf5+bVry5Wzp7QC8THh9GndkB8kKRVCCMvc3dJdKTUYU27zXaA7cDGmp/Pk3g6t30Fr63tSa12M1u9YuVWKdAo2pGbx6Bd/cm63Jrz89z6ES0Lqde+8Y+n/z/K2bzfHtm1PvnbHHZCaarYhLatfPzMntLaUMr2lf/2reVffvuWvR0eb3aDatDFbl4a7sZFaejr07Fnp5eROjXhhyRYO5xRK9QchhLDm+Jbuju/vUEoNx6y7cd2ACMzOmHu11i84vt+hlHoFeKXKtyj1HaDRemgF18xwoNaW/xKSpFTw8rdbqRcVxrOX95KE1J+lpJj96qMrqLDRqpUpeO9NZ58Nw4fDv/5V9X0hIdC8OfToYbYqHTas8jJUWlfZUwomKX3+my38tPUgF/VuUYsfQAgh6r4yW7q7ViCqakv3n4GnlVIjgS+BRMwC8+p2TxkCVLbX9aNAKZUsUq9I0CelCQkJLF261O4wbJNfVEqPkGyGnhbFml9/tjucoDFzpvnldfz48Zaf6b16NSQmstbG/17VvfcSVkHMYXl5RO7fT1R6OlGOY8Pffyfyb3/jaJcu7Bo7lkMDB56or+p87tgxziwsZFt2Nnur+Lnu61VC9o51LM3c6ukfSQghAk2YUur3Mt+/pbV+q8z3bm/prrVeoZQagxmuj8bkh98A19YoQqUaOL9y57GgT0oPHz7MkLIldoLMze+v4uftRSy776/ER7sx5CpqZc6cOQDu/bd36BAMGxY4/70WFMB771F/2jR6Tp1qhujvugsaldnLfr/ZyK3DoEF0qOLn+uzfa1i+4RC/TUkmRCpCCCGCW7HWup8nG1RKdcMM1T8BfIUp2fks8CZwjcvN1+KarJph/LLaOI5H3Ikj6JPSYPZn6lEWb9jPXUM7SkLqY2+99Vb1N5WVlwf79p288t6fRUbC+PFmu9K5c00lgBsrWYDpunjLRXKnRnz+v1Q27j9Kl6b1KSoppbCklKLiUkorGzhyER8dTkSYTE8RQtR5NdnS/QHgN631s47v1ymlcoCflFJTtNZ7y9zblvLD9gpIdmnP2Xuwwp3AJSkNYi9/u5V6kWFcP7id3aGI6uzcaY6BlJQ6hYXB2LFw5ZWwfj0UuyzajI2ttjrAWZ2SADj/5WU1DuOURrF8fvuZxEbKH3tCiLpLa12olHJu6f5xmUvDgMo2FYrBJLJlOb+v7Ld5RfnEtKxDwC/AHVZidpI/nYPUxjTTS3rn0I7Ex0gvqa/de++9ADz3nMWdcCsrBxVIQkLg1FNr9GjjelH887JT2X0ol/DQEMLDFBGhIYSHhlgazj+aV8SzX23m5W+38sBpEvQhAAAgAElEQVSIrjWKQQghAsjzwPtKqd8wi5huxmVLdwCttXNo/gtgplLqFk4M378IrNZa7y7XstaPAY9hGirFrL73yDCUJKVBytlLeoP0ktoiLy/PvQec5aACsafUQ67o16pWz+86lMPby3Ywqm9LOjet56GohBDC/7i7pbvWerZjS/fbgelAFvAdcF81r7rOk3ErrS1OyKqjYmNjdU5Ojt1h+NTGtKOc99JP3Dm0I5OG1bKouvCNiRPh7bfh6NHKyyuJKh3OKeSv05fSqXE95t10Bkr+OQohApBSKldrHWt3HAAo1Q5oBRxE641lzncFGgF70HqH1eZk1n8QeuU76SUNONu3m15SSaRqLCE2ggfO68JvOw/zyaq91T8ghBCiOq8B3wOnu5zv5zj/qjuNSVIaZPYczmXhH/u5bnBbmUtqo4kTJzJx4kTrD6SkBPZ8Uj9x+WmtOK1NQ55ZtIkjOYXVPyCEEKIqzu39FrmcX4xZ/OSy/V/VJCkNMjsPmakKZ3ZsVM2dwm+UlpqkNIjnk3pKSIjiyYt7kJVXxD+/2mR3OEIIEegaOo75Luedv/UnuNOYLHQKMgezCwBIipM9xO304osvWr95/37Iz5ek1EO6NqvP9YPbMvOnHVzm6DkVQghRI0cwc0cvB94uc/7SMtctk57SIJNxzCSljepF2hyJsKwulIPyMxPP6USz+Cim/mc9wb7YUwghauEXzDD9DJR6G6XuQal/Aa9japj+4k5jkpQGmYxjhUSGhRAnBcRh61ZYuNCWV992223cdttt1m6WclAeFxsZxq1nd2Bj2lFSMoKr+oYQQnjQS5jkMwwYB/wTUyYq3HH+BXcak6Q0yGRkF5AUFynlcADuvhtGjTJ7tPtYdHQ00dHR1m5OSTGF59u0qf5eYdmgUxIB+H3nYZsjEUKIAKX198BEoAjTY+r8FAJ3o/UP7jRnqbtMKQZoza9uhir80MFjBTJ0D5CZCV9/DUVFsGYNnHGGT19veScnMD2lrVpBhMwD9qT2SbEkxkbw244jjO7fuvoHhBBCnEzrV1DqM2A40ARIBxaj9T53m7I6hrtCKf4A/gV8oLV7E1eF/ziYXUDLhjF2h2G/L74wCSnA8uU+T0rdIuWgvEIpRb+2DVkpPaVCCFE7JgF9u9r7quHO8H0PzD6o+5TiQ6U4u7YvF76XcayQRvWkx41PPoGWLaFtW5OU+tiECROYMGGCtZudhfOFx/Vvm8Duw7mkH3WtZiKEEMISpRJQ6k6Ueg2lZrl83EpUrfaUPg9cAbQEooAxwBilSMFkxrO1Zr87Lxa+V1KqOZxj5pQGtaNH4auv4JZb4OBB+P570NqnuyUlJiZau/HYMThwQJJSLzm9nSmh99uOw4zs1dzmaIQQIsAodQqwDGhc0VXMYqcbrDZnqadUa+7VmtbAXzDL/A86XnYK8BSwWyk+VoreVl8sfO9wTiGlWspB8d//msVNl10GgwZBairs3u3TEJ555hmeeeaZ6m/c4dgyWIbvvaJbs/rERITKEL4QQtTMo5h5pKqCj9vcWn2vNcu05jagP1B2RVUYMAr4VSkuqkkgwvucNUqDvqf044+hWTMYONAkpWDLEL4lUg7Kq8JCQ+jbuiErd8o0eSGEqIFkTG/oU47vNXAhpj7pFuA8dxpzKylVimFK8SmwDdNrCiYbXgOkYOpSPVXJ48JmkpRihsMXLYJLLzVllnr0gLg4nyel1113Hdddd131N0rhfK/r3zaBTfuPkpVXZHcoQggRaJo4jtOPn9H6S+DvQCfMinzLLCWlSjFZKbYCi4GLOVEU9T9AstacBvQGjjqCEH5IthjFFMvPzzdD9wBhYTBggM+T0latWtGqVauqb9Iali6FhARoKFthekv/dg3RGlbvkt5SIYRwk3OP+6OASTKUaglkO85f5U5jVhc6/QOThCrHi2cBL2vNTucNWpOjFPuBju4EIHxHthjFrLpv0gTOPPPEuYED4ZlnTC9qXJxPwnj88cerv2nOHFO66umnvR9QEOvTqiFhIYrfdh7m7C4VzdUXQghRiQygNZAA7AXaAwtxJqjgVsLhzvD9DuBuoKXWTCqbkJbxV0dAwg8F/RajublmkdOoURAaeuL8oEFQUgIrV9oXm6vdu+G220xs//d/dkdTp0VHhNKjRTwrd8hiJyGEcNMGx7Ez8A2m87I70BfTmbnMncasJqWXAB215iWtOVbZTVqTqjW73AlA+E7QbzG6aJFJTJ1D907OwvnuDOFrbeZ7al2jUMaOHcvYsWMrvlhaCuPGmUT5/ffLJ9DCK05vl8C6vVnkF5XYHYoQQgSS54EpQC7wGLCRE6vvNwF3utOY1aR0KdBKKZLKnlSKJKVorRTx7rxU2OPgsQKSgn3oPikJ/vKX8ucbNoRu3awnpTt2wAUXmMVHt99ukkg3de7cmc6dO1d88aWXTO3UF1+UVfc+0r9tAoUlpazbm2V3KEIIETi0/g6t/4HWa9A6HbPRUh+gJ3AqWqe405zVpHQWZvj+SpfzYxzna721lPC+g9kFNArWlfd5efDll3DJJWZxk6tBg2DFiqoTzMJCM/e0e3f48UcYORJmzICbbjK9mm546KGHeOihh06+sH49PPAAXHghXH+9W22KmuvXxiwkk3qlQghRA0olotQlwI2YaZz70drtoSerSekAx/FTl/OfYbpoByD8XlBvMfr112Yhk+vQvdOgQXDkCGzeXPH1n36CPn1gyhQ47zzYuBEWLICpU+Ff/zLD7cXF1uMpKYEilxJEBQUwdizUrw8zZ/p0h6lg1zA2go6N4/hN5pUKIYR7lHoUs8jpE+ANx3EvSj3iblNWV7w0chwzXc5nuVwXfiqotxgtLYXXXzellc4+u+J7yhbR79q1/LUPPzTJYps2ZjX8BRecuPbEExAVZZLTggJzb3h45XH8+CPMmcOY2bOhqIi5MTEQH28+WpukeMECaCyrwH2tf7sEvlibSkmpJjREfiEQQohqKTUZeLiCK5HAwyh1DK2nV3C9QlaT0mygIXAuML/M+XMdx0oXPwn/4NxiNOiSUq1h0iSz1/0LL1SeMHbqZJLW5cvhhjLb9KakwC23mBJSixdDbOzJzz74oElM773XJKYTJ5a/XlhoemrnzYN9+yA2lt7du5v5rb16QVaW+WRmwo03mqF74XOnt01gzq+72Zh2lB4tZJq8EEJYcJvjmIfJD3djSkRdAkQDd1C2sH41rCalq4FzgFlK0R2zuqorMAmz5H+V1RcKewRtjdLnnjMLhyZOhLvuqvw+pUxvadnFTsXFcNVVZuenDz+sOCF1uucek5jefjt8/vnJ18PDYcQImD4dRo7k/piYmv9Mwiv6t0sA4PedhyUpFUIIa5pg8sCL0HrJ8bNKDQO+Atwa9rOalL6BSUrrY5b8H3+tI5g33Hmp8L2g3GL0gw9Mjc/Ro00yWN0czUGDzGKow4dNr+kTT8Avv8DcudC6dfXvu+02+Otf4cCB8ueVgp49ZVcmP9eiQTTN46NYufMI4wa3szscIYQIBBuBXpi97sta4Tiud6cxS0mp1nymFM9jekZdTde63JC+8EMnktIgWej09ddw3XVmDum775rezuo455X+8ouZ4/nkk3DNNSaptapr15PnpFbg0ksvBeDTT13XDgo79W+XwPLth9h2oPoZSZFhIbRKkB5vIURQmwp8DtwK/LPM+VuBIkwNU8ssb+2jNfcqxTzgQkx3bTrwudb40TY4ojLH972vC8P3RUWQnW3mYR49enI5prQ0GDPG1B6dPx8iLf7M/fubQvWLFpkFTW3bwiuveDx8gIEDB3qlXVE7Z7RPZMHaVM55/gdL9089vys3niW1ZIUQQWsyZtH7Myh1O7AHaOn4HASmoJQzMdVoPbSqxpSu4Y40dUVsbKzOycmxOwyve3rhRt5dvpNNTwwPzB2dcnLMHvXbtpmao9Vp08bMD23e3L339OsHq1aZ5HTZshO7PYmgUFBcwvebDlJYUv2GCB//voeVOw/zzd3J0mMqhPAZpVSu1rqKRQ4+pFQpZhpntXdiktIqtyi03FOqFGHACMz+ptGu17XmcattCd8L+C1GFy2CP/4wBeXbtjXD6/XrQ716Fa+oHzgQGtWgUtmgQSYpfeQRSUiDUGRYKMN7NLV0b782DRn2/A88vGA9s8b1D9z/t4QQonY89oefpaRUKRpjthqtZF9EAElK/VnAbzE6f74pofTmmxXvyOQpt95qFiQ98ID33gFc6Cj79HlFK/VFQGjeIJq7h3Xiyf9uZNH6/Yzo2czukIQQwre0troJkyVWG3sM6ILJhiv6CD8X0FuMFhaaVfEXXujdhBSgSxd47DGvv2fo0KEMHVrl1BoRAMYNaku3ZvV57IsNZOcXVf+AEEKISln9m/dczJyB2cB1jq/vwhRF1cA0bwQnPCfjWCF9WjewO4ya+e47s6DpkkvsjsRj7qqqZqoIGGGhITw9qieXzPiZ6V9v4dELu9sdkhBC+JZSIZjt5ltjdnIqT+v3rDZlNSlt4Tjej0lK0ZpXleJ74A/MKivhpwJ+i9H58yEuDs45x+5IhDhJ71YNGDugDe+t2MmlfVvSs6UU3hdCBAmlugILgFMquUMDlpNSq8P3zpo7hzB1p1CKRsAux/kJVl8ofC+gtxgtKTF7wY8YYXZMqiPOO+88zjvvPLvDEB4yeXhnEuMimTL/D0pKg7uiiRAiqMwAOlD59E63pnha7Sk9hOktjQf2Y3pGPwTyHddlqxo/FtBbjP7yC6Sn16mhe4CRI0faHYLwoPpR4Tx8QTfu+PcaXv52K3cP62R3SEII4QunYXpD/wMsBgpr05jVpHQzJik9BfgRuApwrtLQwOraBCG8K6C3GP3sM4iIMD2ldcitt95qdwjCwy44tRnfbzrAS99uJTI8hFuHdLA7JCGE8LZ0oD0wDq2za9uY1eH7mcBbQBRmJf5BTnTLZgATaxuI8J6A3WJUazOfdOhQU5NUCD+mlOLZy3txUe/m/HPxZl5fut3ukIQQwtuexuSC96JUrXu+LPWUas1HwEfO75WiI3A2UAz8rDWZtQ1EeE/AbjG6bh3s2AFT3No6NyCc41i0tWTJEpsjEZ4UGqKYfnkvtIZ/LN4EwC1DKpv/L4QQAU7rd1DqImAqcB9KHcDkhsfvQGvLfwhWm5QqRSTwp+Pb87Vmk9Ycxay2EgEg41ghkWEh1Iv0co1PT5s/H0JCTH3SOmb06NF2hyC8JCw0hOev6IXGJKZKwc3JkpgKIeogpR4ALsRM5YzgRLUmcG4t6oZqsxStKVCKRKAekOJO48I/BOwWo599BoMHQ+PGdkficePHj7c7BOFFYaEhvHBFLwCmLdrE4vX7iQit2cYnSfUiuHVIB3q0kFJTQgi/c4fjqFyONWL1T0nnGGOv2rxM2CMgtxjdvt3sdV/HVt2L4OFMTMef1Y7o8FBCQ1SNPiu2H+KCV5Zx57/XsOdwrt0/lhBClBWH6Q29BIhB6xCXT6g7jVkdz30RSAb+rRQPAmuBvLI3aM1ud14sfOdgdgEtG0bbHYZ75s83xzqalA4ZMgSApUuX2hqH8K6w0BAePL9brdo4ml/Emz9s5+1lO1i0Po2xZ7Th9rM7kBiI1TSEEHXN58DfgZVonV/dzdWxmpT+iMmEE4A5FVzXbrQlfCwgtxidPx/69IG2be2OxCvGjRtndwgiQNSPCmfy37pwzcC2vLhkC+8u38ns5TsJreF0nNAQxeMXdWd0/9YejlQIEYQ+wWxFvwilXgJ2Un6hE2j9o9XGlNbVz0FVitJqbtFa41YXrb+IjY3VOTk5dofhNSWlmo4PLuS2sztwz7md7Q7HmpQU6NABHnsMHnrI7miE8CvbDmTz+f/SKCmt7o/liv20NYOdGTn89H9/JT4m3MPRCSG8TSmVq7WOtTsOAJQqperFTBqtLXdaWr3xXasNCv9yJDfAthjNzYXLL4d69eDqq+2OxmuKiooACA+XpEC4p0PjekwaVq/Gz5/f8yjnv/ITr/+wnfvP6+LByIQQQcpjq6it1im9zlMvFL51vEZpICSlWsP48bBmDXz+eZ0dugcYNmwYIHNKhe91a16fi3u34J2fd3DtoDY0iw+w+eZCCH/i0U5LmQdaxwXUvvfTp8OcOfDkk3DBBXZH41U33nij3SGIIDZpWCe+XJfKS0u2Mu3SU+0ORwgRqLT2aKelpaRUKWZVc4vWmhs8EI/wML/aYnTrVpg5E/7+d7OIqayvvoL77oPLLquTOzi5Gjt2rN0hiCDWKiGGsWe04d3lO7nxrPZ0aBxnd0hCCOHWQqfKblS4udBJKXUrMBloBmwAJmqtf6ri/gjMFlZXA82BdOA5rfXLjuvjgHcqeDRaV1OioK4vdJr5YwpPLdzIukfPpX6UzfMXJ0+G554zX591Ftx1F1x0EezcCf37Q6tWsHw5xNX9vyBzc029yZiYGJsjEcHq0LECkp9dyuAOibx5dT+7wxFCWGT7QielZmEWMN3g+Loq5j6L3Bm+98hEVqXUaOAl4FZgmeO4SCnVTWtdWa3TuUBLYAKwFWgCuE6EygXK7eVXXUIaDA4eK/CfLUZXr4aePeHaa+HVV02vaOvWEBZmthNdsCAoElKAESNGADKnVNgnMS6S8We154UlW1i9+wh9Wze0OyQhRGAYB5QCNzi+rq5303JSanVHp3Yun47A3zBJZQ7wV6svBCYBs7XWM7XWG7XWdwBpwC0V3ayUOhcYCozQWn+jtd6ptf5Va73U5Vattd5f9uNGTHWW32wxqrVZwHTGGXDPPbBtm6lF2r497NsHH30E7drZG6MP3XLLLdxyS4X/yQvhMzee1Y6kuAj+sWgTVkbNhBDCQbl8XdnHLVZX3++q4PR2pVgBZGASyh+qa8cxDH8a8JzLpa+BQZU8djGwEpiklLoGs5PUImCK1vpYmfuilVK7gFDMjlMPaa3XVBdTXec3W4zu2gVHjkDfvub70FC4+GLzKSqCICuNNHr0aLtDEILYyDDuHNqRhxdsYOnmg5zdpbHdIQkh/F+7Sr6uNas9pZUJw3TbDrd4fxImaUx3OZ8ONK3kmfbAmUAv4FLgdsf7Zpe5ZzNwPXARZrurfOBnpVTHihpUSk1QSv2ulPq9uLi4olvqjIxjhTTyh0VOaxy/H7gucIKgS0gBsrKyyMrKsjsMIRjTvzVJcZEsWLvP7lCEEIFA611ovavc11V93FCb1fdRwGAgEvDm364hmMT3Sq11lolH3Q58pZRqorVO11qvAFaciFctx/SW3gHc6dqg1vot4C0wC528GLvtDmYX0LtVvN1hmPmkoaFwqpSfAbjooosAmVMq7BcRFkLrhGgOOip1CCGEXayufhlHxRNZnfMFFlpsJwMowSxUKqsJUNkc0DRgnzMhddjoOLbm5F5XtNYlSqnfMXNfg1ZJqeZwToF/FM5fswa6dIFoKdQNcOedJ/2uJIRtkuIi2XUo1+4whBBBzp3h+4omsBZghtEnWmlAa10IrAKGuVwaBiyv5LGfgeZKqbLLsjs5jhV2CyuzqudUTEIbtPxqi9HVq0/MJxWMGjWKUaNG2R2GEIBZiZ8hPaVCCJtZ7SmtaCJrgdaV9m5W5XngfaXUb5iE82ZM7dE3AJRS7wFora9x3D8HeAh4Ryn1KNAAU1LqE631AcczjwC/YMpF1ccM2Z9KJSv6g4XfbDG6fz+kpVU8nzRIZWRkAJCUlGRzJEJAo7gIDucWUlxSSlhobZcaCCFEzdRm9X2NaK3nKaUSMcXwmwHrMeWenO9o7XL/MaXUOcArmFX4R4D/APeXua0BZo5oU8z81jXAX7TWv3kq7kCUftSUabV9i1HnIifpKT3usssuA2ROqfAPSfUi0RqO5BbZ/+eFECJoWV3oNBw4HVijNV+UOX8h0Bv4TWsWW32p1noGMKOSa0MqOLcZOLeK9u4G7rb6/mCQU1DMtEWbqB8VRke7txB0JqW9e9sbhx+555577A5BiOOcoykZxwokKRVC1JxS0WidV9PHrQ7fPwwMAM5zOX8MeBSz8t1yUiq8S2vNvR//jy3p2bxz3ek0jLW5JNTq1XDKKRDvB1UA/MTIkSPtDkGI48ompUII4RZTfvNZzPqgSCAMpV7ETKecjtYbrDZlNSnt4jiucDnvHB7vavWFwvte/W4bi9bv58ERXUnu1MjucExPqQzdl7N/v5mO3bRpZeV5hfCdREctY0lKhRBuUaoNJjdsiFkA76zUVARci1lw/qDV5qzOaI9xHF3Hgeu5XBc2+3rDfqZ/s4VL+rTgxrP8YNvOzExISZGk1MWYMWMYM2aM3WEIAZTpKc0utDkSIUSAeRRIwCShZX2CSVLPcacxqz2laZgFSA9idlRymuI4prrzUuEdW9KzuXveWk5tGc8zo3rav989wNq15igr78u5//77q79JCB+pHxVGRGgIGTnSUyqEcMu5mN7RvwHflzn/h+PYxp3GrCalS4AbgFuU4lzMtp6dgVMcwSxx56XC8w7nFDL+vd+JiQzjrav7ERUeandIxurV5ihJaTnDh1vdmVcI71NKkRQXIT2lQgh3OecIutaad/aKNXSnMavD99Mwi5rAJKIjHEcF5DiuCxtorVn4RxrnvvAjaZn5vDG2L03jo+wO64Q1a6B5c2jiuolXcNuzZw979uyxOwwhjkuqJwX0hahLlFK3KqV2KKXylVKrlFJnVXN/hFLqccczBUqp3Uqp6rYfPOw4tnU5f6HjeMidmK3WKd3u6CF9m/KLmv4EbtSaFHdeKjxjf1Y+Dy1Yzzd/ptOjRX1mX9efHi38bIW77ORUoauvvhqQOqXCfyTGRnAgW5JSIeoCpdRozEZDtwLLHMdFSqluWuvdlTw2F2gJTMBsRtQEqG5v8BWYBHROmZe/CVyDGUlf5k7cVofv0ZpfgO5KcYoj0HSt2e7Oy4RnlJZq5vy2m38s2kRRaSlTRnTh+sHt/G8nltxc2LQJLr3U7kj8ztSpU+0OQYhykuIi2ZiWbXcYQgjPmATM1lrPdHx/h1JqOGanywdcb1ZKnQsMBU7RWmc4Tu+08J5/ABcAfTmx8v5GzEh6CTDdnaAtJ6VOjkRUklEbzfwphWcWbWJwh0SevqQnbRJj7Q6pYuvWQWmp9JRW4Jxz3FqQKITXJdWL5FBOAVpr/1gkKYSoEaVUBHAa8JzLpa+BQZU8djFm18xJSqlrgDxgETBFa32skmdA619Q6irMhkgJZa4cAW5D61/did3qjk4fAmOAR7XmiTLnH8KUA5ijNVe782J/kZCQEHBDqNGHc3mgTymdm+Sz44+V7LA7oEo0X7CATsCKggIKAuyfsbelppqCFc2bN7c5EiGMU4oLubNbEd99v5TQEElKhfBjYUqp38t8/5bW+q0y3ycBoUC6y3PpVF6iqT1wJlAAXIrZvv0VoDlwWZXRaP0RSn0BDAYaAweA5Wida+mnKcNqT6kzs/7A5fz7wGOYHyQgHT58mCFDhtgdhlsufu1nYiNDuWn0GXaHUrUPP4SEBAZecQVIz0s5zv/mAu0XIlF3LVi7j+nL1rJkWD86NK5X/QNCCLsUa637ebjNEMzw+5Va6ywApdTtwFdKqSZaa9cE11BqFqDR+gZcKzGZHlfQ+j2rQVhNSps5jvtdzjuDlG1pfCg1M88/dmqqjnMnJ0lIT/LYY4/ZHYIQ5TgL6B/MLqRDY5uDEULURgZmPqdr2ZsmnJzHOaUB+5wJqcNGx7E1J/e6Oo3DJLM3VHBtNlAKWE5Kra6MyXccB7qcH+hyXXhZYXEpB48V0KxBdQvibFZUBH/8IfVJK5GcnExycrLdYQhxnDMpPSQF9IUIaFrrQmAVZi/6soZxcj1Rp5+B5kqpsjt3dnIcd7kdhFLOnT7d6pWy2lP6B2auwGylmILJnrsCT2Ey5D+qeFZ4UPrRfLSG5v5Ui7Qif/4JhYWyyKkSmzdvBqBz5842RyKEkRQXAUCGlIUSoi54HnhfKfUbJuG8GTM/9A0ApdR7AFrraxz3zwEeAt5RSj2KmVP6EvCJ1vpAuZaVugi4yOXcLJf3d3Qcj7oTtNWkdDYmKW0BvFs2DExSOtudl4qaS83MA6C5v/eUyk5OVbrpppsAmVMq/EfDmAhCFGQck12dhAh0Wut5SqlEYCpmCuZ6YITW2tnr2drl/mNKqXMwi5tWYlbP/weoaE/s3pwYtgeTC15bURjAWnfitlo8/22lGI5ZkeXqE61xzZCFl6RlmZkSzRv4cU9pWhq88grEx0PHjtXfH4Sefvppu0MQopyQEEVCrOzqJERdobWegSnVVNG1IRWc24zZy94qZ8ek82tXG4CJbrTnVvH8y5XiCmAkjuL5wOda87E7LxS1k5plekqbxftpT+maNTByJGRmwr//DSF+VtDfTwwaVFmpOCHskxQXIUmpEKI6L2JGyBWQgklM25W5roFDaJ3jbsNuFc/Xmo+Aj8qeU4o44FKtyw3rCy9Jy8ynflQYsZFu73vgffPnw9ixkJgIy5ZB7952R+S31q9fD0CPHj1sjkSIExrVi5TheyFE1cwKfbNKX6nHMSWh3F8MVYEaZTZKEQIMB67G9JxGgSSlvpCamed/80m1hmnTYMoUGDAA/vMfaCpVwqpy++23AzKnVPiXpLhIdmS43bkhhAhWWj/qyebcSkqVoj8mER2N2TEAys8pEF6WmpXvf0npE0/AI4/A3/8Ob78N0X4Wnx969tln7Q5BiJM4h+9lq1EhhGVKjQUmAZ0xnZRlabS2nGtWe6NStAPGAldxYol/2T+t8jArtIQPpGXl0bd1A7vDKG/xYjjjDLODk/xFZkn//v3tDkGIkyTGRZJfVEpOYQlx/jhFSAjhX5S6AlMcX+NmTdKKVPqnjlLcjElGyxbMd32hBppozbHaBiKql1tYTGZukf/1lO7dC2efLQmpG9auNVUyesu8W+FHnAX0M7ILJHArcmAAACAASURBVCkVQlhxm+OYB8Rg8sLDQCKQ6fhYVtXS6BmYhFQ5PkXAQsxWUkOcN0lC6jupmX5YDqqkBFJToVUruyMJKBMnTmTiRLcqZQjhdc4C+rKrkxDColMxieg5x89o3Qh4BJM3jnSnMSu/CmtgFjBZa5PxKkV3d14iPCPNH8tB7d9vEtOWLe2OJKC8+OKLdocgxEmcPaUHs2UFvhDCkljHcTXO9UVKhQLTgceAl4GhVhuzOj5zPTBSKeYDnwIZVl8gPCfN0VPawp+G7/fuNUfpKXWLDNsLf9SonmP4XmqVCiGsOQo0xIyoZwP1gPNwloyCAe40VtXw/TRgDyeG7xsDE4CvgGVuhSw8Yl9mHkpBk/p+NHy/Z485Sk+pW1auXMnKlSvtDkOIchJizfC9JKVCCItSHcfGwEbH1wuApY6vD7vTWKVJqdZM0Zq2mPmj/8Jkvc4E1TmZFaXYqxTT3HmpqJm0rDyS4iKJCPOjXZKkp7RGJk+ezOTJk+0OQ4hywkNDaBATLkmpEMKqNZi8cABmFb4q8wE3a9hXO3yvNT8CPyrF7ZgJq2OBEUC445bmwGTgfndeLNyX5o81SvfsMXVJGza0O5KA8uqrr9odghAVSoqL5JDs6iSEsOZW4P+AbLTORal4TC37YmA+8A93GrNc80NrCjHzST9ViobAGEztUtnE20dSM/Po1KSe3WGUt3evGbqXclBuke1Fhb9yFtAXQogqKRUJXOr47msgF62nQc1Hz2s0Dqw1R7Tmda05E+gAPFrTAIQ1WmtSM/P9a+U9mJ5SGbp32/Lly1m+fLndYQhxkqS4SDKkp1QIUR2tCzDTO2dhyj/VWq2rI2tNCvCEB2IRVcjKKyKvqMS/apTCicL5wi1TpkwBYOnSpfYGIoSLpLhIMrKlp1QIYUkKZrfPUk80Jlt2BIgThfP9qKdUCufX2Jtvvml3CEJUKCkuguyCYvKLSogKD7U7HCGEf5sOvAncA0ytbWOSlAaIE4Xz/ainVArn11jnzp3tDkGICjkL6B/KKfSvmshCCH80CDgEPIBSo4D/YbYcddJofYPVxiQpDRCpmebfsV/1lEo5qBr74YcfAEhOTrY5EiHKcyalGdkFkpQKIapzLc6dnKCz4+NKktK6JjUrn/BQRSPHXxh+QQrn19gjjzwCyJxS4X+SZFcnIYR7qiq/o6u4dhJJSgNEWmYeTepHERLiR6WXpKe0xmbNmmV3CEJUKFF2dRJCWNfOk41VmpQqxV/cachRZF94SWpmPs39sRyUFM6vkfbt29sdghAVanS8p1TKQgkhqqH1Lk82V1VP6VKsd7vqatoStZSalUe/Nn6W/O3da3pJpXC+25YsWQLAOeecY3MkQpQXFR5KXGSY9JQKIaz7//buPD6uut7/+OvTNGnSJd1S2lJayiaLoEUWBVEqm4CyI+31YoX7wwoFBBEUuCA7VFkuRUWkCmXxXioFBJSyShEB2UEKGKBNW1q6703SNsvn98f3DJmGLDPJzJxJ5v18POZxZs75nnM+M5O0n3xXswGEqaE+W3vmnnKlZXuJpLKNPNDY6Cxdt5Hh+Tbo4OOP1Z+0g66++mpASankp7Cqk2pKRaQdZsXAbcAEWl6QKa1Ky7YK3tXs9WHAMOAFYCGwDfBVYAXw11RvKOlbsWETdQ3O1vk0HRSEmtKDDoo7ii7pnnvuiTsEkVYN1gT6IpKa84FTM3WxVpNS96abmPGfhCx4nDszkvafBPwfIVGVLFm0JjFHaR7VlCYmzldNaYeM1OAwyWMVfUuYu7w67jBEJP+NJ9SGvg2MiZ4/BBxJqMD8RzoXa6mqtSWJWfofb7b/MUIT/wXp3FTSs3htHq7mpInzO+Xxxx/n8ceb/zqJ5IeKvr1YWa3mexFp1w7R9sRP97ifCHyHMDL/0XQulmpSOjraTmq2/8xou206N5X0NE2cn0fN95oOqlMmT57M5MmT4w5DpEUVfXuxumYz9Q0ZWc5aRLqv4mg7H2gAwKwMeBooAq5I52Kpdj79ANgduM6MnwCLgeFABaGq9oN0birpWbx2I2XFRfQvK26/cK5o4vxOue++++IOQaRVFf164Q6rqjezVXke/TEsIvlmNTCEMOp+FSEvvBTYEB3fMZ2LpZqU/jehj0BRdMOKaL8BjcDF6dxU0vPJmlq2HlCK5dPUS6op7ZRhw4bFHYJIqyqiCfSXb9ikpFRE2jKXkJSOAN4Avgn8LDrmQFU6F0up+d6dvwCHAy9HN7Fo+0/gMHeNvs+mT9ZuzK/+pKCJ8zvp0Ucf5dFH0+pqI5IzFZpAX0RS8xShtXwX4AZCRaXRNKXolelcLOW5o9x5BnjGjN7AQGC1OzXp3Ew6ZvGaWnbeeUjcYWxJE+d3yo033gjAUUcdFXMkIp9V0TckpSs1gb6ItMX9MuCyT1+bHUgY9FQP/Bn3tGZnSmsVJjN6EvqWDnZnZjrnSsdsrm9k+YZN+VlTqv6kHTZjxoz2C4nEpKJvaL7Xqk4ikqZX0k1Ek6U6+h4zvgMsAl4iGuJvxjNmzDXjsI4GIG1bum4j7uTfuveJmlLpkIqKCioqKtovKBKDvr160qtnDzXfi0j7zHbDbAZma4GNmK3B7H7Mdkv3UiklpWZ8jTBJfgVb9hX4K2G6qBNbPlM6KzEd1PB8mg5KE+d32oMPPsiDDz4YdxgiLTIzKrSqk4i0x2xfwnij44B+hPywHDgeeBmzfdK5XKrN9xcREth/EzqzJvyV0LF1v3RuKqnL64nzVVPaYbfccgsAxx9/fMyRiLSsom8Jf31nMS9XrWq37LaDe3PP//syRT3Ux1ykwNwE9Ime1wErgcGE+Uv7RMe/lurFUk1Kv0IYbX8U8GHS/rnRdkSqN5T0JJYYzavm+8R0UKop7bCHH3447hBE2nTmN3bkyfeWtltuxYZNzKpczqvzVvGV7QfnIDIRySN7EfLDXwEX4V4bTZ7/C+Cs6HjKUk1KE1nwgmb7B0TbPMqYupfFa2sZ0LuYspKiuENpoonzO61///5xhyDSpsM+P4zDPt/+fLrVm+r50lVPMfOdxUpKRQrPamAocCnuoRYtJKaXEJLSlelcLNWBTouibfNm+vOj7cJ0biqp+2jZBrYZmGc5vybO77Tp06czffr0uMMQ6bQ+vXoyduchzJy9hMZGjzscEcmtu6PtLs327xxt70jnYqnWlD4B/BD4c2KHGf8GdiJU2z6Rzk0lNWtr6nh13momfn37uEPZkibO77Tf/va3AIwbNy7mSEQ678g9hvPEu0t5Y8Fq9h49KO5wRCR35hBqQx/FbCqhRX0UcBqhQnM+ZhM+Le1+d0sXSUg1Kb2aMMJ+MCEJhZCQWhTMdanHL6l6tnIZDY3OobsNjTuULWni/E577LHH4g5BJGMO2mUrSop68Ng7S5SUihSW39GUF7a05PzUpOdOU81qi1JdZnQR8FXgSZqWkGqMXn8tOi4Z9tR7SxnSrxdjthnQfuFc0sT5nda7d2969+4ddxgiGdGvtJivf66Cx2cvxl1N+CIFxtJ4tCmdZUY/AA43oxQYBKxyZ2PaoUtKNtU3MKtyGUeP2Zoe+TbNysKFcNBBcUfRpd17770AnHzyyTFHIpIZR+w+nKffX8bbC9cyZmSe/SEtItlyaiYvllJSakZ/oD9Q484K4JNofwXQG1jrztpMBlboXpqzkurNDfnXdK+J8zPi97//PaCkVLqPQ3YdSnGRMfOdxUpKRQqF+12ZvFyqo+/vAKqA7zbbPz7a/4dMBiWh6b53SRH775BnS1Fq4vyMeOqpp3jqqafiDkMkY/r3LuarO1bwmJrwRQqXWQ/MxmF2AWZj0j091aT0y9H2gWb7HyT0EfgyaTCzSWZWZWYbzex1M2tztn8zKzGzK6NzNpnZAjP7UbMyJ5jZe9Hx98zsuHRiyieNjc5T7y3lwM8NobQ4j+YnBU2cnyHFxcUUFxfHHYZIRh25+3A+XlXLu5+sizsUEckFs8mYLcPssmjP/cD/ApOBVzE7OJ3LpZqUDom2a5rtX9vseLvMbBwwBbgW2BN4EZhpZqPaOO0+4HBgImHuq+8A/0q65n7AdOCPwJhoe7+ZpZUs54t/LVrLsvWbctd07w5Tp8KTT4Za0LYkJs5XTWmnTJs2jWnTpsUdhkhGHbrbUIp6GI+9szjuUEQkNw4kzMz0d8xGAMfRNKipCLgwnYulmpSuj7aHNdufeL0hjXueB0xz96nu/r67nw0sBs5oqbCZHQYcDBzp7k+5+zx3f9ndZyUVOxd41t2via55DTAr2t/lPPXeEop6GAftslVubvjWWzBxInzzm7DttnDxxVBZ2XJZ1ZRmhJJS6Y4G9ilhv+0H89g7asIXKRA7RNt3gX2i5/fSNABqz3QulmpS+gYh673DjEvMOMGMSwh9SR14PZWLmFkJYR3UJ5sdehLYv5XTjgVeBc4zs4Vm9qGZ3WJmfZPK7NfCNZ9o45p57an3lrLP6IEM6F2Smxs+8kiYc/SOO2DMGPjlL2GXXWC//eC222DVqqaymjg/I2bNmsWsWbPiDkMk447YYxjzVtbw7yXr2y8sIl1dYs3sVYRVnRx4lNCED1CezsVSTUpvS7r4FcCfou2AZsfbU0Gozl3abP9SoLVFlrcHDgC+CJxAWEv1cGBaUplh6VzTzCaa2Wtm9lp9fX2KoefG/JXVfLB0A4fu1v6a0xnzyCOw//5w6qnwl7+ExPP662H9ejjjDBg+HI4/Hv78Z5g7VxPni0irDtttGD0MZqoJX6QQJGqtjqOp9fwDoF/0PK0O5qlOnv8gcBMtT4J6ozsPpXPTNPUgZN7fjZrtnyAkpieYWYc6Xbr77e6+t7vv3bNnylO15sRT74Xc+rBc9SdduBDeeAOOPrpp3/DhcP758M474dikSfDCC3DccSExVdN9p02dOpWpU6e2X1CkixnSrxf7bjeIx2YviTsUEcm+t6PtfYT+peuB2YQKRQjLjqYs1ZpS3DmfMMr+GuD30fbL7vw0jfutABqA5hnXUKC1f8EWA4vcPXke1PejbWJw1JI0r5m3nnxvKbsM68fIQTla7efRR8M2OSlNMIM994T/+R9YtAgeewwmTIDTTstNbN3Y9OnTmT59etxhiGTFt/YYzkfLNvD8h8vjDkVEsmsysJGmyspf4t4AfDs6/mI6F7Ncd0Y3s5eBt919YtK+D4AH3P2iFspPBG4GtnL3DdG+g4GngaHuvszMpgMD3f2wpPOeBFa6+3+0FU+fPn28uro6E2+t01ZVb2bvq5/irG/syHmH7Zybmx5xBMyZEwY2qUleRDJgY10Dh9/8dwAeP/fr+Te1nUgXZmY17t4n7jg+ZTaSMMipCvc3o327Elb/nIt7yn15Um67NqMfcCSwLVDa/Lg7V6Z4qZuAe8zsFeAF4HRga6J+qWZ2d7ieT4jK/y9wKXCnmV1O6Mc6BZjh7suiMlOAv5vZhcCfCX0bvkHoi9plPPP+Uhqd3PUnXb8e/vY3OPtsJaQikjGlxUVcc9we/OfvX+Y3z37ET3L1R7aI5J77x8DHzfa933LhtqW6zOg+wGOErLc1KSWl7j7dzAYDlwDDCX0PjnT3+VGRUc3KbzCzQ4BfEUbhryYknhcmlXnRzMYDV0dxzAHGufvLqcSUL55+fynD+5ey+4i0Bqt13BNPwObNcMwxubmffOrWW28FYNKkSTFHIpIdX92xguP3HMFtz83h6C9uzU5D+7V/kojkP7NQaeh+96fP2+J+d8qXTqX53owXCNMutX5Lp0u2z+RT8/2Xr32ar+5YwU0npb0yV8dMmBD6iS5ZAnk24Ku7O+KIIwCYOXNmzJGIZM/KDZs4+Kbn2GmrvkyfuB89eqhFRqSzYm++N2sEGnHvGT1vK5F03FNOMFIt+IXops8RlhqtbicISdP6jXUsXbeJHbfq237hTKivh7/+Fb79bSWkMVAyKoVgcN9eXHzkrvx0xr/402sfM37fthbuE5EuxFp53impZiNrgN7A8e6fWWpUMqBqRait3WFIjpLSF18Mk+K3NOpeRCRDvrPXNsx4fSHXPvY+B+86lCH9esUdkoh0zqmtPO+0VJPSuwl9OHcH/pHJACSYszys1LrDkBzVyD/yCJSUwGHNV46VXJgyZQoA55xzTsyRiGSXmXHtcXtw5JTnueov73HlMZ/P+j379upJz6KUZzwUkXS439Xi8wxINSmdB6wFHjbjD0AlUJdcwJ2UO7LKZ81dXk1RD2PUoBwkpe7w8MNw0EHQT4MP4vDMM88ASkqlMOy4VV/OGLsDU575kEfe/iTr99tn9EDuP71LrjItUtBSTUp/R1Mf0p+0cNxBSWlnzFm+gVGDelPSMwd/3VdWwkcfwXnnZf9e0qJHHnkk7hBEcuqsg3Zk28G9WVtb137hTpg5ewkfLduQ1XuIFDSzuWmUdtx3SLVwOiNcNGwyi+Ysq85t0z3AUUfl5n4iUvCKi3pw/Jeyv0Txig2beGP+atwd0/zLItkwmi0Huyd+0ZoPgLcW9rUp1aQ0ox1ZZUsNjU7VymoO3HlIbm748MOw115awz5GN9xwAwDnn39+zJGIdC/lpcXUNzq1dQ30LtHMIiJZ0tJffJ3+KzCl31h3MtqRVba0aHUtm+sbc1NTunw5vPQSXH559u8lrXrppZfiDkGkWyovKwZgbW2dklKRbHBv6mdotgPw9+jx38BCYBvgWsLKmgemc2n9xuaBOStC/6ftczEd1LvvhoFOB3SpFVi7nQceeCDuEES6pfLSkJSuq61neP+YgxHp/m4BhgFn4J6YMnQuZqcDq4CbgcNTvVjKo2rMONmMN8yoNqOh2aM+nXcgW5qzLDEdVA6S0oULw3bkyOzfS0Qkx8rLQl3Luo3ZHVAlku/MbJKZVZnZRjN73cy+luJ5B5hZvZnNTqF4oiZ0+2b7d4y2adWApVRTasZJhNH1jgY8ZdzcFdUM7F3MoD4l2b9ZIikdMSL795JWTZ48GYALL7ww5khEupemmlIlpVK4zGwcMAWYRJhffhIw08x2c/cFbZw3kJDvPQOkkihsAMqAmZjdQ1Pz/feSjqcs1eb7M6NtLWFlJydUyw4mrPakVZ46Yc6yDblpuoeQlA4aBL175+Z+0qK33nor7hBEuqVEn1LVlEqBOw+Y5u5To9dnm9nhwBnARW2c9wfgLkIF5Ikp3OcewlShFcCPk/YnRt6nNV1oqs33X4gufkhihztDgMsIk+hrbqFOmLM8h9NBLVyoUfd54L777uO+++6LOwyRbqe8NGq+r1WvMilMZlYC7AU82ezQk0Crq0qY2SRgKHB1Gre7mJB4WrMH0f6L07hWyjWliYzpDaI5p8woAm4EriB0dD04nRvni0GDBjFr1qzY7t/gzve3q2ZYaQOzZq3O+v32ev99Ng8cyDsxvmcRkWxx4Cd71FO2spJZs+bFHY5INvQ0s9eSXt/u7rcnva4AioClzc5bSlLlYjIz24NQ0fgVd29IeY5f9zrgFMyuI4y2HwysAGbhXpnaRZqkmpSuAwYSst/1QD/gCMLSowBfTvfG+WLVqlWMHTs2tvu/uWA1Nz7xIlMnjGHsbkOzf8O1a+HAA2N9zwJXXXUVAJdeemnMkYh0P5MufZyTvzKSk8buFncoItlQ7+57Z+piZtYLmA6c7+5VHbpISEDTTkKbSzUp/YSQlG4FvA/sCzycdHxVZwMpVHOXVwOwfS6a7zdvhqVL1XyfByorO/27KyKtKC/rqeZ7KWQrgAZCU3yyocCSFsoPB3YF7jSzO6N9PQAzs3rgSHdv3hWgZWEJ0rSWFk2WalL6JrA7oUb0bj5bM6rJ9TtozvIN9OxhjBqUg4FHn3wStkpKY3fvvffGHYJIt1VeWqyBTlKw3H2zmb0OHArcn3ToUKClSbIXAXs02zcpKn8cMC+N248mzaVFk6WalE4Cfgqsd6fGjP7AOKAeeAj4RUcDKHRzl1czanBviotSnjK24xLTQSkpFZFurLxMSakUvJuAe8zsFeAF4HRga+A2ADO7G8DdJ3joF7rFnKRmtgzY5O6pzFWaMakuM1oNVCe9ngxMzlZQhWTO8g25mTQflJTmkZ///OcAXHnllTFHItL9lJf2ZMWGzXGHIRIbd59uZoOBSwjN87MJzfDzoyKjYguuDa0mpWbpBexOq5OxSsvqGxqZt7Kag3fNwQAnUFKaRz7++OO4QxDptsrLipm7orr9giLdmLvfCtzayrGx7Zx7OXB5xoNqR1s1pfNIvV+At3MtacHC1bXUNXhuBjlBSEr79YPy8tzcT1p15513tl9IRDqkvLRYKzqJxMG9U30R20sktaRoFs1ZnsM17wEWLVItqYh0e+VlPVm3sR53J+X5FkUkdm0lpRpRn2WJ6aC0mlPhueiisMrbddddF3MkIt1PeWkxDY1OzeYG+vRSI55IzoX+rMuBRtxT/iVstaA7p2YiLmndnOUbGNynhAG9S3Jzw4UL4ZAWF3OQHFu5cmXcIYh0W+VlxQCs21inpFQkXmk1Vei3NUZzl1fnrj9pfT0sXqya0jxx++23t19IRDqkvDRKSmvrGd4/5mBEuhuzSSmU6lByk3JSasbOwA+BnYGyZofdnYM7EkAhm7N8A4fmYmlRCCs5NTQoKRWRbq+8LPzXprlKRbLi13Rigvy2pJSUmrEXMAtoadkhI0vBdWdrajazsnpzbkfeg5LSPHH++ecDcMMNN8QciUj301RTqqRUJIsyPoow1ZrSi+lgVay0bM6ng5w0cX4hqq2tjTsEkW4ruU+piGTcZqCYsDrU0lbK9AYuSPfCqSal+xNqQycBv42efxG4GtiFsOSopCHn00EpKc0rv/nNb+IOQaTbKi+Nmu9r62OORKRbegvYB3gW9/tbLBFG36edlKY6yengaPvHxA53ZgMTgc8BP073xoVu7vJqiouMbQY2756bJQsXQmkpDBqUm/uJiMSkn5rvRbLpZULT/ZczfeFUa0prgb7Axuh5aTTwaUN0/OhMB9bdzVm+gdGD+9CzqFOLH6QuMUepJpLOC+eeey4AN998c8yRiHQ/JT17UFZcpOZ7key4CrgDWNNGmVXAduleONWkdBkhKR1EWH50F+BZINE20pjujQvd3OUb2HGrHDXdgybOF5GCUl7WU833ItngvgJY0U4ZB+ane+lUk9J3gO2BLwB/AXYFEnMZOfBkujcuZA2NzoJVNRy627Dc3XThQjjggNzdT9qkGlKR7CovLVZNqUgXk2rb8RXAdwm1pFcTktBEO/AzwDkZj6wb+2RNLXUNzujBLc2wlQWNjWHd+xEjcnM/EZGYlZcpKRXJCrPLMBuQRvkBmF2WStGUklJ33nZnujsfubPencMJTfn93TnMneUpBycsWFUDwKhBOUpKV6yAujo13+eRM888kzPPPDPuMES6rfJSNd+LZMllwHzM/oDZYZh9dspQsz7RsTsIzfg/T+XCnVlmtASo7sT5BWv+yigpzVVNqaaDyjtlZTmadUGkQPUvK2buCv0XJZIF7wG7AadEj0bM5tHUz7QCGE1TxacB76Zy4TaTUjO+BIwHSoE/u/M3M04DriPUlG4y41Z3zk/9vcj8VWE6qOH9czgdFCgpzSNayUkku8rLijUllEh2fAH4f8D5wE5AEbADYewRbLnS0xzgeuD3qVy41aTUjAMI/UUTZc4043rgp4TBTUZIVn9sxkfu3Jbquyl0H6+qYeTA3hT1yNH0TEpKRaTAhIFO9bg7pqnwRDLHvRGYCkzFbCzwTcJk+onR20uBV4EncH82nUu3VVN6AWEZqeb7ICSkKwhVtAZ8D5SUpmr+yprcNd1DSEp79oSttsrdPaVNEydOBOD222+PORKR7qm8rCcNjU7N5gb69OpMTzURaZX7LGBWpi7X1kCnvQk1ok8QlhedSUhAHfgPd7YC/jMqu1umAuru3J0FK2vYNleDnCAkpSNGQI8cTdQv7Ro8eDCDBw9uv6CIdEh5YlUnjcAX6TLa+vOxItqOc2edGf8HrI72PRhtHyAsPdovS/F1O6tr6li/qZ6RuU5K1XSfV6677rq4QxDp1srLEkuN1jO8f8zBiHRXZvsSKjHfx/1ZzA4FbgFGAY8DE3BPecRhW1VnxQDurIu2axMH3KmLtpsTYaXzHgrZ/JXhu9l28GdnUMgaJaUiUmBUUyqSEz8FfgV8DrNiQkXl54Ay4FjC9FEpa7ejjdln55ZqaZ+kJjFH6ba56lPqHpLSo4/Ozf0kJaeeeioAd955Z8yRiHRP5WXhv7e1NUpKRbJoz2j7N2AvQiv7YuCT6PUxhMQ1Jan0/k7Ocr2FfZKGBStzPHH+6tVQW6vVnPLMyJEj4w5BpFtTTalITiSWnP8YODB6Phn4EyE53Tadi7WXlKpZPsPmr6phaHkvSouLcnNDTQeVl6688sq4QxDp1pr6lCopFcmiRGVlb2D36PW7NI1BakjnYm0lpVekHZq0K4y8z3F/UlBSKiIFpV9p+O9t3UYtNSqSRYsIE+g/StNMTO8CW0fPV7R0UmtaTUrdlZRmw/xV1Ryw45Dc3VBJaV46+eSTAbj33ntjjkSkeyou6kHvkiLVlIpk10PAz4CvEFrXX8F9KWYnRcf/lc7FNKNwDm2sa2Dpuk25G+QEsGhRmJ902LD2y0rO7LzzznGHINLthVWdlJSKZNEVQDnwNaAKOC/aP4qwKuh96VxMSWkOdWjkfWNjePTs4Fe1cGFISIubL84lcbr00kvjDkGk2ysv68m6WjXfi2SN+0bgzBb23wDckO7llJTm0PyOjLz/7nfho4/gxRehpCT9m2qOUhEpUKopFckRsz2AQwhTQq0AnsJ9drqX0bqTOdRUU5riQKc334Tp0+H11+GGdv7gePNNOOccWL58y/1KSvPS+PHjGT9+fNxhiHRr5WVKSkWyyqwnZtOAtwg1oxdG27cxuxOztKYaUlKaQwtWVtOvV08G9k6x5L3RvgAAIABJREFUKf3aa6G8HI44Aq66CubMabncypVwzDFwyy3wpS/BSy81HVNSmpfGjBnDmDFj4g5DpFsrL1XzvUiWXQ1MIAxyav6YEB1PmZLSHJq/qoaRg3pjlsL0r++/Dw88AGedBVOnhj6hkyaFFZqSNTbChAmwdCnceWco9/Wvw5QpsG5deCgpzTsXXnghF154YdxhiHRrqikVyboJhLlJlwHXAqdH22WExPSUdC6mpDSHFqysSX2Q0+TJUFYG554bVmO6+mp48snQnJ/shhvgscfgppvglFNCU/+RR4bzvv3tUEarOYlIASovLWZdbR3e/I95EcmU/tH2W7hfgvvtuF8CHBXtL0/nYkpKc6Sh0fl4dQ2jUklKq6rgj3+EiRNhSDSn6Zlnwl57hWRzzZqw7x//gIsvhu98J9SiAgwcCA89BL/4BbzwQtinmtK8c8IJJ3DCCSfEHYZIt1Ze1pNGh+rNaS0qIyKpey3afthsf2W0fSWdiykpzZHFa2upa/DUVnP65S+hqAjOP79pX1ER3H57GMh08cVhO348jB4dmveTuwT06AE//Sn87W9w2mmw994Zfz/SOfvttx/77bdf3GGIdGvlpVpqVCTLfgxsAK7GrBQg2l4FrKNp3tKUaEqoHEl5jtJPPoE77ghN8c2b3b/0JTj77DCg6eWXQ2L6z39C//4tXooDDwwPyTvnJ//BISJZUV4WJaUb69iaspijEekmzOY230OYq3QiZiuBwUAxUA3MAHZI9dJKSnNkQapzlN54IzQ0wM9+1vLxq66CGTPgjTfg1lthzz0zHKmISPfQVFOqEfgiGTSaMLgp0USbeF4CDE/a1xdIcQ7MQElpjsxfVUPPHsbw/qWtF1qxAm67Df7jP2D77Vsu068fPPhgmPbp9NOzE6xk3dFHHw3AI488EnMkIt1XeVn4L07N9yIZtYCQdGZcLEmpmU0CLiBk1O8C57r7862UHQs828KhXd3931GZU4A7WyhT5mEJrNgtWFnDNgPL6FnURjfeKVOgpgYuuqjti+27b3hIl3XwwQfHHYJIt/dpTammhRLJHPfR2bp0zpNSMxsHTAEmAf+ItjPNbDd3X9DGqZ8HViW9brZ0ETU067eQLwkpwPxV1YxqbyWn+++Hww+H3XbLTVASm3POOSfuEES6vU/7lKqmVKRLiKOm9DxgmrtPjV6fbWaHA2cAbVURLnP3FW0cd3dfkqkgM8ndmb+yhj1HDmy9UGNjmArqmGNyF5iISDfWrzRqvt+oPqUiWWPWEzgS2BlaGFHofmWql8ppUmpmJcBehHVRkz0J7N/O6a+ZWS/gPeBqd2/epF9mZvOBIsIarJe6+5sZCLvT1tbWsX5jfdsj7z/5BDZvhu22y11gEpsjjjgCgJkzZ8YciUj3VVzUg94lRaopFckWs62AWYSEtDX5mZQCFYSkcWmz/UuBQ1o5ZzGhFvVVwsiu7wHPmNmBSf1QK4H/At4G+gHnAC+Y2RfdvfmErpjZRGAiQElJSafeUCrmRyPvR7Y18r6qKmyVlBaEo446qv1CItJp5aVaalQki64AdmnjeFoDovJ+9L27V9K0MgDAS2Y2mjBQ6vmozEvAS4kCZvYiobb0bOBHLVzzduB2gD59+mR9/bn5qcxROjea9qu1UffSrUxKrMAlIllVXtZTU0KJZM9hhMRzGnBq9PwcQv7lwOR0LpbrFZ1WAA3A0Gb7hwLp9Ad9GdiptYPu3kBY+qrVMrm0YGU10M4cpVVVYVWmUaNyFJWISPenmlKRrEqs8nPhp3vcfw0cD3wOSGud85wmpe6+GXgdOLTZoUOBF9O41BhCs36LzMyAL7RVJpfmr6xhSL9e9C5po2K6qiqs4NSrV+4Ck9gccsghHHJIaz1WRCRTysuUlIpkUUO0XQmEXzSzIcD8aP/EdC4WR/P9TcA9ZvYK8AJwOrA1cBuAmd0N4O4TotfnAvMI85mWACcDxwInJC5oZpcB/wQ+BMoJTfZfIPRFjd38VTVs295KTnPnqum+gIwbNy7uEEQKQnlpTz5apuZ7kSxZSagt7U9o8d4G+COQmJKzjWmHPivnSam7TzezwcAlhMnzZwNHunsiq27efl0CXE94o7WE5PRb7v5YUpkBhD6iw4C1wJvA1939lay9kTQsWFnD/jsObrtQVRWo5qxg/OAHP4g7BJGCoJpSkayqJCSlOwB/B/4TSKwO48Ab6VwsloFO7n4rcGsrx8Y2e/1L4JftXO/HwI8zFV8mbaxrYMm6jW33J924MUwJpZH3IiIZ1b+smHW1dbg7oWeXiGTQVOAjoJQwEv8wYEh0bDlwbjoXy/vR913dpvpGvveVbdl39KDWC82fD+5qvi8gY8eOBWDWrFmxxiHS3ZWXFtPoUL25gb699F+eSEa5/wn406evzXYCvgHUAy/gviady+k3NMv6lxVz1bG7t11Ic5QWnFNOOSXuEEQKQnlZtKpTbZ2SUpFsc18HPNzR0/Ubmg+UlBYcJaUiuVFeWgzAuo11bN3CCogikj9yPU+ptGTu3DAV1PDhcUciOVJXV0ddnQZfiGRbeVmUlGoCfZG8p5rSfFBVBaNHQw/9jVAoDj00TNWrPqUi2fVpTWmt/ggUyXdKSvNBVZWa7gvMaaedFncIIgXh0z6lmhZKJO8pKc0Hc+fCV74SdxSSQyeffHLcIYgUhERN6VrVlIrkPbUXx23NmvBQTWlBqampoaamJu4wRLq9fqWJ0ffqUyqS71RTGjeNvC9IRx55JKA+pSLZ1rOoB31KitR8L9IFKCmN29y5YauJ8wvKGWecEXcIIgWjPFrVSUTym5LSuKmmtCCNGzcu7hBECkZ5abFqSkW6APUpjVtVFQwYEB5SMNauXcvatWvjDkOkIJSX9VSfUpEuQDWlcauqUtN9ATrmmGMA9SkVyYXy0mKWrNsYdxgi0g4lpXGbOxd23z3uKCTHfvSjH8UdgkjBKC8r5oNl6+MOQ0TaoaQ0To2NMG8eHHVU3JFIjh1//PFxhyBSMMpLe7K6uo5/fLgiY9fs0QO+NGogpcVFGbumSCaZ2STgAmA48C5wrrs/30rZ44HTgT2BUuA94Bp3fyRH4QJKSuO1ZAls2qTm+wK0YkX4z7GioiLmSES6v6H9S9mwqZ6T//ByRq974RG7cPqBO2T0miKZYGbjgCnAJOAf0Xamme3m7gtaOOVA4G/AJcAq4D+Bh8xsbGuJbDYoKY1TYjoojbwvOCeeeCKgPqUiuXDaAdvz5e0G0+iesWv+4O7XmL+yOmPXE8mw84Bp7j41en22mR0OnAFc1Lywu5/TbNcVZvYt4FhASWlB0HRQBesnP/lJ3CGIFIySnj3Ya9uBGb3miAFlLFmrwVOSf8ysBNgLuKHZoSeB/dO4VD9gdabiSkXBJ6WDBg2KrbZq27/9jdFmPD9/Po2LF8cSg8SjX79+gGpKRbqqk0bUsLlhrX6HJQ49zey1pNe3u/vtSa8rgCJgabPzlgKHpHIDMzsT2Aa4pzOBpqvgk9JVq1YxduzYeG4+bRpsvTVfP+yweO4vsVmyZAkAw4YNizkSEemIix96h5nvLubNk8bGHYoUnnp33ztbFzezE4DrgXHuPj9b92lJwSelsaqqUtN9gRo/fjygmlKRrmp4eSmra+rYWNegEfiSb1YADcDQZvuHAkvaOtHMTgTuBia4+6PZCa91SkrjVFUF3/hG3FFIDC688MK4QxCRThjavxSAZes2MWpw75ijEWni7pvN7HXgUOD+pEOHAg+0dp6ZnQTcBXzf3WdkN8qWKSmNy6ZNsHChakoL1OGHHx53CCLSCcPKQ1K6eG2tklLJRzcB95jZK8ALhDlItwZuAzCzuwHcfUL0ejyh/+j5wN/NLNG3bLO7r8pV0EpK47JgAbgrKS1QH3/8MQAjR46MORIR6YjhUU2pli+VfOTu081sMGHe0eHAbODIpD6io5qdcjohJ7w5eiQ8B4zNbrRNlJTGJTEdlCbOL0jf+973APUpFemqEs33S5WUSp5y91uBW1s5Nrat13FRUhoXTZxf0C655JK4QxCRTujXqye9S4pYsnZT3KGIdBtKSuNSVQUlJbD11nFHIjE45JCUpooTkTxlZgzrX8qSdbVxhyLSbfSIO4CCVVUFo0dDD30FhWju3LnMTdSWi0iXNKy8VKs6iWSQakrjMneumu4L2H/9138B6lMq0pUNKy/l5aqcDUwW6faUlMalqgr22SfuKCQmV1xxRdwhiEgnDetfytJ1G2lsdHr0sLjDEenylJTGYeVKWLUKdtop7kgkJgceeGDcIYhIJw3rX0p9o7OiehNb9SuNOxyRLk8dGuNQWRm2O+8cbxwSm8rKSioTPwci0iUNjSbQX6oR+CIZoZrSOCgpLXg//OEPAfUpFenKEhPoL15byx7b9I85GpGuT0lpHCorobg4jL6XgnTttdfGHYKIdFJiqVFNoC+SGUpK4/DBB7DjjtBTH3+h2n///eMOQUQ6aXDfXhT1MC01KpIh6lMah8pK+Nzn4o5CYjR79mxmz54ddxgi0glFPYyh/XqxWHOVimSEqupyraEBPvoIvv3tuCORGJ111lmA+pSKdHVDo2mhRKTzlJTm2rx5sHmzBjkVuOuvvz7uEEQkA4aVl/LB0vVxhyHSLSgpzbUPPghbJaUFbR8tnCDSLQzrX8rfP1gedxgi3YL6lOZaYjoo9SktaG+99RZvvfVW3GGISCcNKy+lenMD6zfWxR2KSJenmtJcq6yEgQOhoiLuSCRG5557LqA+pSJd3bD+TdNC9Sstjjkaka5NSWmuVVaGpnvTOsmF7Oabb447BBHJgMRcpUvWbmLHrfrFHI1I16akNNc++AAOOSTuKCRmY8aMiTsEEcmAYUmrOolI56hPaS5t2ACLFqk/qfDqq6/y6quvxh2GiHTSUK3qJJIxqinNJY28l8gFF1wAqE+pSFdXWlzEwN7FWtVJJAOUlOZSYuS9ktKC9+tf/zruEEQkQ4aWl7JEqzqJdJqS0lz64IMwwGnHHeOORGK2++67xx2CiGTIsP6lqikVyQD1Kc2lykoYPRpKS+OORGL24osv8uKLL8YdhohkwPD+pSxZuynuMES6PNWU5lJlpQY5CQAXX3wxoD6lIt3B0PJSVmzYxOb6Rkp6qq5HpKOUlOaKe2i+P+CAuCORPPC73/0u7hBEJEMSc5UuW7+RbQb2jjkaka5LSWmufPJJmBJKg5wE2Fk/ByLdRvKqTkpKRTpO7Qy5oumgJMlzzz3Hc889F3cYIpIBTRPoa7CTSGeopjRXEtNBqU+pAJdddhmgPqUi3UHTUqNKSkU6Q0lprlRWQu/eMGJE3JFIHrjjjjviDkFEMqR/WTGlxT20qpNIJykpzZXEyPse6jEhsP3228cdgohkiJkxrLyUJes0LZRIZyhDypUPPlB/UvnU008/zdNPPx13GCKSIWFVp9q4wxDp0lRTmgubNkFVFXz3u3FHInni6quvBuCQQw6JORIRyYTh/Ut5fcHquMMQ6dKUlObCnDnQ2KiaUvnUPffcE3cIIpJBQ/uXsnTtJtwdM4s7HJEuSUlpLiRG3isplcjIkSPjDkFEMmhYeSmbGxpZVb2ZwX17xR2OSJcUS59SM5tkZlVmttHMXjezr7VRdqyZeQuPXZqVO8HM3jOzTdH2uOy/kxQl5ijVdFASefzxx3n88cfjDkNEMmR4NFfpEo3AF+mwnCelZjYOmAJcC+wJvAjMNLNR7Zz6eWB40uPDpGvuB0wH/giMibb3m9mXM/4GOqKyEoYNg/LyuCORPDF58mQmT54cdxgikiFDy5tWdRKRjomj+f48YJq7T41en21mhwNnABe1cd4yd1/RyrFzgWfd/Zro9TVm9o1o/39kIuhOqaxU071s4b777os7BBHJIK3qJNJ5OU1KzawE2Au4odmhJ4H92zn9NTPrBbwHXO3uzyYd2w/4VbPyTwBndSLczKmshBNOiDsKySPDhg2LOwQRyaAhfXvRw+DNBWvYYUjfuMORmAzoXcwuw9Qq2lG5rimtAIqApc32LwVamxtnMaEW9VWgBPge8IyZHejuz0dlhrVyzfj/51+5MjxUUypJHn30UQCOOuqomCMRkUzoWdSDkYN6M+P1hcx4fWHc4UhMxu48hGmn7ht3GF1W3o++d/dKoDJp10tmNhq4AHi+pXPaY2YTgYkAJSUlnYywHf36wfPPw6j2usxKIbnxxhsBJaUi3cn//uArzF9ZHXcYEqMBZVnOKbq5XCelK4AGYGiz/UOBJWlc52VgfNLrJelc091vB24H6NOnj6dx3/SVlMABB2T1FtL1zJgxI+4QRCTDRgwoY8SAsrjDEOmycjr63t03A68DhzY7dChhFH6qxhCa9RNeysA1RXKmoqKCioqKuMMQERHJG3E0398E3GNmrwAvAKcDWwO3AZjZ3QDuPiF6fS4wD3iX0Kf0ZOBYIHnk0BTg72Z2IfBn4DjgG4CqKCUvPfjggwAcf/zxMUciIiKSH3KelLr7dDMbDFxCmG90NnCku8+PijTvfFkCXA9sA9QSktNvuftjSdd80czGA1cDVwJzgHHu/nJW34xIB91yyy2AklIREZEEc89ul8p816dPH6+uVsd0ya21a9cC0L9//5gjERGRrsTMaty9T9xxZEPej74X6Y6UjIqIiGwp58uMighMnz6d6dOnxx2GiIhI3lDzvZrvJQZjx44FYNasWbHGISIiXUt3br5XUqqkVGJQU1MDQO/evWOOREREupLunJSqT6lIDJSMioiIbEl9SkVicO+993LvvffGHYaIiEjeUPO9mu8lBupTKiIiHdGdm++VlCoplRjU1dUBUFxcHHMkIiLSlXTnpFR9SkVioGRURERkS+pTKhKDadOmMW3atLjDEBERyRtqvlfzvcRAfUpFRKQjunPzfcEnpWbWCNRm+LI9gfoMX1MyQ99NftP3k7/03eQ3fT/5K9PfTZm7d8uW7oJPSrPBzF5z973jjkM+S99NftP3k7/03eQ3fT/5S99N6rplpi0iIiIiXYuSUhERERGJnZLS7Lg97gCkVfpu8pu+n/yl7ya/6fvJX/puUqQ+pSIiIiISO9WUioiIiEjslJSKiIiISOyUlKbJzCaZWZWZbTSz183sa+2UPzAqt9HM5prZ6bmKtRCl8/2Y2fFm9qSZLTez9Wb2spkdnct4C0m6vztJ5x1gZvVmNjvbMRayDvzbVmJmV0bnbDKzBWb2o1zFW0g68N1818zeMrMaM1tiZvea2bBcxVtIzOzrZvaImS0yMzezU1I4Zw8ze87MaqPzfm5mloNw856S0jSY2ThgCnAtsCfwIjDTzEa1Un474LGo3J7AdcCvzOyE3ERcWNL9foADgb8B34rKPwY8lGqyJKnrwHeTOG8gcDfwTNaDLGAd/H7uAw4HJgI7A98B/pXlUAtOB/7f+SpwD3AX8HngWGA34I85Cbjw9AVmA+eQwkI8ZlYOPAUsBfaJzrsAOC+LMXYZGuiUBjN7GfiXu/8gad+HwAx3v6iF8r8Ajnf3nZL2/R74vLvvl4uYC0m6308r13gFeN7df5KlMAtSR78bM3sQeBsw4ER33z3rwRagDvzbdhhwP7CDu6/IXaSFpwPfzfnA2e6+bdK+U4FfuXvfXMRcqMxsA3CWu09ro8wZwC+Aoe5eG+27BDgD2MYLPClTTWmKzKwE2At4stmhJ4H9WzltvxbKPwHsbWbFmY2wsHXw+2lJP2B1puKSjn83ZjYJGApcnb3opIPfz7HAq8B5ZrbQzD40s1vMTElPBnXwu3kBGG5mR1lQAYwntARJ/PYjVHwk16o+AWwNjI4lojyipDR1FUARoco92VKgtb46w1op3zO6nmROR76fLZjZmcA2hKYvyZy0vxsz2wO4DDjZ3RuyG17B68jvzvbAAcAXgROAswhN+dOyE2LBSvu7cfeXCEnoH4HNwHJCS8P3sxempKG1vCBxrKApKRUBon6+1wPfdff5ccdTyMysFzAdON/dq+KOR1rUA3DC78vL7v4EITE9wcyGxhtaYTOz3YBfAVcRalkPJyQ7v4szLpFU9Iw7gC5kBdBAaE5MNhRY0so5S1opXx9dTzKnI98PAGZ2ImEwzQR3fzQ74RW0dL+b4cCuwJ1mdme0rwdgZlYPHOnuzZszpeM68ruzGFjk7muT9r0fbUfx2Zog6ZiOfDcXAa+4+/XR63+ZWTXwvJld7O4LsxOqpKi1vCBxrKCppjRF7r4ZeB04tNmhQwmjIVvyUivlX3P3usxGWNg6+P1gZicRmutPcfcZ2YuwcHXgu1kE7AGMSXrcBnwUPW/1+5T0dfB35wVg62Z9SD8XbdXSkCEd/G56ExLZZInX+j8/fi8BXzOz0qR9hwKfAPNiiSifuLseKT6AcYQ+OqcRanKmABuAbaPjdwN3J5XfDqgGbo7Knxadf0Lc76U7Pjrw/YwH6ghTcgxLegyK+710t0e6300L518OzI77fXTXRwd+d/oCHxNG4H8e+CphWpz7434v3e3Rge/mlOjftTMIfX+/ShiU9nrc76U7PqLfhcQfzzXAz6Pno6Lj1wHPJJXvT6gRvQ/YHTgeWAf8JO73kg8PNd+nwd2nm9lg4BJCE+NsQlNiomZgVLPyVWZ2JPA/hH8gPgF+5O4P5DDsgpHu9wOcTujCcnP0SHgOGJvdaAtLB74byaEO/Nu2wcwOIfRdfJUwY8WfgQtzF3Vh6MB3M83M+hH6+N4IrCXMx/yz3EVdUPYGnk16fUX0uIvwB8JwYIfEQXdfa2aHAr8BXiP87twI3JSjePOa5ikVERERkdipf4mIiIiIxE5JqYiIiIjETkmpiIiIiMROSamIiIiIxE5JqYiIiIjETkmpiIiIiMROSalInjOznczs12b2vpltMLP1ZvZvM5tqZl9JKjfPzNzM5sUYbiKWaVEsbmajk/YPNbM/mtliM2uIjt9sZqOTyk/LYlwDzOzy6HFsqnHnipmNTbp/e4/Lo3MSr2flOt72ZPN7Tee7ava5ZjQOEckcTZ4vksfM7FTgt0CvZod2jh5DgM8kV3lsCmGFmrgMAC6Lnt9FmPBdRETygJJSkTxlZgcBvye0aDhwDfA7YBmwLXAiTeuN5xV3P4Wwmklze0XbNcB27r4m6ZhlOax2tRF3ru4/i6TPwcxOAe6MXt4VxZdxZlbq7huzcW0RkVSp+V4kf11H0+/oLe5+qbsvdPfN7v6hu18H/KCtC5jZGDN70Mw+MrN1ZlZnZkuifXs3K7udmd1tZgvMbKOZrTGz2VEz6VZJ5X5gZq+Z2Soz22Rmi8zsKTP7flKZLZpWE82nwI5RkQHA6uj4KW0185rZl8zs/6L7bDazFWb2rJntGx3va2Z3mdk7ZrYyeo9rzOzvZjYu6TqXA1VJl/5+83u20e2gj5ldYWbvmlmtmdWY2Ztmdp6Z9Uwqt8X7MLMJ0WdYa6H7xffJIjM7yMz+Gd1vjpn91MySk9zLk+I7zsz+YGYrgNqkMrua2T1Jn/cyM5thZl9odq+Ufl6anXOSmf2rrc/DzL5mZo+Y2fKkn9f7mt+/jc9g6yjeDdHPw2+Bfq2UTfs9iEgWubseeuiRZw9gK0LtaOIxIoVz5kVl5yXtG9/sOsmPamDXpLLvtlF296jMd9ooMyPpWtOS9o8GxrZx3ilRmcTraUnXOQ6oa+28qMywNq7twISo3OVtlJnWUtzRvj7A622c+xjQIyqb/D5Wt1L+gDR+Dk5p6XNpViZxfEUrn9XJSWUvb1b+03LR8QOAmlbirgW+lubPS/LnsaS9zwM4GWhopdxGYGxrP2PRvjLg/RbO/aSlzzGV96CHHnrk7qGaUpH8NDrp+Tp3X9TB67wBfBMYTuiXWg6cER3rDfwQwMwGA7tF+28hJGKDgH2AS4G10bGvR9sNhD6tvQhdCU4CHm8tCHef5e4GzI92zXd3ix7TWjrHzMqAqTR1M/o5MBSoICTHc6P96wn9VEdH76kU2J+QXAGcF8VwObBd0i3uSorhlNZiB84FvhQ9f4LwWW5P+GwBjiAk/80NACYB/YFfJO3/Xhv36ozBwC+BgcBZKdzPgMMJn9ke0b6phMRuPqGrRS9gT2A54XP9DaT185JsKG18HmbWB/gVoXWgnvAHSTlwelSuF6H7SlsmALtEz/8JbEOonV/TvGAH34OIZJH6lIp0b0uA/wfcTEjaypod3znarib8xz2AkGStJ9Q4ve3uVyeVr4q2fYBLCDWI7wNPunum/xP/KiHRApjl7lclHZuR9LyGkKhOB3YlNNUm90/dmc75VtLzi9x9CYCZXUnTQKkjgf9tdt7r7v7bqOy9wM+i/dt2Mp7WLAV+7u4NZnYX8Ot27nejuz8RPZ9tZjvRlNBtS/hum9vDzIYR+jWn8vOSrL3P46vR9QAec/fEZ/s7MzsdGAN8zsx2dPePWrnHQUnPr0v8MWdmNxL6ZydL9WdeRHJENaUi+Wle0vNyM9u6g9f5E/BTQrLWPCElsc/dGwk1VguBnYD/Bu4F3on6ao6Myt8K3A8kyt9MqD1camYXdjDG1gxNev5eG+V+RqjB+zKhZq35gKnSTsYxJOn5gqTn85Oet9T/sDLpeXUG42nNHHdvSON+bzZ7nWofysFp/Lwka+/zaO1zhvY/609jS3q+sJXnQFo/8yKSI0pKRfKQuy8DXknadUFL5ZIH2bRwbCCh6R5CLdrngSKgxQEj7v4XYBShZvFo4EpC/77dCbWiuPtGdz+J0Mx5APBfwMuEptVrzWxEau8wJUuTnu/aRrnkpvNjgV5RV4GVLZT1DsSxLOn5qFaeJ5dJqOvkfdP16f3cPZX71TZ7nfwenk7q2vDpg9B39t3oHu3+vLQWHy1/Hq19zs1ft/RZJ6xIer5NK8+bgkj/PYhIFikpFclf/02okQT4UTRyemszK7Ywof7FhD6Aramn6T//emAdoZn7qpYKm9lKTGQ+AAACoUlEQVSvgIMJ/UUfBx4ANkWHR0VlTjCzs4ARwNuEWtO3E5eglf/8O+gFmhLLb5jZxWY2xMwGmtmxZpbo31qfdM4aoNjMLmXLWrOE5ER1p6gfY3v+mvT8GgsLAIwm9HFtqUyX5O4fAh9ELw82s3MtLDZQamZfNLOfA/clyqfy85KmFwlN6gBHmNnRFmZW+AGhXytAZRtN9wDPJj2/0MxGmNkOwE9aKpyF9yAinaCkVCRPufvThIFIdYTf1cuARcBmQvJwDWFQS2vnrweeiV6OAD4m1D7u1sopZwBPJd3jbcIgGAhN9BBqLH9FaE5fHz0mRscWA/9K4y22yd1rCVNeJZLOawi1ZKuAhwiDjYieJ8wiJBg/ooXBLe6+gTDiGsJgqA3R9EintBHKFLYc1LSE0Lc2MefqTEJ/1u5gImGUuwH/Q0gSa4G3gCvYsktFKj8vKXP3auBswh9ixcDDhJ+v26Mim2ga9NSau4F/R8/3IzTNf8SWXQOSZfQ9iEjnKCkVyWPu/ntCc/uthES0ltAfrxL4AzC5nUucTEiYVhNGE99L6ysqTQb+QUj86gkDiN4gJHhTojLPEAb0fERI/hoIyeh9wIFRIpkx7v4Qoa/ofYRpfeoJSelzNPUz/QVwLSGxqI2OHUTro6e/B/ydUHOcSgzVhFkHriQMhNlESNzeItTAHR31T+zy3P05QrJ9NyGhqyN83v8i/DFycVLxVH5e0r3/HwnTh/2FUKtdT/hD6k/Avh4WF2jr/FrgEOBBwu/JGsLiA63N55vx9yAiHWepdT0SEREREcke1ZSKiIiISOyUlIqIiIhI7JSUioiIiEjslJSKiIiISOyUlIqIiIhI7JSUioiIiEjslJSKiIiISOyUlIqIiIhI7JSUioiIiEjs/j8O3UQsMmkwrAAAAABJRU5ErkJggg==\n", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], - "source": [ - "fig, ax1 = plt.subplots(figsize=(10,7))\n", - "ax1.plot(class_thresh_arr, bal_acc_arr_orig)\n", - "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", - "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", - "ax1.xaxis.set_tick_params(labelsize=14)\n", - "ax1.yaxis.set_tick_params(labelsize=14)\n", - "\n", - "\n", - "ax2 = ax1.twinx()\n", - "ax2.plot(class_thresh_arr, np.abs(1.0-np.array(disp_imp_arr_orig)), color='r')\n", - "ax2.set_ylabel('abs(1-disparate impact)', color='r', fontsize=16, fontweight='bold')\n", - "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", - "ax2.yaxis.set_tick_params(labelsize=14)\n", - "ax2.grid(True)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "```abs(1-disparate impact)``` must be small (close to 0) for classifier predictions to be fair.\n", - "\n", - "However, for a classifier trained with original training data, at the best classification rate, this is quite high. This implies unfairness." - ] - }, - { - "cell_type": "code", - "execution_count": 15, - "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" - } - ], - "source": [ - "fig, ax1 = plt.subplots(figsize=(10,7))\n", - "ax1.plot(class_thresh_arr, bal_acc_arr_orig)\n", - "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", - "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", - "ax1.xaxis.set_tick_params(labelsize=14)\n", - "ax1.yaxis.set_tick_params(labelsize=14)\n", - "\n", - "\n", - "ax2 = ax1.twinx()\n", - "ax2.plot(class_thresh_arr, avg_odds_diff_arr_orig, color='r')\n", - "ax2.set_ylabel('avg. odds diff.', color='r', fontsize=16, fontweight='bold')\n", - "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", - "ax2.yaxis.set_tick_params(labelsize=14)\n", - "ax2.grid(True)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "```average odds difference = 0.5((FPR_unpriv-FPR_priv)+(TPR_unpriv-TPR_priv))``` must be close to zero for the classifier to be fair.\n", - "\n", - "However, for a classifier trained with original training data, at the best classification rate, this is quite high. This implies unfairness." - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Train classifier on transformed data" - ] - }, - { - "cell_type": "code", - "execution_count": 16, - "metadata": {}, - "outputs": [], - "source": [ - "scale_transf = StandardScaler()\n", - "X_train = scale_transf.fit_transform(dataset_transf_train.features)\n", - "y_train = dataset_transf_train.labels.ravel()\n", - "\n", - "lmod = LogisticRegression()\n", - "lmod.fit(X_train, y_train,\n", - " sample_weight=dataset_transf_train.instance_weights)\n", - "y_train_pred = lmod.predict(X_train)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Obtain scores for transformed test set" - ] - }, - { - "cell_type": "code", - "execution_count": 17, - "metadata": {}, - "outputs": [], - "source": [ - "dataset_transf_test_pred = dataset_orig_test.copy(deepcopy=True)\n", - "X_test = scale_transf.fit_transform(dataset_transf_test_pred.features)\n", - "y_test = dataset_transf_test_pred.labels\n", - "dataset_transf_test_pred.scores = lmod.predict_proba(X_test)[:,pos_ind].reshape(-1,1)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Predictions from the transformed test set at the optimal classification threshold" - ] - }, - { - "cell_type": "code", - "execution_count": 18, - "metadata": {}, - "outputs": [ - { - "data": { - "text/markdown": [ - "#### Predictions from transformed testing data" + "source": [ + "display(Markdown(\"#### Predictions from original testing data\"))\n", + "bal_acc_arr_orig = []\n", + "disp_imp_arr_orig = []\n", + "avg_odds_diff_arr_orig = []\n", + "\n", + "print(\"Classification threshold used = %.4f\" % best_class_thresh)\n", + "for thresh in tqdm(class_thresh_arr):\n", + " \n", + " if thresh == best_class_thresh:\n", + " disp = True\n", + " else:\n", + " disp = False\n", + " \n", + " fav_inds = dataset_orig_test_pred.scores > thresh\n", + " dataset_orig_test_pred.labels[fav_inds] = dataset_orig_test_pred.favorable_label\n", + " dataset_orig_test_pred.labels[~fav_inds] = dataset_orig_test_pred.unfavorable_label\n", + " \n", + " metric_test_bef = compute_metrics(dataset_orig_test, dataset_orig_test_pred, \n", + " unprivileged_groups, privileged_groups,\n", + " disp = disp)\n", + "\n", + " bal_acc_arr_orig.append(metric_test_bef[\"Balanced accuracy\"])\n", + " avg_odds_diff_arr_orig.append(metric_test_bef[\"Average odds difference\"])\n", + " disp_imp_arr_orig.append(metric_test_bef[\"Disparate impact\"])" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "pCQHgBCF6QV8" + }, + "source": [ + "#### Display results for all thresholds" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "metadata": { + "id": "uQ0FsHep6QV8", + "outputId": "5bc18b7f-c8e7-463c-9e7e-ab85fe31d4be", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 451 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "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\n" + }, + "metadata": { + "needs_background": "light" + } + } + ], + "source": [ + "fig, ax1 = plt.subplots(figsize=(10,7))\n", + "ax1.plot(class_thresh_arr, bal_acc_arr_orig)\n", + "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", + "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", + "ax1.xaxis.set_tick_params(labelsize=14)\n", + "ax1.yaxis.set_tick_params(labelsize=14)\n", + "\n", + "\n", + "ax2 = ax1.twinx()\n", + "ax2.plot(class_thresh_arr, np.abs(1.0-np.array(disp_imp_arr_orig)), color='r')\n", + "ax2.set_ylabel('abs(1-disparate impact)', color='r', fontsize=16, fontweight='bold')\n", + "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", + "ax2.yaxis.set_tick_params(labelsize=14)\n", + "ax2.grid(True)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "qeW0Z5-C6QV8" + }, + "source": [ + "```abs(1-disparate impact)``` must be small (close to 0) for classifier predictions to be fair.\n", + "\n", + "However, for a classifier trained with original training data, at the best classification rate, this is quite high. This implies unfairness." + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "metadata": { + "id": "wN0MFLte6QV8", + "outputId": "14c52a45-1437-4c42-ac62-c6a1b2f1cb65", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 451 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "iVBORw0KGgoAAAANSUhEUgAAArEAAAGyCAYAAADzppweAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAAsTAAALEwEAmpwYAACNVklEQVR4nOzdd5hU5fXA8e/ZXtlO70gVFLuoKEYwSqKxRTCWqFHsJZYEjYm9JJafGluwoaIRRWMXFBUbqIiiUqXv0tkK29v7++OdcYdly8zuzNwp5/M889zZe+/ce7awnH3nvOcVYwxKKaWUUkqFkxinA1BKKaWUUspXmsQqpZRSSqmwo0msUkoppZQKO5rEKqWUUkqpsKNJrFJKKaWUCjtxTgfgtJiYGJOcnOx0GEoppZRS7aqsrDTGGB2ERJNYkpOTqaiocDoMpVq0detWALp37+5wJEoppUKBiFQ5HUOoiPokVqlQNnnyZADmzZvnbCBKKaVUiNEkVqkQNnXqVKdDUEoppUKSRPuKXampqUbLCZRSSikVDkSk0hiT6nQcoUALg5UKYQUFBRQUFDgdhlJKKRVytJxAqRB29tlnA1oTq5RSSjWnSaxSIeymm25yOgSllFIqJDlSTiAil4rIOhGpFpFFIjK2jXOni4hp4VHhcc64Vs4ZFpzPSKnAGD9+POPHj3c6DKWUUirkBD2JFZFJwEPAXcB+wHzgfRHp28pLrgJ6NHusBV5p4dy9m523yq/BKxVka9euZe3atU6HoZRSSoWcoHcnEJGvgR+NMRd67FsFzDLG3ODF6w8HvgAON8bMd+0bB3wC5BljCn2JR7sTqFA2btw4QGtilVJKWdqdoElQa2JFJAE4ALiv2aEPgMO8vMyFwFJ3AtvMtyKSCCwD7jDGfNLhYJUKAbfeeqvTISillFIhKdgTu3KBWGBbs/3bgHYL/0QkAzgdaD5iuwW4BFgIJABnAx+JyFHGmM9buM4UYApAQkKCj5+CUsFz1FFHOR2CUkopFZLCrTvBWdg63hc8dxpjVgIrPXYtEJH+wPXAHkmsMWYaMA1sOUGgglWqs1autD/WQ4cOdTgSpZRSKrQEe2JXIdAAdGu2vxuw1YvXXwi8Zowp9uLcr4HBvoWnVGi56KKLuOiii5wOQymlVJjxpROU6/yjXOdVi8haEbk4WLF2VFBHYo0xtSKyCJgAvOpxaALwWluvFZGDgX2Bq7283WhsmYFSYeuuu+5yOgSllFJhxqMT1KXYyfCXYjtBjTDG5Ldw/gDgPeAZ7LveRwCPicgOY0yb+ZmTnOhOMAlbDnAp8CVwMfAnYG9jzAYReR7AGHNOs9c9BRxpjBnSwjWvBtYDS7E1sWcBU4FTjTGvtxWPdidQvqqpb6C0so7qugb6ZqcgIk6HpJQKtu3b7UMpT2lp0L9/QG/hTXcCXztBicg/gVOMMYM99j2Fzc3G+C96/wp6TawxZqaI5AA3YXu5LgEmGmM2uE7Zo1+siKQDk4HbWrlsAnAv0BuowiazvzHGvOfn8FUU2VVdx1dri/li1Q6+yy+lqLyGkso6quoafjnn4P7ZTJ04jP37ZgUkhiVLlgAwcuTIgFxfKdUBdXUweDDs3Ol0JCrUHH88vOds6tHBTlBjXMc9zQH+KCLxxpg6/0bpH45M7DLGPAY81sqxcS3s2wWktXG9fwH/6kgs2dnZ2oNT/aK+0VBcUUt5dT2VtQ0YDF1FOKVHLPGxMcTGCHExccTGCI3GsGNXIV9+/hlLkuPp3iWJhDj/lplfffXVADz44IN+va5SquOStm7l0J072XTSSZSOHu10OCqE1GZnUxb4nCJORL71+Hiaa8K6W0c6QXUH5rZwfpzreiFZnhlu3Qn8rri4+JeG8iq6LdpQwiUzFrGjvJ6RPTMYOziXIwbnckC/LBLjYlt8TUVNPU9+vpZ7PltLTX01kw/qw6heGb8cd1caCG2UHAgIICLExQiJcTEkxseQEBvLn2+9j5gYIa3/PtQ3GhobDfWNhrz0RIb36OLHz14p5bXPbdObXpdcQq9jj3U4GBWF6o0xBzodRCiI+iRWKYCXvs7n5reW0D0jiXevGMuInt4liKmJcVw9fghnHtKPhz9axUvf5PPi13vUzHfelwt2+zA+VvjmxvFkpWqfY6WCrqDAbvv0cTYOpVrWkU5QW1s5v951vZCkSayKajX1Ddz85lJeXljAkUPyeHjyaDJTfE8M89ITuf2kkVz366FU1tYD4J4z2dbUSWMMnnMr6xoaqW1opKaukZr6Rpb8uJgGA8P2HkVsjBArwobiSv4y60e+WlvE8aN6+ByrUqqTNIlVIayDnaAWACc32zcB+DZU62FBk1gVxQqKK7niv9+zuKCUy44exDUThhIb07lOAxnJ8WQkx/spQvjLeXYuo2fd9v79srj1raV8uaZQk1ilnJCfD5mZdia6UqHpAeAFEfmGpk5QPYEnAFroBPUEcLmIPAj8BzgcOBc4I6hR+0iTWBV1tpZV89i81bz8TQHxscITZ+3PcSNDMxlsaUJXfGwMhwzMYf7qouAHpJSyI7E6CqtCmK+doIwx60RkIvB/wCXAZuDKUO4RC5rEqiiyfVc1T8xby4yvN9DYaPj9gX24/Fd70Ssz2enQWjW6lZnPhw3K4eMV29lSVkWPjNCNX6mIpEmsCgMd6AT1KbB/gMPyK01iVVjbtrOa4opawLMG1VBWVcf2nTVs3VnN1jL7mPfzduoaDKfu34srfjWYPtkpDkbunYULFwJw0EEH7bb/8L1yAfhydRGnHdA76HEpFdUKCuCQQ5yOQqmop0msCju7qut4/6etvPbdRr5eV9zu+WmJcXTrkshv9+nJZUfvxYDcNhc6CSnXX389wB69jId2SycnNYH5qws1iVUqmCoroagI+u6xLo9SKsg0iVVhY/6aQl5ZWMDspVuprmtkQG4q104YwuBunpMr7MSs9KQ4unVJontGEmmJ4ftj/sgjj7S4PyZGGDMohy/XFGKM0aVvlQqWjRvtVssJlHJc+P7vrqJGaWUtN7+1lDcXb6ZLUhyn7t+bUw/ozX59MiM+eWtrudnDBuXyzo9bWLOjgr266ixppYJC22spFTI0iVUh7ZOV25n62o8UlddyzYQhTDlyIEnxLa+eFYnmz58PwGGH7bnc9eF75dhz1hRqEqtUsGgSq1TI0CRWhaTymnrufHcZ//2mgCHd0nj6jwcx0mM512hx4403AnvWxAL0zU6hV2YyX64u5Jwx/YMbmFLRyp3E9tZadKWcpkmsCjnrCys4+5mv2VhSxUVHDeSaCUNIjIue0VdP//nPf1o9JiIcvlcOc5Zuo6HRdHqhBqWUFwoKoGtXSEx0OhKlop4msSqk7NhVwznPfEN5dT2vXjSGA/tnOx2So4YOHdrm8cP3yuWVbzeybPNORvWOvpFqpYJOe8QqFTJinA5AKbeKmnr+9NxCtu+q5ulzD4r6BBbg008/5dNPP231+JhBti72yzWFwQpJqeimSaxSIUOTWBUS6hoaufTF71iyqYxHztif/ftmOR1SSLj55pu5+eabWz3eNT2JId3S+HK1JrFKBYUmsUqFDC0nUI4zxnDD6z/x6c87uOeUUYwf0c3pkELGM8880+45hw3K5eWF+dTUN0Rt7bBSQbFzp31oEqtUSNCRWOW4Bz78mVmLNnL1+MFMPlhXwfE0cOBABg4c2OY5hw3Kobquke/zS4MTlFLRSttrKRVSNIlVjqlraOTOd5fx749Xc8bBfbjqmMFOhxRy5s6dy9y5c9s855CBOcQIzNeSAqUCS5NYpUKKlhMoR2wpq+Lyl75n0YYSzj60HzefMCLiV9/qiDvuuAOA8ePHt3pORnI8o3pn8uWaIq4JVmBKRSNNYpUKKZrEqqD79OcdXP3y99TWN/LwGftx4r49nQ4pZL3wwgtenXf4oBymfbaWsqo6MpLjAxyVUlEqPx9iYqCn/s5SKhRoOYEKmoZGw/0frOTcZ7+hW5ck3rriCE1g29GnTx/6eDHqM3FUDxqM4Z73VwQhKqWiVEEB9OgBcTr+o1Qo0CRWBc19H6zk3x+v5vQD+vC/Sw9nUF6a0yGFvNmzZzN79ux2zxvZK4MLxw7kv9/k88nK7UGITKkoVFAAfXXyqVKhQpNYFRQrtu7kyc/W8vsDevPP0/YhOUFbQXnjnnvu4Z577vHq3GsmDGFItzT+OutHSitrAxyZUlFIe8QqFVI0iQ0WY5yOwDGNjbYPbEZyPH/7zXCnwwkrL7/8Mi+//LJX5ybFx/LA6aMprqjlH28uDXBkSkUZYzSJVSrEaBIbaNu2Qf/+4EXT+kj10jf5fJ9fyk2/HU5mSoLT4YSV7t270717d6/PH9krgyuPGcxbP2zm3R+3BDAypaJMURFUV2sSq1QI0SQ20HJzYft2WBqdI2Pbd1bzz9krOHyvHE4a3cvpcMLO22+/zdtvv+3Tay4dN4h9e2dw0xs/sX1XdYAiUyrKaHstpUKOTrEMtNhYGDYMli1zOhJH3PbOMmrqG7njpFHaB7YD7r//fgBOOOEEr18TFxvD/aeP5jcPf87U135i6vHDdjve0GjYWlZNQUkl+UWVFJRUsqWsmsuO3otf7+39qK9SUUWTWKVCjiaxwTBiBHz+udNRBN0nK7fzzo9buGbCEAbkpjodTliaNWtWh163V9c0/nLcMG5/Zxkfr2i9W0FSfAx9slLYurOaGV9t0CRWqdZoEqtUyNEkNhhGjIAXX4RduyA93elogqKqtoG/v7GEQXmpXHTUQKfDCVu5ubkdfu35h/dnePd0SirrdtsvAt26JNInO4W8tEREhFvfXspLX+dTXddAUrx2jlBqDwUFEB8PXbs6HYlSykWT2GAYMcJuV6yAgw5yNpYAamg0LC4oZd7K7XywdBsbS6qYOeVQEuM0Keqo119/HYBTTjnF59eKCIft5V0SfOSQPJ79cj1fryvmqCF5Pt9LqYhXUAC9e9sVu5RSIUGT2GBwJ7HLlkVkErti604en7eGT3/eQWllHTECB/TL4t7T9uGQgTlOhxfWHn74YaBjSawvDh2QQ0JcDJ/9vEOTWKVaou21lAo5msQGw8CBkJAQsZO7/vraT6zdXs6xe3fn6GF5jN0rj4yUeKfDighvvvlmUO6TnBDLwf2z+eznHUG5n1Jhp6AADj/c6SiUUh40iQ2GuDgYOjQik9id1XX8tLGUy4/ei2uOHep0OBEnIyMjaPc6ckgud723gs2lVfTMTA7afZUKeY2NsGmTjsQqFWK0uCdYRoyIyCR24bpiGg0cOkjLBgJh5syZzJw5Myj3OtJVRvD5Kh2NVWo327ZBXZ0msUqFGE1ig2XECFi3DiornY7Er+avKSIhLob9+2Y5HUpEevzxx3n88ceDcq+h3dLp1iWRz34uDMr9lAob2l5LqZCk5QTBMmKEXXt75UrYbz+no/Gb+WuKOKBvlrZlCpD33nsvaPcSEcYOzuPDZdtoaDTExujiFEoBkJ9vt5rEKhVSdCQ2WIYPt9sIKikoqahl+ZadHKalBAGTkpJCSkpK0O535JA8yqrq+GFjadDuqVTI05FYpUKSJrHBMniwXYI2gpLYr9YWATBGk9iAmTFjBjNmzAja/cbulYsI2qVAKU8FBZCSAtnZTkeilPKgSWywJCTYRDaCktgFa4tISYhln96ZTocSsZ566imeeuqpoN0vKzWBfXplaBKrlCd3j1jREhulQonWxAbTiBGwZInTUfjN/DVFHNQ/m4Q4/VsoUD788MOg3/PIIXk8+slqyirrtN+vUqALHSgVojT7CKYRI2D1aqipcTqSTtu+q5rV28u1lCDA4uPjiY8PbiJ55JA8Gg18uUa7FCgFaBKrVIjSkdhgGjHCNs1etQpGjnQ6mk5ZsMbWw+qkrsCaPn06AOeee27Q7jm6TybpiXF89vMOJo7qEbT7KhVQFRX296+v6uthyxZNYpUKQZrEBtOIEXa7bFnYJ7FfrS0iPSmOvXsGb0WpaOREEhsfG8Nhe+Xw2c87MMYgWgeowt1TT8GFF3buGv36+ScWpZTfaBIbTEOGQExMREzumr+miEMG5Ggv0QCbN2+eI/c9ckgec5ZuY82Ocvbqmu5IDEr5zU8/2e4Ct93WsdcnJsLvf+/fmJRSnaZJbDAlJ8PAgWGfxG4qrWJDUSXnjOnvdCgqQI4cbJegfeXbjVwwdgBd05McjkipTiguhm7d4NprnY5EKeVHmsQG24gRYZ/Eaj1s8Dz55JMAXNjZt0J91Cc7heE9ujDts7VM+2wtuWmJDO+RzogeXeiekYQ34+9ZqQn8eu/uupqbcl5xMeTo7yulIo0mscE2YgS8/z7U1UGQZ537y4I1RWSlxDO0m77NHGgzZ84Egp/EAsy6eAw/bixj+ZadLN+yk2VbdvLsl+upbfB+ckx2agJnHdqPc8b0IzctMYDRKtWG4mJdqECpCKRJbLCNGGET2DVrYNgwp6PxmTGGBWsKGTMohxithw24uXPnOnbv1MQ4xgzK2a2NWl1DI+XV9V69fsXWXTz9xVoe/mgVT3y6hpNH9+KCsQMYrH/8qGArKoL+/Z2OQinlZ5rEBtvw4Xa7bFlYJrH5xZVsLqvmkoH61lw0io+NISs1watz3Qnwmh3lPPPFOmYt2siriwr4v0mj+d3oXgGOVCkPOhKrVETSxQ6CzZ24hmld7HxXPeyYQbkORxIdHnvsMR577DGnw+iUQXlp3HnyKBbccAwHD8jmzzMX886Pm50OS0WLxkYoKdEkVqkIpElssKWl2X6DYZrELlhTRNf0RAblpTodSlR4++23efvtt50Owy+yUxN4+o8HcWC/bK56eTHv/bTF6ZBUNNi50yaymsQqFXG0nMAJYdqhYGtZNR8s28qJ+/bUBvhB8v777zsdgl+lJsbxzHkHce4z33Dlf78nRoTjRnZ3OiwVyYqL7VaTWKUijo7EOmHECFixAhoanI7EJ/d/sJLGRrjiV4OdDkWFsbTEOJ497yBG9c7g8pe+48Nl2zDGUFXbwI5dNawvrGDVtl00NhqnQ1WRQJNYpSKWjsQ6YcQIqKmBdetgr72cjsYryzbvZNZ3G7lw7ED6ZKc4HU7UeOihhwC46qqrHI7Ev9KT4nnu/IM5++lvmPLCtwjQPGe955RRTD64ryPxqQiiSaxSEUuTWCeMGGG3y5aFRRJrjOGu95aTkRzPZeNCP95I8tFHHwGRl8QCdEmK5/nzD+aZL9bR0GhITYwjLTGWlIQ4/m/uz8xeulWTWNV57iRWFztQKuJoEusEd5ut5cvhxBOdjcUL837ewRerC/nHb0eQkRKeCzSEq7feesvpEAIqIzmeP08Yssf+pZt3MuPrDVTW1pOSoL+mVCcU2Y4qOhKrVOTRmlgnZGRAbq4tJwhx9Q2N3PXucvrnpHDWof2cDkdFifHDu1Jb38jnqwqdDkWFO/dIbFaWs3EopfxOk1in5OY2jRCEsFe+3ciq7eVMPX4YCXH64xJs9913H/fdd5/TYQTdQQOySU+M4+Pl250ORYW74mJITw/bZb6VUq3T9+mckpsLhaE9ylReU88DH/7MQf2z+PXe2gbJCQsWLHA6BEfEx8Zw5NA8PlqxncZGo0scq47T1bqUiliODK2JyKUisk5EqkVkkYiMbePc6SJiWnhUNDvvKNe1qkVkrYhcHPjPpBNyckJ+JPY/n66hsLyGGycO176wDnnttdd47bXXnA7DEeOHd6WwvIYfN5U5HYoKZ5rEKhWxgp7Eisgk4CHgLmA/YD7wvoi0Ng35KqBHs8da4BWPaw4A3nNdaz/gbuDfInJqgD6NzgvxkdhFG4r5z6drOWHfnuzXV2vJVPCNG9KVGIGPl29zOhQVzjSJVSpiOTESew0w3RjzpDFmuTHmCmALcElLJxtjyowxW90PYBAwEHjS47SLgc3GmCtc13wSeA64LrCfSie4k1gTeg3dC4ormfL8InpmJnHbiXs7HU5Uu+eee7jnnnucDsMRWakJHNAvi7laF6s6Q5NYpSJWUJNYEUkADgA+aHboA+AwLy9zIbDUGDPfY9+YFq45BzhQRPao5heRKSLyrYh8W19f7+Vt/SwnB+rqoLzcmfu3Yld1HX96biF1DY08fe5BZKUmOB1SVFu8eDGLFy92OgzHHDO8G8u27GRzaZXToahwpUmsUhEr2COxuUAs0Pz9wW1AuzOHRCQDOJ3dR2Fxvbala8a57rkbY8w0Y8yBxpgD4+IcmtuW6worhEoK6hsaueK/37NmRwWPn3UAg/LSnA4p6r388su8/PLLTofhmPHDuwLw8QodjVUdYIxNYnWhA6UiUrj1TDoLG/MLTgfSae4kNoQmd9353nLmrdzB7b8byeF77ZH7KxV0g/LS6JudwkdaF6s6YtcuqK/XkVilIlSwk9hCoAHo1mx/N2CrF6+/EHjNGFPcbP/WVq5Z77pn6HGPDITISOyMrzbw7JfrOf/wAfzhEF3qM1Tcfvvt3H777U6H4RgR4ZjhXflyTRGVtQ6V/qjw5V7oQJNYpSJSUJNYY0wtsAiY0OzQBGxngVaJyMHAvuxZSgCwoJVrfmuMqetYtAEWQuUEa3aUc8tbSzl6aB5/+81wp8NRHlauXMnKlSudDsNR44d3o7a+kS9Xh867FipMaBKrVERzoiD0AeAFEfkG+BLbWaAn8ASAiDwPYIw5p9nrpgCrjDHzWrjmE8DlIvIg8B/gcOBc4Az/h+8n7pHYECgneOCDn0mIi+He3+9LrDaVDykzZsxwOgTHHdTfrt710fJtTBjR/A0XpdqgSaxSES3oSawxZqaI5AA3YXu+LgEmGmM2uE7Z471sEUkHJgO3tXLNdSIyEfg/bKuuzcCVxpjQ7RKfmQkxMY6PxP60sYx3f9rClccMJjct0dFYlGpJQlwMRw7J42NdvUv5SpNYpSKaI1PzjTGPAY+1cmxcC/t2AW1OlTfGfArs74/4giI21v5idTiJ/decFWSlxHPh2AGOxqFa9o9//AOA225r8e+3qHHM8K68+9MW5q8pYmBeKg2NhrqGRuobTYutlvvlpJAUHxv8QFVo0SRWqYjmUH8pBTi+9Oz8NYV8vqqQv00cTnrSHu10VQgoKChwOoSQMG6oXb3rrKe/9ur8Yd3TeevyI0iIC7cGLMqv3Elslq46qFQk0iTWSQ4uPWuM4V+zV9IjI4mzx/RzJAbVvmeffdbpEEJCdmoCz5x7EAUlVcTFCHExQnxsDHGxQozsXl6wsaSSu95bwbTP1nD5rwY7FLEKCUVFkJICSUlOR6KUCgBNYp2Umwvr1jly6w+XbWNxQSn3nDJK33ZVYWHc0K5en/tDQRkPf7ya3+zTkwG5qQGMSoU0XehAqYim77U5KSfHkZHYhkbDvXNWMjA3ldMO6B30+yvv3XDDDdxwww1OhxF2bj5hBImxMdz0xk+YlopmVXTQJWeVapeIJIrIv0WkUEQqROQtEWkzORCRI13nbRIRIyLnBinc3WgS6yR3OUGQ/5N94/tNrNpezrXHDiUuVn8EQllRURFFIdCGLdx07ZLEX44fxperi3hj8Sanw1FO0SRWKW88CJyKbUs6FugCvCMibb1Nm4btLnUVUBXoAFuj5QROysmB2lqoqIC0Npsv+E1tfSP/N/dnRvbqwvEjuwflnqrjpk2b5nQIYevMg/vy+ncbuf2d5Ywb0pWs1ASnQ1LBVlwMw3UBF6VaIyIZwJ+A84wxH7r2nQ1sAMYDc1p6nTHmPeA91/nTgxJsC3QYzkkOrNq1dHMZG0uquOjIQdpvU0W0mBjh7lNGsbOqjrvfX+50OMoJOhKrVHsOAOKBD9w7jDEFwHLgMKeC8lbUj8RmZ2czb948R+6ds2ULo4BFc+awa+jQoNxzV3U9146qJ277cubN+zko91Qd9/jjjwNwySWXOBxJ+Lrj0Bh27FrPux8UkpqgkxijhjEcWVTExvJy1jr0O16pAIkTkW89Pp5mjOno23bdgQag+WjaNtexkBb1SWxxcTHjxo1z5ubxtjfrAf37Q5BieGVhAffP+5HP/zKGPtkpQbmn6rhXX30VwLmf0QhQVdvAsQ9+SlphPO9fNdbpcFSwVFRAXR19R4+mr/77UZGl3hhzYFsniMgdwN/auc7R/gvJGVGfxDrKgXKCoopawPbdVKHv0UcfdTqEsJecEMsfx/TnjneXs6Wsih4ZyU6HpILBPSFSywlUdHoQmNHOOfnAoUAskAvs8DjWDfg8IJH5kSaxTnInsUGcfV5cUUNiXAwp+raqiiJjBtleoQvWFHHK/tpWLirokrMqihljCtmzRGAPIrIIqAMmAC+59vUGhgPzAxmjP+jELidlZoJI0Edic1ITENFJXeHg6quv5uqrr3Y6jLA3vHsXMlPiWbBG25VFDXcSq4sdKNUqY0wZ8DTwLxEZLyL7AS8APwJz3eeJyAoRudzj4zQRGS0io7G5ZF/Xx32DGb8msU6KjbWjBEFMYosraslO01ICFV1iYoRDB+QwX5PY6KEjsUp562rgf8BM4EugHDjBGNPgcc5QbMmB24HA965HMnCr6/ltQYj3F1pO4LScnCCXE9SSnZoYtPupznnwwQedDiFijBmUw+ylWykortRJjdFAk1ilvGKMqQGucD1aO0eafTwPcPwtXR2JdZp71a4gKSq35QRKRRt3Xez8NcFf6lk5QJNYpSKeJrFOC3ISa0diNYkNF5dddhmXXXaZ02FEhMFd08hNS9C62GhRXAxJSZCs3SiUilSaxDotiOUEVbUNVNU1aBIbRpKTk0nW/4T9QkQ4dGAOC9YWYYxxOhwVaLpal1IRz6uaWBEOMYavAx1MVHKPxBpjOxUEUHGl7RGr5QTh47777nM6hIhy2KBc3vlxC2sLKxiUl+Z0OCqQioo0iVUqwnk7ErtAhB9EuEKErIBGFG1yc6GmBiorA36r4nJd6EBFN89+sSrC6UisUhHPl3KCkdgVIDaJ8KJI+C9XFhLcPQyDUBdbVFFjb6kttsLGlClTmDJlitNhRIz+OSl075LEgrWaxEY8TWKVinjeJrEPAJuw7RSSgMnAXBFWiTBVhO6BCjDiBXHp2eJflpzVFlvhIicnhxxt1u43IsJhg3L4ao3WxUa84mJd6ECpCOdVEmsM1xlDX+BI4HHs+roCDALuBPJFeFWE0YEKNGK5f8kGYXLXL0lsio7Ehou7776bu+++2+kwIsqhg3Ioqqjl523lToeiAklHYpWKeD51JzCGL4zhMuAg4FOPQ3HAKcDXIvzOj/FFviCOxBZV1BIXI3RJ1jUuVPQ6TPvFRr6qKqiu1iRWqQjnUxIrwgQRXgNWY0dlwY7Ifg+sBeKxI7PKW8EsJyivJSs1AQlwFwTlP+eddx7nnXee02FElN5ZKfTJTtbJXZFMFzpQKip422LremAKMNC9C2gE3gT+zxg+FyEVWzc7JBCBRqysLNtaKwjlBEUVulpXuOnTp4/TIUSkMQNzmLN0Gw2NhtgY/aMu4mgSq1RU8PZ95X8CBpu87gSeAR42hvXuE4yhQoStwGB/BxnRYmNtIhuUiV012l4rzNx2221OhxCRDhuUyyvfbmT5lp2M7JXhdDjK39yDAprEKhXRfCknWAf8GehtDNd4JrAefkXTaK3yVm5u0CZ2aRKrlPaLjXg6EqtUVPA2iT0ZGGwMDxlDq1N6jWGzMWzwT2hRJCfHt5HY6mrYts3n2xRrOUHYOeusszjrrLOcDiPidOuSxMC8VJ3cFak0iVUqKnhbTjAP6CNCpTH88ltfhFwgBSgzhrIAxBcdcnMhP9/782+5BZ56CjZsgNRUr15S19DIzup67REbZoYOHep0CBFrzMAcXvomn5E3z2n33JSEWP556j4cPaxrECJTnaZJrFJRwdsk9hngJGw5wcMe+ycDDwH/A07za2TRJCcHvv/e+/O/+86WH8ycCeef79VLStw9YnW1rrDy97//3ekQItaUIweSmhhHQ2P7ix58ubqQi2Ys4qlzDuTIIXlBiE51SnExJCR4/Ue+Uio8eZvEHuLavtZs/+vYpPYQVMfl5vpWTvDzz3Y7bZrXSWyRK4nVcgKlrH45qdw4cbhX55ZW1nLGk19z4fPf8uy5B3HYXrkBjk51inuhA20nqFRE87Ym1j30UNpsf1mz46ojcnNtnWtlZfvnVlXZ0oOePeHrr+GHH7y6RdOSs5rEhpPJkyczefJkp8OIepkpCbx4wSH0z0nlT899y9drdUJYSNPVupSKCt4msbtc22Ob7Xd/rOs3doZ76VlvRmNXrwZj4MYbITERnnyy9XNLS+GRR6CxUUdiw9To0aMZPXq002Eo7B+AMy44hJ6ZSZw3fSGLNhQ7HZJqjSaxSkUFb5PY77A9Yp8R4SYRThXhJuBpbP/YRYEKMCr4smrXypV2O2YMnHYavPBC6yO4V14JV1wB339PcXkNAFmaxIaVqVOnMnXqVKfDUC556Yn898JD6dYliT8+s5DFBaVOh6RaUlSkSaxSUcDbJPYJ17YLcCvwimub2ey46gh3EutNr1h3PeyQIXDRRbBzJ7zyyp7nzZ1rE1yAHTsorqhFBLJSNIlVqjO6dknipQsPITs1gXOe/pqlm7UxS8jRkVilooJXSawxvA48gB2N9XwA3G8M/wtMeFHCl3KClSttPWxaGhxxBAwbZid4eaqqgosvtiuBARQVUVRRS2ZyvC6xGWZOPfVUTj31VKfDUM30yEjmxQsOIS0xjrOf/oaft+1q/0UqeDSJVSoqeL1ilzFch+1CcCfwlGt7iDH8JUCxRQ9fygl+/hncvUNFYMoUWLAAfvqp6Zzbb4c1a+A///nlurpaV3gaM2YMY8aMcToM1YI+2Sm8eOGhxMUIZz71NesKK5wOSUHTJFlNYpWKeL4sO4sxLDSGvxvDFNd2YaACiypZWTYhba+cwBg7EjtkSNO+c86x/RDdE7x++gnuvRf++Ec45ZRfrltUUUuOLnQQdq677jquu+46p8NQrRiQm8qLFxxCQ6PhzCe/oqDYiw4jKrBKSuxWk1ilIp63fWIRIQ6YCAwFkpsfN4bb/BhXdImLg8zM9kdii4rsL2jPVZxycpomeN19t62TzciA++6D2Fj7i7yoiJLMWgblpQX001AqGg3uls6MPx3CGU9+xR+e+orLxu21R3vSQwbk0D9XG+8HhXu1LneZllIq9In0/eW5MV4vYepVEitCV+zSs22tgalJbGfk5rY/EuvuTOA5Egu2pOCll+DEE21pwXPPNZUo5OTYcoLutRw0QMsJws2JJ54IwFtvveVwJKotI3p24fnzD+bsp79m6us/7XE8OT6WW04cwekH9kG0AX9g6ZKzSoWj9dhuVwYfBli9PfFWYFgbx9tft1G1zZVstsndmWBos78ljjzS7vv4YzjmGDj77N2ua4qKKKms1R6xYeiYY45xOgTlpX37ZPLVjcdQWlm32/6KmnpueXspf33tJz5fVchdp4yiS1K8Q1FGAU1ilQpXPv+F720Seyw2UZ0OnOd6fhVwhev5Pb7eWDWTmwsbN7Z9zsqVEB8P/fvvvl/E9oT961/h8cd3X2oxN5eG9RtoNLpaVzi66qqrnA5B+SAlIY6UhD1/rb5w/iE8/ukaHvjwZ37YWMrDk/djv75ZDkQYBdzvaGkSq1Q4+YwODIh6m8T2cm2nYpNYjOERET4BfgJ6+3pj1UxubvtLyP78MwwaZGtom7vkEjj3XEhJ2X1/Tg5m0XeAJrFKOSUmRrjs6L04dGAOV/73e37/xAJO2Lcn3bokkZ0aT3ZqItmp8QzKS6NfjtbOdoqOxCoVWkQeAAzGXIvIOQAY8/xu5xgzriOX9jaJbQDigSKgDogTIQ/Y4Do+BbijIwEoF2/KCVau3LOUwE1kzwQWIDeXGNcvde1OEH6OP/54AN5//32HI1H+cEC/LN67aiy3vr2UL1fb1nd1DbsPPhzcP5tJB/Vh4qgeJCfEOhRpGCsutpNa09OdjkQpZV0NNALXYt/RbwSeb+N8r3mbxBZhR2MzgK3YkdcXgWrXcX1frLNyc+0iBZWVLSejDQ2wejX85je+XTcnh9jqKpLqqnUkNgydcMIJToeg/CwjOZ4HTh8NgDGG8pp6iitqKaqo5Zt1xcxcWMC1r/7ALW8t5Xf79eS4vXuQFO9TN8QOGdGzS4ulEH5TVgZTp0JFgPvpfvONHYXVCXRKhYpGQBDp4vrYb/84vf2NtRKbxA7C1i2cCbhnnBjgO38FFLU8l55tKYndsAFqa1sfiW2Nq81MVtUuTWLD0KWXXup0CCqARIT0pHjSk+Lpl5PK/n2zuOjIgXztSmZf/XYjM77yuttMp5x1aF/uOGlU4G7w9tvwxBPQt68dKQ2kk08O7PWVUr7YDnQD1v2yR2RtK+cajBnk7YW9TWKfBFYDSdhOBccCea5jO7BDxaozPJee7dNnz+PuzgTN22u1x5UcZ1ftJCtVZ0QrFepEhEMH5nDowBxuOWFvftpUhglwA5h/vLmULaXV7Z/YGV99ZZfLXrs28EmsUiqUfAKcQdO79gL0b+Vcn37ZeZXEGsMrwCvuj0UYDBwN1ANfGkOpLzdVLfAciW2Ju0dsB0die9ZXkBin/3GEm/HjxwMwd+5chyNRTshIieeIwbkBv0/PzCRKq+raP7EzFiyAgw/WBFap6PNnIBbYH9gLm6j65S2mdpNYERKBZa4Pf2MMK4xhJ/CmPwJQLp4jsS35+We7qldeXsvHW+NKjns1VnU8NuWYSZMmOR2CigKZKQks37IzcDeoqLDdV6ZODdw9lFKhyZjtwGQARBpd+wb449LtJrHGUCNCDpAOtFbDoDrLPRLbWhK7cqUtJfB1ssIvI7HlnQhOOeXCCy90OgQVBTKT4ymrDOBI7Lff2smphx4auHsopUKTZ4utprUG/MLbKa/u9zL39deNVTPunoatlRP8/LPvpQQe1+1aF+AZwUqpsJWVkkBpVR3GBKj29quv7FaTWKWi0dXYBbIAngWe8deFvU1iHwSKgf+KMEmEoSL09Xz4K6CoFRdnywVaGomtqICCAt8ndQHEx7MrKZXc6l2dDlEF37hx4xg3bpzTYagIl5kST0OjYWd1fWBusGABDB7c9I6TUiqaON5iy70cWDbwUgvHjQ/XUq3JzYWtW/fcv3q13XZgJNYYQ0lSOllVmsSGo3PPPdfpEFQUyEyx7ffKKuvISPZzFxNjbBJ73HH+va5SKlw43mIL/Jg5q1YcfjjMnAmbNkGvXk373Z0JOjASu6umnuLkLuRUlvkpSBVMmsSqYMh0Ja4llbX0zWmhT3VnrFsH27drKYFS0cvZFlvAc75cVHXQzTfDSy/BrbfCtGlN+909YgcP9vmSxeW1lCSn029XqX9iVEFVV2cn28THa49fFTjuHtIBabPlrocdM8b/11ZKhQPPFlvuUdbgtNgCMIbz/HEz1Y4BA+Dii+Gxx+Daa5vKB1autAsgtLSSVzuKKmopSe5CUulqPwergmHChAkAzJs3z9lAVETLSLblBKWVtf6/+IIFkJoKI0f6/9pKqdC3Z4stE7QWWyrIbroJnn3Wbl991e5bubJjnQmAYlcSm7C+xI9BqmC54IILnA5BRYGsFNdIbCDabLkXOYjT/26UUhztz4t59VtFpN12CMYY/uSHeFTXrnYU9tZbYeFCOPBAW05w1lkdulxxRQ0lyenEVpRDbS0kJPg5YBVIZ3Xw+66ULzI8amL9qrLSLnLwl7/497pKqfAhYjtYGZOPe3KXe19L7Hle8fZP43NpvdhWXMc0ifWXa66BRx+1q9u89BKUlXWsvRZN5QT2gyLo0cOPgapAq6ysBCClA6UkSnkrLjaG9KQ4/4/ELloE9fU6qUup6LYe22YrzvW8rclbPnW78rZPLNhktaWHz0TkUhFZJyLVIrJIRMa2c36CiNzmek2NiOSLyJUex88VEdPCI6kj8TmuSxdbTvDxx/DII3ZfR8sJymspT8uwH7S2GpgKWRMnTmTixIlOh6GiQGZKvP9rYhcssFtNYpWKdtLseVsPr3mb7TYvwI0DBgJ/B/YDfuvtDUVkEvAQcCnwhWv7voiMMK0PIb8M9AamAKuw/caSm51TSdOsNwCMMdXexhVyLr4YHnwQ7r7bftzBkdjiiloas+3Ss62uBqZC1iWXXOJ0CCpKuFft8qsFC2CvvSAvz7/XVUqFk+dpGn31fN5p3nYn2NDC7jUiLAAKgUuAT7285zXAdGPMk66PrxCR41zXuKH5ySJyLHAMMMgY4x5KXN9imMa0sFJAmEpMhNtug3POsc/7dmxRtOLKWtJzXEmsjsSGnUmTJjkdgooSGcnxlPiznMC9yMGxx/rvmkqp8GPMuS0+9wNfyglaEofNqL1aikVEEoADgA+aHfoAOKyVl50ELASuEZGNIrJKRB4WkbRm5yWLyAbXOe+IyH5txDFFRL4VkW/r6wO0zKI//OEPMGoUjBgBsbEdukRxRS2xea6lHnUkNuyUlZVRVqYLVajAy0pJoMyf5QQbNsC2bVpKoJQKmM50J0gCDgcSAW//l83FNrzd1mz/NmB8K68ZCBwB1ACnApnAv4GewGmuc1YC5wM/AOnAVcCXIrKvMWZV8wsaY6YB0wBSU1P9Nqztd7Gx8MEHUN3xqoii8lqG9XK9ladJbNj53e9+B2ifWBV4mSl+Hol118PqIgdKRTeR9jpceTIY43WjgM52J3AX4L7n7Q07IMZ17z8YY8oARORyYI6IdDPGbDPGLAAW/BKUyHxgMXAFcOWelwwj3bt36uXFFbV0yepum41rOUHYufLK8P7xVeEjMyWBndV1NDQaYmP8sMq4e5GDUaM6fy2lVDg7F+/qYH3uduVL9+mWfqvVAP8FrvbyGoVAA3ZilqduQGv1rFuATe4E1mW5a9uXPUd1McY0iMi3gO/rtEaQqtoGquoayE5LgJwcHYkNQ6eccorTIagokZkcjzGws6qOrFQ/9JP+6is46CBd5EApBR3sZtWejnYnAKgxptXEs0XGmFoRWQRMAF71ODQBeK2Vl30J/F5E0owx5a597qn6LU04Q0QE2AdbXhC1iipqAMhJTYDcXB2JDUOFru9Zbm6uw5GoSJeV6lq1yx9JbFUVfP89XH+9HyJTSoU5z1W60oH/AKXA/cBGbPepa7Elpxf6cuHOdCfoqAeAF0TkG2yCejG2vvUJABF53t7TnOM6/yVsK69nReQWbE3sQ8AsY9fjRURuBr7Ctt/qgi0h2Afb8SBqFVfYSRrZqYk6EhumTjvNln1rTawKtMxkm7iWVNYygFT4/HOYPbtjF9u+XRc5UEpZxjR1rxJ5DOgOHIEx6zz2f4rN4U4A3vL20t5O7DoOOBj43hje9th/IjAa+MYYvPptZ4yZKSI5wE1AD2AJMNEY406U+zY7v1xExmMncy0ESoA3gKkep2ViJ2p1x04y+x440hjzjTcxRartO+1IbLZ7JHbdunZeoULNtdde63QIKkpkptiR2DL35K4bb4Qvvuh4OUCPHnDEEX6KTikVIU53baua7Xd/fAo+jMZ6+9vpH8AhwPHN9pcDt2AnVXn9J7sx5jHgsVaOjWth30qg1WaDxpg/A3/29v7RoLHR8Pina8hKiWdItzQ7EqvlBGHnhBNOcDoEFSUyU5pGYgH7zs2pp8KsWQ5GpZSKMImu7WuI3E1TOYF7YDLel4t5m8QOc20XNNvvHukc7stNVeD9d2E+izaUcN/v9yU9Kd6OxJaW2rf4dKJF2Ni61Zadd+9klwql2pPlGoktdY/ElpRAVpaDESmlItAc7GjrocCbzY4Z13GvebvYQYpr23yBgfRmx1UI2L6zmnveX8GYgTmcun8vu9O9aldxsXOBKZ9NnjyZyZMnOx2GigLpSfGIQKl7JLa0FDIznQxJKRV5rsD29pcWHivxsS2qt0NyW7C1qn8DLvfYf6Nru9mXm6rAuvWdZdTUN3LnySOxjRpoSmKLiqBrV+eCUz6ZOnVq+ycp5QexMUJGcjylVXV2gZXqah2JVUr5lzFbsCuqngP8CsjBtl/9BHgeY3xa3cnbJHYutvnsJSIci82WhwKDsMO/c325qQqcT1Zs590ft3DNhCEMzPMYOM/VpWfD0XHHebWis1J+kZnsWrWrpMTu0CRWKeVvNlH9ZeXUzvC2nOAe7CQusInrRNdWgArXceWwytp6bnpjCYPyUrnoqIG7H3SPxOrkrrBSUFBAQUGB02GoKJGZkmDLCTSJVUqFAa+SWGNYg+0OsILd6xeWAccaw9qARai89tDcVWwqreKuk0eRGBe7+0EdiQ1LZ599NmeffbbTYagokZkSbyd2lZa6dmQ6GY5SSrXJ62nqxvAVsLcIg7DLxG5zJbcqBCzfspOnvljHpAP7cMjAnD1P0JHYsHTTTTc5HYKKIlkpCazZUQ4lrpaNOhKrlAphPvdaciWumryGmOcXbCAxLoYbJg5r+YSUFEhM1JHYMDN+/HinQ1BRJCM5ntIKrYlVSoUHr8oJRHhRhAYR/t5s/99d+18ITHjKG/UNjcxesoXxw7v90rB8DyK2pECT2LCydu1a1q7Vah0VHFkpCeyqqaehyNWKT5NYpVQI83Zi12Gu7Yxm+1/A1sbq2oIOWrC2iJLKOiaO6tH2ibpqV9g5//zzOf/8850OQ0WJrFS74EF1oeuPXa2JVUoFmoggkteRl3pbTuDOjrY227/NtdXlhBz07o9bSE2IZdzQdn4GdCQ27Nx6661Oh6CiSEayTWLrdhRBairE+7QCpFJKtU3keOBo4CuMeR2Rs4HHgBREvgcmYsx2by/n7Uisu/nsmGb7xzQ7roKsrqGROUu3Mn5EN5LiY9s+WUdiw85RRx3FUUcd5XQYKkpkucqR6ouKtZRAKRUIlwLXAqmIJAOPAqnYd/X3A27z5WLeJrE/uW4wXYSzRDhAhLOAZ7GLHfzky02V/yxYY0sJftNeKQHYJFZHYsPKypUrWblypdNhqCiRmWJHXhuLSzSJVUoFwj6u7efAwUAasBx4B5tn/tqXi3mbxE53bXsBzwHfuLZ9mh1XQfbuj1tIS4zjyCFelJPk5kJxMTQ2Bj4w5RcXXXQRF110kdNhqCjhHomltFTrYZWKEiKSKCL/FpFCEakQkbdEpHc7r7lBRBaKyE4R2SEib4vISC9u505WNgEjXM8fBP7oet7Tl9i9qok1hqdFOA44tYXDs4zhGV9uqvyjrqGR2Uu3Mn541/ZLCcCOxDY22v+gsrMDHp/qvLvuusvpEFQUyXCNxMaWlULvIc4Go5QKlgeB3wFnAEXAA8A7InKAMaahldeMw9ayLsSOoN4GzBWREcaY4jbuVQskArnYUVmDXUir2uO413xZ7OD3IpwOnIBrsQPgLWN41ZcbKv/5cnUhZVV1/GYfL/9w8Vy1S5PYsHDYYYe1f5JSfpKeGEdsjBC/s0zLCZSKAiKSAfwJOM8Y86Fr39nABmA8MKel1xljdnvb3/WaMuBw4O02bpkP7A18gR11NcAS7Dv9AF5P6gLvywkAMIZXjOFsYzjWtX1VhDSRX4aBVRC9++MW0hPjGDs417sX6KpdYWfJkiUsWbLE6TBUlBARMpPjSdylSaxSUeIAIB74wL3DGFOArVP1ZRQlHZtTlrRz3kvYkdsB2BHZjzGmhKZWrd/5cE/fV+wCECEGOA44Gzsym4StkQ072dnZzJs3z+kwfGaAAfU7+evoeL768nOvXpO+fj0HAD/Nm0dRTU1A41P+cfXVVwPw4IMPOhqHih4XD6oksbqSdaWlbAjD341KRYE4EfnW4+NpxphpHbxWd6ABaD66tQ3f2qc+BCwGFrRz3j9d9xsLrKOpG0Ec8DTwmg/39C2JFeEgbOI6CVvPADajNr5cJ5QUFxczbtw4p8Pw2Scrt/Ov2Qt5+o/7M254N+9e1LcvAKN69oQw/Jyj0ZNPPgnAQQcd5HAkKlpMv/c9LgQG7LcfA/T3hFKhqN4Yc2BbJ4jIHcDf2rnO0f4IRkQewI6kHtFGDa1ljAHudT089z8FPOXrvdtNYkUYAJwFnAkMdu/2OKUKeMPXG6vOeffHLaQnxXGEt6UEoOUEYUiTVxVsPRur7BMtJ1AqnD3InqusNpcPHArEYgcmd3gc64Ztg9UmEfk/YDJwtDEm6Gukt5rEinAxNnn1XOBAmp1mgG7GUB6A2FQrauvtAgcTRnQjMc6LrgRuXbpAXJz2ig0jixcvBmD06NGOxqGiR9cGTWKVCnfGmEL2LBHYg4gsAuqACdh6VVzttYYD89t57UPYd+aPNsasaOPEtkdnd2cwxusqgbZOfAybpLoT11pgLrZeYQ0wD0AT2OD7YvUOdlXX89t9vFjgwJOIrtoVZtw1seFYt63CU9d6VxKrfWKVinjGmDIReRr4l4hsp6nF1o/YnA8AEVkBPGKMecT18aPY8tKTgBIRcdfPlhtjmueFzQdA/cabbNcAzwDXG0MpgAh7Byog1b5PV+4gNSGWI/byYoGD5nTVrrCiE7pUsGXXVQBQm96FBIdjUUoFxdVAPTATSAY+As5pVt86lKa5UGCXj8V1rqdbgVua7ctn97lTOdiVuuqwSXMOtkNCJT622PJ2yPZ84AQR/ocdidWhPAetL6pkQF4qCXE+dUizcnM1iQ0jWkaggi2rxiaxO5PS8KHiXikVpowxNcAVrkdr50hbH7dzg/6/PBc5APgYuB/4O8ZUI5IE3AlchJ1/5bW2sqB7gALsMLAAXYEp2Ma3X/hyE+VfBcWV9M1O6diLtZwgrCxcuJCFCxc6HYaKIl1q7DuBxQlpDkeilIpAD2JHYW/DGLtKl93eAqQA9/lysVaTWGO40Rj6Y5cWewq7EoM7oU3BNTQswkYR7vHpU1Ad1tBo2FhSRZ+OJrE6EhtWrr/+eq6//nqnw1BRJL2qgprYOEoaO/BOj1JKte0A1/bgZvsPcW338+Vi7ZYTGMNnwGciXI5d2OAsYCK2fgHssmHXA1N9ubHqmG07q6ltaOz8SKwxdqKXCmmPPPKI0yGoKJNSsZOdSWmUVNU7HYpSKvLsAHoDbyPyHrDR9fFE7ODojjZeuwev2xgYQy22HvY1EbKwfcHOxLdlyVQn5RdXAnQuia2vh127bMstFdJGjhzpdAgqyiRX7GRTYhplVbVOh6KUijyPA3dhl5w92WO/e+GsR325WIfeLzKGEmN43BiOAPZiz5loKkDcSWy/7NSOXSDXNVVD62LDwvz585k/v81WfUr5VcLOMsqS0iiprHM6FKVUpDHmHuxSszU0lagKUI2tk/2XL5fzadnZluNhLXB7Z6+jvFNQXElsjNAjM6ljF3Cv2lVUBAMH+i8wFRA33ngjoH1iVfDE7CyjPDmNUk1ilVKBYMwt2JW+xmDbaxUCX2FMma+X6nQSq4Irv7iSnplJxMd2cNKFeyR22zb/BaUC5j//+Y/TIagoIyUlVKX1p7RSywmUUgFiE9bZnb2MJrFhJr8z7bUARo6E5GR47z347W/9F5gKiKFDhzodgoo2JSXU9NhHR2KVUv4h8g+fzjfmNm9P1SQ2zBQUVzJhRLeOXyA9HU48EV55BR56COLj23+NpylTYMQIcC2HqgLr008/BeCoo45yOBIVFRobobSU+i6ZlOhIrFLKP25h9xW72uN1EquNAMNIRU09heW1He8R63bmmbYmds4c3163YAE8+SS8807n7q+8dvPNN3PzzTc7HYaKFjt3gjE0ZmRQVqUjsUopvxEvHz7RkdgwUlDSyfZabr/+NWRnw4sv+lZScMcddrvDpzZuqhOeeeYZp0NQ0aS0FACTlaUjsUopfznP43k8cCs2YX2Kpj6xFwCxwE2+XLjVJFaEI325kGtRBBVA+UV+SmITEuD00+G552y/2PT09l+zaJGto01M1PZcQTRQO0ioYCopASA2O1trYpVS/mHMc788F7kD6A7sjzE/eOz/H7AIGOzLpdsaiZ2H9zUMpp1rKT/o9EIHns48E554At58E846q/3z77gDMjPhjDPgqad0xa8gmTt3LgDjx493OBIVFdxJbG42NVsbqaptIDkh1uGglFIR5HzXNr/Z/g2u7dnAX7y9WHs1sd7WMGg2EwQFxZWkJ8WRkezjZKyWHHYY9OtnSwra89NP8MYbcNVVMGgQ1NXZ2jkVcHfccQd3uMs4lAo0VxKbmJsNQKmu2qWU8q9M1/ZJREYikonISOBJ136flhJta/T0uWYfH4sdAv6SphqGw7FNat/15aaqY9zttcQfI6AxMXZU9d57Yft26Nq19XPvvBPS0uDKK5smde3YARkZnY9DtemFF15wOgQVTVw1sUl5ucB2Sirq6JGR7GhISqmI8gUwHrvk7MnNjhnXca+1OhJrDOe5H8BcbAI7yRiONIY/GMORwBlALjaxVQHW6R6xzZ15JjQ0wMyZrZ+zYoVtx3X55XYyWF6e3a+Tu4KiT58+9OnTx+kwVLRwjcSmdrP/znUkVinlZ1cAO2j5Hf0dwJW+XMzbFlvu2WLNV1d4z3Xj6325qfJdY6OhoKTKv0nsyJGwzz7w0kutn3P33ZCUBNdcYz/WJDaoZs+ezezZnV7URCnvlJRATAxdumUB6OQupZR/GbMSGAn8E/gGWAN8DdwDjHId95q3k7H6u7aXAv/y2H+Za9vPl5sq323fVUNtfWPne8Q2d+aZ8Ne/wpo1tt7V09q1tmb2yiubklf3srWaxAbFPffcA8Bxxx3ncCQqKpSUQGYmmSmJgCaxSqkAMGYHcIM/LuXtSOzPru3dImwTYbEI24C7sDUMP7f+UuUPfu1M4OmMM+y2+WhsbS3cfDPExcF11zXt15HYoHr55Zd5+eWXnQ5DRYvSUsjKIjPFTh7VXrFKqYAQORSR2xB5wrU9uCOX8TaJ/RvQiC0dyAVGubaCTWJv7MjNlfcClsT26QNHHmlHXI2B776zXQh69YIZM+DSS6Fnz6bzU1MhOVmT2CDp3r073bt3dzoMFS1KSiAri6T4WJLiY3TVLqWU/4k8jp1L9TfgQtd2ASKP+nopr5JYY3gHOA5bt2BoSl6/Ao41RrsTBFp+cSUxAj0zAzBT+MwzYeVKGDoUDjjA9o89+mjbieDee/c8Py9Pk9ggefvtt3n77bedDkNFC1cSC5CVkkBJhY7EKqX8SORc4CJanth1MSLn+HI5rxcoMIaPgI9ESAGygBJjqPTlZqrjCoor6ZGRTEKct4PnPvj9723pQHY2PP44TJr0y39kLdIkNmjuv/9+AE444QSHI1FRoaQE+vYFICM5nuVbd/LKtwUdulROagLHDO/mz+iUUuFvimu7Afg/17Yv8Gfs/KuLgOe9vZhPq2yJEIedVZZjDO/78lrVOX5vr+UpKwu2bPH+fE1ig2bWrFlOh6CiiasmFmBgXirv/bSVv8z6scOXm3vNkezV1YtlrZVS0WIk9p38EzBmyS97RT4BfnQd95rXSawIvwcewdbCGiBOhI+AAcDFxvCBLzdWvskvruRXQ9tYkCCY8vJg+XKno4gKue5uEEoFmjG7lRM8NHk/bpxY3aFLbS6t5vT/LODb9SWaxCqlPCW4thub7d/Y7LhXvEpiRRgL/Jc9l5h9F7gPOA00iQ2UqtoGduyqoW9OgEZifaUjsUHz+uuvA3DKKac4HImKeJWVdklpVxIbHxtD76yO/c7plZlMZko83+eXMvngvv6MUikV3gqAgcB9iFyHMaWIZAD3ehz3mrcFlje4zm3ehNY9oWuMLzdVvikosaXHfu8R21F5efY/vEotiQ60hx9+mIcfftjpMFQ0cK3WRWZmpy8lIuzXJ5PvC0o6fS2lVER5BzsYeh5QhEgZUAycj32X36eZzN4msYfirmHY3VrXtpcvN1W+yS8KUHutjtJesUHz5ptv8uabbzodhooGpaV229akTh/s3zeLVdvL2VmtbbqUUr+4A8in6Z39dI/nG4A7fbmYt0lsqmub32x/pmsbgL5Pyi1gPWI7SpPYoMnIyCAjI8PpMFQ0cI/E+imJ3a9vFsbADwWlfrmeUioCGFMEHAI8DWwB6oHNwJPAGIwp9uVy3k7s2oRdWrZ52YB7KafmBbrKj/KLK0lLjCPLtYqO4zSJDZqZM2cCMGnSJIcjURHPz0nsvn0yEIHvNpQydnCeX66plIoAxmzDLnLQad4msXOwvbvecO8QYQUwGFtmMMcfwaiW5RdX0ic7BRFp/+Rg0CQ2aB5//HFAk1gVBH6siQVIT4pnSNd0rYtVSgWMt0nsHdgOBDnYpBVsAitAEXC3/0NTbvnFlQzKS23/xGDRJDZo3nvvPadDUNHCzzWxAPv3y+S9n7bS2GiIiQmRP8KVUhHD22VnNwGHY9toNWKT10bXx2Ndx1UANDYaCgK50EFHdOkC8fGaxAZBSkoKKSkh9L3vjPJyeOUV249UhR4/j8QC7Ncni7KqOtYVVfjtmkop5eb1GqbG8LMxHIedSdYbSDeG44xhRcCiU+wor6GmvjG0klgRyM3VJDYIZsyYwYwZM5wOwz+mTrVLGq/QXxkhqaTE/oEaG+u3S+7XNxOA7zZoSYFSyv+8SmJFyBChrwi5xlBtDJuNoVqEXNd+nT4dIO7OBCHTI9ZNFzwIiqeeeoqnnnrK6TA6b8UKeOIJ+3zrVmdjUS0rKfHrKCzAoLw00pPi+F47FCilAsDbkdhngHXAH5rtn+za/7Q/g1JNQq5HrJsmsUHx4Ycf8uGHHzodRuddf31TGYH+3ISm0lK/1sMCxMQIo/tk6kisUiogvE1iD3FtX2u2/3Vsfewh+EBELhWRdSJSLSKLRGRsO+cniMhtrtfUiEi+iFzZ7JxTRWSZ6/gyETnZl5hCVX5xJSLQKyvEWvHm5UFhodNRRLz4+Hji40OktVpHffQRvPMOXH21/ViT2NBUUuL3JBbsogc/b9tFeU2936+tlIoQIn1/efjA2yTW3eSvtNn+smbH2yUik4CHgLuA/YD5wPvSduAvA8cBU4ChwO+BHz2uOQaYCbwIjHZtXxURn5LrUJRfXEmPLkkkxvmvTs0vdCQ2KKZPn8706dOdDqPjGhrg2muhXz+49Va7T39uQlOAktj9+mbSaOBHLSlQSrVuPfad/bXtnLcbb5PYXa7tsc32uz8u9+Ge1wDTjTFPGmOWG2OuwK7acElLJ4vIscAxwERjzIfGmPXGmK+NMfM8Trsa+MQYc6frmncC81z7w1Z9QyOfrypk3z6ZToeyp7w8KCuD2lqnI4loYZ/EPv88/PAD3HMPpKVBdrYmsaEqADWxYDsUAFoXq5Rqj3v5Wa952yf2O2A88IwIewPLgeHYhNQAi7yKTiQBOAC4r9mhD4DDWnnZScBC4BoROQeoAt4HbjTGuJPnMcC/m71uDnC5N3GFqi/XFFFYXsPvRvdyOpQ9uXvFFhZCz57OxhLB5s2b53QIHVdeDn/7Gxx6qO1KAPbnZvt2Z+NSLQtATSxARko8g/JS+T5f62KVUq36jKZ1CLzmbRL7BDaJ7QLc6rFfXDd9wsvr5AKxwLZm+7e5rt+SgcARQA1wKpCJTVh7YhdgAOjeyjW7t3RBEZmCLU0gISHBy9CD783vN9ElKY6jh4Xgko2eCx5oEqtact99sGULzJpl27IBdO2qI7GhqLYWKisDksQC7Nc3i49XbMcYEzorDyqlQocx4zryMq+SWGN4XYQHsCOvzd1vDP/ryM29FINNlP9gjCkDEJHLgTki0s3YNXh9YoyZBkwDSE1NDcnO61W1DcxZupUT9u0ZevWwoKt2BcmTTz4JwIUX+mWZ6c4xBpYsgTVr2j+3pgb+9S84/XQ4zONNlrw87RMbitwLHQQoid2/bxazFm0kv7iSfjkhtPqgUiqseTsSizFcJ8JM4ESgG3ak8y1jWOjD/QqBBtfrPXUDWmseuQXY5E5gXZa7tn1dcWz18Zoh78Pl26iobQjNUgLQJDZIZs6cCTiYxDY2wldfwf/+B6+/Dmt9qLlPS7O1sJ7y8uDzz/0bo+q8AKzW5emXRQ/ySzSJVSqaiRzZxlEDFGHMMm8v53USC+BKWH1JWpu93tSKyCJgAvCqx6EJ7Nm+y+1L4PcikuZRAzvEtd3g2i5wXePeZtec39FYnfbG95vokZHEIQOynQ6lZZrEBsXcuXOdu/kDDzSVBMTHw/jxcMMNcMABTeUBbenVq+nnxC0vD4qKbHIc4/WCgSrQSkvtNkAjsUO6pZOaEMv3+aWcvF/vgNxDKRUW5tFe7avIJuASjHm3vYt5ncSKkA5MBPoBSc2PG8NtXl7qAeAFEfkGm6BejK1vfcLeR5631zPnuM5/Cfg78KyI3IKtiX0ImGWMcc8QeQj4TESmAm8AJwNHY2tpw05xRS2f/byDP40dQExMiNaPZWXZJEST2Mj1wAO2m8D998PEiZDhh4X58vJsAltcbJcuVqEhwOUEsTHCvn0y+U4ndyml2u9A0Bt4HZGDMObHtk70KokV4SDgPaCtYUGvklhjzEwRyQFuAnoAS7Dts9yjqn2bnV8uIuOxk7kWAiXYRHWqxznzRWQycIcrjjXAJGPM197EFGre/XEz9Y2Gk0K1lADs+uraLingHnvsMQAuvfTS4N64sRG2bYOzz4YzzvDfdd0js9u3axIbSgKcxIItKXji07Ws3r6L+NjAjsLnpSeSkuDTG41KqeB4DvtOeU/su+X5QB/gcGAz8D12on8Cdh7WuW1dzNt/5Q8COW0c92lylDHmMeCxVo6Na2HfSvbsUdv8nFnALF/iCFVvLN7MsO7pDO/RxelQ2qYLHgTc22+/DTiQxJaUQH09dG+xwUfHde1qt/pzE1oCXBMLcGC/bBoa1zD+gc8Cdg+3fXpn8NblYflGnFKR7iPgHOAMjJn5y147EPkidoDyUezA6VHtXczbJHYfbKL6KbZ2tYIO9PNS7csvqmTRhhL+etwwp0NpnyaxAff+++87c+NtrqYf3ZrPl+wkraUOTQGuiQU4akgeT5y1P5W1DQG7B8D/vt/EjxvL2j9RKeWEm1zb5vWu72DLDG7EmBGIlNFKm1RP3iaxpUAKcIoxeyw9q/zozcWbADhxdBj0Xs3Lg2VeTyJU4WSrq7GHJrHRoaQEkpMhMTFgt4iJEY4b2SNg13dbX1jBF6sLaWw0oTunQKno1c+1vQqRuzDGPSB6sWs7wLXdhRc5qreFSc+7tiO9PF91gDGGNxZv4uAB2fTKTHY6nPbpSGzAPfTQQzz00EPBv7F7JNbf5QTuOlj9uQktJSUBHYUNpi7J8RgDu6rrnQ5FKbWnla7tbcB2RBYjsg34J/Yd/pWIxGLbpG5u72LejsSuB8qAN0V42hVEnecJxvyS6KoOWrJpJ2t2VHDB2IFOh+Idd7ukhgY70Uv53UcffQTAVVddFdwbB6qcID7e1l3q0rOhpaQkoPWwwZSZYldhLKuqIyMl3uFolFLN3Ai8iV29NZumhgEC1AM3AL8C4rEdrNrkbRL7H5pqYK9t4bgBTWI7643Fm0iIjWFiEN5y84u8PLuKU3Hxnv1AlV+89dZbztx461abcAZidE6Xng09paURMxKbkWwT19KqWvqS4nA0SqndGPMeIhOAO4FDsBUBjdh+/3/DmE8RiQPSgZr2LudLDxItLgqweSu3c/heOeEzeuBZ36hJbGTZts2OwgZinXstQ/FedTX8+KP9YzGQNm+GwYMDe48gyXT9/iyrqmvnTKWUI4yZBxyOSAqQBRRjTJXH8XrsqGy7vE1iz/MxROWjxkZDQXEV44f7+e3bQNJJOgF33333AXDdddcF98Zbt/q/lMAtLw9Wrw7MtSNFVRVMmwb//KddMS0Yjmq3m01Y+GUktlKTWKVCjsg84GlgFsZUApWduZxXSawxPNeZm6j2bdtVTW1DI32yw+jtL01iA27BggXO3HjbNugRoLKWvDxw6vMKdZWV8MQT8K9/2e/BuHHw0EOQlhb4ex96aODvEQSZyToSq1QIOxIYC/wbkZnAM3RiYSpd0iRE5BfZP0b6ahKrPLz22mvO3HjbNthvv8BcOy8PCgvtqmAxgV25KeQVF8PSpbBkiX3MmmUnvf3qV/DKK3DkkU5HGHa6aBKrVCirxa7G1QW4ALgAkeXAM8ALGONTQuH1/yAinCXCdyJUiNDQ7KG9TDopv9gmsf1ywiiJ1XZJkcm95Gygygm6drUdLdyrREWbhga47jro1QtycmyieumlMGMGHHggfP45fPSRJrAdlBQfS2JcjCaxSoWmbsCfsCt3NWLnW40A7gU2IvK6LxfzaiRWhNOx3QcMOsErIPKLK4kR6BkO/WHd3O2SNIkNmHvuuQeAqVOnBu+mxcU20QpkTSzYn5uctlazjkDV1XDmmfD663DKKTBmDOy9N4wcCb17B2YiXRTKTImnTGtilQo9xpQBzwLPItINmAScge1UEA/8zpfLeVtOcJlrW4VducsAxUAOdjWvUl9uqvaUX1xJz8xk4mPD7O1VnWkeUIsXLw7+TQO10IGbZxI7LAyWV/aXnTvhpJPgk0/gwQch2L1/o0hGcjylVbVOh6GUals5NpcsARqwvWN94m0Suw82cR0PzAcwhjwR/g5cDpzg643V7vKLK8OrHtYtN1eT2AB6+eWXg3/TQC056xaNtdTbt8Pxx9t2WS+8AGed5XREES0zOUHLCZQKRSLxwETgD8BvAPfbz+63oT7z5XLeJrGpru13uBY9ECEWuB+4FXgYOMaXG6vdFRRXhld7Lbe8PFi/3ukolD8FarUuN38ksQ0NsGiRHdVs6Tq//jVMmNDx63dURQWUle2+r7AQTjsNNm6EN9+EiRODH1eU6ZIcz8aSTnXuUUoFxjYgw/XcnbhuAp4DnsWYNb5czNskdie2Ia0Au7ArKRyPXYoWbC2D6qCKmnoKy2vpG06Tutzy8mDhQqejiFi33347AH//+9+Dd9NAlxN0dELgunXwzjt20tO8eU3JYmrq7udVV9vkNthJbGmpXTCgsHDPY5mZ8OGHcPjhwY0pSmWmxLNss47EKhWCMl3bWuzys88AH2A6tqKLt0nsZmwS2xVYDhzsurlbcUdurqyCkjBsr+XmbpdkjE5KCYCVK1cG/6Zbt0JCgk28AiExETIy7Fvs3qqvh9GjbV3pwIFw+ulwzDFw9NG224GnKVPAieV6H3vM/lv45z/3/NqNH2/jVkFha2I1iVUqBC3GTux6EWM6nTt6m8R+D4zEjrg+z54jr7oYQidsCMcesW55eVBXZ0fFApX0RLEZM2YE/6aBXHLWzdcJgevX2wT2kUfgssvaPrdnT5sg19XZDhrBUFkJ//d/tu71L38Jzj1VqzKS46msbaCuoTH8JssqFcmM2d+fl/M2ib0U+AuwyxgqRcjAtkWoB/4H/NOfQUWbguIwT2LBJiSaxEaGQPaIdfM1iV2+3G739+L3X69e9p2BLVugb9+Oxeerp56yo7A33hic+6k2ZaY0LXiQm5bocDRKqd2IxGEndw2laWJXE2Nu8/ZS3i47WwFUeHx8D3CPtzdRbcsvriQ9Ke6XNb/DimcSO3iws7FEoH/84x8A3Hab1/+mO2/rVpsIBpKvEwJXrLBbb1pyuWPfvDk4SWxtLdx7L4wdC0ccEfj7qXa5f5eWVmoSq1RIEekKzMMmsK3pfBIrgk+//Y0h35fzVRN3ey0Jx5rSaGyXFEQFBQXBv+m2bXDAAYG9h68TApcvt6PDWVntn9uzp91u2tSx2Hw1Y4btPPDkk8G5n2pXhi49q1SouhVoazTCpwlebY3ErvfhYqada6k25BdXMqx7utNhdIwmsQH17LPPBveGjY22njTQ5QRdu9qfGW8nBK5YAcOHe3dtz5HYQGtogHvugf32s229VEhoSmJ1wQOlQsyx2JxxOnCe6/lVwBWu5z69y99exbv48FAd0NBo2FhcRZ9wrIeF1pNYY+x/8Cq8FBUFdslZt7w823GgtLT9c42xSay3q3vl5NgJXcEYiX39dVi1ytbChuM7KREqMyUB0JFYpUKQu1ataS11Yx4BTgGGAL19uVhbo6facSAItu2sprahMTwndQEkJ9s+nZ5J7Icfwp//bP9T/+kn52KLADfccAMAd999d3BuGOgesW6ef/y0VyKwfTuUlHg/EhsTY0sKAp3EGgN33QVDh8LJJwf2Xsonv4zEVmoSq1SIaQDigSKgDohDJA/Y4Do+BbjD24u1msQaw3mdCFJ5KT+cOxO4uWear1wJ111nG9LHxtoRvWC2OYpARUVFwb1hoFfrcvNMYocMaftcXyZ1ufXqFfhygtmzYfFiePZZ+/OuQkaXJPtfm/aKVap9IpII3Aecge0W8BFwqTFmYxuvuQy4COjv2rUUuMMY8247tyvCjsZmAFuxI68vAtWu415MfGiidawOi5gk9r334OWX7cjsv/5lG9pfdZUdRQv0TPcINm3atODecOtWuw1mEtsed3stX5LYnj0D/y7AXXfZ7gdnnhnY+yifxcXGkJ4Yp+UESnnnQeB32CS2CHgAeEdEDjDGtFYXuBH4K7AKW5r6R+AN12t+bONeK7FJ7CDgM+BM4BjXMQN850vgXiexIgzFZt0t9fUyxvwShPJBQXElsTFCz8w9W6WFjf797Tr2F14It91mJ+286VrQbcsWTWLDiRPlBO1ZscKWrPT2oVSqVy+YM6djsXlj1iz44gv497/1nYYQ1SU5XssJlGqHiGQAfwLOM8Z86Np3Nvbt/fFAi79IjTFvNtv1NxG5BBgDtJXEPgmsBpKwnQqOBVz/IbADuNqX+L1KYkU4ANvXq6XhQsHHlgihJDs7m3nz5jl2/64VVfxl3wa+/Pwzx2LorPg//IHYk0+mukcPWLYMli0jffNmDgB++uADisrLnQ4xbD3++OMAXHLJJUG538CFC+kdH89n338f0IlKMbW1HAms/fpr8tvpL7zP/PnE9+zJos+8/zfSp6qKQbt28fl779GQ4t93OeJLSznowgupGTKE74YPxzj4+0O17ryBlcTHVjNvXqnToSjlb3Ei8q3Hx9OMMR192+4AbI3qB+4dxpgCEVkOHEYrSawnEYkFfg+kAfPbPNmYV4BXPF48GDgau3jWlxhT6kvw3o7E3gik+nLhcFFcXMy4ceMcu/9Jj35JamIsF0061LEYAmLQILj0Ukbl5oKDX99w9+qrrwIE72f02Wehe3fGHX104O+Vns7AtDQGtve5bdsGY8f69jXYtAmmTWPswIG+lSF4Y9IkqKgg4bXXOGrkSP9eW/nNtCe/oqa+kddOPczpUJTyt3pjzIF+ulZ37GSrwmb7t7mOtUpERgELsKOq5cDJxhjf6riM2Qk0H9X1mreLSh+GHW11DwcZYB/gLeBnwK9r4UaTAtdCBxHHXVPprrFUHfLoo4/y6KOPBu+G27YFvpTAzZulZysqID/f90Q0UL1iZ82CV16Bm28GTWBDWmZKvNbEqqglIneIiGnnMa6Tt1kJjAYOAR4HnhORoP5i9DaJzXFtX3TvMIYl2FYIQ4A/+zmuqFBeU09RRW349ohtS0KC7de5ZYvTkShfbNsW+Eldbt4ksStX2q237bXcArFqV2EhXHqpXc3sr3/133VVQGQkaxKrotqDwPB2Ht9gOwTEArnNXt/NdaxVxphaY8xqY8wiY8wNwGKCnA96W05Qha11qHY9T3JN9HIXO54YgNgiXkEkdCZoS48emsR20tVXXw3Agw8+GJwbbt0a+CVn3fLy7HKtbelIey1oGon1ZxJ7+eV2cYaPPoI4bewS6twTu4wx4bmkt1KdYIwpZM8SgT2IyCJsv9YJwEuufb2xSW7b9a17igESfXxNp3g7Ervdtc3GLkcL8Am2FgKg0Y8xRQ13e61+2RFZbmyTWC0nCB+NjXZkNFjlBO6lZ9uyfLntwbrXXr5dOzUVMjL8V07w2mswcyb84x8wapR/rqkCKjM5gdqGRqrr9L8npVpjjCkDngb+JSLjRWQ/4AVsh4G57vNEZIWIXO7x8T0iMlZE+ovIKBG5GxiHxzv2weDtcMJPwEBsHew72Azd/Z6jwWNWm/JeflGEj8R27970drDqkKCNwELwlpx1y8uzfYSNab0TwooVMHCg7Tvsq/ZW7VqyxPY09mZ55DlzYL/9tIwgjLhX7SqtqiU5IYxbGCoVeFdjuwPMpGmxg3Oa9Ygdyu4lB92BGa5tGTbpPd4YE8DehnvyNom9FfvJrccuB7YvtreXwX6yVwUiuEiXX1xJl6Q4MlIitM+keyS2rSRFhY5gLXTglpdnV3TbudOOmrZkxYqOdxdob9Wup56Cl16CAQPav1b//jB9uvaEDSOZrt+rZVV19MjQJFap1hhjaoArXI/WzpFmH58b4LC84lUSaww/AD947DpOhEyg3hi0CWgH5RdX0jcnQkdhwSaxtbV23fvsbKejCUuXXXYZQHA6FARroQM3zwUPWkpi6+vh559h4sSOXb9nT/jkk9aPL11qR1cXLuzY9VVI+2UkVhc8UCpieVsT25IEoMZfgUSjiG2v5eZOhnRyV4clJyeTnBykUSQnRmKh9brY9evtH0GdGYndssXW+rZk6VLYe++OXVuFPHcSqx0KlIpcbY7EirA/MBnbyPYNY/hYhAuAu7GTvGpEeMwYrgt8qJGlodGwsaSKY/cO0qiXE3r0sNstWzRZ6KD77rsveDdzj8QGK4nt2tVuW0tily+3W1/ba7n16mVHc3fs2PNzKi7Wn8sIp0msUpGv1SRWhCOw9a7ucy4T4V7gL9haWMEmt38WYbUxPBHoYCPJtp3V1DY0RvZIrDuJ1Q4F4WHbNjuBqrX6VH9rbyS2o+213Dx7xTZPYpcutVtNYiOWe65BmZYTKBWx2ionuB67nq54PK53HROa+o8JcHagAoxUGyK9MwHsPhKrOmTKlClMmTIlODfbutUme8GahOdOYrdvb/n4ihW2JCUzs2PXb2vVLncSq6tuRaz0xDhiY0RHYpWKYG2VExxIU/usN4ETgONd+/5gDDNFOAPbE2xEoAONNBG/0AFAWhqkpGgS21HGkBMb2/GksrgYJk+GXbt23x8XBw8+uOeiBsFcrQsgOdn2c22rnKCjo7DQ9qpdS5dCejr06dPx66uQJiJ0SYqjtKrW6VCUUgHS1kisux/YJFepwB88jr3u2r7m2qb7O7BIl19cSWyM0CMzyelQAkdEFzzwljE2iZw7F+64A044Abp14+4nnuDuF1/sWNP+r76CDz+034cuXZoeixfDP/+55/nbtgWvM4Fba0vPGmNHYjtaDwv2c4mJaT2JHTFCW79FuMyUBMqq6p0OQykVIG2NxMYDxhh2AhhDmfv3vTHUuba1rn36P4GP8osr6ZmZRHxsZxpEhIFoW3p23Tr7KClpepSW2glGnoyxiwts2AD5+fZR49HsY9gw21pqn33gxhvhmmvg5Zd9i2X9ert97bWm0g6Aq6+Gxx6DwkLI9ehdvXUrHHSQb/forNaS2O3b7deuMyOxcXF2ZLm1coLf/rbj11ZhoUtyPKWVOhKrVKRqt0+sCP/wZp/yTX5xZeQuN+upe3f46SenowiOzZthyJA9E9bY2Jab5GdmQr9+tlfp735nnw8dahNJVx3oeeedByNG8OzMmXDeefDrX3sfz4YNkJCwZ4nAn/4EDz0EL74IV7nWKWloaHkWf6B17dryHznuSV2dGYmFllft2rHDJsk6qSviZSTHU6ZJrFIRy5vFDm72eG5a2Kc6IL+4kl/vHeSEwQk9eti3tKPB7Nk2gX3xRZsgZWXZR1pah9+27tOnj/0aVlbCpZfapVK97Ru7YQP07WvfUvc0ahQceCA8/TRceaWNrajI9lN1opzghx/23O9ur9WZkViwk7vcI9Ju2pkgamQmx5NfVOF0GEqpAGkvidUygQAoKq+huKKWQXlpTocSeD16QFkZVFV5n3yFqzlz7MjfGWf4rdbytttus08mTIBf/Qruugtuv927F69fb5dLbcmf/gSXXALffmtHfoO90IGbu5yg+dLEK1bYSV+9e3fu+j17wpdf7r5Pk9iokZEcT6l2J1AqYrWVxN4atCiizOrtdqXevbpGSRIL9i3jgQOdjSWQGhrsiPNJJwVmstDRR8PZZ9sJWWee6d0I5YYN8JvftHzsjDPgz3+GZ56xSWywl5x1y8uztcDl5bZbgNuKFfZz7OzXslcvO8pcXQ1JrkmUS5faXrjuFlwqYmWmxLOzqo7GRkNMjI7JKBVpWk1ijdEkNlBWuZLYwd2ioKmD59KzkZzELlxoJyL5UrPqhbPOOguAGTNmwH33wdtv2xHUjz9uO8Grrrajq/36tXw8IwNOOw1eegnuvz/4q3W5eS544JnELl8OY8d2/vruRHXLFhgwwD53LzernQkiXkZyPI0GdtXU/7KCl1IqckT41PjQtHp7OakJsfTMiOD2Wm7RsmrXnDk2KRo/3q+XHTp0KEOHDrUfdO1qR2LnzYMZM9p+YX6+3bZWTgC2pGDnTtu9wKlyAvfSs1Onwrnnwu9/D8cfb+Pv7KQu2LNXrDFNSayKeO7EdaeWFCgVkbyZ2KX8bNX2XezVNQ2JhpGgaFm1a/ZsOPhgyMnx62X//ve/777jggtg+nS49lq7kEFLXQ/AlhJA6yOxAEcdBYMG2QleBx1k327v0sUvcXtt5EibyH70kZ0Al5pqtxMm2NKMzmq+atf27ba8QJPYqOBOYksr6+iT7XAwSim/0yTWAau2lTN2cJ7TYQRHbq5tMRXJSWxJCXzzDdx0U+DvFRMDF18Mf/yj7Uc7ZEjL53mTxIrA+efD3/4GtbXBXXLWrW/fplKGQGg+Ertkid1qEhsV3EmsLj2rVGTScoIgK6usY/uuGgZ3i4JJXWCTrm7dIrucYO5c257Kz/WwAJMnT2by5Mm773RP6nL3Um3J+vX2j4f2Ji/98Y/2e7RgQfBLCYIhK8uOMLuTWO1MEFUyUxIATWKVilSaxAbZ6h12Hfsh0ZLEQuSv2jV7tl2c4OCD/X7p0aNHM3r06N13umtk20piN2yw7ani2nmzpVcvOO44+zzYnQmCQcR+ju5ygqVLITs7Mj9XtYdfygmqdMEDpSKRlhME2aptrs4EXaOgM4Fb9+4tr18fCYyxk7rGj28/YeyAqVOn7rkzI8N+TdtLYtsqJfD0pz/Be+9F5kgs7L5ql3YmiCqZKVpOoFQk05HYIFu1vZyk+Bh6ZUZ4439PPXpEbjnB0qU2QQpAKUGbhg2DlStbP97WQgfN/fa3sO++cOih/ogs9PTqZb9H2pkg6iTFx5IQF0NZpSaxSkUiTWKDbNX2cvbqmhZdjbd79LCzwhsanI7E/+bMsdsAJbGnnnoqp5566p4Hhg2zvVSN2fNYXZ1N2rwdiU1IgMWL7SSvSOQuJ9iyBUpLNYmNMpnJ8ToSq1SE0iQ2yFZv2xVdpQRgk9jGRpvIhrqyMrugQGOjd+fPmQMjRkCfPgEJZ8yYMYwZM2bPA0OH2q4IhYV7Htu40cbvbRIb6Xr2tMsef/GF/ViT2KiSkRxPqY7EKhWRNIkNol3VdWwuq46O5WY9uSfRhENJwRNPwIknwimn2IS2LZWV8NlnAS0luO6667juuuv2PODuUNBSSYG7vZa35QSRzt2h4YMP7FaT2KiSoSOxSkUsTWKDaM2OCgAGR1sSG04LHuTn2wUE3nnHdhtYvrz1cz/9FGpqmmb3B1Nbbba86REbTdy9Yj/80C5z614lTEWFzBRNYpWKVJrEBtGqbba91uBuUVhOAOGRxG7ebN+q//hjWz958MHw+ustnztnju1BOnZswMI58cQTOfHEE/c80LevvXdLSez69XYboBKHsOMeic3P11HYKNRFR2KViljaYiuIVm8vJyEuhj5ZUdSZAJpaN4VDOcHmzTbpPvJIWLQITjsNTj0VrrkGjj/e9l7t3dsujTp7tl26NTlw389jjjmm5QMxMXa1rtZGYnv2hMTEgMUVVtwjsaBJbBTKTE7QJFapCOXISKyIXCoi60SkWkQWiUirQ1kiMk5ETAuPYR7nnNvKOUnB+Yy8s2p7OQNzU4mLjbIB8KQku3JSOIzEbtnSlPT07m1LBi68EB54ACZMgOHDIT3dLm6wcmXASwmuuuoqrrrqqpYPttZmy5cesdEgKckucACaxEahjOR4ymvqqWvwcrKmUipsBH0kVkQmAQ8BlwJfuLbvi8gIY0x+Gy/dGyj2+HhHs+OVwCDPHcaY6s5H7D+rtu9idJ8sp8NwRjis2tXYuHsSC3Y0c9o0uOkm+zb9xo32UVBgJ3794Q+OhcuwYTBrlq3L9Rx1Xb8+IKuHhbVevaC4WJPYKORe8GBnVR05afruhFKRxIlygmuA6caYJ10fXyEixwGXADe08brtxpgW+gn9whhjQvb96sraejaWVPH7A6K0TrF799AvJygshPr6phpeT3372keQHX/88QC8//77ex4cOtQm3qtXNyVnjY02wT799CBGGQZ69oSfftIkNgq5l54t0yRWqYgT1Pe1RSQBOAD4oNmhD4DD2nn5tyKyRUQ+EpGjWzieLCIbRGSjiLwjIvu1EccUEflWRL6tr6/37ZPooLU7KjAmCjsTuIXDSKw7Ps+RWIedcMIJnHDCCS0fbKnN1pYtdrEDLSfY3dChMGAA5OQ4HYkKMncSW6p1sUpFnGCPxOYCscC2Zvu3AeNbec0W7CjtQiABOBv4SESOMsZ87jpnJXA+8AOQDlwFfCki+xpjVjW/oDFmGjANIDU1tYUlj/xv1XZ3Z4IoT2KNCd116zdvttsQSmIvvfTS1g8OGWK3npO73J0JNInd3Z13wg1tvdGjIlVGStNIrFIqsoR8dwJjzEpskuq2QET6A9cDn7vOWQAscJ8gIvOBxcAVwJXBirUtq7aVExcj9MtJdToUZ/ToAdXVsHMnZGQ4HU3L3ElsS+UEoSgtzU4+80xidaGDlqWl2YeKOr+UE+iqXUpFnGBPky8EGoBuzfZ3A3wpmPwaGNzaQWNMA/BtW+cE26rt5QzITSU+2joTuLlX7QrlkgJ3bCGUxI4fP57x41t7kwJbUtBSEutA/a5SoSgzWUdilYpUQc2ojDG1wCJgQrNDE4D5PlxqNLbMoEUiIsA+bZ0TbKu3l0dvKQGEx4IHmzfbmskQ6q86adIkJk2a1PoJ7jZbxlUVs3495OZCapSO+CvVTBd3TayOxCoVcZwoJ3gAeEFEvgG+BC4GegJPAIjI8wDGmHNcH18NrAeWYmtizwJOAk51X1BEbga+AlYBXbAlBPtga2kdV13XwIaiCk7YN3RqLYMuXJLYEBqFBbjwwgvbPmHYMFuisXWrjX3DBi0lUMpDfGwMaYlxOhKrVAQKehJrjJkpIjnATUAPYAkw0Rjjeh+U5u+DJgD3Ar2BKmwy+xtjzHse52RiJ2p1B8qA74EjjTHfBOrz8MXaHRU0RnNnAmgqJ2jeZquqyi4ocOyxdhUqJzXvERsOhg612xUrmpJYbSOl1G4ydOlZpSKSI1mDMeYxY0x/Y0yiMeYAY8xnHsfGGWPGeXz8L2PMYGNMsjEm2xgztlkCizHmz8aYfq7rdTXG/No12SskRH1nArCTuZKSdh+JramBk06yy7nefHPbr9+wAY44Aj77rO3zOmPz5pBLYseNG8e4ceNaP8GzzZYxulqXUi3okhxPWVWt02Eopfws5LsTRILV28uJERiQG8V1iiK794qtr4czzoAPPoAxY+COO2DECLuvuaIi+PWvbaI2ezYceaT/42tstKPEIZbEnnvuuW2f0KuXrX9dsQJ27LAj21pOoNRuMnUkVqmIpElsEKzaVk7/nFQS42KdDsVZ7lW7Ghvh/PPhf/+DBx+ESy6B8ePhvPNg0KDdl0ytqIDf/tZOWMrO3r2xvz+1tVqXg9pNYkVsScGKFU2dCXQkVqndZCTHs2ZHudNhKKX8LEr7PQXXqu272Cua62Hd3COxV1wBL7wAt98OV10FCQnw2mv2+EknwaZN9vy6Opg0Cb75Bv77X1tO8PPPgYktBBc6AKirq6Ourp0RJHebLV3oQKkWZaboSKxSkUhHYgOstr6R9UWVHDeyu9OhOK9HD3j9dVi2DK6/Hv72t6ZjeXnw1ltw2GE2kf30U7j8cnj3XXj8cTj5ZFiwAObMgYYGiPXzqHaIJrETJthudPPmzWv9pGHDbJK/fLn9WJNYpXaTkRLP9l01HHrXRx16fUJcDI/+YX9G9Q7RhVqUilKaxAZYRU09E0f14MD+2U6H4jx3gnjxxfDPf+65/OyoUfDiizaJHTUK1q6Ff/zDng92mdWaGigo8H/dZwgudABwwQUXtH/SsGF2UteHH9oJdJmZAY9LqXBy6v692VlVT2Oj76uM1zU08vr3m/guv0STWKVCjCaxAZaVmsC/z9jP6TBCwx//aBvxX3DBngms24knwt13w9SpMGUK3HJL07EhQ+x25Ur/J7HukdjuoTViftZZZ7V/krvN1oIF2l5LqRYM6ZbO3aeM6tBrGxoNbyzeRGF5jZ+jUkp1ltbEquDp1csmpu31g/3LX2DxYnjssd2TXXeyFoi62M2bbYIdQqt1AVRWVlJZWdn2SYMH269TQ4N2JlDKz2JjhOzUBArLtUWXUqFGk1gVekRg3333rHvt2hW6dAlMh4ItW0KulABg4sSJTJw4se2TkpObkleth1XK73JSEynSkVilQo6WE6jw4W4nFaiR2BCb1AVwySVerpw8dCisW6dJrFIBkJOWoOUESoUgHYlV4WXIkMCMxIZoEjtp0iQmTZrU/onulbu0nEApv8tNS6SoQssJlAo1msSq8DJ0KOTn25Wp/MW9WlcIlhOUlZVRVlbW/onuJFZHYpXyu5y0BAp36UisUqFGywlUeHF3KFi1CvbZxz/X3LHDTooKwZHY3/3ud0A7fWIBJk+G8nLYf//AB6VUlMlNS6SitoGq2gaSE6J85UWlQogmsSq8eHYo8FcSG6ILHQBceeWV3p2YkQHXXhvYYJSKUrlpCQAUVdTQOyHF4WiUUm6axKrwMniw3fqzLtadxIZgOcEpp5zidAhKRb2cVNt6r7C8lt5ZmsQqFSq0JlaFl9RU6N3bvx0K3Kt1heBIbGFhIYWFhU6HoVRUy023Say22VIqtOhIrAo//u5QEKKrdQGcdtppgBc1sUqpgMlJdZUT6IIHSoUUTWJV+Bk6FP77XzCm9eVrfeFerSshofPX8rNrtc5VKcflptmR2B06EqtUSNEkVoWfIUOgtBQKCyEvr/PX27IlJEsJAE444QSnQ1Aq6iUnxJKaEKsjsUqFGK2JVeHHs0OBP4ToQgcAW7duZevWrU6HoVTUy0lLpKhCR2KVCiWaxKrw4+4V66+62M2bQ7IzAcDkyZOZPHmy02EoFfVydelZpUKOlhOo8NOvH8TH+2cktqEBtm0L2ZHYqVOnOh2CUgo7EltQXOl0GEopD5rEqvATFwd77eWfkdgQXq0L4LjjjnM6BKUUdiT2+/wSp8NQSnnQcgIVnoYM8c9IbAgvdABQUFBAQUGB02EoFfVy0xIprqilodE4HYpSykWTWBWehg6F1avtKGpnhPBCBwBnn302Z599ttNhKBX1clITaDRQWqkdCpQKFVpOoMLTkCFQWwsbNsDAgR2/jnskNkST2JtuusnpEJRS2JpYsEvPup8rpZylSawKT+42WytX+ieJ7dat8zEFwPjx450OQSlF04IHdunZdGeDUUoBWk6gwpW7zVZn62K3bLELJoTgal0Aa9euZe3atU6HoVTUy02zvyMKK7ScQKlQoSOxKjzl5UFmZuc7FITwQgcA559/PgDz5s1zNhClopx7JLZwl/aKVSpUaBKrwpOIfzoUhHgSe+uttzodglIKyEiOJzZGdNUupUKIJrEqfA0dCp980rlrbNkC++7rn3gC4KijjnI6BKUUEBMjZKcmUFSu5QRKhQqtiVXha8gQ2LgRKiqa9m3cCJdcAi+/3P7rGxpg69aQHolduXIlK/21vK5SqlNy0xJ16VmlQogmsSp8uTsUrF4NVVVwxx123xNPwKOPtv/67duhsTGkk9iLLrqIiy66yOkwlFLYyV2FOhKrIoyIJIrIv0WkUEQqROQtEentw+tvEBEjIo8EMs6WaDmBCl/uDgWPPgoffgjr18Opp0JdHXzxBRhja2db417oIERX6wK46667nA5BKeWSk5rAusKK9k9UKrw8CPwOOAMoAh4A3hGRA4wxba4oJCKHAlOAHwMdZEt0JFaFr732stsnn4S0NPjoI5g1C371Kyguhh072n59iC90AHDYYYdx2GGHOR2GUgpbTqA1sSqSiEgG8CfgemPMh8aY74CzgX2ANhuVu177InA+UBLoWFuiSawKX6mpcOed8Pjj8P33NnkFGDHCbpcta/v1YZDELlmyhCVLljgdhlIKu2pXVV0DlbX1ToeilL8cAMQDH7h3GGMKgOVAeyMo04BZxphOzrDuuKgvJ8jOztYenOHMPUr5xRe/7EosLWUM8PMbb7C5jZf2W7CA/iJ8tmIFZvXqgIbZUVdffTUADz74oKNxKKWge2Ud146q54vPPyMhVseAlGPiRORbj4+nGWOmdfBa3YEGoLDZ/m2uYy0SkQuBvYCzOnhfv4j6JLa4uJhx48Y5HYbyJ2MgPZ0hDQ0Maet7+9xz0K0bR4Xw0q5PPvkkAAcddJDDkSilPlmxnT9PX8jrY/dj/75ZToejole9MebAtk4QkTuAv7VznaM7cnMRGQrcBRxhjKnryDX8JeqTWBWBRGxJQXvlBN9/D6NHByWkjtLkVanQkeNaelbrYlUYeBCY0c45+cChQCyQC3hOJOkGfN7K68a4zl8qTZOnY4EjReRiINUYE5RedJrEqsg0fDjMmdP68epqWLoUfvOb4MXUAYsXLwZgdIgn20pFg1+WntVesSrEGWMK2bNEYA8isgioAyYAL7n29QaGA/NbedkbwLfN9j0LrMKO0AbtrzxNYlVkGjECpk+HkhLIauFtv59+gvp6OOCAoIfmC3dNrNZtK+W87FT3SKwmsSoyGGPKRORp4F8isp2mFls/AnPd54nICuARY8wjxphSoNTzOiJSARQbY4I6E1mTWBWZ3B0Kli9vmvzl6bvv7Hb//YMXUwfohC6lQkdSfCzpiXG64IGKNFcD9cBMIBn4CDinWY/YodgSgpCiSayKTMOH221bSWxWFvTrF9y4fKRlBEqFltx0XXpWRRZX/eoVrkdr57SxchAYY8b5OSyvaI8QFZn69YPk5NYndy1aZEdh21rRKwQsXLiQhQsXOh2GUsolJzVBJ3YpFSJ0JFZFpthYGDq05SS2ttbWxLrqTUPZ9ddfD2hNrFKhIjctkbWF5U6HoZRCk1gVyUaMgC+/3HP/smU2kQ3xeliARx55xOkQlFIectIS+Ga9jsQqFQq0nEBFrhEjYMMGKG82arJokd2GQRI7cuRIRo4c6XQYSimXnLRESiprqW9odDoUpaKeJrEqcrknd61cufv+776D9HQYNCj4Mflo/vz5zJ/fWqs+pVSw5aUlYAwUV+porFJO03ICFbncbbaWLdu9H+x339lR2JjQ/xvuxhtvBLQmVqlQkeNa8KCovJau6UkOR6NUdNMkVkWuQYMgLm73yV319fDDD3Dxxc7F5YP//Oc/ToeglPKQk6pLzyoVKjSJVZErPh6GDLG9Yt1WrICqqrCohwUYOnSo0yEopTzkpuvSs0qFitB/P1WpzhgxYveR2DBZqcvt008/5dNPP3U6DKWUS26qJrFKhQodiVWRbfhweP11qKmBxESbxKak2B6yYeDmm28GtCZWqVDRJTmO+FihqELLCZRymiaxKrKNGAGNjfDzzzBqlE1iR4+2iyGEgWeeecbpEJRSHkSEnNRECnfpSKxSTtNyAhXZPDsUNDbC99+HTSkBwMCBAxk4cKDTYSilPOSkJehIrFIhQJNYFdmGDLGttJYvh1Wr7MIHYZTEzp07l7lz5zodhlLKQ25aIkVaE6uU47ScQEW2pCQYONCOxIbZpC6AO+64A4Dx48c7HIlSyi0nLYHV28vbP1EpFVCaxKrIN3y4TWL79bOTu9wlBmHghRdecDoEpVQzuWmJFJbXYIxBRJwOR6mopUmsinwjRsDs2ZCdDfvsY/vHhok+ffo4HYJSqpnctARq6hspr6knPSl8fp8oFWm0JlZFvhEjoK4OvvgirEoJAGbPns3s2bOdDkMp5SEntWnpWaWUcxxJYkXkUhFZJyLVIrJIRMa2ce44ETEtPIY1O+9UEVkmIjWu7cmB/0xUWBg+3G6NCbsk9p577uGee+5xOgyllAf3ql1FFTq5SyknBb2cQEQmAQ8BlwJfuLbvi8gIY0x+Gy/dGyj2+HiHxzXHADOBm4HXgVOAV0XkcGPM137+FFS4Gebx906YJbEvv/yy0yEopZrJSU0AYMcuHYlVyklijAnuDUW+Bn40xlzosW8VMMsYc0ML548DPgHyjDGFrVxzJpBtjJngsW8usMMYc0Zb8aSmppqKioqOfCoqnPTtC1u22BZbiYlOR6OUCmNby6o59O6PmHLkQH41rKvT4agQkpkSz7DuXQJ6DxGpNMakBvQmYSKoI7EikgAcANzX7NAHwGHtvPxbEUkElgF3GGM+8Tg2Bvh3s/PnAJd3IlwVSQ48ELZuDbsE9u233wbghBNOcDgSpZRbTloCSfExTPtsLdM+W+t0OCqEjBuax/TzDnY6jKgR7HKCXCAW2NZs/zagtUaYW4BLgIVAAnA28JGIHGWM+dx1TvdWrtm9pQuKyBRgCkBCQoKPn4IKS888A/X1Tkfhs/vvvx/QJFapUBIfG8N7V45l685qp0NRISYzWXOKYAr5FlvGmJXASo9dC0SkP3A98HmLL2r/mtOAaWDLCTobowoDmZlOR9Ahs2bNcjoEpVQLBualMTAvzekwlIpqwU5iC4EGoFuz/d2ArT5c52tgssfHW/1wTaVCTm5urtMhKKWUUiEpqC22jDG1wCJgQrNDE4D5PlxqNLbMwG2BH66pVMh5/fXXef31150OQymllAo5TpQTPAC8ICLfAF8CFwM9gScAROR5AGPMOa6PrwbWA0uxNbFnAScBp3pc8yHgMxGZCrwBnAwcDRwR4M9FqYB6+OGHATjllFMcjkQppZQKLUFPYo0xM0UkB7gJ6AEsASYaYza4Tunb7CUJwL1Ab6AKm8z+xhjznsc154vIZOAO4DZgDTBJe8SqcPfmm286HYJSSikVkoLeJzbUaJ9YpZRSSoUL7RPbxJFlZ5VS3pk5cyYzZ850OgyllFIq5OhIrI7EqhA2btw4AObNm+doHEoppUKDjsQ20SRWk1gVwiorKwFISUlxOBKllFKhQJPYJiG/2IFS0UyTV6WUUqplWhOrVAibMWMGM2bMcDoMpZRSKuRoOYGWE6gQpjWxSimlPGk5QRNNYjWJVSGsrq4OgPj4eIcjUUopFQo0iW2iNbFKhTBNXpVSSqmWaU2sUiFs+vTpTJ8+3ekwlFJKqZCj5QRaTqBCmNbEKqWU8qTlBE2iPokVkUagys+XjQPq/XxN5R/6vQlt+v0JXfq9CW36/Qld/v7eJBtj9J10NIkNCBH51hhzoNNxqD3p9ya06fcndOn3JrTp9yd06fcmcDSTV0oppZRSYUeTWKWUUkopFXY0iQ2MaU4HoFql35vQpt+f0KXfm9Cm35/Qpd+bANGaWKWUUkopFXZ0JFYppZRSSoUdTWKVUkoppVTY0SRWKaWUUkqFHU1ifSQil4rIOhGpFpFFIjK2nfOPcp1XLSJrReTiYMUajXz5/ojIKSLygYjsEJFdIvK1iJwYzHijia//djxed4SI1IvIkkDHGM068LstQURuc72mRkTyReTKYMUbTTrwvfmDiCwWkUoR2SoiM0Ske7DijSYicqSIvCUim0TEiMi5XrxmlIh8KiJVrtf9Q0QkCOFGHE1ifSAik4CHgLuA/YD5wPsi0reV8wcA77nO2w+4G/i3iJwanIiji6/fH+Ao4GPgN67z3wP+521ypbzXge+N+3VZwPPARwEPMop18PvzMnAcMAUYCvwe+DHAoUadDvy/czjwAvAcsDdwEjACeDEY8UahNGAJcBVerP4pIl2AD4FtwEGu110PXBPAGCOWdifwgYh8DfxojLnQY98qYJYx5oYWzv8ncIoxZrDHvqeAvY0xY4IRczTx9fvTyjW+AT43xlwboDCjUke/NyLyOvADIMBpxpiRAQ82CnXgd9uxwKvAIGNMYfAijT4d+N5cB1xhjOnnse884N/GmLRgxBytRKQcuNwYM72Ncy4B/gl0M8ZUufbdBFwC9DaalPlER2K9JCIJwAHAB80OfQAc1srLxrRw/hzgQBGJ92+E0a2D35+WpAMl/opLdfx7IyKXAt2AOwIXnerg9+ckYCFwjYhsFJFVIvKwiGiS5Ecd/N58CfQQkRPEygUmY99pUs4bgx0o8Ry1nQP0BPo7ElEY0yTWe7lALPYtAE/bgNZqjbq3cn6c63rKfzry/dmNiFwG9Ma+Faf8x+fvjYiMAm4GzjLGNAQ2vKjXkX87A4EjgH2BU4HLsaUF0wMTYtTy+XtjjFmATVpfBGqBHdh3Mv4YuDCVD1rLC9zHlA80iVUKcNUp3wv8wRizwel4opmIJAIzgeuMMeucjke1KAYw2H8vXxtj5mAT2VNFpJuzoUU3ERkB/Bu4HTuKexw2OfqPk3EpFQhxTgcQRgqBBuzbm566AVtbec3WVs6vd11P+U9Hvj8AiMhp2MlD5xhj3g5MeFHN1+9ND2A48KyIPOvaFwOIiNQDE40xzd9eVR3XkX87W4BNxpgyj33LXdu+7DnSpDqmI9+bG4BvjDH3uj7+UUQqgM9F5EZjzMbAhKq81Fpe4D6mfKAjsV4yxtQCi4AJzQ5NwM4WbcmCVs7/1hhT598Io1sHvz+IyOnY8oFzjTGzAhdh9OrA92YTMAoY7fF4Aljtet7q91P5roP/dr4EejargR3i2uo7GX7Swe9NCjbx9eT+WP/Pd94CYKyIJHnsmwBsBtY7ElE4M8bow8sHMAlbY3QBdqToIaAc6Oc6/jzwvMf5A4AK4EHX+Re4Xn+q059LJD468P2ZDNRhW5x093hkO/25RNrD1+9NC6+/BVji9OcRqY8O/NtJAwqwHQr2Bg7Hthl61enPJdIeHfjenOv6vXYJtnb5cOwkvEVOfy6R+HD9WxjtelQC/3A97+s6fjfwkcf5GdgR15eBkcApwE7gWqc/l3B8aDmBD4wxM0UkB7gJ+5bnEuxbm+6Rh77Nzl8nIhOB/8P+QtkMXGmMeS2IYUcNX78/wMXYkpoHXQ+3T4FxgYw12nTge6OCqAO/28pFZDy29nIhtqPHG8DUoAUdJTrwvZkuIunYGuX7gTJsP+y/Bi/qqHIg8InHx7e6Hs9h/6DoAQxyHzTGlInIBOBR4Fvsv537gQeCFG9E0T6xSimllFIq7Gh9jFJKKaWUCjuaxCqllFJKqbCjSaxSSimllAo7msQqpZRSSqmwo0msUkoppZQKO5rEKqWUUkqpsKNJrFIhTkQGi8gjIrJcRMpFZJeIrBCRJ0XkUI/z1ouIEZH1DobrjmW6KxYjIv099ncTkRdFZIuINLiOPygi/T3Onx7AuDJF5BbX4yRv4w4WERnncf/2Hre4XuP+eF6w421PIL+vvnyvmn1d/RqHUso5utiBUiFMRM4DHgcSmx0a6nrkAScFOazOeAi7ApFTMoGbXc+fwzboV0opFYY0iVUqRInIr4CnsO+YGOBO4D/AdqAfcBpN69WHFGPMudjVapo7wLUtBQYYY0o9jklAg/JCG3EH6/7z8Pg6iMi5wLOuD59zxed3IpJkjKkOxLWVUipQtJxAqdB1N03/Rh82xvzdGLPRGFNrjFlljLkbuLCtC4jIaBF5XURWi8hOEakTka2ufQc2O3eAiDwvIvkiUi0ipSKyxPW2bVeP8y4UkW9FpFhEakRkk4h8KCJ/9Dhnt7d63W/nAnu5TskESlzHz23rbWcR2V9E/uu6T62IFIrIJyJysOt4mog8JyI/iUiR63MsFZHPRGSSx3VuAdZ5XPqPze/ZRhlEqojcKiJLRaRKRCpF5HsRuUZE4jzO2+3zEJFzXF/DKrHlIH8kgETkVyLylet+a0TkLyLimRTf4hHfySLytIgUAlUe5wwXkRc8vt7bRWSWiOzT7F5e/bw0e83pIvJjW18PERkrIm+JyA6Pn9eXm9+/ja9BT1e85a6fh8eB9FbO9flzUEqFEGOMPvShjxB7AF2xo6/uRy8vXrPede56j32Tm13H81EBDPc4d2kb5450nfP7Ns6Z5XGt6R77+wPj2njdua5z3B9P97jOyUBda69zndO9jWsb4BzXebe0cc70luJ27UsFFrXx2veAGNe5np9HSSvnH+HDz8G5LX1dmp3jPl7YytfqLI9zb2l2/i/nuY4fAVS2EncVMNbHnxfPr8fW9r4ewFlAQyvnVQPjWvsZc+1LBpa38NrNLX0dvfkc9KEPfYTuQ0dilQpN/T2e7zTGbOrgdb4Dfg30wNbVdgEucR1LAS4CEJEcYIRr/8PYxC0bOAj4O1DmOnaka1uOrclNxJY2nA7Mbi0IY8w8Y4wAG1y7NhhjxPWY3tJrRCQZeJKmsqd/AN2AXGwyvda1fxe2zra/63NKAg7DJmMA17hiuAUY4HGL5zxiOLe12IGrgf1dz+dgv5YDsV9bgOOxfyw0lwlcCmQA//TYf3Yb9+qMHOBfQBZwuRf3E+A47NdslGvfk9hEcAO29CMR2A/Ygf26Pgo+/bx46kYbXw8RSQX+jX33oR77B0wX4GLXeYnYcpq2nAMMcz3/CuiNHf0v3eOT79jnoJQKIVoTq1Rk2wr8CXgQm+QlNzs+1LUtwf5Hn4lNynZhR7R+MMbc4XH+Otc2FbgJO0K5HPjAGOPv//QPxyZmAPOMMbd7HJvl8bwSm9jOBIZj3zr2rK8dSuf8xuP5DcaYrQAichtNE8MmAi81e90iY8zjrnNnAH917e/XyXhasw34hzGmQUSeAx5p5373G2PmuJ4vEZHBNCWA/bDf2+ZGiUh3bF12Ke3/vHhq7+txuOt6AO8ZY95wPf+PiFwMjAaGiMhexpjVrdzjVx7P73b/8Sci92Pryz15+zOvlApROhKrVGha7/G8i4j07OB1XgH+gk3umiewuPcZYxqxI2IbgcHA34AZwE+uWtM+rvMfA14F3Oc/iB2d3CYiUzsYY2u6eTxf1sZ5f8WOEB6CHblrPkEsqZNx5Hk8z/d4vsHjeUv1kys9nlf4MZ7WrDHGNPhwv++bfextDWiODz8vntr7erT2dYb2v9a/xObxfGMrzwGffuaVUiFKk1ilQpAxZjvwjceu61s6z3NSUQvHsrClBGBH6fYGYoEWJ8gYY94B+mJHLk8EbsPWJ47EjrpijKk2xpyOfdv1COB84GvsW713iUgv7z5Dr2zzeD68jfM838o/CUh0lS4UtXCu6UAc2z2e923luec5bnWdvK+vfrmfMcab+1U1+9jzc5jrUWrxywNb+7vUdY92f15ai4+Wvx6tfZ2bf9zS19qt0ON571aeNwXh++eglAohmsQqFbr+hh3xBLjSNbO8p4jEi10A4UZsDWNr6mlKFuqBndi33W9v6WQR+TdwDLbedTbwGlDjOtzXdc6pInI50Av4ATsq+4P7ErSSLHTQlzQlokeLyI0ikiciWSJykoi463PrPV5TCsSLyN/ZfVTOzTOxHeyqw2zPux7P7xS7YEN/bI1uS+eEJWPMKuBn14fHiMjVYheHSBKRfUXkH8DL7vO9+Xnx0XzsW/wAx4vIiWI7T1yIrcsFWNlGKQHAJx7Pp4pILxEZBFzb0skB+ByUUkGkSaxSIcoYMxc78aoO+2/1ZmATUItNNu7ETuJp7fW7gI9cH/YCCrCjmyNaecklwIce9/gBO+kHbMkA2BHRf2Pf3t/lekxxHdsC/OjDp9gmY0wVtoWYO0m9EzsKVwz8Dzu5Ctdzt3nYhORKWpjMY4wpx85IBzv5q9zVburcNkJ5iN0ncW3F1ga7e96+j63HjQRTsF0ABPg/bFJZBSwGbmX3Eg9vfl68ZoypAK7A/uEWD7yJ/fma5jqlhqZJXq15Hljhej4GWyqwmt1LFTz59XNQSgWXJrFKhTBjzFPYt/8fwyauVdh6wpXA08D/t3f/KA0EYRjGHy9lK4i1JzAXsbSQWHoFUQgWprG1iFiLWHuH2EgQN4zFu2Ca1RWC+MHza/f/sgvvDPPNTH84xREJWEtSbX3J8IpZU+CBBMWOFEw9kkB43u9zRwqYXkhYXJPwOgP2+uC5Na21ORnrOiPTJHUkxC74Gid7BpySILLqt+0zXF0+Ae5Jz/SYe3gjszKckMKfdxL0nkgP32E/vrK81tqChPMLEgA/yPt+Jo2X443dx3wvv73+FZmO7Zb0mnek4XUN7LYsBvHd8SvgALgh/8krWSxiaD7lrT+DpL+zM27olCRJkvR/2BMrSZKkcgyxkiRJKscQK0mSpHIMsZIkSSrHECtJkqRyDLGSJEkqxxArSZKkcgyxkiRJKucTF35RWLnZcdsAAAAASUVORK5CYII=\n" + }, + "metadata": { + "needs_background": "light" + } + } ], - "text/plain": [ - "" + "source": [ + "fig, ax1 = plt.subplots(figsize=(10,7))\n", + "ax1.plot(class_thresh_arr, bal_acc_arr_orig)\n", + "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", + "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", + "ax1.xaxis.set_tick_params(labelsize=14)\n", + "ax1.yaxis.set_tick_params(labelsize=14)\n", + "\n", + "\n", + "ax2 = ax1.twinx()\n", + "ax2.plot(class_thresh_arr, avg_odds_diff_arr_orig, color='r')\n", + "ax2.set_ylabel('avg. odds diff.', color='r', fontsize=16, fontweight='bold')\n", + "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", + "ax2.yaxis.set_tick_params(labelsize=14)\n", + "ax2.grid(True)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "p_5-jU7S6QV8" + }, + "source": [ + "```average odds difference = 0.5((FPR_unpriv-FPR_priv)+(TPR_unpriv-TPR_priv))``` must be close to zero for the classifier to be fair.\n", + "\n", + "However, for a classifier trained with original training data, at the best classification rate, this is quite high. This implies unfairness." + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "wz3keki56QV8" + }, + "source": [ + "### Train classifier on transformed data" ] - }, - "metadata": {}, - "output_type": "display_data" }, { - "name": "stderr", - "output_type": "stream", - "text": [ - " 13%|█▎ | 13/100 [00:00<00:00, 127.35it/s]" - ] + "cell_type": "code", + "execution_count": 20, + "metadata": { + "id": "lCKBpvOR6QV8" + }, + "outputs": [], + "source": [ + "scale_transf = StandardScaler()\n", + "X_train = scale_transf.fit_transform(dataset_transf_train.features)\n", + "y_train = dataset_transf_train.labels.ravel()\n", + "\n", + "lmod = LogisticRegression()\n", + "lmod.fit(X_train, y_train,\n", + " sample_weight=dataset_transf_train.instance_weights)\n", + "y_train_pred = lmod.predict(X_train)" + ] }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "Classification threshold used = 0.2674\n" - ] + "cell_type": "markdown", + "metadata": { + "id": "zRVHSOfn6QV8" + }, + "source": [ + "#### Obtain scores for transformed test set" + ] }, { - "name": "stderr", - "output_type": "stream", - "text": [ - " 41%|████ | 41/100 [00:00<00:00, 98.54it/s] " - ] + "cell_type": "code", + "execution_count": 21, + "metadata": { + "id": "b6IW5-hf6QV8" + }, + "outputs": [], + "source": [ + "dataset_transf_test_pred = dataset_orig_test.copy(deepcopy=True)\n", + "X_test = scale_transf.fit_transform(dataset_transf_test_pred.features)\n", + "y_test = dataset_transf_test_pred.labels\n", + "dataset_transf_test_pred.scores = lmod.predict_proba(X_test)[:,pos_ind].reshape(-1,1)" + ] }, { - "name": "stdout", - "output_type": "stream", - "text": [ - "Balanced accuracy = 0.7128\n", - "Statistical parity difference = -0.0906\n", - "Disparate impact = 0.7625\n", - "Average odds difference = -0.0266\n", - "Equal opportunity difference = -0.0518\n", - "Theil index = 0.1294\n" - ] + "cell_type": "markdown", + "metadata": { + "id": "D5AVt4If6QV8" + }, + "source": [ + "### Predictions from the transformed test set at the optimal classification threshold" + ] }, { - "name": "stderr", - "output_type": "stream", - "text": [ - "100%|██████████| 100/100 [00:01<00:00, 97.34it/s]\n" - ] - } - ], - "source": [ - "display(Markdown(\"#### Predictions from transformed testing data\"))\n", - "bal_acc_arr_transf = []\n", - "disp_imp_arr_transf = []\n", - "avg_odds_diff_arr_transf = []\n", - "\n", - "print(\"Classification threshold used = %.4f\" % best_class_thresh)\n", - "for thresh in tqdm(class_thresh_arr):\n", - " \n", - " if thresh == best_class_thresh:\n", - " disp = True\n", - " else:\n", - " disp = False\n", - " \n", - " fav_inds = dataset_transf_test_pred.scores > thresh\n", - " dataset_transf_test_pred.labels[fav_inds] = dataset_transf_test_pred.favorable_label\n", - " dataset_transf_test_pred.labels[~fav_inds] = dataset_transf_test_pred.unfavorable_label\n", - " \n", - " metric_test_aft = compute_metrics(dataset_orig_test, dataset_transf_test_pred, \n", - " unprivileged_groups, privileged_groups,\n", - " disp = disp)\n", - "\n", - " bal_acc_arr_transf.append(metric_test_aft[\"Balanced accuracy\"])\n", - " avg_odds_diff_arr_transf.append(metric_test_aft[\"Average odds difference\"])\n", - " disp_imp_arr_transf.append(metric_test_aft[\"Disparate impact\"])" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "#### Display results for all thresholds" - ] - }, - { - "cell_type": "code", - "execution_count": 19, - "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" + "cell_type": "code", + "execution_count": 22, + "metadata": { + "id": "rnTyxVhc6QV8", + "outputId": "51cb8bc1-8e1e-483b-e3ac-500abc92b4c5", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 231 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "" + ], + "text/markdown": "#### Predictions from transformed testing data" + }, + "metadata": {} + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Classification threshold used = 0.2872\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + " 33%|███▎ | 33/100 [00:01<00:03, 17.80it/s]" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Balanced accuracy = 0.7134\n", + "Statistical parity difference = -0.0705\n", + "Disparate impact = 0.7785\n", + "Average odds difference = 0.0188\n", + "Equal opportunity difference = 0.0293\n", + "Theil index = 0.1401\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + " 62%|██████▏ | 62/100 [00:02<00:01, 35.27it/s]divide by zero encountered in double_scalars\n", + "invalid value encountered in double_scalars\n", + "100%|██████████| 100/100 [00:03<00:00, 28.45it/s]\n" + ] + } + ], + "source": [ + "display(Markdown(\"#### Predictions from transformed testing data\"))\n", + "bal_acc_arr_transf = []\n", + "disp_imp_arr_transf = []\n", + "avg_odds_diff_arr_transf = []\n", + "\n", + "print(\"Classification threshold used = %.4f\" % best_class_thresh)\n", + "for thresh in tqdm(class_thresh_arr):\n", + " \n", + " if thresh == best_class_thresh:\n", + " disp = True\n", + " else:\n", + " disp = False\n", + " \n", + " fav_inds = dataset_transf_test_pred.scores > thresh\n", + " dataset_transf_test_pred.labels[fav_inds] = dataset_transf_test_pred.favorable_label\n", + " dataset_transf_test_pred.labels[~fav_inds] = dataset_transf_test_pred.unfavorable_label\n", + " \n", + " metric_test_aft = compute_metrics(dataset_orig_test, dataset_transf_test_pred, \n", + " unprivileged_groups, privileged_groups,\n", + " disp = disp)\n", + "\n", + " bal_acc_arr_transf.append(metric_test_aft[\"Balanced accuracy\"])\n", + " avg_odds_diff_arr_transf.append(metric_test_aft[\"Average odds difference\"])\n", + " disp_imp_arr_transf.append(metric_test_aft[\"Disparate impact\"])" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "pcZpe7I36QV8" + }, + "source": [ + "#### Display results for all thresholds" + ] + }, + { + "cell_type": "code", + "execution_count": 23, + "metadata": { + "id": "_CXynuBr6QV9", + "outputId": "d8111f05-4fda-4a95-ce18-236790be8a51", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 451 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "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\n" + }, + "metadata": { + "needs_background": "light" + } + } + ], + "source": [ + "fig, ax1 = plt.subplots(figsize=(10,7))\n", + "ax1.plot(class_thresh_arr, bal_acc_arr_transf)\n", + "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", + "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", + "ax1.xaxis.set_tick_params(labelsize=14)\n", + "ax1.yaxis.set_tick_params(labelsize=14)\n", + "\n", + "\n", + "ax2 = ax1.twinx()\n", + "ax2.plot(class_thresh_arr, np.abs(1.0-np.array(disp_imp_arr_transf)), color='r')\n", + "ax2.set_ylabel('abs(1-disparate impact)', color='r', fontsize=16, fontweight='bold')\n", + "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", + "ax2.yaxis.set_tick_params(labelsize=14)\n", + "ax2.grid(True)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "mu4UDE7V6QV9" + }, + "source": [ + "```abs(1-disparate impact)``` must be small (close to 0) for classifier predictions to be fair.\n", + "\n", + "For a classifier trained with reweighted training data, at the best classification rate, this is indeed the case.\n", + "This implies fairness." + ] + }, + { + "cell_type": "code", + "execution_count": 24, + "metadata": { + "id": "wr9NS9ty6QV9", + "outputId": "e6b055f6-24b3-4e3d-ef2f-5333b8b63555", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 451 + } + }, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "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\n" + }, + "metadata": { + "needs_background": "light" + } + } + ], + "source": [ + "fig, ax1 = plt.subplots(figsize=(10,7))\n", + "ax1.plot(class_thresh_arr, bal_acc_arr_transf)\n", + "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", + "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", + "ax1.xaxis.set_tick_params(labelsize=14)\n", + "ax1.yaxis.set_tick_params(labelsize=14)\n", + "\n", + "\n", + "ax2 = ax1.twinx()\n", + "ax2.plot(class_thresh_arr, avg_odds_diff_arr_transf, color='r')\n", + "ax2.set_ylabel('avg. odds diff.', color='r', fontsize=16, fontweight='bold')\n", + "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", + "ax2.yaxis.set_tick_params(labelsize=14)\n", + "ax2.grid(True)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "jeZXlbAn6QV9" + }, + "source": [ + "```average odds difference = 0.5((FPR_unpriv-FPR_priv)+(TPR_unpriv-TPR_priv))``` must be close to zero for the classifier to be fair.\n", + "\n", + "For a classifier trained with reweighted training data, at the best classification rate, this is indeed the case.\n", + "This implies fairness." + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "id": "uuxBnP7T6QV9" + }, + "source": [ + "# Summary of Results\n", + "We show the optimal classification thresholds, and the fairness and accuracy metrics." + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "fS9pME8n6QV9" + }, + "source": [ + "### Classification Thresholds\n", + "\n", + "| Dataset |Classification threshold|\n", + "|-|-|\n", + "|Adult||0.2674|\n", + "|German|0.6732|\n", + "|Compas|0.5148|" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "F_-hvWly6QV9" + }, + "source": [ + "### Fairness Metric: Disparate impact, Accuracy Metric: Balanced accuracy\n", + "\n", + "#### Performance\n", + "\n", + "| Dataset |Sex (Acc-Bef)|Sex (Acc-Aft)|Sex (Fair-Bef)|Sex (Fair-Aft)|Race/Age (Acc-Bef)|Race/Age (Acc-Aft)|Race/Age (Fair-Bef)|Race/Age (Fair-Aft)|\n", + "|-|-|-|-|-|-|-|-|-|\n", + "|Adult (Test)|0.7417|0.7128|0.2774|0.7625|0.7417|0.7443|0.4423|0.7430|\n", + "|German (Test)|0.6524|0.6460|0.9948|1.0852|0.6524|0.6460|0.3824|0.5735|\n", + "|Compas (Test)|0.6774|0.6562|0.6631|0.8342|0.6774|0.6342|0.6600|1.1062|\n", + "\n" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "_SMtlDf66QV9" + }, + "source": [ + "### Fairness Metric: Average odds difference, Accuracy Metric: Balanced accuracy\n", + "\n", + "#### Performance\n", + "\n", + "| Dataset |Sex (Acc-Bef)|Sex (Acc-Aft)|Sex (Fair-Bef)|Sex (Fair-Aft)|Race/Age (Acc-Bef)|Race/Age (Acc-Aft)|Race/Age (Fair-Bef)|Race/Age (Fair-Aft)|\n", + "|-|-|-|-|-|-|-|-|-|\n", + "|Adult (Test)|0.7417|0.7128|-0.3281|-0.0266|0.7417|0.7443|-0.1991|-0.0395|\n", + "|German (Test)|0.6524|0.6460|0.0071|0.0550|0.6524|0.6460|-0.3278|-0.1944|\n", + "|Compas (Test)|0.6774|0.6562|-0.2439|-0.0946|0.6774|0.6342|-0.1927|0.1042|" + ] } - ], - "source": [ - "fig, ax1 = plt.subplots(figsize=(10,7))\n", - "ax1.plot(class_thresh_arr, bal_acc_arr_transf)\n", - "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", - "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", - "ax1.xaxis.set_tick_params(labelsize=14)\n", - "ax1.yaxis.set_tick_params(labelsize=14)\n", - "\n", - "\n", - "ax2 = ax1.twinx()\n", - "ax2.plot(class_thresh_arr, np.abs(1.0-np.array(disp_imp_arr_transf)), color='r')\n", - "ax2.set_ylabel('abs(1-disparate impact)', color='r', fontsize=16, fontweight='bold')\n", - "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", - "ax2.yaxis.set_tick_params(labelsize=14)\n", - "ax2.grid(True)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "```abs(1-disparate impact)``` must be small (close to 0) for classifier predictions to be fair.\n", - "\n", - "For a classifier trained with reweighted training data, at the best classification rate, this is indeed the case.\n", - "This implies fairness." - ] - }, - { - "cell_type": "code", - "execution_count": 20, - "metadata": {}, - "outputs": [ - { - "data": { - "image/png": "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\n", - "text/plain": [ - "
" - ] - }, - "metadata": {}, - "output_type": "display_data" + ], + "metadata": { + "kernelspec": { + "display_name": "Python 2", + "language": "python", + "name": "python2" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 2 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython2", + "version": "2.7.10" + }, + "colab": { + "provenance": [] } - ], - "source": [ - "fig, ax1 = plt.subplots(figsize=(10,7))\n", - "ax1.plot(class_thresh_arr, bal_acc_arr_transf)\n", - "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n", - "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n", - "ax1.xaxis.set_tick_params(labelsize=14)\n", - "ax1.yaxis.set_tick_params(labelsize=14)\n", - "\n", - "\n", - "ax2 = ax1.twinx()\n", - "ax2.plot(class_thresh_arr, avg_odds_diff_arr_transf, color='r')\n", - "ax2.set_ylabel('avg. odds diff.', color='r', fontsize=16, fontweight='bold')\n", - "ax2.axvline(best_class_thresh, color='k', linestyle=':')\n", - "ax2.yaxis.set_tick_params(labelsize=14)\n", - "ax2.grid(True)" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "```average odds difference = 0.5((FPR_unpriv-FPR_priv)+(TPR_unpriv-TPR_priv))``` must be close to zero for the classifier to be fair.\n", - "\n", - "For a classifier trained with reweighted training data, at the best classification rate, this is indeed the case.\n", - "This implies fairness." - ] - }, - { - "cell_type": "markdown", - "metadata": { - "collapsed": true - }, - "source": [ - "# Summary of Results\n", - "We show the optimal classification thresholds, and the fairness and accuracy metrics." - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Classification Thresholds\n", - "\n", - "| Dataset |Classification threshold|\n", - "|-|-|\n", - "|Adult||0.2674|\n", - "|German|0.6732|\n", - "|Compas|0.5148|" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Fairness Metric: Disparate impact, Accuracy Metric: Balanced accuracy\n", - "\n", - "#### Performance\n", - "\n", - "| Dataset |Sex (Acc-Bef)|Sex (Acc-Aft)|Sex (Fair-Bef)|Sex (Fair-Aft)|Race/Age (Acc-Bef)|Race/Age (Acc-Aft)|Race/Age (Fair-Bef)|Race/Age (Fair-Aft)|\n", - "|-|-|-|-|-|-|-|-|-|\n", - "|Adult (Test)|0.7417|0.7128|0.2774|0.7625|0.7417|0.7443|0.4423|0.7430|\n", - "|German (Test)|0.6524|0.6460|0.9948|1.0852|0.6524|0.6460|0.3824|0.5735|\n", - "|Compas (Test)|0.6774|0.6562|0.6631|0.8342|0.6774|0.6342|0.6600|1.1062|\n", - "\n" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "### Fairness Metric: Average odds difference, Accuracy Metric: Balanced accuracy\n", - "\n", - "#### Performance\n", - "\n", - "| Dataset |Sex (Acc-Bef)|Sex (Acc-Aft)|Sex (Fair-Bef)|Sex (Fair-Aft)|Race/Age (Acc-Bef)|Race/Age (Acc-Aft)|Race/Age (Fair-Bef)|Race/Age (Fair-Aft)|\n", - "|-|-|-|-|-|-|-|-|-|\n", - "|Adult (Test)|0.7417|0.7128|-0.3281|-0.0266|0.7417|0.7443|-0.1991|-0.0395|\n", - "|German (Test)|0.6524|0.6460|0.0071|0.0550|0.6524|0.6460|-0.3278|-0.1944|\n", - "|Compas (Test)|0.6774|0.6562|-0.2439|-0.0946|0.6774|0.6342|-0.1927|0.1042|" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - } - ], - "metadata": { - "kernelspec": { - "display_name": "Python 2", - "language": "python", - "name": "python2" }, - "language_info": { - "codemirror_mode": { - "name": "ipython", - "version": 2 - }, - "file_extension": ".py", - "mimetype": "text/x-python", - "name": "python", - "nbconvert_exporter": "python", - "pygments_lexer": "ipython2", - "version": "2.7.10" - } - }, - "nbformat": 4, - "nbformat_minor": 2 + "nbformat": 4, + "nbformat_minor": 0 } \ No newline at end of file