{ "cells": [ { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "# ECE 046211 - Technion - Deep Learning\n", "---\n", "\n", "#### Tal Daniel\n", "\n", "## Tutorial 01 - Machine Learning Basics Refresher\n", "---" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Agenda\n", "---\n", "\n", "* [Supervised Vs. Unsupervised Learning](#-Supervised-Vs.-Unsupervised-Learning)\n", "* [Evaluation: Train-Test Split](#-Evaluation:-Train-Test-Split)\n", " * [Train-Validation-Test Split](#-Train-Validation-Test-Split)\n", "* [The Bias-Variance Tradeoff](#-The-Bias-Variance-Tradeoff)\n", "* [Feature Scaling: Normalization and Standartization](#-Feature-Scaling:-Normalization-and-Standartization)\n", " * [Normalization - MinMax Scaling](#Normalization---MinMax-Scaling)\n", " * [Standartization](#Standartization)\n", "* [Linear Regression](#-Linear-Regression)\n", " * [Linear Regression Cost Function](#-Linear-Regression-Cost-Function)\n", " * [Closed-Form Least-Squares Solution](#-Closed-Form-Least-Squares-Solution)\n", " * [Gradient Descent Solution](#-Gradient-Descent-Solution)\n", "* [Regularization](#-Regularization)\n", " * [Ridge Regression](#-Ridge-Regression)\n", " * [LASSO Regression](#-LASSO-Regression)\n", "* [Recommended Videos](#-Recommended-Videos)\n", "* [Credits](#-Credits)" ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "# imports for the tutorial\n", "import numpy as np\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "from mpl_toolkits.mplot3d import Axes3D\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import MinMaxScaler, StandardScaler\n", "from sklearn.linear_model import Ridge, Lasso\n", "%matplotlib ipympl" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "## Supervised Vs. Unsupervised Learning\n", "---\n", "* We roughly divide ML algorithms to 2.5 groups:\n", " * **Supervised Learning** - along with our data (features), we are also given **labels** which match the features. For example, a datasets that contains images of cats and dogs, and each image is accompanied by a label \"dog\" or \"cat\".\n", " * Typical tasks: classification, regression.\n", " * **Unsupervised Learning** - we are only given features from some data distribution, and try to learn only from them, without any known labels. For example, a dataset that contains samples from a mixture of Gaussians, and we try to learn the parameters of the disribution, or discover patterns in unlabeled medical data using clustering.\n", " * Typical tasks: pattern discovery, clustering, dimensionality reduction, density estimation.\n", " * **Semi-Supervised Learning** - we are given features, but also very few known labels. It is usually the case that we can't employ supervised learning algorithms, due to the risk of overfitting. For example, a dataset containing samples of regular data, and we are also given very few labels of anomalies in order to build better outlier detectors.\n", " \n", "* Note: there is also *Reinforcement Learning*, but we will not address it in this course." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n", "\n", "* Image Source" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n", "\n", "* Image Source" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "## Evaluation: Train-Test Split\n", "---\n", "\n", "It is important to evaluate the classifier **generalization** performance in order to:\n", "* Determine whether to employ/distribute the classifier.\n", "* Compare classifiers (even compare the same type of classifiers, but with different parameters).\n", "* Optimize the classifier to perform better on unseen data." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* **Train-Test Separation** - The *naive* approach is seprating the data into train set and test set, that is, taking a portion of the data for training the model (usually about 80% of the dataset) and save another portion, that the model **has not seen** in order to test the model's performance. This is called the test set (usually about 20% of the dataset).\n", " * **Shuffling** - Shuffling the data serves the purpose of reducing variance and making sure that models remain general and overfit less. The popular case where shuffling is very important is when the data is sorted by their class/target. By shuffling we add randomness that assures that the training/test/validation sets are representative of the overall distribution of the data.\n", "\n", "* Note: Scikit-learn has a function we can use called `train_test_split` that makes it easy for us to split our dataset into training and testing data." ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "total samples: 569\n", "total positive sampels (M): 212, total negative samples (B): 357\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
iddiagnosisradius_meantexture_meanperimeter_meanarea_meansmoothness_meancompactness_meanconcavity_meanconcave points_mean...texture_worstperimeter_worstarea_worstsmoothness_worstcompactness_worstconcavity_worstconcave points_worstsymmetry_worstfractal_dimension_worstUnnamed: 32
61858981B8.59820.9854.66221.80.124300.089630.0300000.009259...27.0462.06273.90.163900.169800.090010.027780.29720.07712NaN
334897374B12.30019.0277.88464.40.083130.042020.0077560.008535...28.4684.53544.30.122200.090520.036190.039830.25540.07207NaN
494914102B13.16020.5484.06538.70.073350.052750.0180000.012560...28.4695.29648.30.111800.164600.076980.041950.26870.07429NaN
5585759902B11.52018.7573.34409.00.095240.054730.0303600.022780...22.4781.81506.20.124900.087200.090760.063160.33060.07036NaN
2238812877M15.75020.25102.60761.30.102500.120400.1147000.064620...30.29125.901088.00.155200.448000.397600.147900.39930.10640NaN
477911673B13.90016.6288.97599.40.068280.053190.0222400.013390...21.80101.20718.90.093840.200600.138400.062220.26790.07698NaN
517916838M19.89020.26130.501214.00.103700.131000.1411000.094310...25.23160.501646.00.141700.330900.418500.161300.25490.09136NaN
423906878B13.66019.1389.46575.30.090570.114700.0965700.048120...25.50101.40708.80.114700.316700.366000.140700.27440.08839NaN
345898677B10.26014.7166.20321.60.098820.091590.0358100.020370...19.4870.89357.10.136000.163600.071620.040740.24340.08488NaN
195875878B12.91016.3382.53516.40.079410.053660.0387300.023770...22.0090.81600.60.109700.150600.176400.082350.30240.06949NaN
\n", "

10 rows × 33 columns

\n", "
" ], "text/plain": [ " id diagnosis radius_mean texture_mean perimeter_mean area_mean \\\n", "61 858981 B 8.598 20.98 54.66 221.8 \n", "334 897374 B 12.300 19.02 77.88 464.4 \n", "494 914102 B 13.160 20.54 84.06 538.7 \n", "55 85759902 B 11.520 18.75 73.34 409.0 \n", "223 8812877 M 15.750 20.25 102.60 761.3 \n", "477 911673 B 13.900 16.62 88.97 599.4 \n", "517 916838 M 19.890 20.26 130.50 1214.0 \n", "423 906878 B 13.660 19.13 89.46 575.3 \n", "345 898677 B 10.260 14.71 66.20 321.6 \n", "195 875878 B 12.910 16.33 82.53 516.4 \n", "\n", " smoothness_mean compactness_mean concavity_mean concave points_mean \\\n", "61 0.12430 0.08963 0.030000 0.009259 \n", "334 0.08313 0.04202 0.007756 0.008535 \n", "494 0.07335 0.05275 0.018000 0.012560 \n", "55 0.09524 0.05473 0.030360 0.022780 \n", "223 0.10250 0.12040 0.114700 0.064620 \n", "477 0.06828 0.05319 0.022240 0.013390 \n", "517 0.10370 0.13100 0.141100 0.094310 \n", "423 0.09057 0.11470 0.096570 0.048120 \n", "345 0.09882 0.09159 0.035810 0.020370 \n", "195 0.07941 0.05366 0.038730 0.023770 \n", "\n", " ... texture_worst perimeter_worst area_worst smoothness_worst \\\n", "61 ... 27.04 62.06 273.9 0.16390 \n", "334 ... 28.46 84.53 544.3 0.12220 \n", "494 ... 28.46 95.29 648.3 0.11180 \n", "55 ... 22.47 81.81 506.2 0.12490 \n", "223 ... 30.29 125.90 1088.0 0.15520 \n", "477 ... 21.80 101.20 718.9 0.09384 \n", "517 ... 25.23 160.50 1646.0 0.14170 \n", "423 ... 25.50 101.40 708.8 0.11470 \n", "345 ... 19.48 70.89 357.1 0.13600 \n", "195 ... 22.00 90.81 600.6 0.10970 \n", "\n", " compactness_worst concavity_worst concave points_worst symmetry_worst \\\n", "61 0.16980 0.09001 0.02778 0.2972 \n", "334 0.09052 0.03619 0.03983 0.2554 \n", "494 0.16460 0.07698 0.04195 0.2687 \n", "55 0.08720 0.09076 0.06316 0.3306 \n", "223 0.44800 0.39760 0.14790 0.3993 \n", "477 0.20060 0.13840 0.06222 0.2679 \n", "517 0.33090 0.41850 0.16130 0.2549 \n", "423 0.31670 0.36600 0.14070 0.2744 \n", "345 0.16360 0.07162 0.04074 0.2434 \n", "195 0.15060 0.17640 0.08235 0.3024 \n", "\n", " fractal_dimension_worst Unnamed: 32 \n", "61 0.07712 NaN \n", "334 0.07207 NaN \n", "494 0.07429 NaN \n", "55 0.07036 NaN \n", "223 0.10640 NaN \n", "477 0.07698 NaN \n", "517 0.09136 NaN \n", "423 0.08839 NaN \n", "345 0.08488 NaN \n", "195 0.06949 NaN \n", "\n", "[10 rows x 33 columns]" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# let's load the cancer dataset, shuffle it and speratre into train and test set\n", "dataset = pd.read_csv('./datasets/cancer_dataset.csv')\n", "# print the number of rows in the data set\n", "number_of_rows = len(dataset)\n", "print(\"total samples: {}\".format(number_of_rows))\n", "total_positive_samples = np.sum(dataset['diagnosis'].values == 'M')\n", "print(\"total positive sampels (M): {}, total negative samples (B): {}\".format(\n", " total_positive_samples, number_of_rows - total_positive_samples))\n", "num_train = int(0.8 * number_of_rows)\n", "dataset.sample(10)" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "train size: 455, test size: 114\n" ] } ], "source": [ "# split to train-test\n", "x = dataset[['radius_mean', 'texture_mean']].values\n", "y = dataset['diagnosis'].values == 'M' # 1 for Malignat, 0 for Benign\n", "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2, stratify=y)\n", "# stratify=y: same proportions of class labels in train and test sets, we provide the labels y\n", "print(\"train size: {}, test size: {}\".format(len(X_train), len(X_test)))" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* The accuracy/error estimates on the *training* data **are not** good indicators of performance on future data!\n", "* **The reason**: new data will probably not be exactly the same as the training data.\n", "* The accuracy/error estimates on the training data measure the degree of classifier's **underfitting** or **overfitting**.\n", "* A typical learning curve usually looks like this:
\n", "\n", "Image Source" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Train-Validation-Test Split\n", "---\n", "* Let's say you have a dataset which you separated into *train* set and *test* set. \n", "* You want to train a classifier that has **hyper-parameters** (parameters that are not trained, but are user-selected before the training) that you have to tune. \n", " * For example, in the classifier \"K-Nearest Neighbours\" you need to choose $K$, the number of nearest neighbours needed to perform classification, or in \"Stochastic Gradient Descent\", you need to choose the learning rate (which is a continuous value, like $0.0001$). " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* Would it be fair to train the classifier on the train set for each hyper-parameter and then test it on the test set and finally selecting the best hyper-parameters based on the performance on the test set? **NO!**\n", " * It is like taking an open-material exam, but instead of bringing all of the material, you bring only the material relevant to the questions asked in the exam. It is sort of cheating." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "That is why we seperate into 3 sets:\n", "\n", "* **Train Set** - from which the model learns.\n", "* **Validation Set** - on which the hyper-parameters are tuned.\n", "* **Test Set** - untouched samples on which you test the generalization ability of the model. This set has **never** been seen by the model." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Cross-Validation\n", "---\n", "* Separating to validation and test sets is not very data efficient as we now allocate less data for training.\n", "* **Cross-validation** is randomly splitting the data into $k$ groups. One of the groups is used as the test set and the rest are used as the training set. \n", "* The model is trained on the training set and scored on the test set. Then the process is repeated until each unique group has been used as the test set. This method is also called **\"K-Fold Cross Validation\"**.\n", "* Scikit-Learn: `cross_val_score` - takes in a classifier, training data, $k$ (number of folds) and the scoring technique (\"accuracy\" for example). Read the Doc Here\n", "\n", "
\n", "\n", "Image Source" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### The Bias-Variance Tradeoff\n", "---\n", "An important theoretical result of statistics and **classic** ML is the fact the **the model generalization error** can be expressed as a sum of 3 different errors: **bias, variance and irreducible error**.\n", "* Given a **true** (but unknown) function $F(x)$ with noise $F(x) = f + \\epsilon$, we seek to estimate it based on $n$ samples from a set $\\mathcal{D}$. We denote the **regression function** as $g(x; \\mathcal{D})$.\n", "* The error of the regression model is given by: $$ MSE = \\mathbb{E}_{\\mathcal{D}} \\big[(F(x) - g(x; \\mathcal{D}) )^2 \\big] $$\n", "* The total error can be decomposed into 3 terms: $$ E[(F - g )^2] = E[(f + \\epsilon -g)^2] = E[(f + \\epsilon -g +E[g] - E[g])^2]$$ $$ = ... = (f - E[g])^2 +E[\\epsilon^2] +E\\big[(E[g] - g)^2\\big]$$ $$ = Bias[g]^2 +\\sigma^2 +Var[g]$$\n", "\n", "\n", "* Full derivation on Wikipedia " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* **Bias** - $(f - E[g])$: It is the difference between the expected value and the true value. This part of the generalization error is due to wrong assumptions (e.g. assuming the data is linear when it is actually quadratic). A *high-bias* model is most likely to **underfit** the training data.\n", "* **Variance** - $E\\big[(E[g] - g)^2\\big]$: This part is due to the model's sensitivity to small variations in the training data. A model with many degrees of freedom (such as high-degree polynomial model) is likely to have **high variance**, and thus tend to **overfit** the training data. Low varaince - the estimate of $F$ does not change much as the training set varies.\n", "* **Irreducible Error** - $\\sigma^2$: This part is due to the noisiness of the data itself. The only way to reduce this part of the error is to clean up the data (e.g., fix the data sources, detect and remove outliers...)." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n", "\n", "* Note: this is the classical curve, and as you will see later in the course, there is actually something special that happens in deep learning, and it is different than what you see in the curve above.\n", "\n", "Image Source" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "## Pre-Processing - Feature Scaling: Normalization and Standartization\n", "---\n", "* Feature scaling is a fudemental part of the data pre-processing stage.\n", "* It can improve the performance of some machine learning algorithms, but may also harm others.\n", "* It is especially important for **Gradient Descent-based** algorithms such as linear regression, logistic regression, neural networks, and etc.\n", "* The range of feature also significantly affects **distance-based** algorithms such as KNN, SVM and K-Means." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* Let's take at the general formulation of Gradient Descent for linear regression: $$ \\theta_j^{(t+1)} \\leftarrow \\theta_j^{(t)} -\\alpha \\sum_{i=1}^n (h_{\\theta}(x^{(i)}) - y^{(i)})x_{j}^{(i)}, $$\n", "where $x, \\theta \\in \\mathcal{R}^m$ and $h_{\\theta}:\\mathcal{R}^m \\to \\mathcal{R}.$\n", "* The presence of feature value X in the formula will affect the step size of the gradient descent!\n", " * The difference in ranges of features will cause different step sizes for each feature. \n", "* To ensure that the gradient descent moves smoothly towards the minima and that the steps for gradient descent are updated at the same rate for all the features, we scale the data before feeding it to the model.\n", "* **Having features on a similar scale can usually help the gradient descent converge more quickly towards the minima.**" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Normalization - MinMax Scaling\n", "---\n", "* Normalization is a scaling technique in which values are shifted and rescaled so that they end up **ranging between 0 and 1**. \n", " * It is also known as Min-Max scaling. $$ X_{scaled} = \\frac{X - X_{min}}{X_{max} - X_{min}} $$\n", " * $X_{max}$ and $X_{min}$ are the maximum and the minimum values of the feature respectively.\n", "* Normalization is good to use when you know that the distribution of your data does not follow a Gaussian distribution. \n", "* This can be useful in algorithms that do not assume any distribution of the data like K-NN and Neural Networks." ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [], "source": [ "# fit scaler on training data (not on test data!)\n", "scaler = MinMaxScaler().fit(X_train)\n", "\n", "# transform training data\n", "X_train_norm = scaler.transform(X_train)\n", "\n", "# transform testing data\n", "X_test_norm = scaler.transform(X_test)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Standartization\n", "---\n", "* Standardization is a scaling technique where the values are centered around the mean with a unit standard deviation. \n", "* This means that the mean of the features becomes zero and the resultant distribution has a unit standard deviation. $$ X_{scaled} = \\frac{X - \\mu}{\\sigma} $$\n", " * $\\mu$ is the (empirical) mean of the feature values and $\\sigma$ is the (empirical) standard deviation of the feature values. \n", " * Note that in this case, the values are not restricted to a particular range.\n", "* Standardization can be helpful in cases where the data follows a Gaussian distribution. \n", "* However, this does not have to be necessarily true. \n", "* Unlike normalization, standardization does not have a bounding range. So, even if you have outliers in your data, they will not be affected by standardization." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n", "\n", "Image Source" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [], "source": [ "# fit scaler on training data (not on test data!)\n", "scaler = StandardScaler().fit(X_train)\n", "\n", "# transform training data\n", "X_train_norm = scaler.transform(X_train)\n", "\n", "# transform testing data\n", "X_test_norm = scaler.transform(X_test)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Feature Scaling In Practice\n", "---\n", "* At the end of the day, the choice of using normalization or standardization will depend on your problem and the machine learning algorithm you are using. \n", "* You can always start by fitting your model to raw, normalized and standardized data and compare the performance for best results.\n", "* It is a good practice to fit the scaler on the training data and then use it to transform the testing data. \n", "* This would avoid any data leakage during the model testing process. \n", "* Scaling of target values (or labels) is usually not required." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "## Linear Regression\n", "---\n", "A linear model makes a prediction by computing a weighted sum of the input features, plus a constant called the *bias term* (also called *intercept* sometimes).\n", "\n", "We denote:\n", "* $\\hat{y}$ - the predicted value\n", "* $n$ - number of features\n", "* $x_i$ is the $i^{th}$ feature value\n", "* $\\theta_j$ is the $j^{th}$ model parameter, where $\\theta_0$ is the bias weight ($\\theta_0 \\cdot 1$)\n", "\n", "The *Linear Regression* model prediction: $$ \\hat{y} = \\theta_0 + \\theta_1 x_1 + ... + \\theta_n x_n $$\n", "In *vector* form: $$ \\hat{y}=h_{\\theta} (x) =\\theta^T \\cdot x $$\n", "\n", "* In fact, we can look at the linear regression problem as one application of a *single neuron* as we will discuss in the next lesson." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Linear Regression Cost Function \n", "---\n", "How do we train a linear regression model?\n", "\n", "* Training a model means tuning its parameters so that the model best fits to the training data (train set)\n", "* The most common performance measure on a regression model is the Root Mean Square Error (RMSE), thus, we need to find the value of $\\theta$ that minimzes the RMSE. In practice, it is simpler to minimize the Mean Square Error - **MSE** (the same values that minimize the RMSE, also minimize the MSE).\n", "* The MSE Cost Function: $$ MSE(X, h_{\\theta}) = \\frac{1}{m}\\sum_{i=1}^m (\\theta^T \\cdot x^{(i)} - y^{(i)})^2 $$\n", " * Pay attention to notations: $m$ is the number of samples, where $n$ is the number of features. That is, each sample has $n$ components." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Closed-Form Least-Squares Solution\n", "---\n", "To find the value of $\\theta$ that minimizes the cost function, there is a *closed-form solution* - a mathemtical equation that gives the result directly. It is also called the **Normal Equation**. We will now derive it.\n", "\n", "* We wish to find a solution for $\\hat{y} = X \\theta$\n", "* The parameters $\\theta$ are obtained by minimzing the *sum of squared* errors or residuals (SSE): $$ SSE(\\theta) = \\sum_{i=1}^n (\\theta^T x_i - y_i)^2 = ||X \\theta -y ||_2^2 = (X\\theta -y)^T(X\\theta - y) = \\theta^TX^TX\\theta -\\theta^TX^Ty - y^TX\\theta +y^Ty$$\n", "* Minimizing w.r.t to $\\theta$: $$\\nabla_{\\theta}SSE(\\theta) = 2X^TX\\theta - 2X^Ty = 0 \\rightarrow \\theta^{*} = (X^TX)^{-1}X^Ty $$\n", "* The matrix $(X^TX)^{-1}X^T$ is the *Pseudo Inverse* of $X$." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Gradient Descent Solution\n", "---\n", "* The general idea is to tweak parameters **iteratively** to minimize a cost function.\n", "* It measures the local gradient of the error function with regards to the parameter vector ($\\theta$ or $w$), and it goes down in the direction of the descending gradient. Once the gradient is zero - the minimum is reached (=convergence).\n", "\n", "* **Learning Rate** hyperparameter - it is the size of step to be taken in each iteration.\n", " * Too *small* $\\rightarrow$ the algorithm will have to go through many iterations to converge, which will take a long time.\n", " * Too *high* $\\rightarrow$ might make the algorithm diverge as it may miss the minimum.\n", "\n", "
\n", "\n", "Image Source" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "#### Stochastic Gradient Descent (Mini-Batch Gradient Descent)\n", "---\n", "* **Pseudocode**:\n", " * **Require**: Learning rate $\\alpha_k$\n", " * **Require**: Initial parameter $w$\n", " * **While** stopping criterion not met **do**\n", " * Sample a minibatch of $m$ examples from the training set ($m=1$ for SGD)\n", " * Set $\\tilde{X} = [x_1,...,x_m] $ with corresponding targets $\\tilde{Y} = [y_1,...,y_m]$\n", " * Compute gradient: $g \\leftarrow 2\\tilde{X}^T\\tilde{X}w - 2\\tilde{X}^T \\tilde{Y}$\n", " * Apply update: $w \\leftarrow w - \\alpha_k g$\n", " * $k \\leftarrow k + 1$\n", " * **end while**" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "slideshow": { "slide_type": "skip" } }, "outputs": [], "source": [ "# helper functions\n", "\n", "def plot_3d_lls(x, y, z, lls_sol, title=\"\"):\n", " # plot\n", " fig = plt.figure(figsize=(5, 5))\n", " ax = fig.add_subplot(111, projection='3d')\n", " ax.scatter(x, y, z, label='Y')\n", " ax.scatter(x, y, lls_sol, label='Xw')\n", " ax.legend()\n", " ax.set_xlabel('Radius Mean')\n", " ax.set_ylabel('Area Mean')\n", " ax.set_zlabel('Perimeter Mean')\n", " ax.set_title(title)\n", " \n", " \n", "def batch_generator(x, y, batch_size, shuffle=True):\n", " \"\"\"\n", " This function generates batches for a given dataset x.\n", " \"\"\"\n", " N, L = x.shape\n", " num_batches = N // batch_size\n", " batch_x = []\n", " batch_y = []\n", " if shuffle:\n", " # shuffle\n", " rand_gen = np.random.RandomState(0)\n", " shuffled_indices = rand_gen.permutation(np.arange(N))\n", " x = x[shuffled_indices, :]\n", " y = y[shuffled_indices, :]\n", " for i in range(N):\n", " batch_x.append(x[i, :])\n", " batch_y.append(y[i, :])\n", " if len(batch_x) == batch_size:\n", " yield np.array(batch_x).reshape(batch_size, L), np.array(batch_y).reshape(batch_size, 1)\n", " batch_x = []\n", " batch_y = []\n", " if batch_x:\n", " yield np.array(batch_x).reshape(-1, L), np.array(batch_y).reshape(-1, 1)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "train size: 455, test size: 114\n" ] } ], "source": [ "# split to train-test\n", "x = dataset[['radius_mean', 'area_mean']].values\n", "y = dataset[['perimeter_mean']].values\n", "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2)\n", "N = len(X_train)\n", "print(\"train size: {}, test size: {}\".format(len(X_train), len(X_test)))" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "total batches: 28\n" ] } ], "source": [ "# multivaraite mini-batch gradient descent\n", "# scaling (done maually, but you can also use sklearn)\n", "X_train = (X_train - X_train.mean(axis=0, keepdims=True)) / X_train.std(axis=0, keepdims=True)\n", "y_train = (y_train - y_train.mean(axis=0, keepdims=True)) / y_train.std(axis=0, keepdims=True)\n", "batch_size = 16\n", "num_batches = N // batch_size\n", "print(\"total batches:\", num_batches)" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "iter: 0 batch: 0 w = \n", "[[0.]\n", " [0.]]\n", "iter: 1 batch: 0 w = \n", "[[0.42729943]\n", " [0.41258379]]\n", "iter: 2 batch: 0 w = \n", "[[0.49676955]\n", " [0.47094993]]\n", "iter: 3 batch: 0 w = \n", "[[0.51174938]\n", " [0.47548105]]\n", "iter: 4 batch: 0 w = \n", "[[0.51838691]\n", " [0.47187408]]\n", "iter: 5 batch: 0 w = \n", "[[0.52370328]\n", " [0.46708033]]\n", "iter: 6 batch: 0 w = \n", "[[0.52876727]\n", " [0.4621567 ]]\n", "iter: 7 batch: 0 w = \n", "[[0.53374208]\n", " [0.45726337]]\n", "iter: 8 batch: 0 w = \n", "[[0.53865315]\n", " [0.4524241 ]]\n", "iter: 9 batch: 0 w = \n", "[[0.54350493]\n", " [0.44764194]]\n" ] } ], "source": [ "num_epochs = 10\n", "alpha_k = 0.001\n", "batch_gen = batch_generator(X_train, y_train, batch_size, shuffle=True)\n", "# initialize w\n", "w = np.zeros((2, 1))\n", "for i in range(num_epochs):\n", " for batch_i, batch in enumerate(batch_gen):\n", " batch_x, batch_y = batch\n", " if batch_i % 50 == 0:\n", " print(\"iter:\", i, \"batch:\", batch_i, \" w = \")\n", " print(w)\n", " gradient = 2 * batch_x.T @ batch_x @ w - 2 * batch_x.T @ batch_y\n", " w = w - alpha_k * gradient\n", " batch_gen = batch_generator(X_train, y_train, batch_size, shuffle=True)\n", "\n", "lls_sol = X_train @ w" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "w:\n", "[[0.54829871]\n", " [0.44291675]]\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "fb5e0eede05d4f72a04822d9f64dc5c6", "version_major": 2, "version_minor": 0 }, "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfQAAAH0CAYAAADL1t+KAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOydd3xb1fn/P1dbtmx5jzgeseM4zk6cZaeQBGgpo1AoBQpfdku/paWTFkr7g5YOCqWMLqCMQCktfCktq6wwQyCBDO947z3kqT3u+f3hnsuVLEtX0pVlO+f9euXVIkvnHl1J53Oe5zyDI4QQMBgMBoPBWNQoYj0BBoPBYDAYkcMEncFgMBiMJQATdAaDwWAwlgBM0BkMBoPBWAIwQWcwGAwGYwnABJ3BYDAYjCUAE3QGg8FgMJYATNAZDAaDwVgCMEFnMBgMBmMJwASdwWAwGIwlABN0BoPBYDCWAEzQGQwGg8FYAjBBZzAYDAZjCcAEncFgMBiMJQATdAaDwWAwlgBM0BkMBoPBWAIwQWcwGAwGYwnABJ3BYDAYjCUAE3QGg8FgMJYATNAZDAaDwVgCMEFnMBgMBmMJwASdwWAwGIwlABN0BoPBYDCWAEzQGQwGg8FYAjBBZzAYDAZjCcAEncFgMBiMJQATdAaDwWAwlgBM0BkMBoPBWAIwQWcwGAwGYwnABJ3BYDAYjCUAE3QGg8FgMJYATNAZDAaDwVgCMEFnMBgMBmMJwASdwWAwGIwlABN0BoPBYDCWAEzQGQwGg8FYAjBBZzAYDAZjCcAEncFgMBiMJQATdAaDwWAwlgCLQtA7OzvBcZzwT6FQIDk5GaeffjrefPNN2a+3Z88e7NmzZ9b1n3jiCdmvFYiCggJwHOc1FzF//etfhXvy3nvvzevcFhO1tbXgOA5qtRoDAwOxno5ffvaznwnf7fb29ll/t1gsSExMBMdxuPrqq+d/ggucPXv2eK0Rer0eGzduxP333w+e52W7TqzWAsqvf/1rvPDCC/N6zZN1HSooKMC5554b8DlXX301DAZD0LE+/vhjXHDBBcjLy4NWq0VmZibKy8vxgx/8IOhrQ1kbFoWgU2688UYcOnQIH3zwAe655x60tLTg7LPPxoEDB6J63ezsbBw6dAjnnHNOVK/jj4SEBBw4cABtbW2z/vb4448jMTFx3ue02Hj00UcBAG63G3/9619jPJvAGAwG7Nu3b9bjzz33HFwuF9RqdQxmtTgoLCzEoUOHcOjQITz77LPIycnB9773Pfz4xz+W7RqxXAuA2Ag6wNahSPjPf/6DiooKTE1N4e6778abb76JBx54ALt27cKzzz4reRwpa0NUBd1qtco6Xl5eHnbu3Ildu3bhuuuuw9/+9jd4PB489thjsl7HF61Wi507dyI9PT2q1/HHZz7zGeTk5ODxxx/3erytrQ0HDhzAJZdcMu9zWkw4HA48/fTT2Lhxo9/7GAibzRbFmfnnkksuwZNPPjnLqnzsscdwwQUXQKPRzPucFgt6vR47d+7Ezp07cd555+HFF19EYWEh/vjHP8LlckU0tsfjgcPhiOlaEC3oewsEW4fC5+6778aKFSvwxhtv4NJLL8Xu3btx6aWX4p577kF3d7fkcaSsDYo77rgDpaWl0Ol0SE1Nxd69e/HRRx8JTyaE4M9//jM2bdoEvV6P5ORkXHTRRbNM/z179mDdunU4cOAAKioqEBcXh2uvvTayOxGErVu3AgCGhoa8Hv/Tn/6EU089FRkZGYiPj8f69etx9913z/pRE0Jw9913Iz8/HzqdDlu2bMFrr7026zr+3GxXX301CgoKZj2XukfEPPfcc9ixYweMRiPi4uJQWFgo+d4oFApceeWVsz7Ixx9/HLm5uTjjjDP8vu7o0aM477zzkJKSAp1Oh82bN+P//u//vJ4zMjKCG264AWvWrIHBYEBGRgZOO+00fPDBB37f/z333IN7770XK1asgMFgQHl5OQ4fPhxw/tXV1eA4zu+m67XXXgPHcXjppZeE+Vx//fXIzc2FVqtFeno6du3ahbfeekvSvfLHCy+8AJPJhK9+9au46qqr0NzcjIMHD856HnWv/etf/8LmzZuh0+nw85//HAAwODiIr3/961i+fDk0Gg1WrFiBn//853C73V5j/PznP8eOHTuQkpKCxMREbNmyBY899hgIIZLne+2116Knpwf79+8XHqNznus7MzU1hZtuugkrVqyARqNBTk4Ovvvd78JisXg9T+rvgv6Wjxw5glNOOUX4zv7mN78J6r7evHkzTjnllFmPezwe5OTk4MILLxQee/DBB7Fx40YYDAYkJCRg9erVuPXWW4PeI6mo1WqUlZXBarViZGQEgLTPkn7f7777bvzyl7/EihUroNVq8e677/pdC+hvvqamBl/+8pdhNBqRkpKC73//+3C73WhqasLnP/95JCQkoKCgAHffffesuUr5DDmOg8ViwZNPPim4uMVu8EjfWyAW+zoUS0wmE9LS0qBSqWb9TaGQblNLWhtUKhW56aabyKuvvkpeeuklcuutt5J//OMfhPK1r32NqNVq8oMf/IC8/vrr5O9//ztZvXo1yczMJIODg8Lzdu/eTVJSUkhubi75wx/+QN59913y/vvvEzno6OggAMhvf/tbr8fr6uoIAHLjjTd6Pf69732PPPjgg+T1118n77zzDrnvvvtIWloaueaaa7yed/vttxMA5LrrriOvvfYa+ctf/kJycnJIVlYW2b1796zr79u3T3jsqquuIvn5+bPmSsekfPTRR4TjOHLppZeSV199lbzzzjtk37595Iorrgj6vvPz88k555xDWltbCcdx5NVXXyWEEOJ2u0lOTg657bbbyHPPPUcAkHfffVd43TvvvEM0Gg055ZRTyLPPPktef/11cvXVV896D42NjeQb3/gGeeaZZ8h7771HXnnlFXLdddcRhULhNR59/wUFBeTzn/88eeGFF8gLL7xA1q9fT5KTk8nExETA97F582aya9euWY9ffPHFJCMjg7hcLkIIIWeeeSZJT08nf/nLX8h7771HXnjhBXLbbbeRZ555Jui9movPfvazRKvVkrGxMeE+Xn311bOel5+fT7Kzs0lhYSF5/PHHybvvvks++eQTMjAwQHJzc0l+fj55+OGHyVtvvUV+8YtfEK1WO2ucq6++mjz22GNk//79ZP/+/eQXv/gF0ev15Oc//3nQedLvzcjICDnllFPIxRdfLPzt5ptvJgUFBYTneRIfH0+uuuoq4W8Wi4Vs2rSJpKWlkXvvvZe89dZb5IEHHiBGo5GcdtpphOd54blSfxe7d+8mqamppLi4mDz00ENk//795IYbbiAAyJNPPhnwfTzwwAMEAGlubvZ6/NVXXyUAyEsvvUQIIeQf//iH8Nt98803yVtvvUUeeugh8u1vfzvovfLH7t27ydq1a2c9vmXLFqJSqYjVapX8WdLve05ODtm7dy/55z//Sd58803S0dHhdy2gn11JSQn5xS9+Qfbv309+9KMfEQDkW9/6Flm9ejX5/e9/T/bv30+uueYaAoA8//zzwuulfoaHDh0ier2enH322eTQoUPk0KFDpL6+nhBCZHlvc7FU1qFQoe87EFdddRWJj48P+JyvfvWrwnf98OHDxOl0hjSPUNYGPPLII3MOdOjQIQKA/O53v/N6vKenh+j1evKjH/1IeGz37t0EAHn77bdDmqwU6Ad51113EZfLRex2O6mqqiLl5eUkOzs74JfR4/EQl8tF/vrXvxKlUknGxsYIIYSMj48TnU5HLrjgAq/nf/jhhwSAbIJ+zz33EABhfdnEX6jdu3eTiy66iBBCyH/+8x/CcRzp6Ojw+0NavXo12bx5syCUlHPPPZdkZ2cTj8fj93put5u4XC5y+umne90X+v7Xr19P3G638Pgnn3xCAHhtAP3x+9//ngAgTU1NwmNjY2NEq9WSH/zgB8JjBoOBfPe73w1yV6TT2dlJFAoFufTSS4XHdu/eTeLj48nU1JTXc/Pz84lSqfSaIyGEfP3rXycGg4F0dXV5PU4/V7qg+kK/d3fccQdJTU31ElZ/iH+0+/btI1qtlphMJuJ2u0l2djb52c9+RgghswT9zjvvJAqFghw5csRrvH/+858EgLD4zjU/398FvUcAyMcff+z1mjVr1pAzzzwz4PsYHR0lGo2G3HrrrV6PX3zxxSQzM1P4Tn7rW98iSUlJAccKBSroLpeLuFwu0t/fT2655RYCgHz5y18mhEj/LOn3vaioaNbiG0jQfdfJTZs2EQDkX//6l/CYy+Ui6enp5MILLxQeC+Uz9P38KXK8t7lYKutQqMgl6KOjo+Qzn/kMAUAAELVaTSoqKsidd95Jpqeng84jlLVBEcj1+8orr4DjOPzP//wP3G638C8rKwsbN26cFdGYnJyM0047LbDfYEbtvMbzdV3Oxc033wy1Wg2dTodNmzahrq4OL7/88izXd2VlJc477zykpqZCqVRCrVbjyiuvhMfjQXNzMwDg0KFDsNvtuPzyy71eW1FRgfz8fEnzkcK2bdsAABdffDH+7//+D319fWGNc+211+Kll16CyWTCY489hr179/p1+be2tqKxsVF4X+J7fPbZZ2NgYABNTU3C8x966CFs2bIFOp0OKpUKarUab7/9NhoaGmaNfc4550CpVAr/vWHDBgBAV1dXwLlffvnl0Gq1Xm7Kf/zjH3A4HLjmmmuEx7Zv344nnngCv/zlL3H48OGIzz337dsHnue9XNXXXnstLBaL32CUDRs2YNWqVV6PvfLKK9i7dy+WLVvmdS/POussAMD7778vPPedd97BGWecAaPRKHzvbrvtNphMJgwPD0ue95e//GVoNBo8/fTTePXVVzE4ODhnZPsrr7yCdevWYdOmTV7zO/PMM2dFHUv5XVCysrKwffv2Wfcn2GedmpqKL3zhC16u2fHxcbz44ou48sorBbfj9u3bMTExga985St48cUXMTo6Kvn+zEV9fT3UajXUajWWLVuG3/3ud7j88svxyCOPAAjtswSA8847L6QgRN+I6NLSUnAcJ4wPACqVCitXrvS6j6F8hnMR7fdGWczrkK/mkBCOwiIhNTUVH3zwAY4cOYLf/OY3OP/889Hc3Iwf//jHWL9+fUjf/WBrgyKQD39oaAiEEGRmZgo/FPrv8OHDsyaSnZ0taVLvv//+rPE6OzuDvu473/kOjhw5goMHD+Kee+6By+XC+eefD5PJJDynu7sbp5xyCvr6+vDAAw8IN/JPf/oTgE8DnehrsrKyZl3H32Phcuqpp+KFF16A2+3GlVdeieXLl2PdunX4xz/+EdI4F110EXQ6He677z68/PLLuO666/w+j8YT3HTTTbPu8Q033AAAwud277334hvf+AZ27NiB559/HocPH8aRI0fw+c9/3m9AWGpqqtd/a7VaAMGDx1JSUnDeeefhr3/9KzweDwDgiSeewPbt27F27Vrhec8++yyuuuoqPProoygvL0dKSgquvPJKDA4OSrlFXvA8jyeeeALLli1DWVkZJiYmMDExgTPOOAPx8fF+z/T9fX+Hhobw8ssvz7qXdN70Xn7yySf43Oc+BwB45JFH8OGHH+LIkSP4yU9+IukeiYmPj8cll1yCxx9/HI899hjOOOOMOTeZQ0NDqKmpmTW/hIQEEEKE+Un9XVB8P2tg5vOW8j6uvfZa9PX1CWd9dPMmXniuuOIKPP744+jq6sKXvvQlZGRkYMeOHV7ng6FSVFSEI0eO4OjRo6irq8PExAT+9re/wWg0ApD+WVKkrmeUlJQUr//WaDSIi4uDTqeb9bjdbhf+W+pnGIhovzfKYl6HfOfx5JNPhvbmI2Tr1q24+eab8dxzz6G/vx/f+9730NnZ6TemYi6CrQ0qnufnPJhPS0sDx3H44IMPhJsmxvcx32CwuSgrK8ORI0e8Hlu2bFnQ1y1fvlwIhNu1axeysrLwP//zP7j99tvxxz/+EcBMEJTFYsG//vUvrzdaVVXlNRb9UvgTi8HBQb+7TjE6nc5vZKi/H9/555+P888/Hw6HA4cPH8add96Jyy67DAUFBSgvLw94HUpcXBwuvfRS3HnnnUhMTPQKLhKTlpYGAPjxj38853NKSkoAAH/729+wZ88ePPjgg15/n56eljSnULjmmmvw3HPPYf/+/cjLy8ORI0dmXTctLQ33338/7r//fnR3d+Oll17CLbfcguHhYbz++ushXe+tt94Sduz+xOnw4cM4ceIE1qxZIzzm7/ublpaGDRs24Fe/+pXf69Dv7TPPPAO1Wo1XXnnFawEPN8Xo2muvxaOPPoqamho8/fTTcz4vLS0Ner1+zuh9+n2Q+ruQgzPPPBPLli3Dvn37cOaZZ2Lfvn3YsWOH170GZr4T11xzDSwWCw4cOIDbb78d5557Lpqbm8Pykul0OmF98IfUz5IidT2LFKmfYbAx5uO9LeZ1yFdzVqxYIev4oaBWq3H77bfjvvvuQ11dXUivDbQ2qJ544ok5o2fPPfdc/OY3v0FfXx8uvvjisCfvS0JCQsAfnlQuv/xyPProo3jkkUfwwx/+EPn5+cIXVbzZIIQIbjfKzp07odPp8PTTT+NLX/qS8PhHH32Erq6uoIJeUFCA4eFhDA0NITMzEwDgdDrxxhtvzPkarVaL3bt3IykpCW+88QYqKyslCzoAfOMb38DQ0BB27949a9dPKSkpQXFxMaqrq/HrX/864Hgcx83alNXU1ODQoUPIzc2VPC8pfO5zn0NOTg727duHvLw86HQ6fOUrX5nz+Xl5efjWt76Ft99+Gx9++GHI13vsscegUCjwr3/9S7DQKL29vYKFeM899wQc59xzz8Wrr76KoqIiJCcnz/k8juOgUqm8XIE2mw1PPfVUyHMHgPLyclx77bWYnJzEBRdcEHB+v/71r5GamhpwgZL6u5ADpVKJK664Avfffz8++OADHD16FA8//PCcz4+Pj8dZZ50Fp9OJL37xi6ivr5f12Isi9bOcb6R+hsDcXpL5fG+LdR2SQ3PCYWBgwK9HhB4nSDFmxQRaG1T/+7//i6amJuzduxc8z+Pjjz9GaWkpLr30UuzatQvXX389rrnmGhw9ehSnnnoq4uPjMTAwgIMHD2L9+vX4xje+Ef47lYG77roLO3bswC9+8Qs8+uij+OxnPwuNRoOvfOUr+NGPfgS73Y4HH3wQ4+PjXq9LTk7GTTfdhF/+8pf46le/ii9/+cvo6enBz372M0ku90suuQS33XYbLr30Uvzwhz+E3W7H73//e8GlTLntttvQ29uL008/HcuXL8fExAQeeOABqNVq7N69O6T3umnTJkkW38MPP4yzzjoLZ555Jq6++mrk5ORgbGwMDQ0NOH78OJ577jkAM4vAL37xC9x+++3YvXs3mpqacMcdd2DFihWS4xqkolQqceWVV+Lee+8VdvZioZ2cnMTevXtx2WWXYfXq1UhISMCRI0fw+uuve+3w77jjDtxxxx14++2357x/JpMJL774Is4880ycf/75fp9z33334a9//SvuvPPOgGeJd9xxB/bv34+Kigp8+9vfRklJCex2Ozo7O/Hqq6/ioYcewvLly3HOOefg3nvvxWWXXYbrr78eJpMJ99xzj1/PllSk1Ff47ne/i+effx6nnnoqvve972HDhg3geR7d3d1488038YMf/AA7duyQ/LuQi2uvvRZ33XUXLrvsMuj1+ll5yl/72teg1+uxa9cuZGdnY3BwEHfeeSeMRqMQd9LV1YWioiJcddVVstSakPpZzjdSP0MAWL9+Pd577z28/PLLyM7ORkJCAkpKSub1vS3mdShUBgcH8c9//nPW4wUFBcIGwePx+H0O3aieeeaZWL58Ob7whS9g9erV4HkeVVVV+N3vfgeDwYDvfOc7Ic9rzt/DbbfdRoqLi4lGoyGpqanktNNOIx999JFXlN3jjz9OduzYQeLj44lerydFRUXkyiuvJEePHhWeM1faiBzMlbZG+fKXv0xUKhVpbW0lhBDy8ssvk40bNxKdTkdycnLID3/4Q/Laa6/NisLkeZ7ceeedJDc3l2g0GrJhwwby8ssvk927dweNcidkJhVn06ZNRK/Xk8LCQvLHP/5xVpT7K6+8Qs466yySk5NDNBoNycjIIGeffTb54IMPgr5vKVGW/qJLCSGkurpaSAtTq9UkKyuLnHbaaeShhx4SnuNwOMhNN91EcnJyiE6nI1u2bCEvvPDCrAj+QPcfALn99tuDvhdCCGlubhYiPffv3+/1N7vdTv73f/+XbNiwgSQmJhK9Xk9KSkrI7bffTiwWi/A8en9936+Y+++/nwAgL7zwwpzPeeihh7zShwLd65GREfLtb3+brFixgqjVapKSkkLKysrIT37yE2I2m4XnPf7446SkpIRotVpSWFhI7rzzTvLYY48RAAEzMcTva2RkJODz/EU5m81m8tOf/pSUlJQQjUZDjEYjWb9+Pfne977nlVoq9Xcx1295rsyOuaioqCAAyOWXXz7rb08++STZu3cvyczMJBqNhixbtoxcfPHFpKamRngO/d75i+r2Rer6I+WzDPR9DxTl7vvZzRUB7W+uUj/DqqoqsmvXLhIXFzcrGyfS9zYXS20dkkp+fr6wXvn+o9/Jq666as7n0Lk/++yz5LLLLiPFxcXEYDAQtVpN8vLyyBVXXEFOnDgRdB6hrA0cIfMU6sdgMBgMBiNqLKpa7gwGg8FgMPzDBJ3BYDAYjCUAE3QGg8FgMJYATNAZDAaDwVgCMEFnMBgMBmMJwASdwWAwGIwlABN0BoPBYDCWAEzQGQwGg8FYAjBBZzAYDAZjCcAEncFgMBiMJQATdAaDwWAwlgBM0BkMBoPBWAIwQWcwGAwGYwmgivUEGAwGgyEfHo8HLpcr1tNYEKjVaiiVylhPY95ggs5gMBhLAEIIBgcHMTExEeupLCiSkpKQlZUFjuNiPZWowwSdwWAwlgBUzDMyMhAXF3dSCFggCCGwWq0YHh4GAGRnZ8d4RtGHCTqDwWAscjwejyDmqampsZ7OgkGv1wMAhoeHkZGRseTd7ywojsFgMBY59Mw8Li4uxjNZeNB7cjLEFTBBZzAYjCXCye5m98fJdE+YoDMYDAaDsQRggs5gMBgMxhKACTqDwWAw5h1CCM444wyceeaZs/725z//GUajEd3d3TGY2eKFCTqDwWAw5h2O47Bv3z58/PHHePjhh4XHOzo6cPPNN+OBBx5AXl5eDGe4+GCCzmAwGAyB9hEz3m0aRseoJerXys3NxQMPPICbbroJHR0dIITguuuuw+mnn46rr7466tdfarA8dAaDwWBgwurEt/9RhQMtI8Jjpxan4w9f2QxjnDpq173qqqvw73//G9dccw2+9KUvoa6uDnV1dVG73lKGWegMBoPBwLf/UYUPW0e9HvuwdRQ3/qMy6tf+y1/+ghMnTuC73/0uHn74YWRkZET9mksRJugMBoNxktM+YsaBlhF4CPF63EMIDrSMRN39npGRgeuvvx6lpaW44IILonqtpQwTdAaDwTjJ6RqzBvx7pyn65+kqlQoqFTsFjgQm6AwGg3GSk58SuGRsQWr8PM2EEQlM0BkMBuMkpzDdgFOL06H0KZOq5DicWpyOFWlM0BcDTNAZDAaDgT98ZTN2rUzzemzXyjT84SubYzQjRqhwhPhEQTAYDAZjUWG329HR0YEVK1ZAp9NFNFbHqAWdJgsKUuOXhGUu571Z6LAIBAaDwWAIrEhbGkJ+MsJc7gwGg8FgLAGYoDMYDAaDsQRggs5gMBgMxhKAnaEzGFGEEAKPxwO73Q6VSgWlUgmVSgXOJz2IwWAwIoUJOoMRJQghcLlccLvdcDqdcDqd4DgOHMcJVbGYwDMYDLlggs5gRAGPxwOXywVCCDiOg1KphEKhACEEPM/D5XIJAk+fn5CQAKVSCaVSyQSewWCEDBN0BkNGCCFwu91wu90AAI7jIC71QMVd/Pzp6WlUVlaioqICHMdBoVB4WfBM4BkMhhSYoDMYMkEtb57nAQAKxUzMaaDaTVTgCSFCYwqe5+F0OuFwOJjAMxgMybAodwYjQmjgm9PphMfjEURYKlSc6fk6PVenAk4IgdPphMViwfT0NKampmCxWOBwOOB2uwNuGBiMhYzH40FFRQW+9KUveT0+OTmJ3Nxc/PSnP43RzBYnTNAZjAiggW9OpxOEECgUirCsZ3+iHEjgHQ4HrFYrE3jGokapVOLJJ5/E66+/jqefflp4/MYbb0RKSgpuu+22GM5u8cEEncEIE4/Hg4mJCbhcLsEqD0fMpb5GLPBqtdor0M5X4K1WKxwOBzweDxN4xoKmuLgYd955J2688Ub09/fjxRdfxDPPPIMnn3wSGo0GZWVl+N3vfic8/4tf/CJUKhWmpqYAAIODg+A4Dk1NTbF6CwsGJugMRojQwDebzYYDBw6A5/mIzrR9A+dCeZ1CofAr8Ha7HRaLBVNTU4LA0yMBJvCMgIy2Ai37AVPbvF3yxhtvxMaNG3HllVfi+uuvx2233YZNmzYBAPbs2YP33nsPwMxv74MPPkBycjIOHjwIAHj33XeRlZWFkpKSeZvvQoUFxTEYIUBd7B6PJypjR7oxoK+nrnn6z263C89RKBRQq9WCCz9czwJjiWEdA57/KtD29qePFZ0OXPQYoE+O6qU5jsODDz6I0tJSrF+/Hrfccovwtz179uCxxx4Dz/Oora2FUqnE//zP/+C9997D2Wefjffeew+7d++O6vwWC8xCZzAk4vF4hHNqcfqZFIs3kBUeLTH1teBVKpWXBW82mzE5OYne3l6YzWZmwZ/sPP9VoP0978fa3wP+ed28XP7xxx9HXFwcOjo60NvbKzx+6qmnCqmd77//Pnbv3o29e/fi/fffBwAm6CKYoDMYQZgr8I0KsRQBDJa6JnWcSPAn8ABw/PhxWK1WmM1mTE9PY3p6mrnoTzZGW2csc+LjeSKemcej7H4/dOgQ7rvvPrz44osoLy/HddddJ3zvjEYjNm3ahPfeew/vv/8+9uzZg1NOOQVVVVVoaWlBc3Mz9uzZE9X5LRaYoDMYAaA54eJCMeI0M0A+IZ5v4RSn19Eoeo7jBE8ETZNjAn8SMN4R+O9j7VG7tM1mw1VXXYWvf/3rOOOMM/Doo4/iyJEjePjhh4Xn7NmzB++++y4OHDiAPXv2ICkpCWvWrMEvf/lLZGRkoLS0NGrzW0wwQWcw/CDOLadBb/7OmsMNaPMdYyHgr4gNFXgaZDc9PQ2z2QybzSbcGybwS4DkFYH/nlIYtUvfcsst4Hked911FwAgLy8Pv/vd7/DDH/4QnZ2dAGYE/fXXXwfHcVizZo3w2NNPP83c7SKYoDMYPvi62MVWuS8cxwmV4eS4bqyYKw+eCjyNouc4Tojwp1H0VOBplTwm8IuQtJUzAXCc0vtxTjnzeGpRVC77/vvv409/+hOeeOIJxMfHC49/7WtfQ0VFheB6P/XUUwEAu3fvFn6Lu3fvhsfjYYIugkW5MxgiaPlWj8cjKfpbDut6oVjowaAbG3FJW5rCR3PxffPkaRQ9YxFw0WMzAXDiKPfCPTOPR4ndu3cLx1m+vPHGG8L/NxqNs573xS9+kW0efWCCzmDgUxe72+0Gz/OSU7mkuNylbgoW2+IkVeB969AzgV+g6JOBK/41EwA31j7jZo+SZc6IDkzQGSc9vrnloeRlSz1DlzLeYhN0X+YSeHp8AcBvoxkm8AuM1CIm5IsUJuiMkxratzwUq1wMzeuOhMXicg+VYALPcZzwv4mJiUzgGYwIYYLOOCkR9y2PpKkKgKBBcdPT0+jo6IDBYEBKSgri4uL8XisWFvp8bib8CfzIyAgmJyeFyGVfFz1NpWMwGMFhgs446eB5Hm63G5988gkKCgqQlpYWtmgEeh0hBL29vWhsbER6ejpGR0fR1tYGlUqF5ORk4Z9Opwv3rSxqxAKvUqlACBGCEqnlTgvh+KbSMRiM2TBBZ5w0iAWD9hiXo7GKPwvd7Xajrq4O4+Pj2Lx5MxISEgDMbCampqYwPj6OgYEBNDU1QavVAgCGh4eRkZEh/PfJgNgrIS6nS/82l8CLo+iZwH/KYo/DiAYn0z1hgs44KRC72IFPrUM5zr99x5icnER1dTX0ej0qKiqg0WiEoDClUilY5sCM8NPn9/f3o6WlBXFxcV4WvFqtjmiOixV/9fL9CbxvkN3JKPD0O2K1WqHX62M8m4WF1WoFgJPid8QEnbHkEeeWi89w5QhoE49BCEF3dzeam5tRWFiIwsLCoJsGlUqF1NRUcByHDRs2QKVSYWJiAuPj4+jo6EBdXR0MBoMg7klJSUIN9qWA1A5z4i5y9HXAp6V5HQ7HSS3wSqUSSUlJGB4eBoA54zROJgghsFqtGB4eRlJSkpf3Z6mydFYGBsOHYLnlclV5o5HbdXV1mJiYQFlZGVJSUkIagwq/Wq1Geno60tPTAQAOhwPj4+MYHx9Hc3MzHA4HEhMTBYGn0eEnG4EE3uFwBEyTW6pCl5WVBQCCqDNmSEpKEu7NUocJOmNJIiW3XC4L3WKxoKGhAQaDAbt27YJGowl7zr5otVpkZWUJC5LNZhMEvr+/H263G0ajURD4hISERZX2FWkPeIpY4H17wfsKPD1/p+1kl4rAcxyH7OxsZGRkwOVyxXo6CwL6WZ8sMEFnLDmoGzZYbnmkZ+hULFpaWlBcXIyCgoKoRMuL0ev10Ov1WLZsmeBSpALf09MDnueRlJQkCLzBYFgyghUK4vr7vgJvt9uF5xBCoFQqodfrhRz4xX6/6KaGcfLBBJ2xZKAudhrFHmxxjsTl7nQ6UVtbC6fTiaKiIqxYEaRblQRC3VxwHIf4+HjEx8dj+fLlIITAbDYLAt/R0QGO47wC7Bba2apcFnow5hL4wcFB9Pf3Y+PGjcIZPO0Vv1QEnnHywASdsSQIp3xruC738fFxVFdXw2g0IjExEXFxcUFfE8wbIJfbOSEhAQkJCcjLywPP85iensb4+DhGRkbQ2to6Kwf+ZEqRE0MF3jcP3teC982BZwLPWMgwQWcsesRWeaBWp76E6nInhKC9vR3t7e1YtWoV8vLy8Mknn8iS5ypHCp0vCoUCRqMRRqMRBQUF8Hg8AXPg09PT513g58tCD3b9uSx4nucFgVcoFLOC7JjAMxYSTNAZi5a5csulEorL3eFwoKamBjabDdu3b4fRaAQgT2AdJdoFMPzlwI+Pj6O2thZ9fX1obm6OSQ58rAXdXxDhXALv8Xjg8Xhgt9uZwDMWHEzQGYsSmltOBTmcyG6pVrHJZEJNTQ2Sk5OxefNmrzxwuSzrWIgAzYEHgE2bNkGhUMzKgU9ISBDE3Wg0yp4DH+sqXqHkwftrNEMFnubBUxe9uA49E3jGfMEEnbGooItoT08P0tLSoFarw14wg1nXhBC0trais7MTq1evxvLly2ddS05XeazFLVAOfFNT05LMgQ/X5S+1F7y/OvRM4BnRggk6Y9EgDnyrqanBqaeeGnbONxDY5W6321FdXQ2n04mdO3cKtdj9jbFYLfRgzFcOfKxd7nIFJEoVeHEd+sVUM4Cx8GGCzlgUiMu30rNLOcu2ihkZGUFNTQ3S09NRVlYW0M0sp6Av9Pap0ciBj7VXItLmPHPBBJ4RC5igMxY0c5VvVSgUEZdt9RVRnufR0tKC7u5urFmzBjk5OSGPEQmxFrdQkDMHfilY6MEIJPC1tbXIzMxEamrqrDK1TOAZocAEnbFgCZRbLkcddo7jhLFtNhuqq6vhdrtRXl4Og8EgeYyl6nIPhXBy4PV6/YJJW5tvxAIvznsXd5LjOI4JPCMkmKAzFiQ0t3yu8q1yutyHh4cFK6m0tDSkQK9QBD2QeMTK5R4tpObAKxQK6PV6OByOmBS5ifWGgs6BHiOJLXh6zETrsvsKPI2iZzAoTNAZCwrf3PK58nrlcLkDEARm3bp1yM7ODvn1J6vLPVTm6gPf1taG6elpfPjhh4iPj/dqEzsfOfALQdDpplUMPV+niAXenwUvjqJnnLwwQWcsGHxzywOl+EQqpFarFb29vfB4PKioqEB8fHxY4zCXe3jQHPjR0VGoVCrk5eUJOfDt7e2wWCxRz4EHFo6gB5uDFIFXKBSzguxi/d4Y8wsTdEbMES9OUsu3RmKhDw4Ooq6uDgaDAVqtNmwxB5iFHin0PccqB36hCHqoZ+NSBf5k6QXPmIEJOiOm+Aa+SS28EY6gezweNDY2Ci52u92O8fHxsOZNkSLoHo8HnZ2dglWq1+v9jnOy4u+9z1cO/FylX+cTOeYgFnj6faRthGkVOybwSx8m6IyY4ZtbHmod9lAsWrPZjOrqaigUClRUVCAuLg5dXV2yp775u25VVZWwoLa0tECr1SI5ORkpKSlITk6GRqNZckFxUpH6nqPVB36hWOhyzkFcgx5gAn8ywQSdMe/MlVseCqFY6P39/aivr0dubi5WrVolWENyRMoHEuKBgQHU1dUhNzcXBQUFAGbeOz0r7urqQn19PQwGAxwOB6amppCcnByVs+KFTKifvZw58AtB0KPtJfAn8PSfw+GA0+mEyWSCXq9HUlISE/hFzMm1cjBiTjh9y/0hRdDdbjcaGhowPDyMjRs3IiMjw+vvcljF/vLheZ5HQ0MDBgYGsHHjRqSnpwvBfkqlEqmpqUJTFKfTifHxcTQ3N2NgYADd3d3CWXFKSgoSExPnxSUcK++AXAGF4eTA0+vHOg8+WtXq5sJfJ7mBgQHhOEhswavVatZJbhHBBJ0xb3g8HvT19QEAMjIyIlocggn69PQ0qquroVarsWvXLuh0ulnPkas4jViUrFYrqqqqAEBw7dPn+UOj0SAzMxM9PT3Izc1FYmIixsfHMTY2hr6+vrBdyaHMP9bIPQepOfDJycmwWq0R9QOIFPrdieU5Pv0Oq1QqqNVqLwteXPSGCfzChwk6I+qIc8tHR0ehVCqRmZkZ0ZhzWdeEEPT19aGhoQH5+flYuXLlnIul3C73oaEh1NbWYtmyZVi9enVIizQdx/es2GKxYGxsTHAlKxQKQdxTUlL8BtgtJubDMzBXDvz4+DisViumpqZgMpnmPQceWBiCDnhH2vuz4P0JvG8nOSbwsYcJOiOq8DwPt9stuNiVSqUsBWH8Wehutxv19fUwmUzYvHkz0tLSAo4hp8u9sbERPT09ERWo8feYwWCAwWAQXMnU0hwcHERzczO0Wq0QXEcD7BYb8y0CNNsgNTUVNpsNCQkJiI+Pn/cceABeNRdiCQ1M9cdcAs/zvCDwtModE/jYwgSdERXmyi1XKpVCFbhI8BX0qakpVFVVQafToaKiwq+LPdgY4eB2uzEyMiJcN5Kc9mCbC4VCgaSkJCQlJWHFihWCpTk2NuYVYEet92gKkVzEOrKfuppj1Qd+IVrowQgk8A6HA3a7nQl8jFjYv3bGosS3fKt4AZCrZCu1rgkh6OnpQVNTE1asWIGioiLJi0akFvrIyAi6u7uh1Wqxc+fOiBb5cBY6saUJfBpg5ytE1IKfrwC7UIl1UJrv9ecrBx5YOBZ6OMVtKL61I+jv0uPxwOPxzAqyE9ehj/X7XmowQWfIiji3XNwukiKXoCsUCrjdblRVVWFiYgJlZWVISUkJaYxwg+IIIWhtbUVnZ6cQOR9MzKXmREcCDbCj8Qk2m004f+/t7fUKsEtJSUF8fHzMF9SFYKEHuwfRyoEHPs1Bj/XnQDMw5EDcSQ7wFni32+3lrfOtQx/r+7DYYYLOkAWpueVyCbrL5cLAwACSkpKwa9eusM6Ow7HQHQ4Hqqur4XA4sHPnTphMJknV5oJdJxqFZfR6PXJycpCTk+M3V5sG2AGA3W6XdEwRDRaahR4IOXPg6fUXgtckEgs9GHMJvNvthsvl8hJ4cR36hXBfFhtM0BkRE0pueaSCTghBV1cXhoaGYDQaUVZWFrYghBrlbjKZUF1djdTUVGzZsgUqlQpjY2OLojmLv1ztqakpjI2NAQAqKyuh0+m8hGg+AuwWg4UeiEhy4IHoCmkoBAqKk5u5BN5ms+HAgQPYsWOHYLUzgQ8NJuiMiKDlJKVWfItE0J1OJ+rq6jA1NYXs7OyIK1lJtYoJIWhvb0d7eztKSkqQm5srXHexNmcRB9h1dnZi+/btwlmxOMCOnr8nJSXJ5pL1ZTFZ6MEIJQeeVgWMtZuZCmq0Pt9giF3tbrdbSBmkFjwAv2VqmcDPhgk6Iyyoi51GsUuNYA1X0MfHx1FdXY3ExERUVFSgu7sbNpstnKkLSDlDdzqdqKmpgcViwfbt22E0GmeNsRgs9GAolUqkpaUJqX40wG5sbEwIsKOBYCkpKREFgomJtYUe7SptgXLge3p6MD09DY7j0NzcPO858BT6G4i1QHo8Hi+h9rXgxZ3kPv74Yzz77LPYt29fLKe84GCCzgiZSMq3hirohBB0dHSgra0NxcXFyM/PF3b0kYpBMJc73UQYjUZUVFT4XWjlFPRYi5sYcYAddYfSc2JxgB214CMJsFtKFnowfDMThoeH0dzcDADzngNPWSiCPtfxgz8XfV9fH+rr6+d7igseJuiMkKBW+Xw0VXE4HKitrfVrHcsRXBeo2lxXVxdaWlq8NhGhjBEOsRT0QNfmOA5xcXGIi4ubFWBnMpnQ1tYW8Jw43OvOB7Gu5U5TuVatWgVgfnPgKeKNeSyhFnowOI6D1WqNqObDUoUJOkMSvrnl0WyqAswEoNXU1CApKcmvdSyHoPsbw+Vyoba2FlNTU9i6davgKp2LpeJyDwV/gWDUjUzPiWmAHbXgA7mRTyYL3d/1xUI6nznwlIWSOidV0AHAYrEwQfcDE3RGUHiex8jICABEXJwkmBATQtDW1oaOjo5ZAWhi5CrbSs/nOI7D5OQkqqqqEB8fj4qKCklR3lLnMR956LFCXF8emDknpi1iOzo6UFdX5+VGFgfYxfo9x1rQg53hRzMHXjyHWAXEiQlV0A0GQ5RntPhggs6YE3H51r6+PiiVSiQlJUU0ZiBBt9vtqKmpgd1ux44dO5CYmBjWOFKhCx/P8+jr60NTUxMKCwtRWFg4b9Xm5B5nIaBSqbwC7MRu5MbGRjidTsHKdLvdMT9qiLWgh1JyNVgOPM1ekJIDH84cokkogm61Wpmg+4EJOsMvvoFv0WyqAgCjo6OoqanxyvEOZ5xQ5wIAtbW1GB8fx5YtW4RgJaksBZd7tK8tdiOLA+zGxsZgNpvR0tLi1e1sPivYxVrQIyksE2kOPGU+c9ADEYqgm81m5nL3AxN0xizE5VvpWXm0mqrwPI/W1lZ0dXWhtLQUOTk5khbYcMu2irFYLABmLMiKigpotdqQx2Au99DwDbA7duwYkpOToVQq/QbYpaSkRLWCXawFXc60uVBz4JOTk6HVaheUhS51HhaLJeTN98kAE3SGQKDyrQqFQrDWI4EKOu2tXF1dDbfbjZ07dyIhISGkcSIRwL6+Ppw4cQIAsGnTprDEHFgaFnqsiYuLQ1ZWFvLz870C7Pr7+4UAO3GLWDnztGMt6NEs/RosB/7EiROIj4+HTqcTNvHznQMvJpSzfHaG7h8m6AwAwXPL5WyqAszk39bV1SEzMxOlpaUhB+WEOx+Px4MTJ05geHgYGzduxPHjx0MeQ4wUQfd4PGhubobdbkdKSgpSUlL8biBOBgs9GOIAu8LCQiHAbmxsbFaAHW0RG0lAV6wFfT6tY98ceJfLhYmJCfT19cHpdOKDDz6Y9xx4MewMPXKYoDMk5ZbLdYZOqampwdq1a7Fs2bKwXh+Oy91sNqOqqgoqlQq7du0SRDUSIQ0m6FarFVVVVQCApKQk9PX1oaGhAfHx8YIoJSUlLamguFAIJqiBAuwaGhqEADtqwYeaxrUQBD1W11er1UhPTxeqPa5Zsyb8HHjeDYADFJ/+nbOaoOg/BoXNBF6fCn7ZVpC4uTsiejweyR4ClrbmHyboJzGh5JbL4XIXi9u2bdsiipgP1eU+MDCAuro65ObmYtWqVV6LfiQblUBCPDIygpqaGmRnZ2PlypWCBeJyuYSFs6WlBXa7HWq1Gnq9HpOTk7KVVV0shCJo/gLsaIvY7u5uEEK8zt+DRXnHWtAXQrc1upEPJwdeaRuFsucjKEabAE4BT8YGePLKoTAPQX30YSim+v57FQK+bT9cW78OPnWl33l4PB7J8RJmszmkI7qTBSboJyn0zIyKWbDCEpFa6IODg6irq0N2djampqYiDnSS6nLneR6NjY3o7+/Hhg0bhF7h4nHkttDF/dKpF8LtdgsbIrVajYyMDKGXus1mE6zN6upqABDKqqakpECv1y/ZM/ZI7z0NsKNpXDTKe3R01CvAjlrwvt+7WAvqQghIm2sOgXLgu7u7oXJOoWT6IIz8OFTGLGhUSqjbXodisguKqT4opvvBp60COAVAeChMLVDVPQvnqT+eeczPPEJxuTMLfTZM0E8yxLnl1DqJZlMVj8eDpqYm9Pf3Y926dcjKyhJqgUeCFBe12CNQUVGBuLi4sMYJZR5UlG02m+RAP71eD4PBAIVCgaKiIkxPT2NsbAwjIyNoaWmBRqMRxH2+2prOJ3JtVjiOQ2JiIhITE5Gfny9EeY+NjaGvrw+NjY3Q6/Ve5++xttBj6XIXzyGYkPrLgXedeBWKyXGMq7NgG3PMbLC0BiS3H4TOPQ1P6kpwVLg5BfiEZVCMt4Ob7AFJyp91DalR7oQQFhQ3B0zQTyJ8A99CKfcYjsvdYrEIFqdYUOVw3wfbYAwPD6OmpgbLli3D6tWr51woIk1/Ewv65OQkKisrhWYuoQQU0XHEokRTj2hQmG9bUzmCwpYyvlHeNAhsfHxcaIQCAF1dXUhPT4/JvYxl21JKOF4CjuMQ5xiCIikd+qRcEMLD4XDAarXBMWEHbxnFlDsRmjgHtDotdDodVAolOPAA7/+3z0q/Rg4T9JMEf7nloRCqy72/vx/19fVYvnw5SkpKvBaMaNVhB2beZ3NzM3p6erBu3TpkZ2cHHSdSC53neXR3d6OpqQkrV65EQUFBWFaXv3kolUqvyGSn04mxsTGMjY2hoaEBLpdLCApLSUkJq/RnLJnPQEAaBJaeng5g5qjj0KFDcLlcXvdS3CI22vdyIVjoYReW0SaA8zhBAHCcAjqdHjqdHgpXNhCvg9btgkWphMViwfjYGBJcIyCJOZiyqZHscMzK9AhV0NkZ+myYoC9xAuWWh4JUEfZ4PGhoaMDQ0BA2btwonBOHM1Yg/LnK7XY7qqqq4Ha7UV5eLsklF6nLned5eDwetLa2oqysDCkpc0fxyjEPjUbjFRRmtVqFoLDOzk4h7YsKfDSLsshFrASNHl2UlJRApVLNOiMG4NUiVkoZ1VBZKGfo4XgJ+Iy1UA5UgrMMg8SlAyDgpvpA4lLhKfoc1M0vIckxCOjiQDgLnHFp6M/9LHr7B3GiqUXI9KD1/aUKutPphNvtZoLuByboS5hI+pb7olQqg7rJxWlhFRUVc7bQlMtCFzdWGR0dRXV1NTIyMrBmzRrJC1QkLneLxYKamhoQQlBRURFQPKNRKU58rpmbmwue54UzY1oZTHxmnJycPOcxQKxS5mJdxx349OjJ94xYHMtAy6jS+zhXLYFw5rAQBD2cgjJ8WinchWdA2X0QClMTAIDoU+BeeSb4ZWUgqSuh7DoIztwPkpADLm8XstNKkA3/xx8cx2FgYACEkIA58GazGQCYy90PTNCXKB6PB3a7HQcOHPDKuQ6XYCJMK6/l5eWhuLg44CIlZx12j8eDjo4OdHZ2orS0FMuXLw95nHBEZWhoCLW1tcjMzITNZovYEpbD8qONOWg6oNvtFmqmt7W1wWazISEhQbDeI+2cJxexstDFgu6Lv1gGWmWN1hKIi4sTBD4pKSk8UVwALvewu61xHDyFp4HPXAdusmcm8C2pANDPxCzwaSXg00r8vtT3+MPhcODQoUPgeT5oDvx8C/qdd96JW2+9Fd/5zndw//33z8s1w4UJ+hLDN7fc6XTK2lTFNyrY7XbjxIkTGB0dxaZNm4QfqJSxIp0PABw7dgxOpzPk0rGUUF3uPM+jpaVFOKNPTExEf39/yNeNdB5SUKlUXoum3W4Xzt/7+vrA8zySk5OFfO5YpMctFAs9GEqlUtgIAd4WZltbG6xWq5cASQ2wWygWeiRzIPEZIPGzj9ZCQavVghCCoqIixMXFzZkD/8YbbyAvL0/ICok2R44cwV/+8hds2LAh6teSAyboSwie573ynZVKpWwlW+niJN7NT09Po6qqClqtNqjLWYwccxofHwcws9MvKysLu0RlKC53h8OB6upqYQNhMBhgs9m8XP+BrhOMaIubTqfDsmXLhLxis9mMsbExodOdRqPxOn+fr/S4hWihB8OfhUljGUIJsFsIZ+gLodsaNRbo2uIvB35oaAgfffQR/vCHP8But+O8887DaaedhtNOOw0bNmyQ/T2YzWZcfvnleOSRR/DLX/5S1rGjBRP0JUCg3HIpZ99SELu4FQoFenp60NTUhIKCAqxcuTKkRTESQSeEoL29He3t7QCA0tLSiOpNS3W5j4+Po6qqCikpKV7tXeXMoZ5PxK03Ozo6sHXrViGCnjbuMBgMXuVpo5FeFUsLXVxUKVK0Wi2ys7ORnZ3tVYRlbGxMCLATdzmjAXYLxeUerhgSQmCyuGB2uBGvUSLNoPF6P1anB4QQxGsD/0bFRogvNL6hsLAQL730Et5++23ceOON2LNnD95++23cfvvtuPXWW/HjH/84rPcwF9/85jdxzjnn4Iwzzoi6oPf398NqtUKr1cJgMCAuLi6sY1Im6IscXxe7b2653ILudDpx4sSJsPuH07HCEXSn04mamhpYLBZs374dhw8fjlgQgrm6CSHo6upCS0sLVq1ahby8PK/7S/+/lAIlwf4eS3FTKBReLmWn0ykIEj3TFKfHyZnSFUsLPRrXlhJgp1arkZycDIfDIcvvMxLCPUO3uTz4qH0cnSYrbE4eWrUCK1L1qChMgdU587e2UQsIgBUpepQXpiAr0b9IBRL0Wde12WA0GnHTTTfhpptugsvlgsPhCHn+gXjmmWdw7NgxHD16VNZx/fHuu+/i1ltvRUtLC8bGxrz+Fh8fj+npacljMUFfxIhzyzmO87vLlrPtKQAcPXoU8fHxYfcPp2OFKugTExOoqqpCYmIiKioqoFarZTl3DuRyd7vdqKurw8TExJy158WCHuk8FlJzFo1Gg8zMTGRmZvqtmQ7Ayz0/V0ZDMGJ9hj4fm4lAAXZOpxNNTU3o6enxahE7n13OwrXQj3VPon5gGtmJWmQbVbA4PTgxYIbLQ9A7YUfPuA1p8TMWe2XvNAannbi0bBmS42YHD4ZSH8O305parZa17WtPTw++853v4M0334x62uf09DSuu+46bNq0CXfccQcMBgOcTifsdjusVmvIazcT9EVIKLnlcnRJo1YqAGRmZmL16tURLYShCDq9dnNzM4qLi72KtsiZ/uaL2WxGZWUldDodKioq5jxPllPQFyq+NdN5nhcszsHBQTQ3N0On03mlx4WywC41Cz0Y4gC7kZERrFixAgqFYlY2Ar2fAbucyUA4Z+jTdjdaRyxIM2gEd3q8RomMRC2OdE3A6vSgNCsBSsXM/TXqVWgdseDEwDR2Fc2u1RCKl8BsNvst4ywXx44dw/DwMMrKyoTHPB4PDhw4gD/+8Y9wOByyfR40huWZZ56RJWaFCfoiI9Tc8khd7i6XC7W1tZicnIRKpcKyZcsiXgSlbjJcLhfq6uowOTmJbdu2CSU8KdEqUEM7s+Xn56O4uFhSsJscluZCstADoVAoYDQaYTQasWLFCr89y2nENy1PO5dgnAwWerA50GBEcTYCjfCur68XIrzFLWLlnHc4FrrDzcPp5mHUe2/c9GoFJm1uaJQKQcwBQMFx0KoUGJzy7xoPZVMR7Trup59+Ompra70eu+aaa7B69WrcfPPNsm6utFotzjrrLFRXV2Pbtm0Rj8cEfRHB87yQhibVPRWJoFM3d0JCAnbt2oWPPvpI1hS4QExOTqKqqkpw7/vbvcrtcqc5sH19fXNWufP3emBpW+jB8NeznKbHUUEStzSNj4/3G4cw3ywEQfcnpjqdblaAHT3uoJ4y8f2MNN0wnDN0g1YJg1aFKZsL6QmfHr1N2txI1KngcM/+fTs9BIl6/5ITStlXX5e73CQkJGDdunVej8XHxyM1NXXW45ESFxeH3NxcfOtb38KDDz6I7OxswRsWzjECE/RFAHWx0yj2UCq+hXOGTghBZ2cnWlpavNzccp7HzyXohBAhgr6wsBCFhYUBe7TL5XKnZWN5np+zM5s/6Nzmo3vcYsE34ttisQiC1N7e7tXSVI4NYrgsFEEP5gHyrQZIW8QODw8L3fjExx2hxrZ4bSoIATfdB85lBW/MB1T+x9KplVi3LAEH28bgmXLAoFXC7PDA5vJgV1Eyavqm0TNuQ7ZRBw7A0LQD8RolSjL8C/HJ1piFfvcaGxvxhz/8Aenp6di+fTu0Wi2USiVUKhWmp6fx+c9/Hi+//LLkcZmgL3AiLd8a6hm60+lEbW0tzGYztm/f7hUIJsd5PDB3IJrb7UZ9fT1MJpOkCHq5XO5msxltbW1IT08PqWwsfb1cLBVBF8NxHAwGAwwGA/Ly8sDzPCYnJzE2Nobe3l64XC7U19cjLS1NSI+br4CwhSDooRaWER93iAPsxOmGtEa61PspePwmOqE5dB+UQ7UA7wExZMK1+Wq4i8/2+7p1yxKgUnCoHzDD4nQjQavCjvwkrM4yYJlRh3eaTegyWUEApMSrcerKVOSl+A+eDEXQzWbzvLdOfe+992Qdj37vSktLsX//fgAQovUdDgecTicmJiaQk5MT0rhM0Bcw1CqPpKlKKC73sbExVFdXIykpSYgkFyNXkRqFQiGk2VHERWqklqqN1Kql1uPw8DDWrl0bctlYOgdAHgv9ZIA2j0lOTkZRUREOHDiAnJwc2Gw2NDc3C+lx4oIs0Sp6shAEPdI8dH8V7Oj5e2trq1e5X1rBTnw/hYIubit0+38MxXg7eH0yoFaBm+6D5uDdIJpEePI/M+vaCo7DmuwErMo0wO7yQKtSQK2cGXtNdgIKUuPQN2EHAZBj1AbMRQ/VQs/KygrhLi1cEhIScOqpp8o2HhP0BYhvbnm0m6qIi7X4y7WmyOVyVyqVcDqdwn/TOvAFBQUoKiqSvIBHssGgwX42mw3Lly8PS8wpUj4bnudht9sDuvKXooUeDI7jBOEGIKTHUYsTkPe8WMxCKLsq9xzUajUyMjKE+A8aYDc2NiaUUE1KShLuJ03L0vR8AMVEJ/iEbEA5s5Enaj0Uk71QnfinX0GnqBQcDH7EOk6jRHGGNNd4KIF5Vqt10bvcxTidTvT09MDlcoHjOKG4jFarRXx8fEjfDyboCwx6RqZSqaBQKCL+sQcTYYfDgZqaGthsNuzYsQOJiYlzPlculzsVYnGrVal14MWE2yltamoKlZWVMBgMyMjIiDhdJNg8rFYrKisrMT09Db1eL1hU4nzjWFuKscJ3E6PX65GTk4OcnByvgiz0vFir1Xrlv0eSfxxrC52WDI7mpiJQgF1nZ6fw/i39TdATAii87ydR66Eca4va/CihWujz7XKPFk6nE48++ij+9Kc/oaGhQXiceh+///3v45577pE8HhP0BYK4fOvBgwexffv2gOIqFaVSOcu9TTGZTKiurkZKSgo2b94c9KxNzqA4p9OJw4cPQ6lUBmy1GmycUK3a3t5eNDQ0CAF39fX1US0KQ9u6ZmVlYcOGDYJA0XzjxMREITgs1gFisWIuUfVXkIWmx3V1daG+vt4rX1tqQxRKrAVdztKzUvAXYDc6Ooq6ujqMO1VIdNhhmxyHWq2FSq2CWqUG57KBT10V9bmdbEFx1CPx/PPP44EHHsDNN9+MI0eOoLu7Gz/84Q/x61//GhqNBjfccENI4zJBXwD4Br7JVa51rrF4nkdbWxs6OzuxevVqLF++XHIKnByiQ5uC5OfnY9WqVWFbKKG43MXegM2bNwspVuG2T/Wdhy+EEHR0dKCtrQ2lpaVYtmwZnE6nV3qXuPvZ6OgoCCGora2NuPraYiKUe69UKpGamioESzocDsGdfOLEiVnuZIPBEPB7HWtBp+89Vm5/hUIhuHSX770WmokDUE/1w0GS4LA54XZMQsXxGEjcCsXoaFQDFhdS2tp8QD/7Dz74ABUVFbj22mtx9OhRZGVlYc+ePcjLy8OPf/xjVFZWorCwUPK4TNBjjLh8Kz0rj6ag2+12r45hobQcjdRC53kejY2N6O/vh8FgwOrVq8Mei85HiqBbrVZUVVWB47hZ3oBw3fa+iMdwu91CMZ7t27fDaDT6vYa4+1l/fz/6+vpgMBiE6mti9/x8Rn/PN+GKqlarRVZWFrKysrzcyWNjY+js7BTq0/ueF1NiLejzbaHPNQeFQgHok+E4/RfQfnAX9JPdiOM84BOTMV5wLsYzPoOxlhbY7fZZLWLl2ox4PB5JR180kHWpCPrExIQQ70A3NR6PB4WFhRgZGRHqDkhlaa4Qi4BA5VtVKpVsgi4W4ZGREdTU1CAjIyOslqORBKFZrVZUV1eDEILi4mIMDw+HNY4YKWJM33N2djZWr149awGSQ9DFY0otGesL3citWLFCqL5Grc+W/y6m0WqOEkvkcvX7cydPTU0JwWBNTU3CBokKUqw7ncXaQge8i8qMGFbjnYLfoKq1Byq4sa14Ofauy0PJfwPe/PUoT0pKEu5pMI9IIBZ62prc0M+8sLAQk5OTAIDNmzfjb3/7G2pqaoRaINnZ2SGNywQ9BgTLLY+Ghd7U1ITu7m6sWbMm5NxG8VjhiN/w8DBqa2uRnZ2NkpISjIyMyBZcN5cgEELQ2tqKzs5OrF27FsuWLZtzDDnuNc/zGBoaQm1tLXJzc7Fq1aqQFjffc3iVSuXVb1sc/d3d3Q2O47yCwyJpIhHrjUE0rq9QKJCUlISkpCQUFhZ6bZBo/IJOpwPP85iYmEBiYuK8CyvdUCwEC33c6sK9b7ejZcQKvToOPAEaqyZQZ+LxvdMKoVUpZvUot1gswj3t6OjwSklMTk4OKSMhlCj3pWCh0/d60UUX4eOPP0Z/fz+uuuoqvPjii0IN+Ysvvhh79uwJaVwm6POMlNxyOQXd4/Fgenoabrcb5eXlEf0QaDCbVHieR0tLC7q7u71EVc58dn/jOJ1OVFdXw2azBT1WkKtCW3d3N4aHh7F+/fqwcmSDLXzi6G9xc5SBgQE0NTUhLi7Oyz0fzWYecjJfwXi+GyS73Y729naYTCbU1taC53mv9DjarzyaxNpDAHxaQ/1AqwktI1bkJ+ug+m8uud3lQXXvNI50TeAzPg1VxAWDxBXsxsbGMDQ0hObmZq+MhOTk5IDeKqkWOj1aWeyCTtm0aRM2bdok/Pd//vMffPjhh1Aqldi0aVPITWiYoM8ToeSWyyXoQ0NDaGxsBMdxKC8vj3iRD8VCp2f1Lpdr1kZCLkH35y6n9eeNRiMqKiokRe5HMhen0ykU84h0wyRV3Hybo7hcLkxMTMBkMgm9y6krVEpwWKyJxdx0Oh2MRiOcTic2btwoBGqOjo6ira0NarXa6/xdjk5YviyEPHhqWNT2TUOj5AQxB2bKu3oIj5YRyyxB98X3O0kzEmj9+fr6esTHxwv31DcmRKqg2+12eDyeJSPowEwa7dDQEJxOJwoKCnDKKaeEPRYT9HmABr6Jg2Ci2SFN3GSkoKAAfX19slhsUsWPpmplZGT4LaUqp4VORVBcA37lypVebVYDEYmFTvPZOY5DaWlpRItMJKKmVqsF61Pcu9w3OIz+C7ePfTRYCN3WOI5DQkICEhISkJ+f71VOtbu7GydOnIDBYPASIzl+T+H2IZcTOgedWgl+js9Cowj9u+mbkUBLmYpjQsQd+dxut6R7arFYAGDJCPrk5CR+9rOf4bXXXkNzczNefPFFfOELX8B9990HnufxjW98IyQrnQl6FBHnlosXj2AEyh0PBo3oBoDy8nK43W6h4lakBDtvFp9bl5aWzll9TW6Xu8fjEWrAl5WVCWUwpRBuUBytbldYWIj+/v6IF2a5XP/+epfT4LC+vj40NDQIlhJ1z8eahdZtzbecqtPpFM6KGxsb4XK5ZpWnDec9LASXOw2K255vxPGeSUzb3UjQqUAIwbjVBb1aiU3LjRFfR6PReFWwEwfY1dbWwuVyoaOjAxaLJWCAndlsFr7jixn63bvjjjvwzjvv4NFHH8VFF10kbFQ2bdqEm2++GVdccQUT9IWAb+BbKMEv4VroAwMDqK+vR05ODkpKSqBQKGA2m2UNsJtL/MQV54KdW8vpcnc6nTh06BDUajXKy8tDDg4LVUhp6t3AwIBQ3W5gYGDB9kP3DQ6jxwMmk0kQJ57nMTAwgKysrFmtTaNJrEvdSk1b02g0yMzMRGZm5iwPSFdXl1cwWCj1AxaSy728MBn1g2Z81DaOEbMTBDOlW89em4E12fJbw74BdgcOHIDRaMT4+PisADvxPaXn59H6jj744IN48MEH0dnZCQBYu3YtbrvtNpx11llRud4//vEPPPPMM/jMZz4Dt9stiPfq1avR2toaciYSE/Qo4C+3PBSUSiVcLpfk53s8HkFk1q9fj8zMTK+xPB6PLDm3cwkxbeqSnJwsueKcHIJus9kwPDwcUYGaUArL0BarHo8H5eXlwo9PjuI08yWi4lrfNMDok08+wdTUlHA0I3bPR+Ps2JeFZqEHwp8HhAaD0foBOp3OKz1urvK0C8HlTtcotVKB63floaIwGU1DZig5DuuWJWBVRvQ3eHRTvWzZMsTHxwtepfHxceGeqtVqPPHEEygsLERycnLU5rR8+XL85je/wcqVKwEATz75JM4//3xUVlZi7dq1sl1HKLn7X48EMFO/glroNpsNDocj5OJSTNBlhud5OJ3OiDqkhZKHbjabUV1dDYVC4bePN10w5BJ08bzE1dBKSkqQm5sr6RqRCjqNnh8ZGUFycnJEBWqkutzHx8dRVVWF1NRUrF271uu8T6qVH+zezLfFSnO3FQoFSkpKoNfrZ7XipGfH1D0vpwAtFgs9EL7BYG63Wzgrbm9vD9jtbKG43Ol8lAoOG3MSsTEn8pLToUCPJulvSuxVogF2/f39UKvVeOaZZ9Dd3Y3NmzfjjDPOwOmnn45TTjlFtjP1L3zhC17//atf/QoPPvggDh8+LKugU84880w89dRT+OlPfwqe56HRaOB2u/Hkk09iw4YNTNAXAtSVFs0OacCn57h5eXkoLi72u9jSHwndiUeC2OXu2zfdaJR+zkYt2nAWVIfDgaqqKrhcLuTm5obkyfBHMDEmhKC7uxvNzc0BO9FJESf6HH+vj/XCDsxubSo+O6alVcW573Kldi0mCz0YKpVqVnlfcbczj8cjZCDEulId4F1YJlaIS177Q6lUIjc3Fw8++CBefvll/PKXv8SPf/xjvP322/jmN7+JW2+9FV/96lejMq/nnnsOFosF5eXlso8PADfffDMuvvhiTE9Pw2Kx4IknnkBzczP+85//4JVXXgl5PCboMiNHh7Rggu52u9HQ0IDh4eGgXcrEgh5JZyrgUwudpoYlJib67ZsuZRwg9MWEWskpKSkoKytDT08PHA5HSNf2N5e5xFgcbLd161bBNRbKGFKRKyhOTnzPji0WC8bGxmAymbxSu6TkGfsj1u93PgTVt9uZ+B6Oj48DAE6cOCFslAJlIHBTfVC1vgbOagKfXAT3ys8B2sis6YXi9gfmFnQxFosFSUlJuOSSS3DJJZd4vV4uamtrUV5eDrvdDoPBgH//+99Ys2aNrNegBlZZWRmef/55/PznP8fKlSvx+OOPY/v27Xj77bfD2kQwQY8CkS7OgQR9enoaVVVV0Gg02LVrV9AgMI7jZO+S9sknn6C4uFhyapi/cQDpgk4IQVdXF1paWrysZDnO4udyudNsAYVCISnYbqEGxcmFuJBIXl6eV2qXuPMZFfhQ6nzH0kKfTzHzvYe9vb3o7++HVqsVugCKc7WTk5OF34ey8z3o3v4p4JgWxlNXPwn72X8EScoPe048z8e8R4DH4xF+z8Ewm82zjhXl9jCUlJSgqqoKExMTeP7553HVVVfh/fffl1XUxXPetGkT/v3vf8syLhP0BYg/QSeEoLe3F42NjSgoKEBRUZHkxUiOLmkulwstLS3weDzYsWPHnNaqFMSCHgy32426ujpMTExg27ZtXmlWcli1/qxrmkc/V/13KWOESqwt9FCv7S+1i0Z+19fXC67l1NRUIUrZV7hjvYGJtcub4zhotVoUFRWhqKhIyEAYGxtDc3MzHA4HjEYjUhO0KPngZ4BjGm59GlycGlreDsVEJ7Qf3QP72X8Iew5Sm6JEk1A8dfNR9lWj0QhBcVu3bsWRI0fwwAMP4OGHH5ZlfLPZjLfeegsGgwEqlUroQRAXFweDwYC4uLiQyuaKYYK+APHNQ3e73YLrV9z6UyqRWuhTU1OCV4DWEI8E+kUNJujBGp3IZaGLi9PQIL9Qat7HWowXAhqNxqvzGXUtj4yMoLW1FRqNxss9Lz6mWUpn6KHgGxQnzkAAZrxE4+PjIE2vwTI1hqc9n8X+6U2wEi0K1SZcpvsQn+n9BJx1FCQutDVBPIeFcIYu1TiJRdlXQkjER3ti6uvrceGFFyIjIwOZmZmw2Wyw2WyCp0KtVsNqtaK8vBwvvvhiSGMzQV+AiC10KqZ6vR67du0Kq8pXuHntYq9AYWEhsrOzceDAgYgXQinHAP39/aivr0d+fj6Ki4v9Xk9OQffX8jTUMYI9J9IxFgv+3PM08rujowN1dXVITEwUvC1ypDCGQ6zzwINdn6bHKaez8OtjX8Y77o3QcC6oiRtVjmVocZ6Hn+j/haLBXiTkGsOKkVkoZ+ihdFqLj4+P2lxuvfVWnHXWWcjNzcX09DSeeeYZvPfee3j99ddlu0ZcXBxyc3ORkpKCU045BevXrxc2cdPT07Db7ZicnAyriRYT9Cggxxm62+1GV1cXmpubUVhYiMLCwqhHzYtxu904ceIERkdHsWXLFqSmpgq7VLnSffzdI3HP9I0bNwpf9LnGkKP1qdvtxqFDh0JueUqRI20t1tHO0cS3DKjD4RDqpgPAoUOHZkXPzwcLwUKXIqa1qlJ86LYjiTPDoJz5HScRB/rcCfib53R8ZcgKa+cHXulxUlMM5ch+iZRQBJ0GxUWLoaEhXHHFFRgYGIDRaMSGDRvw+uuv47Of/axs1ygtLcUzzzyDJ598EgcPHsT09DTOPfdc7N27N6Qql/5ggr4AoSld7e3tIZcy9Ueowkdd3RqNBhUVFUJAmJwpcP7mRAu38DzvN6feFzms2vHxcdjtdqxYsSLklqdyzgOI3ZnyfIuaVqtFdnY20tLSMDIygo0bN2JychLDw8NoaWmBVqsVzt6Tk5OjFrS1EARdyvVbrQY41EakuXsAngBQgAOPRIUdPdpVWLN5O7ScZ1aKoZRe5YvNQrfZbHOWlJaDxx57LGpjU1QqFcrLy1FeXo7BwUH86U9/wk9/+lPExcXhiiuuwKWXXhpWx0aACfqCg6aEAcCOHTtksVZCsdBpbnt+fj5Wrlzp9WMPJZgtGL6CbjKZUF1djfT0dL8NXaSMEQqEELS0tKCzsxMqlQolJSVhjQNIE/ShoSG0trYiISEBqampC+YcOZbQe5aQkICkpCShMYpv3/LExETBek9ISJBNgGIt6FJd/gatEkSbCFd8PtS2EXAeF4gmDg5FOuLiDNCpFNCoVF4xDFarVQhSpKVUxd3j6CZ9sQl6tF3u8wEhBB6PB6r/fma/+MUvcOutt+L3v/89vv/97+Pw4cN45plnwvp+MkGPAuEsEoQQdHZ2oqWlBUVFRWhpaZEtWEWKoHs8HjQ0NGBoaGjO3Hb6vuRKgeN5XvBEtLe3B2zoEmiMUKH90u12O9avX4+GhoaQxxATSNDFDWsKCwtht9vR0dGB+vp6QaioK3qpnKGHivj3olQqZxVmocJEmwyJ3fOhVtISE+tKbVLFdHtBEjISNBieNiAjNRVKBQeb0wOb1YlzVqVCo/Ieg1YAjI+PF3qVT05OYnx8HP39/WhqaoJer0dKSgqcTme03p5kQo1yX8yCTkVapVJheHgYra2tqK+vx/vvv4/Dhw9j3bp1uPjii8Menwn6AoBWXZuensa2bduQnJyM1tZWWZuqBBrLYrEIOdcVFRVzLpIcx8mSAgfMiLHD4UBlZSWmp6dDDkSj8wlVBCcnJ1FVVYWEhASUl5fDarVGLeXM5XKhpqYGFosFO3fuFM7mOY4ThMpkMqGnp0cofzkwMLDgWpxGCyn3XafTYdmyZUITD1o3fWhoCM3NzYIw0dK0objnY22h8zwPswt4+sMWtPaPIDkhHmduLkJJpncUt0GrwvdPL8Q9b7VjaNoBEECl5LA9PwmXbwseOCWuAEgb9NAgRafTibq6OhiNxqh4QaQQyhHefKStRQv6fWtra8O7776L1tZWHDlyBA6HA2effTbuuusur0A4lra2CBkfH0d1dbVQdY0u+pH2RBcTKKJ8cHAQdXV1Xh3ago0lh6ATQtDY2Aij0RhWtblw5iJueUqDDMNtnyrGn6DTOAS9Xo/y8nKoVCova0gsVDzPY2hoCI2NjUKL02jWUF9oSF24OI5DYmIiEhMTUVBQ4FU3vbW1FTabbZYwBRo71oLeN+HA/R/0YNShBA8OCkzhX1WD+MHpK/CFbau8nrsl14gHL12HjzsnMG13ozAtDhtyEqEMo1e5Wq1Geno60tPTMTIyglWrVsHtdmNsbAy9vb3geV7wgiQnJ8tW4ncuQnG5xyJtTS7o9+2+++7Dn//8ZxQWFuKGG27AOeecg8TERLjdboyOjsJgMITcNZLCBD0KSPnyi13NxcXFyM/Pn+V6DLcnui/+rGqe59HU1IS+vr5ZHdoCIUfVud7eXlgsFmRmZmLTpk1hLxZSxdhfy1NKNIrCDA8Po6amRqixH8yToFAoBKto69atcLlcgvV+4sSJWUVaFnsvaEqk9923brq4rWl3d7dQM4EKvO8iGWtB//cnjRhxJCIRVii5mflMeeLwwNstqCjNQ6rBe75GvRqfK527zHM48DwPvV6PhIQEwQtiNpu9agjQEr/0XspdiEaqoNP6BotV0Ol3LSsrCyUlJUhISMB9992HO+64Q8hBB2ZqOoyNjeGpp57C5ZdfHtI1mKDHANo73Gq1zulqltNC9x3LZrOhqqoKhBBJ0eS+Y4Vr0YrP6RMSEpCenh7RgipFjOdqeUqRI0KdjkEIQVtbGzo6OrBu3TpkZ2eHNAZFrVbPqqFuMpkwMjIyr1Hg0SZQs5pw0Ov1yMnJQU5Ojldb04GBAa9zY3rf5lXQ7RPgbGMghmxArcfktBm103HQwgXlf50vHMchgbdh0hOHj48ewdl7Ton6tHzP8TmOQ0JCAhISEoQgRVrit7u7e1YHPqPRGHGsj8fjkXzEtBQE/brrrsO5554r1L9wuVxwOp1wOBxwOBxwOp0YHR1ltdwXAyaTCTU1NUhOTg7oag6lhWowlEqlkEM+PDyM2tpaZGVlYfXq1SH/GMO10GltdI7jUFFRgfr6ellyyAONEajlKUUsxpF4CjweD6qqqjA1NYWdO3ciISEh5DH8bSzERVp8o8BbW1tht9tnyoP+V+DnSk9aqERrrr5tTcXnxrSsqlKpBMdxMBqN0btvjiloD94NVctrAO8CtIlwbroa7uzTQAiggPdnznEACOA2j8o/Fx9825b6Y64Sv+Pj42hoaIDL5fI65gjnPoaah75YBZ1Cm/VEAyboUWCu9po02llK73C5z9DdbjeamprQ3d2NtWvXYtmyZWGNFY6FPjIygpqaGq/a6NFsrCK15SkgT794t9uNkZERJCYmory8PGyXpBRPgW8UuM1mg8lkwtjYGDo7O70W4Gi4Rxcr4nNjYGaDWVNTA7vdjuPHjwtpXfSfLEGJhED3xk1QdX2AaYUBZqQh1ToC7aF7kbnNiWKlGvWePOiIG/SrZyUaaOFGWW70e5LT304o8Rm+JX5peVr6/aMBeNQLIiULIZQmTUtB0IHoHfcwQZ8H7HY7ampq4HA4sGPHDiQmBv+xyinoPM/DZDJBq9WivLw8oh9EKEIs3sT4biLkqvJGryNOqZPS8pRCXxeu231kZASDg4MwGAwoKysLO3gt3B+3Xq/H8uXLsXz5ciE9SewepXnvKSkpSExMXFDBdbFM04uLi4NGo0F2djYyMzMxNTWFsbExISiRdj2jQYnhuJUVw3WwdFfiIedF2O/aACdRIUM5jctV7+C82n/giqzT8Yv+DEyQOChBwBMOHAiuNBxFzrpbovCuvQlH0MWI0+Po98/3mEOn03ndR38eSalR7jQjZSkIOvXIyS3qTNCjzMjICGpra5GWloYtW7ZIPu+US9BHR0eF4ik7d+6M+LxVqsud5nrbbDa/Lmg5BZ3u8ENteQp4N4oJZdEWN3JJTU2FXq8PuigFO6+Xo3McTU8qKioSSqyOjY2htrYWhBDZcrjlItaFXWhfgaSkJCQlJQlpXdTqbGxsFNzKdGMUHx8vbd6jzbjd8mV86FkLHeeEhvOgx5OM33ouAGd+ETlbNuIBzcd4sVePGk8+MhST+GJyJ/Z++QZAGXrWR6hEKui++B5ziLMQaJEgcXla2mJXqsvdYrEAwJIQdCA6330m6FGAuoJbWlrQ3d2N0tJS5OTkhPQBRiroYus4OzsbZrNZluApKS53Wu2OpqT5u67cgk7T/6S2PPUdIxQxFbd03b59O4aGhuByuUJ/AyLEngK5fui0xGp2drZXDvfg4KCQw52amgqe52XzBoVCrAvpzHWvxV3PCCFexxodHR2SjzVqbek44lEgUWFFHDfz/YjjnBj2GPB3xyn4lioZOad/Bz/R2qAcbQIftxZ89jcBbn68KLSwTrS8Nr5ZCHa7Xdgo9fX1ged5JCUlwW63w+VyBf3uWywWKJXKsFO6Fhoffvgh1q5dK2tteiboUcBut+OTTz6B2+0O28UdSdoajaKn1rHFYsHU1FRYY/kSSIgJIejp6UFTUxNWrlyJgoKCgGfXcpyhA0BHRwe6urpCannqO4ZUcbFaraisrBTqMWu1WgwPD8sSKR9N/OVwU+ud53kcP35csN5TU1OjnnssnleskLJ54jhO6HomrrpGK9cFivpuU62EkzMjhTdjJpSdAwgPA+wYVGZjjI9DrkIBklQAd1JB9N+wD/Nd9lWn03ltMM1mM8bHxzE+Pi6UYRZnIfjGMdAqcQvp2CgS9u7di/feew8VFRWyjckEPQoolUokJydj5cqVYad0hGuhj42Nobq6GsnJydi8eTNUKhXsdrtsLSrncrmLz66lNJSRQ9Dp6/v6+sKqNAdI780OzGQoVFVVYdmyZV5FeORsfTpfqVQqlUqwQoeGhlBaWgqHwwGTyYT29nYh99hf3Xm5WKgWeiB8jzWcTqdgddKob9oURa9SgdMlwuWyQMPbAUJAOA4OlQH6hBRoFLGtox7LTmvi9Liuri6sX79e8LTRjZI4jkGv10e9jvudd96Jf/3rX2hsbIRer0dFRQXuuuuuiPo8BGL37t2wWq2yjskEPQqo1eqIvwRKpTIkN674TNc3slvunHZf8bNYLKisrIRarZZ8dk0j78OFVmIDgC1btoQl5gCEanHBzra7urrQ0tLit968XLns9FrzDcdx0Ov1yMjIQG5u7qz+5b5154NVYAv12rEilH7o3EQX1E0vg5voBElcDlfJF0BSiqDRaLxqBoibopDRcRhUCgx6spCqc0HLEdg4LexuJc4rzYCaG4h56dmFYO3yPA+1Wi0IeFFRkVccw7Fjx3DZZZdh1apV0Gq1OHLkCMrKymTrdUF5//338c1vfhPbtm2D2+3GT37yE3zuc58TNhdyc+qpp+KOO+5AWloaMjMzodfrodVqodFown5vTNCjgBw/0lDy0GkteLPZ7NdSjWYZ2aGhIdTW1mL58uVYtWpVSGfX4VroQ0NDqKmpQX5+Pjo6OiKODQhUcU7sedi2bZvf8y45BX0h4Nu/3LfuPK3ARoPEwk3xWiwWurL3E2j3/wicbVx4TH3iedhP/yU8BbuFx3ybong8HhhzB/Hbd7oxOE3AEwKNkmBTthrnrzagsyW2ghpqIGg0oJ3HfO+DOI5h1apVOHDgAB555BG88MILOPPMM8FxHE4//XTccsstKCsrk2Uur7/+utd/79u3DxkZGTh27BhOPfVUWa5BcTqdeP3119HQ0IDPfOYzWLt2LRISEqBSqaBUKpGQkIBnnnkm5HGZoEeJSBd5qSJMA9ASEhLmLFQjR7lW8Vhut1sI+uvp6cG6detC7t8bTj47bXlKXXRZWVno7u6OSi12YCbHu7KyUmhaM5dwye1yn2+CiZpv3fnp6WmYTCZZ6s7H2kIPen3eA82Hd4Ozj4PEZ8wErBECzjoM7cG7YV2+E1D5/14olUrsWJWDpwuzcax7AiNTdqSqnUiGBR3NjXA6nWhra0NGRoZQ0nc+78dCsNDpbzfQxkKhUGDNmjXYtm0bGhoacODAARw7dgz79++Pap2FyclJAAh6fBgOhBB88YtfxMUXXwyXy4Xp6WlYrVY4HA5YrVZWy32pEUzQxcVTggWgUfGU43xWqVTCZrPhyJEjcLlc2LlzZ1hBf6E2RRG3PBUHGsoVLe8rpGNjY6iqqkJmZiZKS0sDLnxLzUIPhDg1iaZ4hVt3fiFb6CaLEx0mK1Js3Vg33gWiTfo0+pzjQHTJ4MyDUAzVgs/ZGvA6WpUCFYXeokAIwYEDB5CYmAiTyYS2trZ5iVsQsxAEna5xUtPWDAYDlEoltm/fju3bt0dtXoQQfP/738dnPvMZrFu3TvbxtVotfvjDH8o+LhP0BUogQXe5XELalJTiKUqlMuLyphSHw4Hh4WFkZmairKwsbHd3KEI8OTmJyspKGI1GoXMZRc5a7ID3Rmn16tXIzc0N6fWREmuRC5VgdedpYZHU1FS/7U0XmoXu5gn+8kEXXqrsgt05U2d9PX8Nbk96Azkqu+iZMzVaORJeHAi97rJlyxAfHz8rbqGurk6IW4hWUaBYBsWJ5wBIy4Wfz17o3/rWt1BTU4ODBw9G7RpWqxXvv/8+Dhw4AKPRiJtuugk8z6O/vx9ZWVlhWelM0KNEtFzuU1NTqKqqgl6vx65duyS5nOjuN5IfMA0M6+3tRXx8PDZs2BBxYxUpgu6v5Wk44wSCegs8Hg9OnDiB0dFRSRsl8eulftZzbapiGRQnF75158WFRVpaWmbVnZcr8yJcfD8LZe8n+MfrH+DZoRJo4EQS7HBBjePIx/fGzselifWYIHFYppjEZ8gRGAxp8GRuCPv6NA8cmB234FsUSNzSVK6OewvhDJ0WlZGylsxX2dcbb7wRL730Eg4cODArAFYuHA4HHn74Ydx5551QKpXQ6/W45ZZb0N/fj5tvvhlbtmzBzTffHPK4TNAXKL556IQQ9Pb2orGxEStWrEBRUZFkQaUi7vF4wnLjud1u1NbWYnJyEgUFBZiamorYsgomxOKWp5s3bxaKU4Q6jtS5OBwOfPLJJwAgOVKfcjK53EPBt7CIOAK8s7NT2EgNDg7GpO68WNCVHe+Ae/MneN70TSjhgZGbSSdSwQE3UaDGk4/eiVTEcTM97Z9XrcEtO5YjWx2esFKP2VwbbN+iQOKWpr4d9+YqqRqMheByD2UO0bbQCSG48cYb8e9//xvvvfceVqxYEZVrcByH48eP46GHHsILL7yAxMREXHDBBQBmGrds3rwZBw8eZIK+lBBb6G63W7Act2zZIuzipaJQKEI+s6bQ9DCdToeKigqMjo5iYmIi5HF8CRQUF6zlqRg5xJTnedTX1yMjIwNr164NeZE7mV3uoUALtNC63319fWhvb49Z3XlB0AkPzcd/gsnOY4rEQQfnp88Bh0kY4IECCSoeORoHXMo4tJC1eHgwE7eHeYxFP2cp79FfS1Pq+Whvb4fVavVyzyckJEgad6G43EPptBbq2hcK3/zmN/H3v/8dL774IhISEjA4OAgAMBqNspVJpt+5rq4uGAwGVFRU4KmnnhLWOI7joFQqw15jmaAvUKigm81mVFVVQa1Wo6KiIuzox3BS1/r7+1FfX4+CggKsXLlSKBMph6t0rnGktDyVMo5Uenp64HA4kJubizVr1oS1OEsRdJvNhhMnTkCtViMtLW1WH/OlaKEHQqFQICEhAWq1Gtu3bw9Ydz41NTUq5T7p4spN90Mx2Y1E1YxlbiIJiMNMu2E7NHBCBSV46LVqeDI3QgEgxe7GiUEzhqYdyEoMfW70OxvO5z5XWiGtXgdAUs3+hWChhyro+fn5UZvLgw8+CADYs2eP1+P79u3D1VdfLeu1qIcGmPktUM+Dy+VCZ2dn2N0wmaBHiUgXaJVKBZ7ncejQIeTn52PlypUR/fhCEXTq7u7v78fGjRuRkZEh/E2uFDhfIQ6l5WmgcaTC8zwaGhowNDQEvV6P9PT0iPqhBxJ0GjGfmpoKpVIpNKqg58m01CqwtC10X8Tv1V/deZPJNKvufCTdz3wRzrBVOoBTQKPkcbriGP7q+RxGSSISYYUdmhnrHFYY4z8VRqWCg8fFw+4KbzMZioUeDHFaobhm/9DQEJqbm706nok3kotN0K1Wa9Rd7tGG3u+tW7ciKysLd999N2w2G4AZMd+3bx8OHDiAW2+9NazxmaAvQDweD5qbmwFAyLeOFKlCTN3dPM+joqJilrs7nPzxueZDxwm15amYcNzdDocDlZWV4Hke5eXlqKysjOjHHGgOPT09aGxsRElJCbKysoRF1LfhBz0DHR0dRVZWliyNdBYDcwUI0rrztGsXtUBp9zNaXjWSuvPUQif6NExmleP3TUk46CmGC0qYoccYEqGGGzo4kcaZgfg84XXjVhdyk/XISQrPcxCJhR4IfzX7fTueJSYmIjU1FXa7XZ6+7xEQiqBHu/TrfEEIQXFxMb75zW/i//2//4fBwUGoVCqsWbMGfX19+OlPf4qvfOUrYY19cqwaiwiLxSK0AAUQkrgFQooQm0wmVFdXIyMjA6WlpX5/aHK73GmzE6VSGXIwWjjzmZiYQGVlpZdLP9z4Aoo/QRcH9ZWVlSE5OdmrlK9vH3NaIKinpwctLS3CopuamgqDwbAkXfJSN1HiuvPi8qqR1J2n16b39c+aq7Gf74YTBAQKqOAGDwWU8EDPuWGOW47eSTt0KgVsLh5xaiUu2ZINtTI8C1fu1qVz4RuYaLPZhM3R6OgoOI6Dw+EQLPj57mQWqoW+FFqnchwHp9OJc845BxUVFXjxxRfR2dmJpKQknHPOOSguLg57bCboUSKcBXhwcBB1dXXIyclBSUkJ9u/fL2sN9rnGIoSgvb0d7e3tfmuVi5HT5e52u3Ho0KGQW576jiNVjHt7e9HQ0IDi4mLk5+cLn5G/wjKh4CvoTqcTVVVVcLlcQlBfoPEVCgVSUlKgUCiwceNGcBwnCFZXV5fQrpO6nJdSo5RQfyf+yquGU3deLOgmixMf9HqgSczE2KQNCsJDqwA8hAPP6aGJ08KgVWJVhgEjZgc25MTh7LUZKMsLr38AvX4s3N16vR45OTnIyclBXV0dVCoVtFotBgYG0NTUBL1e7+Wej3ZaW6hR7otd0OkG5l//+hcSExNx9tlnzzqfr6qqQkpKCvLy8kIenwn6AoDneTQ1NaGvr8+rjKrcTVX8jeVyuVBTUzNnHXh/40RqoRNChH7Ia9euDbnlqRgpgi62lv1lCchpoU9PT+P48eNITEzEli1bBNc5bQITDEII4uLihEWXtus0mUzo7OzEiRMnotYoZb6RYxPhGyBmttrwScsA2rvGkdzZA62K8+pdTl3M3oLugsPNg1Mo4eZU0Ko4gOOgIARuD4FOrYCbJ/jfU/JQkBp5/jfgnYMeK+h3LS8vTzjaoA1RxHUD6HctGp4iqRY6LVq02AWdvtcHH3wQ5eXlOPvss2c951vf+hYuvPBCfP/73w95fCboMcZms6GqqgqEkFln1pH0RPfFn2U9NTWFyspKIX1CiuUXqctdnNMOIOxoTkqwM3Sn04nKykqhN72/FLhI087o62nTGHFWQKjj+CJu17ly5UqvRind3d1QKBSC5Z6amhr1cqFyI6dAHOmawC9fb8XApB0EQKJOg+u2ZSJHT2bVnadNdjiOQ2aCBnEaJSZtrpn6bwTgOMBDAAU380+pUCBeI5+1GisLXYyvdaxSqZCeno709HQAMy7u8fFxwVNEPUm+m6NICDXKfbEL+tGjR6HVamGz2eB2u9Hf3w+32w2dTgedTgelUgm73T5n3Y1gMEGPElIWquHhYdTW1iIrKwurV6+e9cWOZttT6n6eqwLbXFD3dDgRsuKc9q1bt+LgwYMRl6MNtMGgJWOTkpIClqmN1OUOzATa1dTURBzEGGwevo1SJicnMTY2hu7ubjQ0NAi53IvBepfFzc+7oew/iv7BYdz0fhIsbg5alRIcgCmbGw8c7Ef+haXYtbVQ6F1uMpnQ0NAAAKivr0daWhr2Fhnx79pRKDnA4eGh4mcE3ahXw+rkcWpxEtIT5AsgWwgR5sEqxdG6AdRTNDU1hbGxMfT19aGxsRFxcXFeTXnCcc+HUuxqMZ+h08/7iSeewEcffYSGhgYMDw/jk08+ASFEqBbX3t4OhUKBDRvCq0DIBD0G0E5l3d3dWLt27ZxWaigtVINBNwcej0dI1wpUgS3QOEDoC9Lg4CBqa2uRn5+P4uJiwfMQ6cI2l7uc5tAXFRVhxYoVAYUtEpe7x+NBe3u74AFITEwMaxw6j1AQW+9FRUVwOBxC5Dxtcyo+e5/vSmxSiGTDwU31Qff2rVCMNuFl82mw2k9DnIIHx8UDnAJ69UwA2zPHBrCrKMWrd7nD4cCHH34Io9GIkZERrHBPYEeaGkdGOAxbCNw8oFMroFFyKMk04PpdoZ9nBmIhuNxD+e0pFAokJSUhKSlJaMpD3fNNTU1wOp2Cez4lJUWye97j8UgKxON5flFb6PQ+n3/++di9ezd+85vfoKysDKWlpbBYLLDZbHC5XKioqMDFF1+MkpKSsK7DBH2esdvtqK6uFgKmAn1B5bbQ7XY7Pv74Y3Ach4qKirCqH4nLyEpJrSKEoLm5Gd3d3V7WKx1H7k5pPM+jubkZvb292LRpk+A+DGUMqdD2qoQQaLXaiMScEonVqtVqvaz3qakpoYe5uBJbamoqEhMThQU3VsISkYVOCLTv3QHFUC1IXBq6rLngwUFBXCCOaUClA6dUg+OATpN1zmHy8/OF8+MtExPoHhhBY/8YeiacUGnjUJhpxKmrs5FikHcztBBc7pFUihP3KyeEeEXPd3Z2CoGc9N9cm0mp9eQtFgsALFpBp3z2s58FgBmv0N69so/PBD1K+FskaVpYWlqapE5lcgq6w+HA6Ogoli9fHnZEORCaEM/V8jTUcYLNh45Br+dwOFBeXi45ZzWcM/Tx8XFUVlYiIyMD2dnZqK2tlXQduecxF2KLilrv9Oy9t7cXAATrPZbFbMLdTByrrcVbrYUYImVY4zHByNmhAAEhPDjCA043CACiiEdeStKs1/umrYnTu7as9647X1dTJVmgpLJQXO5yzIHjuFllfelREN1M0tiFlJQUGI3GkBtGWa0zm7LFLuiEEHg8HuzduxeVlZXYv38/uru78Z3vfAfFxcVCSdhwS9wyQZ8HCCFoa2tDR0cHSktLkZOTI2khk0PQCSFobW3FyMgIkpOTsWbNmojGo+Vfg80rUMtTOk6k0eV0HGqNVlZWIjExEZs3bw6pMEuoQkrjD0pKSpCbm4vJyUlJrw/2nGhayr6V2Kj13tvbC5fLhRMnTiA9PV2w3udDbMLdSDx7rB9/eXcUTsc6EHD4wL4CangAEFighw4OcJgp26rkXbh82QAA757WQtnXOe65P4GigYhy1J1fbC73UPA9CqKxC2NjY2hoaPAqDOR0OiW3TlWr1TEvhBMpHMdBpVLhwIEDuOaaa5CQkICamhpccsklKC4uxr59+9DU1IRHHnkkrM0LE/QoQ4OlbDYbduzYEZJbNlJBpxarzWZDbm6ubBHzwVLXgrU8pcjVKW16ehoff/xxyAF+oc6Dphf29/d7pb8ttuYsHMfBaDTCaDSisLAQH3zwATIzM2G1WoU66tSaSk1NjeoiGupnNTLtwL73G0GcZsTBhSEkg0ABO9RQgEAJHg7MWM+JnB3fUb+A00wEdnzWa5xQgjHFAgXMbm0aTt35heByn6/2qeLYBXFhoLGxMZjNZrS0tGB8fDyg98NsNoddEXAh8qMf/QgXXXQR7rrrLmRnZwvZNxdddBHOO++8sNdFJuhRguM4odFIcnJyyFYjEFnaGq08ZjQaUVFRgd7eXoyPj4c1li9zCaDUlqfBxpEKIQRjY2OYnJzE5s2bvWrOh4IUQfZ154vT3+QS9FgtVhzHITU1FStWrPCqo97f34/GxkbBBUjdpbEUomMtPbDaHUgidvQiDQQcOPBQggMBoIYba7kufDvuTWxUdSLBPQF+enblrUiyK+SoO78QLPRYdFvzLQz00UcfIScnBy6Xy8v7IXbPKxSKqJd9PXDgAH7729/i2LFjGBgYwL///W988YtfjNr1Ghoa8NxzzwGY8T7QeKbU1FQMDAyEXbGPCXqUcDqdOH78OFauXCm50Ygv4VjohBD09PSgqakJK1euREFBgWQ3uVT8jSWuAR+s5al4nHAF3eVyobq6GhaLBWlpaWGLORA8yt1sNuP48eMwGAzYuXOn3+ODxWShB8K3jrrT6fz0LLmuDjzPC5ZopNZ7OO9VaWoEgQI2TgM3UQD/LdVK8N8APxC0k2ysUvbDADvAAZ7M9X6vLYeg+t4vcfR3oLrzS+kMPdI5JCUlCQWtxN+3+vp62O123H///SguLhZqB0QDi8WCjRs34pprrsGXvvSlqF0HmHnPRqMRra2tyM3NBQDhvdXV1cFoNIYdo8EEPUpoNBrs3r07oiYbSqXSq/53MMRNTsrKypCSkuI1lhw12P2NFWrLU0q4gj49PY3KykrEx8cjPz8fZrM55DHEBBLk4eFh1NTUCB3v5momIpeFHmtB90Wj0SArKwtZWVle1igtFRoXFyeIe6jWeyiianV64HTzKEt2wABgksT/V8Q/fT0HAvt/3e1ODwEHG4haD9fmqyO6dij4Rn/7qzsf60BE4NP2nfPhcg+Er9vf9/s2PDyM7du3Y//+/WhsbER+fj4+97nPCf/kEvmzzjoLZ511lixjSeHaa6/Fr371K+j1erhcLphMJoyPj+PWW2/FRRddFPa4TNCjiFqtjuiHG0oeusViQWVlJdRqtd8mJ3JGzFMhDrflqe84oeBbja27u1uWc3jfe0MIQUdHB9ra2rBu3TpkZ2fP+frF7nKXij9rlIpVfX09PB4PkpOTBYGXo9HHyLQDD7zXifdaTPDwBIVJy3Ga5k287NjsPTcQqOGGG0oowSMFU+DTVsOx+yfg02cHgkZL0L3mFKDu/NDQEJxOJ44dOxaTMr7z1RwmGIHc/hzHITMzE7fddhsKCwvx1FNP4ec//znefPNN/OpXv0JKSgrOOOOMeZ5x5CgUCnz3u9/FsWPHcMUVV0Cn0+Gyyy5DZ2cntm3bht/+9rdhj80EfQEjVYSHhoZQW1uL5cuXY9WqVX5/IHILusvlQm1tbVgtT8XjSBVjGq3f2dmJDRs2IDMzE0Dkddj9jeHxeFBXV4fx8XFJgYxSBX0+09bmA7Va7RXsZDabvc6SaSWx1NRUJCUlzfpeUiuxs/o9qFvfQJGtDkhZAVfxOYA+GTZOj2++5kDHiBkq4gQHHs0jGnRwFfiC+iM879oJJ9TgQKAA/9/uaDxSdRz6LnoF2TkFMzVc/TAfgu6LuO68Wq3G1NQU0tLShFoBtBCQnKVV52IhCDo1CqR4CaxWKxITE/HZz34Wn/3sZyMSvYVAUlISXn75Zezfvx+1tbVwuVzYtm0bTjvttIjGZYIeRSJdoIOJMK0419PT49XUxR9ynqEDQGtrK3Q6XVgtT8VzkiLGbrdbOC/fuXMnEhISvMaIVATFY9jtdhw/fhwKhQLl5eWSFlX6OcshEotJ0MVwHIeEhAQkJCSgoKBAOEs2mUw4ceKEYL1Tgdfr9TjWZ8Ff3h/BsEMJkL0oUK7H/xv5B3Y0Xg8oVHjDXYFOx5XQwwElZr4nGuKCBTo0JmxHkW0aHpcdo3wC7NAggbMhUa+FMS0b8Wm5c4o5EPugNEII1Gr1rEJAtLSquO58OMcZwaBrQSwFnc5BiqCbzeZFn4PuD7pBAT7NUY/kGIQJ+gImkKA7HA6vFp3BIkDlOkMfGRnB5OQkkpKSsG3btogWBCmCTuu/6/V6lJeXz6r7LEfqGxVkcbGYNWvWSH5vcglDrAVGDur6p/HPygEMTzuwOtOAi8sKsXr1algsFphMJgwPD6OlpQUmtxa/q3TD5VJDAzcADm2eDNxg/TruU/8Rr3p24mW+HHaoQQDEwQEFN3NqriIedJoV2FW6BicGzCiNJ9DCCTNJw6iVoKIoBYm6wEtbLCx0Mb4bCt/SquK68+LjDGq9Swk6DXZ9hUIR03sQiqAv5rKv/mhpacErr7yC1tZWADM9Gmi53JKSEnzlK18Ja1wm6AuYuQSdBqGlpKRIqjgXaCypiHumJyQkICMjI+LdfTAxpgFpubm5WLVq1ZzdyOQQdLPZjKNHj4YVC0CfezJb6ADwf8f78YvXWqFUcPDwBB+2j+PvR/vx2P9swNrsBBgMBuTn58PtduNnL5+Awz0GgMM0ZlJ2NHDBAi1udH0bHijB/zd63Q4NXFAhCRZwIPBAgRR+HN889VT84f1OtI9Y4PRooddw2FWUhEvLgnfwi7WgB8tD983dNpvNGBsbEzZEOp3O6zgj1ODbhRDhTs/PpXwOFosl4k1MrKHfuY6ODnzjG9/AkSNHsGHDBnAcJ9RyHx4ext69e5mgL0QiXTB889AJIejq6kJLS4tQpUzqNSIRdLfbjZqaGkxPT2PHjh3o7OyUxdqfS4zFm4doB6TxPI+RkRFMT0+jrKwsrLaFYkGfC5rW53Q6kZaW5teNutCD4gIxZnHizjfaAAAefuY+8ASwuTz4+ast+L/rtgjPValUaB5zwUU4iJcgJ2a8Lw5ooIcDGjjhgBoEHDxQwEbU/3W9c/iy9mPkJF2CX36hBCcGpjFhc2OZUYuiNGnFR2It6KEIqvg4g26IJiYmYDKZvPqW09x3KY1RFoKgh1LYxmq1SurLEC5ms1mwlgGgo6NDMJry8uRpzEO/cx9//DFaWlrQ1NQUUWdGfzBBX8CIRVjcR3zbtm0hp2uE2/ZU3PK0vLwcGo1GtvN4f+PQ9zk1NSUpIC3SXPaqqipYLBYkJyeH3YM4mKBPTk7i+PHjSE1NxbJlyzA+Pi64UcU53YstKE7M+61jcPOz584ToGHQjN4JO9INGnAAhqYdmLQFLphkgxY2aKH479k5AQcHNNDBiTOVR3H5xkQQACoFhw05oTfFWQiCHu5ZqbjuPOBdd17cGIUKvL/2pLEoKhPJHMxmMwoLC6M2l6NHj3o1S/n+978PALjqqqvwxBNPyHotQghKS0vDXm8CwQR9AUMFfXp6GlVVVdDpdKioqAir6EA4bU99W57SBVCu83hfMbZarTh+/Di0Wq2weQh1DKnQYjHx8fFYsWIFTCZTyGNQ6H3xNw96D1euXCmU36U5tjQqnOZ0cxyHwcFBaDQar25oiwGXJ/BncP7DR+Byzwi+Ejw8kPbeCBRQ/NcBv5brxO2av2GVwYWRVftgjCCAKNaCLmfp17nqznd1daG+vt5v3fmFYKGHEgBmtVqj6nLfs2dP1DfT9H6fccYZ6OjowKOPPoovf/nLQrdGjUYDpVIZUe0SJugLGJVKBZ7ncfjwYSHvOtxFSNzdKNgXhrY87enp8UoRo8hpoVMRHBkZQU1NDXJycuZMvfNHOFbtyMgIqqurkZeXh+LiYvT19UX0Y/b3mYgb8mzcuBEZGRlwu91eXb58o8KPHj0Kp9OJmpoaEEIEyz3avczlELZdhSngAMx1F51uIvzVLVHM8d9X8FBAAw9uzG7E8uKL0Ji8CyM9o3C29QuuZnEVNknjLgALPRrXl1p3XqPRxHzDGIqgL6WguPT0dOh0OvzgBz/A448/jqKiIiiVSsTFxcFiseArX/kKzj333LDGZoIeRSL5wfA8L5zpiPuIRzIXjuOCCrG45enOnTv9/ogUCgWcTmdE86HjeDwetLe3o62tDWvXrsWyZcEDmnzHCCWXvbOzE62trV7XitTV7Wuhezwe4XjEN81uLtRqNdRqNZYvX46MjIxZvcwTExMF4ZrPAiRSyUnS4eqdy7HvcC84DiAEUHAzLve5ZV46G9M5bL36t+A4DsUAVv63B7fJZBKqsGk0GmEDlJycHHDjuhAEfT4s5Lnqzg8NDcFqteLjjz8WIufnqjsfLUIVdCm/o4UM/cz//ve/409/+hPKysqwevVq2O122Gw2WK1W9PT0YHJyMuxrMEFfgNjtdlRWVgoCEU7RFl+k1HMP1vKUImcZ2eHhYfA8j+3btwv1nENBqqCLy+L6XivS4jTiM3T62XEch507d87KYw90dko3Fr7d0MS9zGkBErH17u+MNBZ877QVKM6IxzPH+tE3bsekzYkIWu+AAweVAjBo1Lhi73ovARb34BZXYTOZTGhtbYXdbkdSUpJwj+Lj471eH2tBj0W3NXGlP61Wi4GBAeTm5s6qO0/vWbS7m4Uq6NFszjIfUEF/6aWXsHXrVjz77LOyX4MJ+gLDZDKhqqoKmZmZKC0txVtvvSVbQZhAQkxbnhYVFWHFihUBf8hypIpZrVYMDQ2B4zhUVFSEXRVLinVNRRaA30I4kRanofdqamoKJ06cQFpaGtauXeu1YNOiEYQQoQc09ZoEW9jFVpb4jLSzs9PLek9LS5slXPMJx3H4wvpMlKSqcPVfq+AmwZcXDvysmuwAoOAU4DggTqPExlwjygsDb2rFVdgAzFlDPTU1FcnJyTEX9FgXtuF5HiqVas66821tbcI9oxZ8JGe7c81BiqATQpaEy53+zktKSmQziHxhgh5FQvnBilO1SktLsXz5cgDylmz1N1aoLU+ByM/Q6aYlLi4O8fHxEZW4DLa5mJiYQGVlpV+RpcgVXV5dXY3i4mKhwx2F53lhjjQugmYc0OtTgQ82D/EZ6cqVK2G32wW3c1dXF1QqlZdwyb0IE0JQP2BG45AZSXo1KgqTEadRgrOMgJvsxsCoCT/6zzimeWlHRMmYxs+Uf8WL/C50kUxMIg5WRQI4lRZJcRqcsToN15bnQq0MzZoVB4qJa6i3tbXBZrNBp9MJ9QdisQmKdVCa7/UD1Z1vb29HfX09EhMTZa07H0qUu9VqXfSCTu/X1q1bce+99+KRRx7B2WefDY7joNfrodVqhaO3cGGCvgBwuVyoqamB2WyelaoVSU90X3wFPZyWp3SccHaY4jz60tJSOByOiDulBRJ06nUoLi5Gfn7+nAtQJC53uhEDZnbe+fn5Xn+jljkwc9/E5+30bzzPw+l0wul0wu12w+12CwtdsAVPp9MhJycHOTk54HlecDtT4aIu1FCDxvxhdXrw/15pwuGOCSHXPDVehbuWH8aOgafwviUPN9quhxnSxFwBHhq4sU7ZifPUhz+9znmPYTTnVOjUCujVkZ/piq334uJi2Gw2tLW1YWJiAkePHoVarfZK85J7E+SPheAhCGQd+94zunEcGxuTre78yRYURz/zjz76CPX19fjBD36AP/7xj0hJSYFCoYBOp8PY2BgefvhhbNiwIaxrMEGPMVNTU6isrITBYEBFRcWs3ZncTVXoWOG2PPUdRyriM2yaR9/R0SFLlTfAe4EUR+lL8TqE63IXN3FRKBRe7WrFLnY6T99Sn8DM5+twOFBXVwe1Wo309HQQQoRNnNh6DybudA4pKSkoLi6G1WqdFTRGi9qEEwD1lw+78WHbOLQqBeLUCvAEME1a8KOJ5bheU4Y7HBfDFcKSkoQpeKDCcVKMFdwwOMLDVXwOPCs/i2QuetarXq+H0WgEIQRr164VNkEdHR2CJUrFTEqRlnBYaBZ6MHw3jrTufG9vb9h156UKusfjgc1mW9SCLj5i2bZtGzIzM6FUKjE1NQWbzQaHwwG3243BwcGI3icT9CgSbCGgP4bCwkIUFhb6fX4oLVSDQTcHXV1daG5uDrnanHicUITYZrOhsrISCoXC67xcjrN4unBQi8PlcqG6uho2m01SjXsgPJe7OPitvLwcBw8eFMYQW97BhJj2dk9JSRHqx1MXPf0n/vwVCoXwLxi+QWO0NnhTUxOcTqfQ6lTKe3e6efynbhhKbuZej9nd8PAEChD0IhV3OC4KScwBwKVMBDg31GlFcKfo4S45D+6SLwBRFHMK3QCKN0HAzHeVniN3dXVJKtIS7vVj3Rgl3Ov7qztPU+NCqTvv8XgkpWNaLBYAWLSCLv6seZ7HhRdeGLVrMUGPAR6PBw0NDRgaGgpqQcptoXd3d8Nut4fd8pSOI1WIx8bGvIL8xIuI3IJOO6XFxcVh586dkhffUOcxNTWF48ePIyUlBevWrRPmIHaxS2l+MTw8jLq6OqxYscLr3F0s2OLzdvr/w7HelUqlUF2MBkCZTCaMjIzA6XSivr4e6enpMBiT8UqLBS/VDGPS7saW5Ym4riIXucl62F0eOFwuOHj6njh4/ruEuBBqnjyBi1MhQafGxktugz0+enn2fq8+h8tbr9d7WaL+AhDlOEdeCEFxcqWoaTQaZGVleRVMklJ3XqqFvtgFneM4PPzww/jiF7+IzMxM7N+/HwCQmJiI+Ph4YeMdFxcHnU4XUc0JJujzjNVqRVVVlRDdrdfrAz5fLkG3Wq2YnJyEWq2OqOUpIM3lTghBd3e34AnwVw9ZTkEfHR1FfX19wEYucxGKhU4rv/lmA9D3IkXMaSxBe3s71q5dO6twj7/3J670F6n1Lg6AysvLw8GDB7Fs2TJYrVbc9K8TqB8jIOCg4IADrWM41DmBP1+yFklqNyZsMnWWA4c4tQIXl8YhdZ7FHJAmqP4CEKn13t3dDYVCIVjutMd5KNePtcs9GumOUurOJyUlISUlRcj2CIbFYhECxhYrf//737Fnzx5kZmbiV7/6FQYGBrw8nRzHQaVSwWKx4NixY2EbW0zQo4jvgkGroWVnZ2P16tWSvsxyCDq9rkajwbJlyyISczqnQELM8zzq6+sxMjIS0BMgh6BTamtrsW7dupAL0wDSBF2chUArv4n/BswU5aHutbnEgmYVjIyMoKysLOTce3/WOz2rD9d65zgOSUlJGOGS0DBugoKbqfk2UxOGh8NFcNdrDcD0EICMgGMFnDs8UIGHSqHExrwUXL5Gh3StPN6nUAknKE2n03n1L5+cnMTY2Bi6u7vR0NAglFiVYr3H2uUup4UeiEB15ycnJ2GxWDA1NRXwSMNsNkc9Jz7a3HfffSgoKAAA/OhHP4LNZgMhBA6HQ/jndDoxPT3NztAXOoQQtLa2orOzM+RqaJEIuliE1qxZg/HxcVnSswJZ6OKc74qKioCbh0gL1NBAOwDYsGFD2NX0gm0sxMFvvpXfqDs8MTERVVVVSExMRHp6OtLS0mYFVNFsBpfLhR07dkS8sfJnvVNxD8d6/6RzAjw+7ZY2w8z8G0adiETMUzQ8Li3moEnKxc6SZViTZUBnZyfsdnvYY0ZCpFHmYuu9qKgIDofDbxT4XKV7Y+1yj1VzFnE64SeffIK0tDTwPC/Unfc90lAoFPNSVObPf/4zfvvb32JgYABr167F/fffj1NOOUW28bds+bTbYGVlJW688cagjafCgQl6lKEdvWw2m+QyoGLCTVvzbXmamJiIqakpWdz3cwkxjZxPS0vDmjVrgloAkaSLiTcOHMdFtKsNZKE7HA4cP35cCH4Tp+eIreMNGzbA6XTCZDJhdHQUHR0dXtaJTqdDbW0t4uPjsXHjxqikRs1lvYtd9PR5/ora9IzbfMRcnmYVBck6PPY/G5GR6L2BiWVnObnTxrRarZf17lu6V2y9JyYmLgiXe6ybs/A8L9TiB7zrztfU1MDpdOLPf/4zCgsLkZSUFLUN0LPPPovvfve7+POf/4xdu3bh4YcfxllnnYUTJ07I3jrV5XLh//2//4fLLrtM+B4As7NgwoUJehTheR6HDh1CQkICKioqwlrEw7HQ/bU8BWYWcjly2mmal3hR7OnpQWNjI1atWoW8vDxJX85wXe7idqRr167Fu+++G3GlN3+vFwe/+ab2+Qt+803tGR8fx+joKBobG+FwOKDT6ZCcnAyn0xn1XOe5rHd/RW0A4Ej3NN5sGBGNEP79zFZMYKN+FGpjFj63czP2lM7tkYqVlRrNPHBxFDi13unZe29vL4AZC9lkMkGj0US18c5cLARB9w2K8607PzAwgFWrVuHtt99GU1MT1qxZg89//vM488wzceqppwaNP5LKvffei+uuuw5f/epXAQD3338/3njjDTz44IO48847ZbmG+Lt200034d1330VBQYHsnwET9CiiUCiwefPmiHJZaSqWVOZqeUrHcjgcYc1DDP0SUrddQ0MDBgcHUVZW5pWLLWWcUAW9v78f9fX1WLlypRAZHulZvD9PwVzBb4B/MfeFBkzZ7Xb09vYKHZVGR0fR0tKCuLg4wXpPSkqK+uI6l/VuMplgdbhw17FeuHkgMqucYLnagh9VpAGumZrhqfpPzwV979NSstAD4StUU1NTOHbsGAYHB9HW1oaEhATBzUzbm0ab+TpDD3cOHMdh2bJluPPOO/HUU0/h6aefxg9+8AO8/vrruP766/HEE0949S8PF6fTiWPHjuGWW27xevxzn/scPvroo4jH98XtdmNychLf/va30d/fj40bN85klxgMiI+PR0JCQkR90pmgRxmj0RiR2EjNQw/W8hSQr6kK/RHabDbU19eD53lJEfu+hNopjb6/TZs2IT09XfhbpM1VxGlnAIS4A9/7SL0S1H0dLJK9paUFfX192Lx5s7DRoZG/1DVfW1sLnueRkpKC9PR0pKamRlQKV+r7Pdo1gcc/aEHjwDQS4vQwme1Qwg13mEtCFky4doMOF519FtRKhXCmLM7nFjeUoR6KWFrosbBQaRQ4AGzatAkcxwnWO21vSvO3o/ldWCgWutQod6PRiAsvvBAXXnihrBvB0dFReDyeWetlZmYmBgcHZbsOZXx8HC+//DJKS0vxhz/8AVarFW63Gx6PB263G/n5+ejo6Ah7fCboCxwpLncpLU+ljiUF+iM8cuSIUCM9nN2+1A2G2+1GdXU1LBaL3/cnV3MVt9uNEydOYHx8fFYJXrGQ02vOJUa0farFYsH27dtnBfSoVCpkZmYiMzNTaGk5MjLidd6alpaG9PT0qLRKfadpBLf+ux52N4FSwWHYZAUPBUJZDhTwQAs3zlAcxfcya5D1ue/Cs3yn8HffM2WaukTrgtPofo1GE5MyqLEsvUq/qwqFAmq12iuHm7Y37e/vR2NjIwwGg7AJklqBTQqxCoqj0N+T1Dx08W8+Gp+bP+9RNK6TnZ2N6upqcBwHjUYDu92O6elpuFwu2Gy2iL0mTNAXOMFEWGrLUyDypiqUvr4+AMDy5ctnufVDQYqFbrFYcPz4cej1epSXl/tNa4nU5S7eoCgUioDBb+Ln+4PWx1epVNi+fXvQ3FlxS8uioiI4nU6Mjo5idHRUyHWmrvnU1NSAn++I2YGXaobQO27D8iQdvrAhCxkJ3haey+3Br185gWknP+Nc5wEgtIVdAze+pn4dl66JR+KZP4ZCrUWgb5VvSVrax7y7uxuTk5M4dOiQV0OZ+XAFx1LQfeMXKOLvwooVK+ByuYTI+bq6OsGTQwU+kiyJWFvo4v4GwaANdKJBWloalErlLGt8eHg4YH2IcOE4Dunp6bDZbOjq6oLRaERRUREAeT4TJuhRJtJFI5Cgh9LyNNhYUuB5Hk1NTejv74dCoUB2dnbEqT++wXViRkdHUV1djZycHKxatWrOL3ukLnfaICYuLg7r168PGvw2F5OTk6iqqkJ6errkOgO+0FoBYst2dHQUbW1tqK2tRXJysiDw4kWuqmcSN/5fLcwODwACngCPfdSNP16yHlvykgDMZFy88O7HGLB4wj4p18CFGwwf4OpLr4MnuRBuAuC/BUKklqTV6/VYvnw5bDYbeJ5HWloaTCYTmpub4XQ6ZzWUiQYLxUIPxFzW+8DAAJqamhAXFyfcp1Ct91ifoYci6FarNWqCrtFoUFZWhv379+OCCy4QHt+/fz/OP//8qFzz6NGjePTRR/Huu+9i5cqV+M9//oPe3l68+OKL2Lt3L9asWRP22EzQFzhytTylY4UrfE6nE1VVVXC5XCgvL8fhw4dlr8NOEXdlW7NmDXJycoKOE67LfWhoCDU1NQCA0tJSv2IerFgMHae+vh6FhYUBO7uFgtiyXbVqFaxWq2C9t7a2QqvVIj09Hckpqfh/L7dh2u4GIYCbnykKM2Fz4+t/r8Fr39qBpv4JvHnwMNoHR0GwIqz5qOHGV+M/xHVXXQcuOR9KGYra0OBB2tWLlqSl71Gv1wt/lzN4cCFa6IHwZ73Ts3dx/XR6r4JZ7wvBQpfy/QBmvHTZ2dlRm8v3v/99XHHFFdi6dSvKy8vxl7/8Bd3d3fjf//1f2a5Bv29tbW346U9/CoVCgV27duHEiRMAZjYWlZWV6O/vx69+9auwr8MEfYHjm4cebstTOlY4FjpN30pKSsKWLVugUqlkOY/3J+i0ytzo6KjQlU3KOKFuLsRFd9avX4+qqiqvTYE4dztY8FtnZyc6Ojqwbt06rwpychMXF4e8vDzk5eUJEeqjo6N47eN6dI95wBMOnv++hZlab4DF6cGZvz8Et9sNDjo4URDyddVw4oK4E/jaKXnI2XILoP70OxdJURtfUfUtSet2u4WGMg0NDXC5XILLWYpoBSLWgh5p3rFarfaKwzCbzTCZTBgcHERzczPi4uK86qf73vtYC3ooHoJot0695JJLYDKZcMcdd2BgYADr1q3Dq6++6tUKOVLo9+2dd94RjpmeeuopNDU1AQAyMjKEDpSRwAQ9ysjpch8bG0N1dXVYLU99x5IKTRObq3Z5JIgFHZgpLFFZWSlsVqQu2KG63MWV38TBb75R7HSOgcq4njhxAmNjY9i2bVvIRYMiQalUIiMjAxkZGZjWj4GrrhEVhZmpxU7/v81NMHNOHtoCrgBBCqbw203D2PGF7wEBvsvhFLUBAv8+VCoV0tPThZayFosFJpMJQ0NDgmiF63KOZelVua8trp9eUFAAl8slbIROnDjh1f2MRs7HuvRsKEF581Ep7oYbbsANN9wQtfGpsTA8PCwEhHZ3d3utGbQJTSQwQV/gUAs90panQGgizPM8mpub0dvbOytNjM4rUgudvgdaF7uyshLJyclYt25dyP3ZpbrcaeU3AF7Bb3RTIDX4jWYW8DyPHTt2RD3VzB/jViee/qQP+xtG4PZTrjWSnHINnDhV1YDrP78N67ZcEPwFIqQWtaFuVynWIq0GaDAYkJ+fL4gWbcrj8Xi8rPdgn8dCsNCjhVqtFjZ74o2QuPsZMBPzkZqaGpOzdKmd1oDoW+jzAf288/Pz8dprr2FkZAQ8zwuCXl9fj7a2tohz65mgL3BoFbP29vaIWp4C0kWYipXD4Zizp7gcFjo9QxsaGkJLS4vk4L5w50KPDvxtGmgGgJTgN7PZjKqqKiQkJIS8+ZCLKbsL33q2Fq3DVig4QK3k4HBHXrY1DjZoFAp8o3gCV5xzOWCI/AjBn/VOxbigoEA4UgolsM5XtMTpXk1NTUK6Fy3WMl9pSVKYT+vYdyPkdrsxPDyMxsZGtLS04MSJE8LZe6De5XITqqDPp/crGtDP+/Of/zxee+01fPvb30Zvby/UajX+/e9/4+6770ZiYiIuvfTSiK7DBD3KRLJoWK1WoV65v3zmUKFBcYEWs+npaRw/fhyJiYnYvHnznGlScqTAURd3S0vLrA5moSBF0GnwW2FhIQoLC2dVflMoFGhvb0dWVhZSUlLmXGxMJhNqamqQm5uLoqKimInCi9WDaBiYBgHg5sVV10IRcoJPrXlADzfUugRkJOpw5tmnAQb5vQ4KhQKjo6OoqalBUVERcnNzvaz3cLvFiQPGnE6nEDBGc37F1rtarY65hR4rd7dKpRKMgvLycqH72cjIiGC9i4MQo7VZDVXQo+1yny/S0tLw6KOP4kc/+hGqqqowNTWFK664Ap///Ofxm9/8BitWhBewSmGCvkChLU+zsrJgNptlqf1Nf0Aej8fveLTcqT/R8zdWJBY6bR5DCIk4mCxQcxVCCDo6OtDW1ua3gh49312/fr1guTidTqSmpno1VgFm6tU3NzejtLQ0rDatcuH08HjiUDfsboJI3Oozny6BAgQqDtCqVCg0cvjazhQkqqNTlrW/vx8NDQ1Yu3at0B1PbL1H2usdmIkYFqd7iRul0DanTqcTdrs9JsIe605rdEOhUCgE610chDg2NoampqZZKYR6vV62eUsVdEIIrFbrona50+/YwMAAPvnkE3Achx/+8If49a9/jcnJSWRlZcnW24EJ+gLDt+VpTk4Oent7Za3w5ivo1Eru7u6WbClH4nK3Wq04fvw4tFottFqtLG1E/c2FtlcdGxvzW/mNBm4BECw4euY4MjKCgYEBoVoXx3GwWCxeZVxjgd3lwT+P92NkKtKa/Dy25RqxuyQTZ5SmAQRwOmxQ2KcwOjqKgwc7ER8fL1SsMxqNES3mNBugs7NzznvozzVPxT0S691oNMJoNKKwsFAoSdvW1obu7m709/f7LUkbTRZChLm/6/sGIfqmEGq1WtkKAIUa5b6YLXSO49DT04Pvfve7eOuttwSBv/vuu3H99dcLBgnrtrbE8NfyFJC/ZKtY/FwuF6qrq2G1WgOWjfU3VjhzMplMqKqqwrJly1BSUoKDBw/KEi3vOwaNmCeEhFT5TXzmuGLFClitVuH+AEBtba3Q73w+A4o6Ri3484FOHOmawJjZiUh65mngxJlFBtx1+Tafv8QBSPWqUjY6OoqqqioA8PJaBKuAJ4YQgsbGRgwPD2Pr1q2SzkPnCqwLt9c7hZakHRgYQE5ODrRaLUwmEzo6OoSStFS04uPjo2JJxzrCXIqY+qYQejweIXK+ubkZDodjVgGgUO5VqFHui/UMnQr173//ezQ3N+P222/H+vXr8dRTT+FnP/sZtm/fjs2bN8t2PSboUUbql3yulqdA+D3R/c1FvDkwm804fvw44uPj5yyrOhehutwJIeju7hZc1suXLwcgX3Cd2OUeKPhNanMVYKb5THV1NTQaDbZu3QqlUomJiQmMjIwIi1pycrIg8HK1cxRDCMGfD3TiycM9sDg8IBH2KP//7Z13fFRl2v6vmfTeGyEhkJBAQnoQEAsICEgJiLDsrsrasIHw+lv1XXVdXXvnxQ4qorsuKolIEUSkCSqQ3kN6T2YmySSTydRzzu+P7HOcSZ3JVMLz/Xz87JIy88wkOdd57ue+r8sFGqRH+uOxzJRRv26wS1lPTw8kEgkaGhp44SPiPlqaIBkRJL72432PxjMWN5pgEFH18/ODn58fYmJioFAo+LP3uro6ODk56e1IzbV7t5eSuzE4ODjwP2+O43j7XuLP7+zszFc5DHmvDC25a7VaqFSqK3aHTgT9xx9/xMMPP8yb1SxZsgShoaFmD4Chgm4HjBZ5Cphvh677WB0dHfxzxsTEWKyzHPh9Xpvs0HQ79c01z04eY6zmN2NtXIODgxEXF8dfAIlzW1xcHORyOSQSCTo6OlBZWWnWErVCw+A/l1qQXdCKhk4FGI6FbgObsbhDiQURzthwYwbSo/zhIDTOpYzke8fExECpVPKOdbrCFxQUpNdQqNFoeMOejIwMs+V+GzoWN1ppfrgSp5ubG59nzzAMHyhTXV0NpVIJX19fvjJjSje4rUvupgazCAQCuLu7w93dnW9qHO69IgI/XKXDUEEntsxX6g6d0NHRgeTkZL2PeXl5jcs1cDSooNsQQyJPAcMjVA1BIBCgoaEBHR0dSExM5BuTjMXQkrtKpUJBQQEYhsG8efOG7NDMJeharRa1tbWoqakZ9nUZ6vwGDNxglZWVISYmZtSZf1KSJHPRg0vURNz9/f0Nrn5wHIfWHiX+8nkBWqVKnR35+P/gZ6IRH25bjyAf8+xyXF1dMXnyZEyePBksy6K7uxtisRiVlZV81cLX1xetra3w9PQc4o9vbkbbvQ9XmidfN1bTJ9mdA+DPk8n5u+55srHd4LbssAfMf0Mx3HtFKh21tbXDVjoYhjHIu4GYrVzJTXHAwHty7tw5vjIUHh4OqVQKqVSK7u5uODo6wtnZ2WQ/CyroFmakP1xDI08B8+3QtVotn8U9d+5ck+56DTkG0LWMHemibg5BBwbO5onJy2jNb2PZuNbW1qKxsRGJiYlDzHRGY7gStVgsHjNURZdz1Z348kIDLjX2QK4BTOlg13lVeOIPN5lNzAej68VOGqlaWlpQW1vLX7xqamoQGBhoVi/20dYDjLx7J7+zurt5Q9Y0eEdKzpMrKiqg0Wj0fNTHOlaw9Q7d0s9P3qvJkyfzu/euri7U1NRAoVDAx8eHf8/GurmRy+Vwc3OzaZCMKZDXFhUVhX/961/Izs4Gx3FwdHSEVqvFW2+9hU8++QROTk5QqVQ4dOiQSddlKug2wJjIU8A8gk5iSAEgNjbW5BLWWEJs6AicqYKuUqnQ1tYGhmEwf/58vY75wc1vo/lnMwyDsrIySKVSgxu3RkK3RE3iQsViMSQSCaqqquDm5safuxOR+6VGgpe+y4VCoYCC9YIpO3Jdpgd5ISM2wiyPNRYCgQAqlQotLS2YNm0aIiIi0NXVBYlEguLiYrAsq9dYZ64S/GgM3pWzLIvGxkao1Wo4OzuPy9RG9zw5NjZ2iBPbWJa0V0JTnLnQ3b3rRufW19ejpaUFIpFIb/c+uJIll8uNbrizJ8i6X3vtNfT09ECpVEKhUECpVGLt2rWQy+VQKpVQqVSQyWR0h34loNu0ZWzkKWC6oItEIt4Mpbu72yx/HCOtieM4VFdXo76+3qAROFMMamQyGXJzc+Hk5ARfX99hxdyQErtKpUJhYSGAAQMfc9u4urm58aEqWq0WXV1daGjtwOmzRRCAQ2qkH774rQmKfi3CBN0QwdibiQHPdTncoIUzIBDAQSiAj5sT/t/iaAitdDEkiXOxsbF806NugEhv78BIXFNTE8rKyuDt7c0Lo5eXl8Uv2gKBAPX19WhubkZaWhq8vb3NYmoz2IltcAraYEvaK7EpzlyQ6FyxWIygoCC4u7vrTRl4e3vrTRlYMgt9JF588UUcOXIEBQUFcHZ2hlQqNfkxly5davrCDIAKupVgWRbl5eVob283KvIUGL+g6860z5o1C2FhYbh06ZLZRuAG76y1Wi2Ki4vR29trcEnfGB92XUQiEQoLC/ndf29vL/85Y8ScTBf4+voiPj7e4jsXR0dHnG9l8PF5OaQKR4Dj8GVNL3rljnCAAHVcoNGF9vWCM3gp5BS+nPkeDtVoIe5TY2aoJ/6YEY7ZUeO3CjaGxsZGVFdXj3hUoTsPHh0dDZVKxTfW1dfX87te0nNg7nlwjuNQXl7Op/gRkTC3qY2jo6OeJS1JQSMZ5h4eHnBwcOCnRGwhrKY2xZlrDU5OTnyTKQC9KYOKigps2bIFCQkJcHd3h1QqNcn22hjUajXWr1+PefPm4ZNPPrHKc5oLKuhWQKlU6qWIGdshO56xNV1xtdRMu+7jELMYZ2fnIWN3Yz2OseNvxPmNNL81NDTwj2FM8xspBUdGRo7pjGcqLVIFdp6qw7nqLkgVGggEgKfLQHOQVMOCw/iqAs7QYNNNidDM/yfWA1h/vXnXPRakItPS0oL09HQ+SWosXFxc+I5ylmX5ccCqqiooFAq9cUBT/cVZlkVJSQlkMhlmz5497Bn3SI11pmS9D5eC1tnZiYaGBshkMpw7d05v926NIwjA9mf4wPBd7rpTBjNmzMDOnTvx73//G83NzQgKCsKcOXOwbNky3H333QgPD7fY2p577jkAwGeffWax57AUVNAtDMdxyM3Nhaen57giTwHjRXg0cTXVsnW4x+nq6kJ+fj7CwsIwY8YMoy4Wxo6/lZSUoLOzE9dccw0vHiSxS3dnNVbzW1NTE6qrqxEfHz/uTn9D4DgOe39rwts/1UKjk4jGcQMBK//9F4w9MxeAhQs0WBrci2nXbjDfgo2AjCN2d3fr7XqNRSgU6o0D9vf3QyKR8PP+bm5ufGnez8/PqN8vrVaLwsJCaLVazJ492yDRtJSpDWmc7Ovrg1arRVhYGDo7O9Hc3Izy8nK9crMljyCseYY+EmONrTk7O+OWW26BWCyGTCbDv/71Lxw7dgxHjx5FT0+PRQX9SoYKuoURCATIyMiAk5PTuP9AHRwcoNFoxv5C/O4BHx4ejtjY2CEXGnPu0EmDUWVlJWbMmIGICOObrwwV9MHOb7rn5eQxyOsarfmNZVlUVlZCJBIhLS0Nvr6+Rq95NAqbe5Bd0IbLHXK4OgkR7OWCY2UiPTEf2r0+0u8FBwEG5s9doYISLnCBFq5CBs5CID3UCVtWLhnH7YDpEFdDtVpt9r4Dd3f3IT0HJCZVq9XqNdaN9rxqtRr5+flwdHREenr6uMv4ljC1cXBw0LOkVavV/FhcU1MTBAKBniWtMaZPY2GvO/ThILav4eHhuOeee3DPPfdYYXVXLlTQrYCrq6tJu2JD5tCJV3Z1dTUSEhJGDA8xR0oaMCCa/f39qK6uRnp6+rj9zQ1ZD2l+G278jVwcu7u7UVhYiODg4BEv9BqNBsXFxVCpVCa5lo3E0VIRXjh6GTKldlA+OTCeEbQBkRZCAA7uUONGj2Y8fMcf0d6jhAvTDydVDypLClDj6Kh3/mzp3Re5uXJyckJGRoZF/c8Hn0nLZDJIJBK0tLSgvLwcnp6efGleNyZVqVTyLojmjLg1h6kNy7JD3jNnZ2eEhYUhLCwMLMvygTL19fV8AyG5kTHVkpacX9sSQ6sE5spCf/bZZ/lS+khcunQJGRkZJj+XLaGCbgVMLZ2NtatmGAbFxcWQSqV6pejxPJYhqNVqVFVV8aNipggjMYUZCdL8NnXq1CFxpWSnFBAQgDlz5uhd6L28vPigCU9PTygUChQUFMDNzQ2zZ882uwgpNAzeOFGNHoUGA1quu28ej5hzcIUaDISY6iTFvWk+WLTgLri6OCM6+Pc5++GMXfz9/XmRMzX4ZjDkOMfHxwcJCQlW3enpxqSSXS1prGtsbORn4r28vNDQ0ICAgADEx8dbtDfCEqY2QqGQH3uMjo6GUqnkd+8NDQ1wdHQ0yZLW1jt08h4Zsob+/n6zdLlv2bJlzKzxqKgok5/H1lBBvwIYTYRJZrqjo+OQEJKRHkutVo97LSQv3c3NDU5OTibvckcquQ/X/Db487qd7KT5iORhk7nvuro63sQhICDA7K5lDMvh19oufHGhGe09qkFe6+M3huEAOAqFuD7KG/+87SZ4uw6/oxps7ELsaHWT4sjuXXcHOx6If8KkSZOGtSi2Ns7Ozpg0aRImTZoElmXR09ODlpYWXL58GcBA13RDQwM/HmXp9RpqakP+lg0VVldX1yENhMSxTqFQGB2SYuszdPL3bqj1qzl26OSIZqJDBf0KYCRBJ8llxjSjmbJDJz7pU6dORXBwMC5evDiux9FlOEEfqfkN+N35bbROdmdnZ/4C2NzcjIqKCvj4+KC3txdnzpzhfcdNMTfpU2mx6+cGfJ3Xgl4leT9Nd3ZzggZOYODt6oS/r0nDjdMDDZ4j152JJl3VZAebl5cHoVDIX9gCAgKM2tlJJBIUFRUhOjoaU6ZMGe/Lsxhk/FEsFmP69OkICQnhG+uIVatuY501BG04Uxvimubv7z8uUxvdBsLp06cPsVklISmjRZzaeodOrj+GltytHVfc2NiIrq4uNDY2gmEY3so5JibG7i1oqaBbAXOX3DmOQ0NDA6qqqvSSywxhPGfoHMehpqYGdXV1vOe8XC43S7f84Dl0tVqNvLy8YZvfdIWcfO9onew1NTVoampCamoqv4OVyWQQi8VobGxEWVkZfHx8+NK8obu4ktZebP6yENJ+0xPw/rtaCMFBAMDbUYtZ4QF4dPksTA827eLh5OSkdy4rlUohkUh4O1p/f39e5EYbDWttbUV5eTkSEhIsOhFgCiKRCMXFxZgxYwbfAR0REcFbtZLGuvLycqjVar3GOnMfSwyHUCiEVCrlb4pIVcEUUxtgqM2qbsSpWq0e1pLW1nPoupMoY2GuM3RjeOaZZ7B3717+3yTe9NSpU1iwYIFV12IsVNCvAHTn0BmGQWlpKTo7OzF79myju7SNHVsbySyG7KxNDZrQvcEg5XwfH59hm99GyjAfDInsJHPH5IKgewZLziYH7+KIuI/kO94pV+H+YcV8fLtzB2jhAA4rvGtwU0YypsfPR4Sfq9nLw7o7u9jY2CGjYe7u7nzVgtiVkmOPhoYGpKamWn2nZCgtLS2oqKhAYmLisM6EDg4O/M+VmL3oHkt4eHjovXZLlOZJhSMuLo6/4TD3WNzgiFMSKEPm+93c3BAQEGDSkZs5IB3uhrzP/f39Vhf0zz777IqcQQeooF8RkB26QqFAfn4+hELhkN2rsY9lCAqFAnl5efz5/OB5dsD08zhyY2BI8xu5eRjLxrWgoABCoRDXXHPNqCV13dQwhmH4ix/xHSdnzwEBAXBwdMTeX5uw59cmdOuJ+fjL7K5QwQ99eGTBNGTeYB1rSMLg0TDy2okFLrnwy2Qyk73tLUl9fT3q6uoMvuHQNXuZOnWq3rGEbkoeOZYwRzd4R0cHSkpKRqxwWMrUhqQBkp9xd3c3JBIJ+vv7cfnyZUgkEn73bo0qBcGYawYZW6MYBhV0K2COkrtWq8Wvv/6K4OBgxMfHj7tkZqigd3V1oaCgACEhIZg5c+aQ59O9AJki6GT8rbCwcNTmN0PEXCaTIT8/H/7+/ka/Rw4ODnrjUb29vRCLxairq0NOYQlOi11xtkkDBa/lpp2Xu0CNhQ4lWLtqNa5LijPpsUzF0dFRz3O9u7sb5eXlUCqV4DgOlZWV/M2NvQRlDHao003YMwbdYwndlLy6ujqUlJTAx8eHf+3jGRdrbW1FRUUFkpKSDErvs5SpjaOjI1+l6OnpweTJk6HVatHe3o7Lly/Dw8ODF3dvb2+LluSNKfn39fXZ7c2kPUIF3c7hOA4dHR3gOA4xMTGIjIw06fEMOUNvampCRUUF4uLiRnw+8gdpykwry7JoaWmBSqXC3LlzjW5+00UkEqGkpARTp05FVFSU0RdeNcPCUSiA8L83DU5unpC5AJeUQhyrbEW9VNfVzXicoIYTGABCzBQ04OOpZ+Hypy8AoX3FQmq1WtTU1MDJyQmzZ88GwzD8Dlb3WGI8rm3mgviyk2Mnc+3gBqfk6R7JkIYzsns3ZN6/qakJVVVVSE5O5rPCjcXcpjbkcTw8PODn58c3T5LGuuLiYnAcZ1FLWkNNZYCBkrup1r9XE1TQ7RhirSkSiQBgRLMYYxjtDJ1lWVRUVKCtrW1MsxhyARlvYxxx8VKpVHBzcxsi5sY0vzU0NKC2thYJCQkICQkxah2X6rvx+YVmFLX0wtFBgOuj/aFlOZyslECpYaFmWPzuETM+MXeBEknCJjg4OOBG/25sWJABl9j7ATvY6epCjnTc3d31ehh0m8s6OzshkUhQUlJikzhU0h8hl8sxe/Zsi5aKBx/JkJK17rw/ee2DxzfJUYA53QjNYWpDvk/3405OTnoVGplMhs7OTj1PB93du6kVGkMFnYxh0h264VBBtwLj+QMggS4AMHfuXJw9exYMw5hsiDJSyV2tVqOgoABqtdrgAJnxus7pNr9FRUWhsrKS/5wxzW/kBkQsFhsVDEK4WN+N/z1Qjs4+NRiOA8sBX+e1DfOV4y+vO0GNva7/B3HaXzEzfmQHP1tDjisCAwNHHIEcfCxBJgZ041DJ7t3T09PspXldX/aMjAyrhZkAQxvOyLx/R0cHKisr+abCgIAAdHV1obm52aSjAEMw1tRG92tH+rvSbRwlng7E1Ka5uRkCgUBv9z6e6pwxO3R6hm4cVNDtkO7ubhQUFCAwMJCP9BQIBBZJSQN+F1hvb2+kpaUZfNMwnqAXks0eFRWF6OhodHd3849hTIldo9HwF/c5c+aMa6f2+W9N6JQPiPnwCa7GCDkLAQRwgBbO0EALR/hAjpdc/oWulEeQnGJcZK416erqQmFhIaZMmYKpU6caJMSDJwZIHOrg8nRQUJBZ5r7N5ctuDoab9+/q6oJYLOZTFQMCAiCXy+Hq6mqVGw9DTW1IVc3Qm62RLGkbGxuHBMoYehNnbFMc3aEbDhV0O4OcX8fGxiIyMpL/AzFXqAoRYdJkRrrLo6KiEBMTY9SuypikNF2veZLNrvsYxmSYy+VyFBQUwMPDAykpKeO6uCs1DEraZGBYDkIBoNXTbuN35E5gkCBoQAB60Q8XJLp1Ys4kJ6iC7gPr4IrS0lK+KckafuuGQjqwdee3x4NuHKpueXrw3HdQUJDRQS5k2sLLywuzZs2yebDIYJycnBAcHIzOzk44OTkhLi4OMpkMDQ0NKC0thbe3N//aLVG5GI7hTG1YloVEIuH/9tVqtdGmNrqWtCqVSs+S1sHBQS9QZqS/S0N36Gq1GhqNxu7NXOwJKuh2AsuyKC8vR3t7O9LS0oY00YwnE304yB8SwzD82fNw3eWGYGjJnWVZlJaWQiKRDHF+001KM0TMu7q6UFRUZJT9qIZhkdMghUimRrivK1IjfMBxgELNgOWg470+XlisEZzDq84fg/Obhr71X6Ooth0ajQZzU1Lg7OzM530P9lsfj8CZi8bGRlRXVxvcgW0ouuXpuLg4yOVyiMVivuObBKoEBQWNGRMql8uRm5uLwMBAzJw50y467AdDfr97e3txzTXXwNXVFSEhIYiJieErFxKJBPX19XD8b5AOaayzRqWBCLZIJOINgtzc3Ew2tXFxcRlivSuRSFBbW4vS0lJ+QmCwJa2hXe5yuRwAqKAbARV0KzDWRYjMTjMMg2uvvXZYf3Rz7tABoKioCL29vZgzZ864z/kMKbmTUinDMMM6v5EbFd2ktJFKlMRAxJjdZHO3Av84XIkaiRxKDQsNw8LD2QGR/m6Qq8xh2cohHBLc7XwcmuseQ1/ag8jPz4ebm5teEpmuqctggRscJGNp0Ro88mVs74Ex6JanyZns4EAVcu4eEBCgt3Mj3vGTJ08e4k1gL7Asi6KiIigUCmRkZAy5OdOtXJAgHYlEgqqqKigUCoPd+kyFVGIGm++YayxOKBTCz88Pfn5+mD59OhQKBb97H2xJq9FoDPZxB0DP0I1AwHHDnx5SzItarcZwb3VPTw/y8vLg7+8/aszjL7/8gujoaKO7uAfT39+Ps2fPwsfHB2lpaSbtDn/99VdMnTp1xN29oc5vJBJTLBajr68Pvr6+enasHMehqqoKLS0tSE5ONtixjOU4bPmqGMUtvdAyHGQqLYakmpoo5tPRhOd8jyJ94a3onrIc+fn5CA4ORlxcnEEXQiJwYrGYL9kSgfP39zd7eZlMTnR3dyMtLc2mF0tiRysWiyEWi6FSqeDn54egoCA4ODigoqLCbr3jgYGdZmFhITQaDdLS0oxuECONdRKJBN3d3XBzc+NL8yM5FY6H9vZ2lJaWjlmJGWxqo3uzbsxY3GAYhuEDZSQSCZRKJVxdXREREcHv3oejoqICCxcuRG9vr90ds9grVNCtxHCC3tLSgrKyMsTExIw5O33hwgVERESY1CXd3d2N/Px8qNVqXHvttSZ34F64cAGTJ08edrdMmt+mTJky5GxedzcwuMSuVCr5C3xXVxfc3d3587+0tDSjym/l7TJs/aoEHMehrVel8xnTf+VdocCfXXOw/S9/hEPAVIi7pCguLsa0adMwZcqUce0mydkzSYrTaDRmCZIhaLVaFBUVQa1WIzU11Wal/pHQrVzI5XK+pGtJS9bxotVqkZ+fD4FAMO4+jsGPRxrrJBIJWJblj2UCAgLG/bNqa2tDeXk5kpKSjG7KHLx7J9cvY0vzunAch6KiIv7xpVIpXF1d+d27r68vf+Ofk5ODP/zhD+jo6LDKz76+vh7PP/88Tp48ifb2dkyaNAm33347nnrqKatOVJgCLbnbAJZlUVlZidbWVqSmGtb9bGrJvbm5GeXl5YiLi+PjJU1luJL7SM1vup8f7byc3LlHRESgr6+P7xpmWRa5ublGNZb1KLTQsiy65MS72hz3rhziUY+/Bf2K9PWPA4GxvIHIeObgdRk8GtXX16c3FmaKa5lKpUJ+fj6cnJz0jgLsCQ8PD3R3d0OpVGLWrFkQCARDLFnJz94clqzjhRwjOTk5ITk52SwNjo6OjqOOBHp5efG/G4bOgpMjnfEa2xgzFmfo7p18ja+vLyIiInhL2s7OTlRUVECj0cDf3x/nz59HQECAVStIFRUVYFkWH330EWJiYlBSUoL77rsPcrkcb7zxhtXWYQp0h24lNBoNWJbVm/dOS0sz+NwsPz+fd3YyBt2bh5SUFAQEBOD06dNITk6Gn5/fOF7J75CjArIm3ea3tLS0IWezut22YzW/9fT08KN7M2fOBAC+PCsSiaBWq/kLvO7uleM4/FbXjez8NlxqkELcpx6UUT5+/AV92DZDhg0Zk8FNmQ9O4ICqqir+vTWXgchw6LqWdXV1GRQkQ5DL5cjPz4ePjw8SEhLssnxJbgTr6+uRkpKi97up23AlFovR398PPz8/PTtaa6FSqZCXl8eb71jjvdTtO+js7DQoBrelpQWVlZUmudSNxuCxOGN27wUFBQgKChpS2SPz/U1NTdi8eTNKSkrg7u6OBx54ALfccgvmz59v9Ru5119/HR988AFqa2ut+rzjxf5u0ycwvb29yMvLg6+vr1Hz3sD4duhqtRqFhYVQqVR6ZjHjNYQZbk1khz5W8xu5uyfPP5qYd3R0oLS0dEj5WrexjOxey2oaUXSuDFoHN4QF+KBWJsDRsk70qhiYY0fuDAWCIcOmG2biTzcuhEAgAIf/OpYVF0Emk+Gaa66xuKgMdi0j5VndIBnyn+5FjzSWGTMVYG1Ij0RbW9uwQTDDNVyR0jRJESM3duY8ex4MGZ/z8fExKU/BWJydnfW6yQfH4JKbG9JY19LSgsuXLyMlJcViCXnjNbUBRh5bIw2UM2fOxM8//4zPP/8cO3fuhEQiwcaNG6FQKHD8+HHMmTPHIq9pOHp6euw2ZXA4qKBbidbWVhQXFyM6Otpg8w5djB1b6+vrQ15eHjw9PTF37ly9mwdzdcyTkTNdY5qkpKRRY09HC1ghu7S6ujrMmjVr2ChM8hheXl74rVmJXRWdaOxyRr9aDZYTm2037gAGCwSFmBEXh+uuXY6k8N/LnKTKIhAIxkx0swSD40BJkEx9fT1KS0v5pkJHR0dUVlaaJQPAUpBxza6uLmRkZBhUYnVzc9NLiht8c0P6DszpQy6Xy5GXl8c76dnqxmikGFyJRILq6mr+OhETE2PRitHgNQH6N/gjmdqQzYQhxxQMwyA8PByffvopWJZFfn4+ZsyYYbkXMoiamhq88847ePPNN632nKZCBd1K9Pb2IiUlZdzzvsaI8GgNaeSxxuvBPvhxZDIZ6uvrR21+M6TETrqvu7q6MHv27DHdocrbZdhxsgadcg361QwYk2fJBxBCC3cBg203TsGfb1gy5POkfO3t7Y2EhASbG8QIBAL4+PjAx8cHMTExUCgUkEgkaG5uRl9fH1xcXKBSqSCVSu2usYxhGBQXF0OhUIzbl33w2TO5uSGmLj4+PvzufTxpacDAzXFubi7CwsLsrsqhG4NbX1+PmpoaBAYGorGxEbW1tXpjcdaISCXiPpypDTlyJA3CWq121LE4uVzON8EKhUKkp6ePa03PPvssnnvuuVG/5tKlS8jIyOD/3draimXLlmH9+vW49957x/W8toAKupWYOXOmSbtiBwcHaDSaUb+G4zjU1dWhpqZm2IY03ccydYdOGnekUimSkpKMbn7ThRwNsCyLOXPmjNnRe6JCjJeOVaGjV/VfCTfPrtwZakT4uGJZyhRsmD90VKq7uxuFhYUIDw832lXPWri6ukKj0UCpVCIlJQUsy0IsFg9pLBvp7NVaaLVaFBQUgGVZZGRkmOVsdPDNjW7fAUmK07WjNaRkTo7JIiIiMG3aNLv8mQNAQ0MD6urqkJGRAR8fH76pUiKR6Bn6EHG31s2drmBrNBqUlJTAzc0Nvr6+Y5ra6Aq6KWzZsgUbN24c9Wt0e5NaW1uxcOFCzJs3D7t27TL5+a0JFfQrBEdHx1FFmKRQdXd3D3FjG4ypgk520729vQgKChoi5sY0v/X19aGgoIC39Ry841VoGPxYLsb5mi6otSz8PZxwokICcZ8a5hJyV6ixdpIU1827DrHhAQj3HWrsQ2Z5Y2NjERERYZbnNTccx6GiogIikUjvLJqkaJGsb3L2SsaihksLsyRqtRp5eXlwdnZGamqqxaocw/UdSCQSlJaWQqvV6iXFDXcTSTIVSC+HvVJXV4f6+nq9MBhyLOXl5aUXsiKRSPhxO93GOks3m5FqDMdxfP/QWKY25gpmIa/TEFpaWrBw4UKkp6djz549dtlAOhq0y91KaLVak0S0sbGRTxUbjFKpRF5eHoRCoUHzxQUFBfD29sa0adOMXodu85u/vz/UajWSkpIADG1+G+28HAA6OztRVFSEiIiIYZ3A1AyL5w5X4qdKCRQaBhwALcPBXEIOcLjWtxcb473h5uaq1zVO1qJ7rp+YmGhWi1RzohsrmpqaOqZA9/f38/P+UqkUHh4e/Os3R0TmSNiDL7vuSKBEIkFvb+8Qv3USWBMbG4vJkydbfY2GUltbi8bGRqSnpxscYqI7NSCRSCCXy81yNDHa8xUUFECr1Y7YDDy4sY7jOCxZsgTe3t44ffq02dYyGq2trbjxxhsRGRmJzz//XO9GczzW2LaA7tCthKl/ICPtqqVSKfLz8xEUFGRw5+14z9DJWaK3tzcSExPR1NQEpVIJwLjmN2AghOby5cuYOXPmiGY5pyvFOFTcAS1v72Y+IZ/tKsbzf1mKyGA/PudbLBajsLAQAPiLG3Fwy8jIsGgUpiloNBoUFBSA4zjMnj3boN2Wu7s7pkyZgilTpkCj0fAXd3JjOJIdqymQRs2goCCbNpbp7l6nTZvGh4yQxkKhUAitVovIyMgRj61sDcdxqK2tRVNTk1FiDgw/NUB+/jU1NXB2duZ//qYm5bEsq+emN9Ixj25jHcdxePnll1FXV2fVkvfx48dRXV2N6urqITdxV8q+l+7QrQTDMCaFq7S3t6Ourg7z5s3jP0ac5gYns41FWVkZHBwcEBcXZ/DzE7HTbX5raGjgZ87JnfVY5hIcx+Hy5ctoa2sbdRae4zgs3vkr2npUMKeQz3Tvw7+3Loery/DdzxzHQSqVoqOjAy0tLWBZFn5+fggNDbVpkMpIKBQK5Ofn83PRporvcHasuqX58TZWkfE5ez+LbmlpQXl5OR9/aq7Xb044jkNNTQ3vxW/O8BLdowmJRAK1Ws031gUFBRn1+onPvVKpRHp6ukE3mhzH4c0338TOnTvx008/ITk52ZSXc9VBd+hXCLo7dJZlcfnyZbS0tBjsNDfSY40Fx3FoaGjgndB0d9Nkp29o85tWq+W7mkeb3VZqGPznTBHaehQATL/wO0KF67yluGPVzZg91Q+Oo9xwCAQCuLm58V3hMTEx6O7u5huLvL29+dK0uUuTxiKTyfjqjLl2vIPHoojfeFtbGx8kExgYiODgYIODZDo7O1FYWGjX43PAgJsimd8mbn2DXz9pLLP00cRIkGCd1tZWg8f8jGHwSCR5/e3t7aisrISHh4deY91IN+8sy6K4uNhoMd+5cyd27NiB48ePUzEfB3SHbiVYlh2zS300SGzo/PnzUVhYCIVCMe5wjaqqKiiVSiQmJo76daT5TSwWIzU1dchca0tLC6qrq5GQkDDqHzcwsJMsKCiAs7MzkpKS9P7AGZZDSWsvfqntws8VbShr74Ua5pgf5nDHtH787/qbIHAx7H0iIhkQEICZM2fqvSYShUnK8MStLTg4eMzXb27IGe+UKVPG5WswHoYLktG1Yx3u9ZNmwvj4eLstXwPgo4QHu9TpQo4myOsnjm3k9Vt6aoBUtzo6OpCenm71YB2NRsM31pFcdd3GOjLzz7Is38+Rnp5ukBcAx3H48MMP8fzzz+PYsWOYO3eupV/OhIQKupUwVdB7enpw6dIlODs7w8PDA8nJyeO+gNTW1vJz8SNBmt9II4tukxVpfpPL5aiuroZEIuHngYcbCZJKpXw86uAUMrWWxYdnapCd2wCREgDMI4o+Lo7Y+YdEzI4y3N6WNOkZIpK65+5isRiA9UbCSBSmMTGy5oZlWb40KxaLhw2SIR739txMSEY9GxsbkZqaanCUrK5jm0QiQX9/v15p2txTAxzHobKykm+Mtabd7Ujr0W2s6+vr4xsLu7u7oVKpkJGRYbCYf/rpp3jqqafw/fff47rrrrPCK5iYUEG3EqYKemNjI8rKyjBt2jSTzS0aGhrQ2dmJtLS0YT9Pmpe8vLyQmJioJ066ZjHAQIma4zh0d3dDJBJBLBaDYRi+LKvVanm3soiICH7dSg2D/EYpPj98Emd6fAGYb3RpeXww3rxtllHfQ7LWx7OT1B0JE4vFUCgU8PPz4/PdzXnu2tjYiOrqarsSSd2ucbFYDJlMBhcXF6jVasTHxyM0NNQuz8x1y9dpaWlGNZYNhkwNkChUd3d3/ubG1OoNGUeUSCTIyMiw6oihoZCZ/9raWqhUKr2Z/9GClDiOwxdffIHHHnsMhw4dwoIFC6y78AkGFXQrwXEc1Gr12F84zPfV19ejqqoKLMti6dKlJl8cm5ub0dbWhtmzZw/5HGl+i4yMHHLjMLiTfbiLFHHrEolEaGlpgUajgbe3N8LDw/mmslqJHC989TNyOoXQmqWNg4MjGEwP8cH910dhycxgg98j0mDU1NRkVNb6aJAYULFYjJ6eHnh5efHnkoaeOw+3zurqar5vwtCdpLXhOA7l5eVob2+Ht7c3enp6jAqSseY6yY7X3LnwGo1GLwoVgJ4drTEz3+T97OrqQnp6ul2KOTCwztLSUvT29iI1NVXvBofk3A+uXnAch6+++gqPPPIIsrOzcfPNN9v4VVz5UEG3EuMRdIZhUFpais7OTiQlJeHSpUtYvHixyeXc1tZWNDY26p1Tjdb8Rj5vaPMbwzAoKyuDVCpFXFwc+vv70d4hwm813dC2XEK2YhakMM8ImLdAjffXTUNafKzR30vS4aRSKVJTU83aLUwg584ikUjv3N0YcSO9DGSd1j47NRTddZIkQd0gGbFYrOe1PjhIxprrLC8vR3d3t8VFUrc0LRaLIZfL4evrqxeDO9r3lpWVobu7GxkZGXbRYT8cZJ1SqRQZGRl6kyAcx+n5zXd3d+PgwYNQq9WIiIjAjh078M033+CWW26x4SuYOFBBtxLGCrpSqUR+fj4AIDU1Fc7Ozjh+/DgWLlxo8uhUR0cHampqcO211wL4/UIsEomQlpY2pPnNGDFXqVT8LHdycjL6tALUNLfj6a9+QTMCYa4zcjcHAf6xcgZWJ4+v0YrMbrMsi5SUFKuMow0nbuRoYqRzd61Wi6KiIqjVaoNMg2wFwzD8iFJaWtqw69T1WifnriRIhhiaWBrSfU0atqz9fpKZb7FYjO7ubri6uvLirnuDR3a8PT09SE9Pt2sx1w3XGWudWq0WBw4cwGeffYZff/0VQqEQK1aswIoVK7B8+fIRA5kohkHH1qyEMWVWYhYTGBiI+Ph4/vxJIBCYLfaUPA5JDtNoNJg3b96wzW/G2Ljm5+cPXKQjovHq0VIUlJTjMhsMFub6QxUgcZInPt+UBhen8Z27m3t221AGjwSNZMVK5n1VKhXy8/Ph5OSEjIwMm3qvjwa5OQIwqi/7SEEyYrEYVVVVcHd358XNElMD5KbDmIYtc+Pm5oaIiAhERETwjZUSiUQvKS4gIIB3cBu847UnyNm+oWIODFhYe3h44MKFC9i7dy+mTZuGI0eO4L333sO5c+ewe/duK6x84kJ36FZEpVKN+TWtra0oLS3F9OnT9bLAAeDEiROYM2eOSc07wO8hIxkZGUY1v40m5uSiFBkZiRBtC+7bX4cKJgQqOMEcs+SAACEeDnjvj8mInzT+8+Oenh4UFBQgJCQEcXFxdtOsRc4cRSIRenp64O7uDpVKBR8fHyQnJ9s81W0kVCoV8vLy4OrqOiQ61xi0Wi0vbpaYGiBhMBzHISUlxSal/tEgYUcikQhNTU3QarXw9vbmGyvH23thKXR7EIxp1Pvpp5/wxz/+Ebt378bGjRv1XpOhsaqUkaGCbkVIbOBwkBlT0pw1XAfzqVOnRp2TNZSenh5cvHgRAoFg3M1vul+bV1GHH/LqoHZwRpD4V0h6ZMhmr4MGpl40OYQ4AymTnLEmKRTXxE026bxTJBKhpKQE0dHRdh22IZFIUFRUxEefOjk58Tt3Q1PCrIFCoUBubi58fHyQkJBgtnUNnhro7++Hn58f/x4Y+zug0WiQn58PBwcHpKSk2K1okOOA/v5+zJo1iz977+zsNKsdq6mQaxUJATL053H27FmsX78e7777Lu688067ukGZKFBBtyIjCbpGo0FRURH6+/tH7bg9e/Ys4uPjjXaG04XjOFRVVaG2thZJSUkmNb+xLIsfLxRh389FqFZ6Qw4XaOEALRxg6q58up8j3t6YhlAPod7O1cvLi593N8apjYx7zZo1y67P6cRiMYqLi3lXNYZh9EYCybn7eDqmzQnx9bdGpYM0VZFzZ2OCZEiym6urq1WPV4xlNJtU8jtA3gO1Wq3XWGjNkjy5frS3tyMjI8Pgefjz589j3bp1ePPNN3HvvfdSMbcQVNCtyHCCLpfLkZeXBzc3NyQnJ496gf7ll18QHR2NkJCQcT0/aX7r6OiARqMZMgI3OJ941E722jNQH3wSf+1di1+QCBUcIIDgv67r4/9jFUKAZ1dMx9rUcDgIB6WvqdX8rq2zsxOurq68uI+U70xKg+3t7XY97gX8PgufkJAwbLqTblMZ6Zg2Zec6XkiPhzVd6gjErYw01um6tQ0OkiEphJ6enjZLdjMEcravVquRlpY26jVAN+dcLBajt7eXt+MNCgqCl5eXxX4eZHSyra3NKKe6CxcuYM2aNXjxxRfx8MMPUzG3IFTQrYhGo9FLOZNIJCgoKEBERARiY2PH/EW/cOECIiIiRkwnGw3d5rfExET88ssvWLJkCX8BNCjDXKuGQ9GX4H56HjnaKHyiWYYzSAVrcuc6Bx9nB6xKDsOji6Ph6jT2eSlpKBKJRJBIJBAIBLywESMLksFMIkVt7a41EsStrKGhwahZ+MERqJ6envx7YKkLOzkOmD59us1z4UcLkvH09ERJSQn8/PwQHx9vtyLCMAwKCwuh1WqRmppqdMWFjEWS0ryDg4OeHau5KhK6gTDGeMjn5eVh1apVeOaZZ7B9+3a7/TlMFKigWxEi6GPNfI9ETk4OQkJCjL6QEuc3T09PPrv8xIkTuOmmm+Dk5DR685u6H8KWHAjKDkBZ9yvYvg68qt6ILNzw3zNyU/5ABVgQ44cnl8chzMd1yI7cUHQv7CKRCBqNBn5+fpDL5XBychpz12NLSKcw8csfb8Ojrs+4RCKxyLl7W1sbysrKRqwg2BIy7ywWi9HR0YHe3l44OTkhIiLC4jvX8cIwDD86mZqaanLj33A3OKSCExgYaFIFp6amBs3NzUaluxUWFmLFihV44okn8Pjjj9vd+z8RoYJuRTQaDbRaLUpLSyGRSIYNPBmN/Px8+Pn5ISoqyuDvIVUA3eY3juPwww8/4MYbb4STk9PIzW/n/g+NP/8b57VxOItUiDlftMIXUhi+5uERwAHAa7fGY/ms8R0fjATHcRCJRCgrKwMwcNEkNqzGxj9aGoZh+BCL1NRUs5XMic86ubBrtVq+JDteMxfiy56UlGRSD4elkclkyM3NRVhYGLy8vPjjGUdHR70bHFufpet23ZtDzAeje4MjkUgglUr53gNiR2uowNbW1qKxsREZGRkGi3lpaSmWL1+Obdu24emnn6ZibiWooFuRvr4+5OTkABgwizFWXIqKiuDh4YHo6Ogxv5bjODQ2NuLy5cuIj48fEuJx/PhxzJ07F66urvol9u4m4Ld3kVNUjv2a2chnp0MEX2ghBAMBjPNc56C/gxfA00mA2+dG4sEbouDkYP4zTZJCFhERgejoaCiVSn7nTsrSwcHBCA4Otmn8KZndJhd0S1UQyDgUeQ+MPXfnOI6/oBt7A2ptyNl+VFQUpk6dyn+cZVl0d3fzNzgajUZv5t/a8+harRb5+fkQCoVW67rX7T3o7OwEYNhYIDkKSk9PN7h6VFFRgeXLl+O+++7D888/T8XcilBBtyKXLl0CACQkJIzrj7i0tBSOjo6Ii4sb9euItWVHRwdSU1OHjLlxHIdTp07Bz88PkyZNgr+vDxxkLXC4tBuORf/GMXUSvmFuRAU3GR3wA2uS/5AAQgDTAt3wzIpYpEVYzsublITj4uIwefLkIZ/Xjf+USCRwcXHhd+6+vr5Wu/DYytiGPDcRtrE6xklDIYnrtIQ1rrno6upCQUHBmGf7ukEyEokEvb29Vs2412q1yMvLs+kIHcuyena0ZCyQCDzpNamvr0d9fb1RYl5dXY1ly5bh9ttvxyuvvGK3jYgTFSroVkSpVAIwzjVOl4qKCrAsi/j4+BG/Rrf5bXDsKfB785tUKoWovhxOFQcg6KpBndoLbQoBnDgtTrGpqOYmoQVBGN8ZOQdXRyECPVww2c8Na1NCcXN8MFwcLXPx0m0qS0xMNKgkPFz8Kck2Hy0dylRI3npQUBBmzJhh093L4HxvXSc7X19flJeX89aj9hoKAvw+6jdjxgyjG0aHy7gnwmbumX8yD+/o6GhXZkG6XutdXV1wd3eHs7Mzent7kZ6ebvBkSF1dHZYvX461a9fi7bffpmJuA6igWxGtVmuSdWtVVRWUSiUSExOH/fzg5rfBzm8cx4HRaiFoy4fj+bdwob4bP6oTcYmbgX64wgd96IAfRPDD+D3XBfB3c8Dzq+OxMM7yZ62kGtHZ2TnupjKO4wZucP47661SqfiLelBQkNnK4eQ4wBbjXmOhW5YWiURQqVRwdHREdHQ0wsLC7LapkGTDz5o1a9zjnARdr32JRAKGYcwWJKPRaJCXlwdnZ2ckJyfbrdiRuOO2tjb+hoN0zY/2HjQ2NmLZsmVYvnw53nvvPbt9fRMdKuhWhGEYfs57PNTV1aGnpwcpKSlDPjfaCNyAkGsgqDgCp5wP0dXRjDdVq3GCTUcPPMBCCAFYAIL/jqAZKzQDX+/qANwUF4SHbpyKqYHuFhcsYshDgkvM0fDGcRzkcjlEIhFEIhEfIEJK8+PdqXZ0dKC0tBRxcXFD+hnsCbKLJJn2nZ2deiEquiVZW9Pa2oqKigqLZMPr9h6IxWL09fXBx8dHrzRvKLrmNklJSXYtdk1NTaiurkZaWhq8vb2HhOmQ94CE6QgEArS2tmLZsmVYsGABdu3aZdevb6JDBd2KmCrojY2NEIvFSE9P1/t4Q0PDiM1vHMeB1agguPA+XHJ2QaFU4G31GnzDLkA/nMHAEb9Pxg8W4NEE+ffP+bg64K+Lp2FFYhhcxxmYYiwkjc7FxWVINcKcDD5zJrPewcHBBvtrE5c6SwiPORnJl500ForFYr4kS25wxnJqsxSk6z4lJcUsGfZjofsekJQ0Iu6jBcmo1Wrk5uby/RL2LHbNzc24fPnysImLwMB7oBum8+STT2L+/PkoLCzE7Nmz8fnnn9vNMcLVChV0K2KqoLe0tKClpQXXXHMNgFGa31gGgu56oKcZDMuC6+uA84V3IJS14qImGk8o70AzgsEBOqYww12UBRjaqT7w6xLgIkSErzP+NHsSbkmZYtULFTmHDgwMxIwZM6z23OTMmWSbOzk56TXVDV4HcdZqaWmxe5e6/v5+5OXlwdfXF/Hx8SO+p8M5tQ029LE09fX1qKurs1nXvVar1RsLBKBn5kLK0iqVCrm5uXbvVAcMXFsqKyuHbaIdDrlcjn379mHPnj2orKyEUCjE0qVLsXLlSqxdu9auf9cnMlTQrQjLstBoNOP+/vb2dtTV1WHevHlQq9UoLCzk7SLdXF2A//4oBXWnIGgpAMtpAZaDQ8tFCMUVKFcH4x+KDcjlosEZfUb++9evig/A1muD0d05cFEnM77BwcHDCps5IU5lU6dORVRUlM3OoQdnm3Mcp5dtLhAIUFZWBqlUitTUVKtkfY8XmUyGvLw8hIaGGuRYSOCbK//be6BWq/Xm3c09Dkbcypqbm/mSsK0ZKUjGz88Pra2tZg+usQTk6MKYakdnZydWrFiB6dOn4z//+Q9KS0tx+PBhHD58GP/+978RExNj4VVThoMKuhUxVdDFYjEqKyuRkpLye/NbdBicq38ARKWAszs49yAIFJ1gfKIA5wERcSjNQlljB/6f/E7UIxhag8fQhl6EMiK8sPcv6fxFn5iYkIs6x3EW6xZvbm5GZWWl3TmVkYs6OXdXqVRwcHCAg4MDUlNT7Xrcq7u7GwUFBYiKijLpBkl3HEwsFkMmk437zHmkx798+TI6OjqQlpZmt+9pf38/2traUF9fD5Zl4eHhwUegGmPmYi3a2tpQXl5ulJhLpVKsXLkS4eHhyMrKskmuPGV4qKBbEVMFvaurC/n5+eA4bqD5zY+D08+vAl1VgIMr4OACaPvBOnmBib8VjLMnOAjQ21SGv+V64QITAwFYqOEMZtQduv7nnIXA1AB33D4nAquSQuHsOPz3Du4WJ6lQ5II23i5h3dJ1cnKyyfGxlkSpVCI3Nxccx8HR0ZFvJCKleXtpKAN+H/eKjY0ddm7fFIY7d9c9czZG2DiO4ycZ0tPT7eo9HIxSqUROTg78/PwQGxtrcJCMLWhvb0dZWRmSk5MREBBg0Pf09vZi9erV8Pf3x4EDB+zKeZFCBd2qcBwHtVo97u+vrKxEXV0dEhMTER7gBeGZlyFsOAv4TAEndAKn6YNA3gWFRoPC4Ftx2TEGWhbokqtxuI6DlHGFC1TogwtUcIbu2bgQQrg7Ak5OjgNn6xwHXzdn3Jocij9eEw4vV+PEmOzYyK5VLpfD39+fFzZDIx8ZhkFpaSl6e3vtvnQtl8uRn5+vV2YdLGxkx2Zrf3FiwmOOca+x0Gq1esImEAgMFjaWZfmff3p6ul0LCMmH9/f3x8yZM/V+tsTMhfwuKJVKPcc+a78uMnVhjJVvX18f1q5dC1dXVxw+fNhm3gTPPvssnnvuOb2PhYSEoL293SbrsScs0xpMMSssy6KiogJtbW0QCAQIDw+HoL0Ywq4awNkbnJP7gB+7sw80yn6c6pmCsnYHeIcJ4CAAiqQuUAkFEMARQrBwYgTQggPzXxtXV0dgerAXIv1dsDA2GNfF+EOpYeHr7gTncdqzCgQCeHl5wcvLC9HR0ejv74dIJEJbWxsqKir4cmxwcPCIOy7SJ8BxHK655hq7Lu319PQgPz8f4eHhiImJ4S/mrq6uiIiIQEREBN9QJhKJ0NDQwAeoWKP3QBfSdZ+SkmLwzswUHB0dERISgpCQEL0AkcuXL0OlUvGz3oNtWElankKhwOzZs+3659/f34/c3FwEBQUNmw8vFAr5s/XY2FjI5XKIxWK0t7ejsrLSKkl5BDK7b4yY9/f3Y/369XB0dMR3331nc6OhhIQEnDhxgv+3rasd9gIVdDuHeH6rVCqkpaXhwoUL4DgOApUMnIMzBABUDKDmHOECBi2CSahmQxDp2AUXZkAop7m4oUoYDEdHJ8g1jhA6Ai4CARQaDgIBEOjpiikBbkgI88bcqX7wdnWCt5k3DO7u7vw5rUql4g1Mqqur+V2r7ihYf38/8vPz+Q5he/6DJaXrmJgYREZGjvh1Tk5OCA0NRWhoqF7vQXFxMViW5S/oo3lrm4JuU5kxDmDmRCgUwt/fH/7+/nrC1tLSgvLycnh7e/P9F1VVVdBqtcjIyLBbYxtgQOxIEqKhTYUeHh7w8PBAVFQU1Go1X8FoaGiAo6MjX8Ewdx+KSCTixdzQEUqFQoGNGzdCq9Xi6NGjdtG/4OjoaFd9NPYCFXQrYuxdt1wuR25uLjw8PDB37lw+FY1hGAicPdHvFoaLYk/8IJmBBiYAzkIWoawYTo4OiIiMAeswUNaO9PWFn1oNUY8S7o6AVuAIlZaBm7MAM0I88edrIjDJxxUxwR7wcLb8r4SLiwsmT56MyZMn642CNTQ0wNnZGd7e3pBIJJg0adKwux17oqWlBRUVFUY36pHz1MDAQHAch97eXohEItTU1KCkpETveMIcO1PdmFZjUrMsiUAggKenJzw9PTF16lT+Rq+jowNVVVUQCoUIDw/njW3s8feA/I2GhobyaYbG4uzsjLCwMISFhek59lVUVPB9KGRywNCjquEgN57G+CGoVCrcfvvt6O3txfHjx+1isgAYcM2cNGkSXFxcMGfOHLz00kuYNm2arZdlc+gZupVRqVQGfd1wzm8sy+L48eNYsGABhBo5fjx5AlmVKkjUjnCFCgwL9MIDzm4eWDNnBoK9B8pi/f39+LGwHt0aRzi5uEKhYeHm7IDrY/yxaW4kPFzs476OYRjU1taioaEBQqGQ9xYnOzZ7Gv3R9Y9PTk42q7kJcaoTi8Xo7e016HhiNFiWRUlJCWQy2bD+/vYEcVVzcnJCeHg47zEOQK+CYQ8Vm76+PuTm5mLSpEl6xyzmgrgWknN33SCZwMBAg42NgIHrSWFhoVE9E2q1GnfccQdaWlpw4sQJqxj4GMLRo0fR39+P2NhYdHR04IUXXkBFRQVKS0utcoRkz1BBtzJqtRpjveWNjY2orKzEzJkzh3Qf//DDD5g3bx66VMDes5UobpLAHzK4QQEIndAhDEa7ygXTgz2wLD4YfbJeVNa3wNXbD5kZ0QPlbLUWkQHuCPEa/92+ueE4Dg0NDaitrUViYiICAgL0OuZJpjeZ87aUM5yhayW73fH6xxuK7vHEeFzaGIbR8yuw53No4lQ32FWNTE/oNpTpxp+asmsdL0TMw8PDER0dbZXqweAgGWdnZ72M95FueDs7O1FYWIj4+HiDq0gajQZ33303qqqqcPLkSYPP2m2BXC5HdHQ0Hn/8cTz66KO2Xo5NoYJuZUYTdN3mt7S0tGFjT0+ePInQ0FDInf3w7zwRWqRKhHkIIIAAcHRGv5qFSKaCr5sjIjw4yGQ9iJoUgmtnTEJ6pC+Edli2ZFkWlZWVEIlESElJGXK2S0rSRNgUCoXZS9KGQhq1+vv7kZqaatXdrlar1Yt/1a1gDHdBJ77sJHfbljdBY0E6xMdyqhtt12qN+FNgwIgnNzcXERERiI6OtuhzjQQxNiK/D+SGl/xH/iaImM+cORNhYWEGPbZWq8XmzZtRVFSEU6dOWXwKwhwsWbIEMTEx+OCDD2y9FJtCBd3KjCTog5vfBpdWOY4DwzAQi8Voa2tDaaMEJ1odIFELEe7vCRdnJwggQK9SC4VagyluGszyVSN+xkxMCfGDv4d97sy0Wi3fyWyoQOqGp8hkMrOEpxgC+RlxHIfU1FSbNmqRs1ZSwSBhKqSCQXK3bZG5bixyuRx5eXm8la8xgkx2raSC4eLiwt/kjOaxPl56e3uRl5fHJ+bZAyMFyXh4eKCtrQ0zZ840OFaWYRg8/PDD+O2333DmzBmDbwJsiUqlQnR0NDZv3oxnnnnG1suxKVTQrYxGowHLsnof021+S05OHjb2lOSYC4VCCAQCyJVq7Pm5Cj9UdEKrYeDrDDg4OaOfFULIqLAk0gmbbrbvuV2VSqWXDz0egVQqlbyokfAU0jFvzt2aQqFAfn6+XQrk4ApGf38/PzaYmJho12fmpHQdFhY27qYywnAZ97rz7qZWKHp6epCXl8fbDtsrSqUSjY2NaGhogEAggJubG3/uPtp4JMuy2LZtG06fPo1Tp06NOrFhS/76179i1apViIyMhEgkwgsvvIAzZ86guLgYU6ZMsfXybAoVdCszWNA7OztRUFCAyZMnDxt7SoQcGOgK1v28SKbCfy4145fabvTIVWBZLdwFGiT7A5kpoYicFGp3zWSEvr4+5Ofnw8/Pb9QSqzGo1Wq98BRXV1de3E1JBSNhMEFBQUbvIK0N2UF6eHjwQk9GwcxhwWpOyOx+ZGSk2fPhdT3WRSKR3rl7YGCg0Te6RMynTZtm96IhlUqRl5eH2NhYhIWF6Zn6sCyrd5NDbqJZlsVjjz2Go0eP4tSpU3ZTfRiOjRs34uzZs5BIJAgKCsLcuXPx/PPPIz4+3tZLszlU0K2MrqCP1vxGSuzkxzNi+hXDokokR35tO9qam5EWHYzEqFB0SgYuZAzD8CVIe+kO7urqQmFhISIjIzFt2jSLCCTDMPz5olgshoODAy9qozUQjbRWUmK1ZzEnvuy6O8jBJWmyWzP1Jsdca7WWQOqeu/f09MDLy4t/H8aq5EilUuTn5yM6Otpud60Estbp06cPe03p6enh/y76+vrw/vvvIy0tDWKxGD/88ANOnTpFg1WuYKigWxmtVguNRjNm8xvDMHol9tFoamrC5cuXh5yV6c43d3R08GlYxFvdFk1Sra2tKC8vN+pcz1R0z5tFIhFv4jLWTU57eztKS0sxY8aMITnz9gYxDImLixtxrboWrGKxmI8+tfZYIGnUsoSHvCEMruSQc/fhYnC7u7t5gYyIiLD6Wo2BVBFiYmIMWqtMJsPOnTtx4MABlJeXIyYmBuvWrcOqVaswZ84cu7j5pxgHFXQrQ7p5x2p+4zhuSIl9MCSBqq2tbczQksHe6v39/VbtFOc4DrW1tWhsbERSUpLN5kWHS0YjJcigoCC+BEnsUY0x4bAV5CbJmBnjwdGnGo1Gr6nOUg1/xBkvPj7eLhquBsfg6jr2CYVCi4XXmJve3l7k5uYaVUXgOA4vv/wydu/eje+++w6NjY04dOgQjhw5ghdffBEPPvighVdNMTdU0K1Mfn4+FAqFwc1vI6HbHZ6SkmK04cjgTnE/Pz/+vNncc70sy6KsrAzd3d12FSdKRqDI+9DX18ffFPX29iItLc0m9qjG0NDQgJqaGqPiLwdDuqSJuJMgHXMHh5CozlmzZiE4ONgsj2lOdM/d29vboVQq4enpicmTJ9skQMVQiJgbc3zBcRzefPNN7Ny5Ez/99BOSk5P5z5Eqoj03U1KGhwq6lVGpVPzumzBW89tgFAoFCgoK4OzsjKSkJJN3UwqFgrfc7Onpgbe3N0JCQhAcHGzyH7VGo0FhYSG0Wi1SU1NtYgJiKHK5HMXFxZDL5WBZVi/21J6ayQB9X/bU1FSz3nj09/fzzWTkvFn3fRjPuXtzczMuX75sVFSnrZBIJCgqKuL7O8j7QCYogoKCjHJpsyRkJp7kJBgCx3HYuXMnXn/9dfz4449IT0+37CIpVoMKupVhGAZarZb/t6HNbwSpVIrCwkIEBwcjLi7O7Oeeg53JyEUsJCTEaFEjo15ubm5ITEy0a2MTrVaLwsJCaDQapKam8hdy4spFHNqCg4NtGnsK6OeDp6WlWfRmY7jJAd05b0PeB+IAmJKSYtdZ9sDvfueDXdXI+0B+H0hSnrFNluaEiLkxM/Ecx+HDDz/ECy+8gGPHjmHOnDkWXqVhvPzyy3jyySexbds27Nixw9bLuWKhgm5ldAXd2Oa39vZ2lJWV8U0vlhYVjUbDi3tnZyfc3NwMFrWenh4UFBRY7MbDnJB5eCcnpyFHIcDvDm0ikQgSiQROTk78+2Dt0BCWZfkqQlpamlXLwGTOm7wPAoGAF7XhUsGI331jY6PZqwiWgJzvj9WLMNy5+3CjYJakr68POTk5/KSIIXAch08//RRPP/00jhw5guuuu87CqzSMS5cuYcOGDfD29sbChQupoJsAFXQrw7IsNBqNUWKu21A2a9YsmzRpkQ7pjo4OSCQSODs786I2eKdGdjnkTM8eSpMjIZfLkZ+fP6blKIFczMnuHYBex7wlb1xIFYEcX9jSl10311wkEkGj0SAgIEBvgqK6uhqtra1IS0uzqN+9OSBinpiYaNT5vq6pD+k/8PPz4290LHEOLZfLkZOTg8mTJxtsPctxHL744gs89thjOHToEBYsWGD2dY2Hvr4+pKWl4f3338cLL7yAlJQUKugmQAXdyrAsC5VKZXDzG8MwKCsrg1QqRUpKil1cGHVFTSQSQSgU8mX5vr4+VFdXIyEhwe49oImxSXh4+LjSskhoCHkfdDvFzT0WqFar9Vz17On4QneCgsw3Ozs7g2EYsyfRWYKOjg6UlpaaZaKB9B+IxWJIpVJ4enry4m6Ooxoi5saEwnAch3379mHbtm04cOAAFi9ebNIazMmmTZvg7++Pt99+GwsWLKCCbiJU0K3Mnj174OnpiYULF8LNzW3UP0iVSoXCwkIAQHJysl02lOnOeLe1tfGe4pMnT7b4jtUUSBUhJibGLGYhup3iZCyQ5FgHBwebtJtWKpW8+5tuCpk9wrIsioqKIJVK4e7ujt7eXrtsJiOQzntLjCdqNBq9MB1Tz937+/uRk5ODsLAwo25As7Ky8OCDD+Lrr7/GLbfcMp6XYhH27duHF154ATk5OXB1daWCbgbs5zb/KqG1tRW7du2CVCrFsmXLsGbNGixZsmTI2BmxRvXx8UFCQoLdmjwIhUL4+vqiubkZTk5OiI+PR09PDyoqKqDRaHhBCwwMtJvX0NLSgoqKCiQkJBgcJzkWAoEA3t7e8Pb2RkxMDO9M1traioqKCr5j3tjJARJcEhAQgJkzZ9qVGA6GnO/39/dj3rx5cHFx4fswxGIx6urq9MJTrN1/MBjys7FU572TkxPCwsIQFhYGlmX5c/fS0lIwDMPf8AUGBo557k7EPDQ01CgxP3jwIB544AF8+eWXdiXmTU1N2LZtG44fP26344BXInSHbgNYlsXFixeRlZWFb7/9Fu3t7bj55puxZs0aLF26FEePHsVXX32FV1991WpZy+NFrVajoKAAAoEAycnJ/E5Ud8fa0dEBpVKJgIAAhISEGHQBswSkSauhocGqpWClUsmfNRsTIEN82cd7JGBNGIZBUVERb5g0XEVipP4D0kxmzRu+lpYWVFZW2mSMbri5/9HO3RUKBXJychAcHDwk72E0vv/+e2zatAl79+7FbbfdZomXMm4OHDiAtWvX6v3MGYaBQCCAUCiESqWymw3AlQQVdBvDsizy8/ORlZWFrKwsNDU1gWVZ3H333fj73/9uU7/tsSANZd7e3mNWEXRd6vr6+niXOlPL0YbCcRwqKiogFouRmppqs14EUoYlneIuLi7DNhcSD3l7T/YCBpr1SKxsSkqKQTdrpP+A3OjoOvbp5nlbAjITb4oZjzkhPhAkMdDDw0PPuTA3NxdBQUGIi4sz+Fpw4sQJ/OlPf8Lu3bvxxz/+0cKvwHhkMhkaGhr0PnbXXXdhxowZeOKJJzBr1iwbrezKhgq6naDVavHII4/gq6++wtq1a3HhwgVcvnwZN910EzIzM7FixQr4+/vbjbh3d3ejsLBwXLvH/v5+Xtx7e3v5PPPg4GCLlN8YhuFLwYZmrlsD3TEw4q1ObnDq6+ut6nc/XjQaDfLz8+Hg4ICUlJRx7aqGc+yzVMZ9U1MTqqqqkJqaapcz8YPP3RmGgYeHB6ZPn25wT8qZM2ewYcMGvPfee7jjjjvs5poxFvQM3XSooNsJmzdvxq+//opDhw4hKioKHMehsrISWVlZyM7ORnFxMa6//nqsWbMGq1atQlBQkM3+UEloSWxsrMmBFYPL0cSVLCQkxGg72+EggiMQCAzePdoC0lxYV1eH7u5uXtztrf9AF7VajdzcXN44yFxrHO6Igpy7m9JU19jYiJqaGqSmpsLX19csa7UUSqUSly5dgoeHB9zc3CCRSPjRwMG5A7qcP38e69atw1tvvYV77rnnihFzgAq6OaCCbidUV1fzkZaDIXPo+/fvx7fffovc3FzMmzcPa9aswerVqxEWFmaVP1yO41BfX4+6ujqLdAWr1Wo9IxsPDw9e1MZzISdOde7u7mYVHEtB3tukpCQ4OjryO1bSf0B2rPZwU0I67728vJCQkGCxzvvhOsXJ+zA4GW00iFvdleDPr1KpkJOTw3sjCAQC/tydlOZJFSMwMBBKpRIzZszAhQsXsGbNGrz00kt46KGHrigxp5gHKuhXGBzHobGxEdnZ2cjOzsZvv/2G2bNnIzMzE5mZmRZzkGNZVu8MergbD3My+KzZ1dWVF3dD+gpkMhny8/MRFBSEGTNm2PXFjeM4VFdXo6WlBWlpaUPe28H9B6SBylJHFGNBEgP9/f2t2nk/2KGN4zi9hLiRbtjIjVJ6errFf29NRaVSITc3l+9LGem9JVWMwsJC/PnPf0Z4eDjkcjluv/12vPHGG3Z/80qxDFTQr2A4jkNrayu+/fZbZGVl4dy5c0hJScGaNWuQmZmJqVOnmuViq9Vq+Q7m1NRUq4vI4LNmR0dHBAUFISQkZNjRJ9JQRjyu7VnMWZZFeXk5urq6DPJlJw1UIpEIUqmUP6IgHfOWRi6XIzc3FyEhIUZ1XJsb3WQ03SoGKUeTpjrisJienm4XpkyjoVarkZOTAy8vL8yaNcvg9/bcuXN47LHH4ODggLq6Ori6umLVqlV47LHHMH36dAuvmmJPUEGfIHAcB5FIhAMHDiArKwunT59GfHw8MjMzsWbNmnFffJVKJfLz882W7GYqZJ6X7FiJn3hwcDD8/f0hEolQWlqKGTNmIDw83KZrHQuGYVBSUjJuX3bdI4quri7eaz8oKMgi0xEkDIRYjtrLjRJpqiPvhUwmg4+PD4RCIXp7e5GRkXFFiHlubi48PDwwa9Ysg48SSktLsXz5cmzfvh1PPfUUtFotfv75Zxw8eBAPPvgg4uLiLLxyij1BBX0CwnEcurq68N133yErKws//fQTYmJisHr1aqxduxYzZ8406IJBytbE1MTeHMo4juNd6oj1KsdxmDJlCqZNm2bXZUcy6sWyLFJSUkwe0yJe++SIglQxiIGLqT87qVSK/Px8REVFGZzsZSuUSiXKysrQ1dUFAPwYmD0k5Q2HRqNBTk4O3+th6M+qoqICy5cvx+bNm/HPf/7T7l4XxfpQQZ/gkNLkoUOHkJ2djR9++AGTJ09GZmYm1q5di6SkpGEvIJ2dnSgqKroiytYcx+Hy5ctoaWlBcHAwpFIpP9dMjGzsyftcrVYjLy+Pr3qYe226VQxy1myKgUtXVxcKCgowffp0k6caLA3pR2htbUVGRgacnZ31bnTsIfZUF41GozcpYOh6qqqqsHz5ctx+++145ZVXbPY6PvjgA3zwwQeor68HACQkJOCZZ57B8uXLbbKeqx0q6FcZMpkMR44cQXZ2No4ePYrAwEBe3NPT0yEUCvHRRx8hKCgI8+bNQ1hYmK2XPCosy6K0tBQ9PT1ITU2Fh4eHXliISCSCXC7X6xK3ZUqZQqHgu8ONKa2OF3JDR94LtVqtl4o21hEK8byfMWOG3c/EcxyHqqoqtLe3Iz09fUhPga79qkgkAsuyejc61r7p02g0/I1dcnKywb8LdXV1WLZsGT+eZsubkkOHDsHBwQExMTEAgL179+L1119Hfn4+EhISbLauqxUq6Fcx/f39OHbsGLKysnDkyBF4eXlh+vTpuHjxIr788ku7SmUaDhInqtFokJqaOmJ4ja5piUwmg5+fH99IZs3Am76+PuTl5dms8364Gx3i2BcUFDTkvejo6EBJScmY+eD2AKnSiEQipKenj+lhQGJPSRVDoVDA39+fF3hL/15otVrk5eXBycnJKDFvbGzEsmXLcMstt+Ddd9+1eYVhOPz9/fH666/jnnvusfVSrjqooFMADESJZmZmIi8vD8HBwejv78eqVauwZs0azJ8/365K1sDAeE9+fj5/QTR0faRLvKOjAz09PfD29kZISIjRoSnGQqJa7amhjDj2icVi/r0gNzpSqRQVFRUW8RswN8TWVyKRICMjY1w/R92mut7eXvj4+PDibu7pASLmJArX0COQ1tZWLF26FDfddBM++ugjuxNzhmHwzTffYNOmTcjPz0d8fLytl3TVQQWdAqlUijVr1kAul+Pw4cPw8/PDqVOnsH//fnz33XcAgBUrVmDt2rW44YYbbFqyBn73kCfGG+O9sKlUKr0ucd3QFE9PT7Ott7OzE4WFhWaLarUEuu9FZ2cnACAsLAxTpkyxu8hTXTiO48f+0tPTzXJTRt4LsViMzs5OuLu7m216QKvVIj8/H0Kh0Cir3Pb2dixfvhxz587Fp59+alcNn8XFxZg3bx6USiU8PT3tLtntaoIKOgUKhQIvvPACnnzyySG7Ea1Wi7Nnz+Kbb77Bd999B6VSiRUrVmDNmjVYuHCh1WfSyU7X3AlkJOaTCBoZATO1M5qUra8EX3Zg4Hy2rq4OU6ZMQV9fHzo7O+Hs7Gw3kae6cByHsrIydHd3IyMjwyK/i4OnBxwcHPj3wtimOoZhkJeXZ7SYi8Vi3HLLLUhKSsIXX3xhd9UytVqNxsZGSKVSZGVl4eOPP8aZM2foDt0GUEGnGAzDMDh//jwf+9rT04Ply5djzZo1WLx4sVm810eDNGhZeqdLLuIdHR2QSCRwdnYeNhFtLEiq15VStq6pqUFzc7OeW93gyNPBc/+2KvtyHMc3Q6anp1vlxpL47ZMbP4Zh9BLiRhNahmGQn58PAEhNTTVYzDs7O7FixQpMnz4d+/bts7kPhCEsXrwY0dHR+Oijj2y9lKsOKuiUccGyLC5cuMCLe0dHB5YuXYrMzEwsW7bMrCVr4Pf86oSEBKs2aA0naETcR9qhEc/7+vp6pKSk2GWqly6koayjowNpaWkj/uxYloVUKuXfC61Wq2e9aq2dI5lskMlkSE9Pt2pjI4E01RFx7+/vH7HBkGEY3nMgNTXV4PdJKpVi5cqVmDx5Mvbv32/zoy5DWbRoESIiIvDZZ5/ZeilXHVTQKSZDMt3379+P7OxsNDU1YfHixcjMzMQtt9xi0rkjx3Goq6tDQ0MDkpOTbZpfTXZopEuczHcTQRMKhbw4tre3Iy0tze4dysgZdGdnp0Hd4brfpytopEvc0qOBLMvy7nrp6el2I3LDNRiScbjq6mpotVqkpaUZLOa9vb1YvXo1AgICcODAAZvctBjCk08+ieXLlyMiIgIymQz79u3DK6+8gmPHjmHJkiW2Xt5VBxV0ilnhOA4lJSX45ptvkJ2djerqatx0001YvXo1Vq5cCT8/P4PFnYiNRCJBamqqXYkjme/u6OjgXeoCAwOhUqmgVCqNEkdbQXa6vb29JpetB48GWiLPnGVZPtfensR8MLqWvBKJBEKhEJMnT0ZISIhBRzZ9fX1Yu3Yt3NzccOjQIYtOX5jKPffcg59++gltbW3w8fFBUlISnnjiCbOJOcMwdtUAaO9QQadYDDJORGJfS0pKcMMNN/CZ7oGBgSNe3BiG4S/eqampdn1RI+JeWloKhUIBgUBgd3GngyHvr0KhMLs4DpdnrhsgM55qDcuyKCoqglKpRFpamt2KOYFlWRQWFkKlUiEyMpI3tBEKhaP2IPT392PdunUAgO+//94qgTv2Csdx/O/Ku+++i7a2Njg7O+PJJ5+0y78pe4AKOsUqkKYrIu55eXm49tpr+Uz30NBQ/o+3p6cHlZWVEAgESElJsfs/Xo1Gg8LCQv6MVKVS6cWdklJ0cHCwXQgROdNlGAapqakWfX/VajUfg9vZ2Wl0DC5Zb1FREdRqNdLS0uz+90H35iM9PZ1f7+AeBFLVkclkiI2NhaenJzZs2AClUoljx47ZVUXK2uiK+aOPPoo9e/bg2muvxS+//IIZM2bgk08+oV30w0AFnWJ1OI5DQ0ODXqb7nDlzkJmZiZSUFNx777146KGH8PDDD9t9uU3Xl304kxBytqprWEKMbGyRZa7RaFBQUMDfLFlzBIphGEgkEn7G25ARMIZhUFhYCK1Wa/GbD3NAjgVI5WOk9XIcB5lMBrFYjMcffxwnTpzAlClToNVq8f3331Ox+i9tbW3YvHkzXn31VcTFxaGrqwuLFi2CQCDAv/71LyQmJtp6iXYFFXSKTSGZ7tnZ2di7dy+Ki4uRmJjI+8vbczAM8WX39vZGQkLCmCNcg0vR1s4yH+vmw5oMbjAkvuqkwdDBwWHc3eG2YrwNe2q1GnfccQdqamrg7++P3377Dampqbj99tuxbds2C6/aftmxYwc+/fRThIaG4osvvuCnW2QyGRYsWACVSoV//etfSElJse1C7Qgq6BS74PTp01izZg0eeOABTJkyBdnZ2Thz5gwSEhL4TPfp06fbjbgTX/bg4GDExcUZvS7dxqnOzk54eHjoudSZ+3WqVCo+b9uYVC9rMDhARqVSISAgAHK5HI6OjkhPT7d7MSfNoDKZjE95MwSNRoO77roL1dXVOHnyJAIDAyGRSHD48GF0dXXh0UcftfDK7ZecnBz84Q9/QFdXF3799VfMmDEDLMtCKBRCoVBg0aJFKC0tRV5eHqKjo229XLuACjrF5vT29iImJgavvfYa/vKXvwD4PdP9wIEDyM7OxokTJxAbG6uX6W4rcSfZ4JGRkZg2bZrJ69BoNPw5s0QigYuLC1+WN9VqFBioJOTm5ppslWsNiLgXFxdDo9GAZVk+TCcoKMgmxxRjQUxuyLSAoSNmWq0WmzdvRlFREU6fPo3g4GALr9R+0T0z16WsrAxLly5FbGws/vOf/yA4OJj/WpVKhaeeegpvvPGGDVZsn1BBtzL19fV4/vnncfLkSbS3t2PSpEm4/fbb8dRTT9lFw5St6OrqGnHGnFzkDx48yGe6R0ZG8mV5a+44JRIJioqKLOZWxzAMbzVKzpl1jWyMFXe5XI68vDwEBgbaJOHNWEhwiYODA1JSUqBWq/mdu26AjCVCU8YDsZ+VSqXIyMgwWMwZhsFDDz2Eixcv4vTp0zaNKX755ZeRnZ2NiooKuLm54dprr+XPrK2Brpi3tbWBYRiEhobyVZny8nIsXboUMTEx+PLLLxEaGjriDcDVDhV0K3Ps2DF89dVX+OMf/4iYmBiUlJTgvvvuwx133EHvNA2kt7dXL9M9ODiYL8uTTHdL0N7ejtLSUsTHx1vlAkzyu4mgAeDF3RDb1b6+PuTm5mLSpElm9b23FBqNBvn5+SOmkBFxHxyaYqrf/nghPgnd3d1GzfGzLItt27bhzJkzOHXqFCIiIiy80tFZtmwZNm7ciNmzZ0Or1eKpp55CcXExysrKLH7TpCvMu3btwvvvvw+5XA6O47Br1y7MmTMHHh4eqKqqwrJlyxAREYEvvvjC5u+ZvUIF3Q54/fXX8cEHH6C2ttbWS7nikMvlepnuvr6+WL16NTIzMzFnzhyzNX41NTWhqqoKSUlJCAwMNMtjGgPHcXpNZAzDDGki04WE2ERGRtp1YyFBo9HoNeyNdbOi1Wr1jimcnJx4cbdGgIxuypsxwTAsy+Kxxx7DsWPHcOrUKURFRVl0neNBLBYjODgYZ86cwQ033GCV5/z000/x6KOP4tVXX8WyZcvwwAMPoKysDM8//zzWrVsHDw8P1NfXIykpCX/4wx+we/duq6zrSsO+O02uEnp6emxqaXol4+HhgXXr1mHdunVQKBT48ccfkZWVhQ0bNsDV1RWrVq3C2rVrce21146rsUrXejYtLQ2+vr7mfxEGIBAI4O/vD39/f8TFxaG3txcikQiXL1+GSqXiPdWDgoIgk8lQUFCAadOmYcqUKTZZrzGQ7ntXV1ckJSUZVGFxdHREaGgoQkNDwbIsf0xRWFgIAHrmLebu5uc4DpWVlejs7DRazJ988kkcPnwYp0+ftksxBwauRwCsdk2qqqrC7t27sWPHDvzlL39BXl4ecnJyEBkZifvuuw8ajQa33XYboqKiUFVVhYCAAKus60qE7tBtTE1NDdLS0vDmm2/i3nvvtfVyJgxqtRonT57kM90FAgFWrlzJZ7obMs98JfiycxyHvr4+fudOypVhYWGIjY21+74MtVqN3NxcuLu7m6UXguM43rxF15I3ODh4zEQ0Qx//8uXLEIlEyMjIMNjBkGVZPPvss/jyyy9x6tQpq51PGwvHccjMzER3dzd+/vlnqzxnbW0tDh06hM2bN6OxsRFLlizB3XffjWeffRYrVqxAQUEBHnvsMdx333120Tdhz1BBNxPPPvssnnvuuVG/5tKlS8jIyOD/3draihtvvBE33ngjPv74Y0sv8apFq9XizJkz2L9/Pw4cOAC1Wq2X6T5cIxPLsnyzU1pamt37sgOASCRCUVERgoODoVAoIJPJ7LpDnIzSeXp6YtasWWbvfSDmLeTcXS6Xm+Tap5tKl5GRYVSQzUsvvYSPP/4Yp06dsmvTmIcffhhHjhzBuXPnMHnyZKs9b2NjIyIjI7FlyxZ0dnZiz549cHV1xQMPPIAjR47A398fOTk5dm8sZGuooJsJiUQCiUQy6tdERUXxF9XW1lYsXLgQc+bMwWeffWbXo0QTCYZhcO7cOT72VSaT6WW6u7m5oa+vD3v27EFGRgbS0tLsNulKl7a2NpSVlSExMZEff1IoFBCLxejo6NDrEA8JCbG5Nz4Rcy8vL4NMecyBXC7nZ/+Jax8R97HeD47jUF1djba2NqPF/I033sA777yDkydPIikpyRwvxSJs3boVBw4cwNmzZzF16lSLPAdpgquqqoJcLodEIsHixYv5z69evRpRUVHYuXMngIHwl61btyI+Pt7uq032ABV0G9DS0oKFCxciPT0d//rXv+ze3nSiwrIsfvvtN17cxWIxbrrpJlRVVcHR0RGnTp2yufAZQnNzMy5fvozk5OQRzxdVKhUvZl1dXXqBKebOrh8LpVKJ3Nxc+Pj4ICEhwSYNe8MFyJBz98HGPiSHoKWlBRkZGQaXfTmOw86dO/H666/jxx9/RHp6uqVejklwHIetW7fi22+/xenTpzF9+nSLPY9AIMCRI0fw6KOPwsXFBd3d3YiOjsZbb72FtLQ0bN++HZ9//jkeeOABFBcXIycnB6WlpbTHyECooFsZUmaPjIzE559/rifmoaGhNlzZ1Q3Lsjh+/Dg2bdoEgUAAhUKBBQsWIDMzE8uXL4ePj4+tlzgsDQ0NqK2tRUpKCvz8/Az6Ho1Go+dS5+bmZrXxL6VSiZycHPj5+SE+Pt4uuu+HM/Yh74ePjw9qa2vR3NxstJh/+OGHeOGFF3Ds2DHMmTPHwq9i/Dz00EP48ssv8d133+md7fv4+Jj9hvbSpUtYvHgxXnnlFdx7770oLCzENddcg08++QR33XUXAOD+++9HWVkZfHx88OGHH1q19H+lQwXdynz22Wf8L+5g6I/CdtTV1WHJkiWYN28ePv74Y73YV5LpnpmZiRUrVozL4MXckO77xsZGpKamjvuGQ6vVorOzEx0dHXrjX4ZmdxsDcazz9/e3qdPfaOgaRvxE6gAAJutJREFU+0gkErAsC47jEBcXh0mTJhl0NMBxHD755BP8/e9/x/fff4/58+dbYeXjZ6Sfw549e3jnRnPxzjvv4NKlS/j888/R0NCAxYsX46abbsJHH32k93UymQyurq70zNxIqKBfZbz44os4cuQICgoK4OzsDKlUausl2QXff/89Tpw4gTfeeEPvok3mjYm4l5aW4sYbb8SaNWuwcuXKUTPdLQXHcaiqqkJbWxvS09PNVjJnGIY3shGLxRAIBHoudaaccysUCuTk5CAoKGhc3ve2oLa2FvX19QgMDIRUKgXDMHod88MdlXEchy+++AKPP/44Dh48iAULFlh/4XZCf38/+vr69Oxa7777brAsi88++wzR0dF8Q7BQKMSuXbsglUrx+OOP23rpVyxU0K8y/vGPf8DX1xfNzc345JNPqKAbAWmMysrKQnZ2NvLz8zF//nw+0z0kJMQqhiYVFRWQSCRIS0uz2BjP4DQ0juPGPdvd39+P3NxcBAcHIzY29ooQ8/r6etTX1yM9PR1eXl7gOI6f/ReJRFAqlQgICODF3dnZGRzHYd++fdi+fTsOHDiARYsW2fpl2Iz+/n6EhYUhKioKhw8f5p3dDh06hJdffhn19fVYtGgRvvjiC17st23bhq6uLnz00UdXxFSJPUIF/Srls88+w/bt26mgjxOS6U7E/cKFC5g7dy4yMzORmZmJ8PBwswuX7ihdenq61Rr2iJd+R0eH3mx3SEgIAgICRp3tlsvlyM3NRWhoqF2l5Y0G6UtIT0+Ht7f3sF/T19fH9yG89dZbaGlpQVJSEr7++mtkZWVh+fLlVl61fVFWVoakpCSwLIuEhAQcPHgQU6dORVlZGbZv346amhq89tprWLduHbq7u/HOO+/g3XffxenTp+16rM/eoYJ+lUIF3XxwHIeWlhZkZ2cjKysLv/zyC9LS0nhxj4qKMlnIWJZFcXExn7Vtq1E63dlukUgEhULB71SDgoL0zjzlcjlycnKuGC95YGAeuqamZlQxH0x1dTV27NiB77//nq+crF27FmvXrsWMGTMsvGL7g+M4KJVKbN26Fb6+vigtLUVJSQlOnjyJ6dOn4+eff8YLL7yA6upquLq6IjAwEDU1Ndi/fz/mzp1r6+Vf0VBBv0qhgm4ZOI5DR0cHvv32W2RlZeHs2bOYNWsWL+7j2aUyDIPCwkKo1WqkpaXZ1TyurktdX18fb9zi7u6O4uJihIeHIzo6+ooQ86amJlRXVyMtLc2oJsPvv/8emzZtwueff44bb7wRhw4dwrfffovKykpUVFRcEa/dXOiGrXzyySd49tlncfDgQfzjH/9AXl4ejh8/jvj4eFy+fBm1tbU4deoU0tPTkZycbLfueVcSVNAnAONxqaOCbnk4jkNnZye+++47ZGVl4aeffkJsbCyfDGdIp7dWq0VBQQE4jkNKSopdd/329/dDJBKhra0NfX19cHV1xZQpUxAcHGx3LnWDIWKemppqlF//iRMn8Kc//Qkff/wxNm7cqPc5hmGuKo+J7u5ueHt7673mDRs2ID09HZs2bcLtt9+O0tJS/PDDD3ZtsHMlQwV9AmCsSx1ABd3aEI/xQ4cOISsrC8ePH8eUKVP4TPfhrE9JnCjJBr8SxKG3txd5eXkIDw+Hq6srb9zi5eXFd8zbmx83MeYxNnznzJkzWL9+Pd5//33ccccdNt2Jnz17Fq+//jpyc3PR1taGb7/9FmvWrLHa88+cORN9fX3YsGEDFi1ahFtuuQUA8O677+Krr77Czz//jL6+Ptx2220oKSnBgQMH9DYYFPNA09YmAIGBgTaJ9KQYjkAggJ+fH+68807ceeedfKZ7VlYWFi9ejJCQEH7nnpaWhra2Nvz973/HQw89dEWJeW5uLqZOnconiUVERECtVvMNZDU1NfDw8NBzqbOlELa0tODy5ctG78zPnTuHDRs2YMeOHTYXc2CgXyE5ORl33XUX1q1bZ9XnrqqqQlNTEz+mdscdd2D16tW4/vrrsWXLFnz88cfYsWMHtm/fjn379mHDhg1YunQp6uvr7TLw6EqG7tCvMhobG9HV1YWDBw/i9ddf5xOVYmJirG4BShlALpfj6NGjyMrKwvfffw9PT0+wLIu4uDjs37//ihjh6enpQV5e3piRrVqtlhd34soWEhKC4OBgeHt7W1UYW1tbUVFRgZSUFKOsRS9cuIA1a9bgpZdewkMPPWRzMR+MQCCw+g49Ly8PmZmZuO666/DII4/gyJEj+Omnn6BQKAAAISEh2L9/P7y8vCCVSlFfX4+UlBSrre9qgQr6VcZf/vIX7N27d8jHT506dVWbYNgLZWVlWLhwIQICAtDe3g43NzesWrUKa9asGXemu6WRSqXIz89HdHQ0IiMjDf4+XVc2sVgMBwcHPSMbSwplW1sbysvLR/W/H47c3FysXr0azz77LB555BG7E3PANoIODIj6zTffjKVLl+Ktt95CSEgIduzYgePHjyMxMRGvvvqqVddzNUIFnUKxE8rLy7F48WJs2LABb731FjQaDU6cOIGsrCwcPHgQQqGQz3S//vrr7aJBrru7G/n5+Zg+fTpvHjIeWJblXepEIhEA8OLu7+9v1jS29vZ2lJWVGS3mhYWFWLFiBf72t7/hr3/9q12KOWA7QQeAoqIiLF26FLNmzcL+/fvh4+MDmUxGS+tWggo6xeK8//77eP3119HW1oaEhATs2LED119/va2XZXfU1dVh//79w4qFRqPRy3TXaDRYuXIlMjMzR8x0tzRdXV0oKChAbGysWQM0OI7Tc6ljGIZ3qQsICDCpn6CjowMlJSVITk42qu+ktLQUy5cvx/bt2/HUU0/ZrZgDthV0YODGdOnSpZg2bRq++uorhISE2GQdVyNU0CkW5auvvsIdd9yB999/H/Pnz8dHH32Ejz/+GGVlZUaVZym/QzLdibjLZDLccsstWLNmDRYtWmQVB7nOzk4UFhZixowZmDRpksWeR9dytaOjAyqVivdTDwoKMuoIQiQSobi4GElJSQgKCjL4+yoqKrB8+XLcf//9eO655+xazAHbCzoAXL58GStXroSfnx+ys7MRHh5us7VcTVBBp1iUOXPmIC0tDR988AH/sZkzZ2LNmjV4+eWXbbiyiQHJdCfiLhaLsWzZMmRmZmLp0qUWGRGTSCQoKirCzJkzERYWZvbHHwmO4/SMbORyuZ5L3WiGO0TMExMTERwcbPBzVlVVYfny5bjjjjvw8ssvm7X0byksJegsyxr1+mtqajB//nzs2rULq1evNutaKMNDBZ1iMdRqNdzd3fHNN99g7dq1/Me3bduGgoICnDlzxoarm3iwLIvc3Fw+Ga6lpQVLlizhM90NtTIdDbFYjOLiYsTHxyM0NNQMqx4/crmcF3eZTAY/Pz9e3HU9F8RiMYqKijBr1iyjyr91dXVYtmwZbrvtNrz55pt2LeZ9fX2orq4GAKSmpuKtt97CwoUL4e/vb5ZKmFar5ash9fX1cHNzg7u7Ox9cM1LVQi6X253vwESGCjrFYrS2tiI8PBznz5/Htddey3/8pZdewt69e1FZWWnD1U1sWJZFUVERL+41NTVYtGgRn+nu6+trdOmYiHlCQoLdnYsqFAqIxWJ0dHSgp6cH3t7eCA4OhpOTEyoqKowW88bGRixduhQrV67EO++8Y9diDgCnT5/GwoULh3x806ZN+Oyzz0x6bF3Huw0bNqC6uhrd3d1YsGABtm7dirS0tFFFnWI9qKBTLAYR9F9++QXz5s3jP/7iiy/iiy++QEVFhQ1Xd/Wgm+menZ2NsrIyLFiwgM90DwgIGPNiPN6StS1QqVQQi8VoaWlBb28vXF1dER4ezhvZjEVrayuWLl2KRYsW4cMPP7R7MbcWy5YtQ2dnJ9577z00NDTgmWeegZubG/7v//6Pb3Klwm5b6G8qxWIEBgbCwcEB7e3teh8XiUR2t8ObyAgEAsTHx+OZZ55Bfn4+SktLsXDhQuzZswfR0dFYuXIldu3ahfb2dgx3f086w5OSkuxezAHAxcUF7u7ukMvlmDlzJqKjo9Hb24sLFy7gl19+QXV1NXp7e4d9re3t7VixYgWuv/56fPDBB1TM/8vOnTvR09ODn376Cddccw0KCgogFovh6+uLhx56COfPnwcAKuY2hu7QKRZlzpw5SE9Px/vvv89/LD4+HpmZmbQpzsZwHIf6+no+0/3SpUuYO3cuVq9ezWe67969G21tbXjooYeM6gy3JWScbnAHvlarRWdnJzo6OiCRSODk5ITg4GDIZDKkpqaiq6sLt9xyC5KTk/H555/bpYmPrTh37hx+/vln/O1vf8Pzzz+P3bt34/DhwxCJRFi/fj2CgoLwxhtv0OY3G0MFnWJRyNjahx9+iHnz5mHXrl3YvXs3SktLR7UIpVgXjuPQ3NyM7OxsZGdn45dffsGsWbNQUVGB1157DXffffcVsfsiRjdxcXGjjkoxDIOuri60tbVh9erV0Gq1CAwMREhICH744Ycrwm7XUozUzd7X14f+/n6sWbMG27dvx4YNG9Dd3Y1bbrkFzs7OuO2227B161YbrJhCoPUkikX5wx/+gB07duCf//wnUlJScPbsWXz//fdUzO0MgUCAiIgIbNu2DadPn8Zrr72G8vJypKen49FHH8UNN9yAN954A1VVVcOWqu0BYkEbGxs75tyzg4MDgoKCkJSUhIsXLyIhIQFCoRA1NTWIiIjAXXfdhdOnT1tn4XYEwzC8mBcXF+PkyZOor6+HVquFp6cnRCIRysrK+M71hoYGhIWF4bXXXqNibgfQHTqFQtFj9+7dePTRR3H48GHccMMNkEgkfKb7yZMnERcXh8zMTGRmZhqU6W4NSDhMTEyMURa0vb29WL16NQICAnDgwAE4OTnht99+Q3Z2NiZPnozt27dbbtF2hm5D21133YXi4mJUV1djzpw5CAgIwMcff4z+/n7cddddkEqluOmmm/Dpp58iMzMT7777ro1XTwGooFMmOLbOib4S2bt3L6ZNmzbEnpdkuh88eBBZWVn48ccfERUVxce+Dpfpbg2ImBsbDtPX14c1a9bA3d0dhw4dsorD3pXAE088ga+++gqHDx9GQkIC1qxZg5ycHJw4cQIzZ87EoUOH8J///AcNDQ249tpr8frrr9t6yZT/QgWdMqE5evQozp8/j7S0NKxbt44Kuhnp7e3F4cOHkZWVhWPHjiEsLAyrV6/G2rVrkZqaahVxJxnsY8W2DkYul2PdunUQCoU4cuSI3Zif2CL3gOzMOY6DTCbDunXrsGXLFmRmZuLTTz/F//zP/+A///kPbrnlFshkMnh4eEAoFKK/v/+q7jWwR6igU64a7MHjeqLS19enl+nu7++PVatWYe3atZg9e7ZJgSojIZPJkJubi6ioKERFRRn8fQqFAhs2bIBKpcLRo0ftJgnMFrkHug1wYrEYbm5uWLRoET755BNcunQJ27Ztw969e7F27VrI5XLs2rULycnJWLBgAR3ps0OooFOuGqigWweFQoEffvgB2dnZOHToEDw8PPhM93nz5pllHIyI+ZQpUzB16lSDv0+lUuGPf/wjpFIpfvjhB/j4+Ji8FnNhy9yDBx54AH19fXjllVewbt06CAQCVFVV4cMPP8T69esBDITU3HPPPfif//kf3HbbbRZdD2V80FssyphwHAeWZW29DMoVgpubG9asWYPPP/8cbW1t+Oijj6BWq/HnP/8Z06dPxyOPPIJTp05Bo9GM6/H7+vqQm5uLyMhIo8RcrVbjzjvvhFgsxtGjR+1KzNVqNXJzc3HzzTfrffzmm2/GL7/8YtHnbmpqwoULF/DAAw9g8uTJePfdd9HY2IikpCSsW7cOfX19aG5uxp133omQkBAq5nYMdU6gjIparYazszPf/cpxHDiOo+U2ikG4urpixYoVWLFiBT788EOcPn0aWVlZuOeee6DVavlM9wULFhiU6U7EPCIiAtOmTTN4HRqNBnfffTcaGxvx008/wc/Pz5SXZXYkEgkYhhnioBgSEjLEadGcvPXWW8jPz0dycjIyMjIADIS77Ny5E3/5y1+QmpoKhmHg5uYGHx8fZGdnW2wtFNOhgk4ZlWeeeQZnz57FwoULcc8992DatGl2MaZEufJwcnLCkiVLsGTJErz77rt8pvvWrVvR19eHFStWIDMzc8RMd7lcjtzcXISHhxsl5lqtFps3b0ZlZSVOnTqFwMBAc74sszL4b8uS3ugMw0CtVmPfvn1644eOjo647bbbMG/ePHz11VcQCoWIjIzErbfeapF1UMwH3WZRRkQsFqOiogJyuRy//vor4uPjMW/ePJw4ccLWS6Nc4Tg6OmLBggV499130dDQgCNHjiA4OBhPPPEEpk6dik2bNuHbb7+FXC4HMGBysmPHDkyaNAnR0dEGixzDMHj44YdRUFCAEydO2K0XvTVzD8jxmYODA7Zu3Yr3338f5eXleOGFF/iv0Wq1CA8Px6OPPort27dTMb9CoIJOGZFLly5BJBJh27ZtOHnyJBoaGjBr1iy88sor6OnpsfXyDKKvrw8FBQUoKCgAMJBxXVBQgMbGRtsujMLj4OCA+fPn4+2330ZNTQ1OnDiBqVOn4tlnn0VUVBRuu+02rFixAs3NzYiJiTFYzFmWxbZt2/Drr7/ixIkTCAsLs/ArGT/Ozs5IT0/Hjz/+qPfxH3/8US96eLzk5OTgtddeg0wm0+uH8fDwwO23346dO3fi5Zdfxj/+8Q8AAzdcLMvarSsgZQQ4CmUEnnnmGe7666/nJBIJ/7Gvv/6a8/f3577++muO4ziOZVn+f7VaLf9ve+HUqVMcgCH/bdq0ydZLo4wBwzDc4cOHOW9vb27atGmci4sLt2LFCm7Xrl1cS0sL19fXx8nl8mH/k8lk3P33389FRUVxdXV1tn4pBrFv3z7OycmJ++STT7iysjJu+/btnIeHB1dfX2/S40qlUi44OJgTCATcNddcw/31r3/lTp48qfc1crmc27VrF+fk5MQ9+eSTJj0fxXZQQacMS2dnJ7d69WrO2dmZu/POO7lTp05x3d3d3EMPPcQJBAKusLCQ/1qNRqP3vfYm6pQrk/r6ei4qKorbunUrxzAMV1JSwj377LNcUlIS5+TkxN18883c+++/zzU0NOiJu0wm47Zs2cJNnjyZq66utvXLMIr33nuPmzJlCufs7MylpaVxZ86cMfkxtVot97//+7/cp59+ymVnZ3Pbt2/n3NzcuPvvv597//339b72008/5QQCAff222+b/LwU60MFnTIsx44d4+bOncvde++93IMPPsh5e3tzTk5OnIODA7d8+XL+6xiG4R588EFu5cqV3BtvvMGJxWIbrto+eemll7iMjAzO09OTCwoK4jIzM7mKigpbL8vu+frrr7ktW7YMuUFkWZarrKzkXnzxRS49PZ1zdHTkFi5cyO3YsYOrqanh/t//+39cWFgYV1lZaaOV2x+fffYZFxYWxjU3N3Mcx3GVlZXc448/zgkEAu6GG27gdu7cyVVVVXEcx3FHjx7lFAqFLZdLGSfUWIYyLM899xyOHz+O3bt3Iz4+Hj/88AMee+wx/OEPf8Ddd9/Nn0c2Njbis88+g4uLC77++muoVCq88847WLhwIf9YnI61pFqthouLC7788kuUlpZiy5Ytdn22aQ6WLVuGjRs3Yvbs2dBqtXjqqadQXFysl1pFGR8cx6Guro7PdL948SJcXV1x6dIlxMfH23p5dsWdd94Jb29vPkhl4cKFUKlUiI+PR21tLU6fPo1du3bh3nvvtfFKKeOFCjplCFKpFPfffz+EQiH+85//8B9fsWIFHnzwQaxcuRIMwwxr57lx40b09/fju+++05tXl8lkehabDz30EC5cuIADBw4gIiICb731FhISErB06VLLv0AbIxaLERwcjDNnzuCGG26w9XImDBzHoaCgAPX19Vi7dq2tl2M3kBvqXbt24ZtvvsGPP/6IW2+9FUVFRThz5gzCw8NRX1+PkydP4tZbb4Wvr6+tl0wZJ7TLnTKEX375BeXl5ZgzZw6AgREWAHq2lA4ODqivr8czzzyDVatW4d5770VOTg5uvPFGaLVa1NfXQygUor29Ha+++ipuvPFGREZG4pNPPkFJSQlqampw/fXXIyIiAizLws3NDb/99hv/XBMZMiHg7+9v45VMLAQCAVJTU6mYD4JMBdx3333o6OiAs7MzCgsLcfToUT43PioqCnfffTcV8yscaixDGUJwcDDmzp3LpzyRMRcXFxdIJBIAwOXLl/H000/j119/xbZt21BcXIyVK1dCJpMhJSUFoaGhAAZ24rm5udi8eTNCQ0Nx7Ngx5Ofng2EYzJ49GwAgFArx4IMP8s/PTWA3Oo7j8Oijj+K6667DrFmzbL0cylUCCWF5+umn8fTTT+O9997D9OnTbb0sipmhgk4ZQkZGBm8DCQzMyAIDd/HNzc1obm5GS0sLb/axbt06AEB+fj42b96MxMREuLm54cyZMzh8+DAOHTrEl9KdnZ2xdetWJCUl8TcML774IsLDw7Fx40a4uLhAIBBMWDe6LVu2oKioCOfOnbP1UihXEeTm+JprroFSqUReXh6WLFli41VRzM3E2wJRLMZ1112HGTNmQCAQIDY2FgqFAh0dHQAGHK2++OILNDc3Y/HixQCAzz77DHPnztU7J54/fz4YhkFKSgoiIyOhVCpx8OBBnDhxAizLQiAQ4OWXX+YfVxeGYazzQi3E1q1bcfDgQZw6dQqTJ0+29XIoVyFRUVF4+umn8eKLLyIvL8/Wy6GYGSroFIOJiorCTz/9hPDwcISGhuKuu+7CY489hvT0dGzfvh179uxBSEgIli1bBgC4cOECbrzxRri5ufGOU+fPn8e0adP48/kzZ86AYRgsWLAA7u7uKC0txVNPPYXCwsIhz6/bhMddQQlwHMdhy5YtyM7OxsmTJ41KCKNcubz44ou49tpr4e7ubldn00uXLkV6ejomTZpk66VQzAwVdMq4cHBwwD/+8Q9IJBI899xz2LhxI5YtW4YpU6bA09MTcrkcSUlJKC0tBfB7Y87x48fh6+vL21meOHECPj4+SEpKAgDs3bsXiYmJSExM5J/rl19+QWxsLMrKyviPCQQCvozIMIxdW1Q+/PDD+Ne//oUvv/wSXl5eaG9vR3t7OxQKha2XRrEgarUa69ev1+sPsQemTJmCo0eP8n0ulIkDPUOnmISbmxtWrlwJYGCsjYRLeHh4YNWqVXj66adx8uRJTJs2DV9//TX279+Pu+66C1OnToVGo0Fubi7i4uKQkJAAAPj666+xfv16BAQE8M/xzTffwM3NjU/gqqqqwg8//IC4uDgsWbJk2PG5wZDdvC0a7chkwIIFC/Q+vmfPHvzlL3+x+noo1uG5554DMHD0ZG+4urraegkUC0AFnWI2HBwc+DEYYMBQ5fz581i6dCmWLl2Kjo4O+Pr68s1wP//8M6RSKZKSkuDh4YHq6mo0Njbi5ptv5hvxAODQoUO47bbbMHXqVBw8eBDPPfcc3NzcUF9fD6VSif/5n//Bli1b4OPjM2RNLMtCoVDY1MDFnqsHFApl4kBL7hSLERAQgPfffx8ajQY7duzA448/jmuuuQYpKSkAgO+//x7+/v5IS0sDMLATnzlzJmbOnMk/RkFBAcRiMd9Yd+bMGXh5eeHcuXNobm7Gxx9/jO7ubj5mk0BEtKysDMuXL8c111yD/Pz8Ydd5pZzFm8IHH3yApKQkeHt7w9vbG/PmzcPRo0dtvSwKhWJGqKBTrEJMTAzWr1+P7777DjNnzgTHcaipqUFQUBBfbu/o6ICbm5veyNqePXsQHR2NuLg4AEB0dDQuXryIAwcOQCaTYc2aNbjvvvuGNPiQxygtLYVarUZOTg4qKysB/N4tT0R/Is67D2by5Ml45ZVXkJOTg5ycHNx0003IzMzkexwohvHss8/yY5Uj/ZeTk2PrZVKuUmjJnWITBAIBvv32W3R0dMDT0xMAsH79euzfvx+nTp3CokWL8MUXX2DPnj3YvHkz7/f+wAMPoLu7G++88w4qKirw2GOP8WI/GLVajaKiIqjVasycOZN3ZiNCfvDgQXz00Ue44YYbsG3bNv6MfiKyatUqvX+/+OKL+OCDD/Dbb7/xN1SUsdmyZQs2btw46tdERUVZZzEUyiCooFNsSkhICP//09LSsHnzZjzyyCO48cYbIRQKwTAMFi5cCHd3d7S2tmLSpEnYvn07oqKi8MQTT6CmpgbvvvsuXFxc+Mch3tV1dXUoKChAXFwcpk2bhvPnz+Pmm2+Go6MjOjo68Ne//hX19fUoLS3F5s2bJ7Sg68IwDL755hvI5XLMmzfP1su5oggMDERgYKCtl0GhDAsVdIrd4ObmhmeeeQbPPPMMurq6cOHCBSiVSsTHx6OpqQlvv/021q1bh/nz5+PPf/4zNBoNnnjiCb6LfDClpaWoqqrC3r178cknn8DPzw8AUFJSgt27d4PjONx6663w8/ODv78/fyMADJyrsywLBweHCeNaV1xcjHnz5kGpVMLT0xPffvstTSSzII2Njejq6kJjYyMYhkFBQQGAgeMnUpWiUMwJFXSKXeLv74/ly5dj+fLlAID29nbI5XLceuutmDNnDhITE7F//34kJyfD0dFRT4wFAgG0Wi3y8vLg5uaGOXPm4PXXX+fF63//938RExODjRs34uLFi/zYHfl+lUoFFxeXCXe2HhcXh4KCAkilUmRlZWHTpk04c+YMFXUL8cwzz2Dv3r38v1NTUwEAp06dGjLCSKGYAxqfSrmiKC0txZ49e1BRUYHMzEwsW7aMT2wTCoW8sFdXV2P79u0ICgrCnj178PLLL6O4uBirV6/G9u3bcfHiRXz44Yc4deoUsrKyMGnSJDQ2NuLzzz/H+fPn0dbWhttvvx3333+/XuwrAN6i9krfuS9evBjR0dH46KOPbL0UCoViBugOnXJFkZCQgDfeeGPIxwfvpsvKytDU1IQ777wTANDb24t9+/bBwcEBL7/8MgIDA1FYWIjY2FhMmjQJnZ2d2LRpExwdHXHfffehsbERWVlZaGtrwxtvvKEn3uS5iP2sIcY29gjHcVCpVLZeBoVCMRNU0CkTCoFAAIZhcObMGfT09PAl+ylTpgAY2JVu2rQJR44cQVtbGzZs2AAA2L17N86cOYP9+/fj1ltv5b926dKl+POf/8zPyn/88cdQKBRYvnw5YmJirhgxf/LJJ7F8+XJERERAJpNh3759OH36NI4dO2brpVEoFDMxsQ4JKRQMiPpNN92Eu+66C15eXuA4Dg888AB6enqwadMmAMBvv/0GR0dHLFq0CMCAh3xMTAxfpr/99ttRVFSEtLQ0fq64trYWv/32G77//nukpaVh0aJFw4bIAL9nutsLHR0duOOOOxAXF4dFixbhwoULOHbsGI3QpFAmEPQMnTLhIeJKSuX19fXYvn07fHx8sHfvXvT29iI8PBxfffUVbrrpJvz888/8Draurg7Hjh3DzTffDJZlwTAMnJyc0NnZiTvuuAMRERH8GbRKpYJUKtUbxQPAn+9TKBSKJaFXGcqERzeZDQDa2trQ1NSE2bNnAwD6+/uRmpqKw4cPw9XVFUuWLMEnn3yCsrIyXLhwgRfzc+fO4f/+7//w3XffISAgAJs2bUJZWRkuX74MAMjJycGqVavw7LPP4vz587h06RKAq8OJbry8/PLLEAgE2L59u62XQqFc8dArDeWqY968eTh37hzuvvtuAEBoaCj+9Kc/IScnBydOnAAwIPJarRazZ89Gb28v/vnPf2LlypU4d+4c/va3v8HX1xcvvfQSWltbeRe76upqNDc347vvvsMXX3yBJUuW4KabbkJjY6PNXqs9c+nSJezatYuPzqVQKKZBBZ1yVeLm5gZ3d3f+33/6058wf/58ZGZmYubMmdi6dSteeeUViMVi1NXVISsrC08//TSys7NRUlKCffv2ob+/H/Hx8fDy8oJUKsVvv/0GhmGwZ88evP/++ygsLERlZSVOnz5tuxdqp/T19eHPf/4zdu/ezRv+UCgU06CCTqEA8Pb2xttvv43u7m4899xzmDRpEq699loEBQXB19cXzc3NmDp1KoRCIYRCIfr7+9Hb28t30V++fBllZWW48847kZKSAqFQCA8PD8yaNQsXL1608auzPx5++GGsWLECixcvtvVSKJQJAx1bo1B0cHZ2xoYNG/hxNgAIDw/H+vXrsXXrVpw9exZ+fn746KOPoFQq+aCO0tJSdHR08CNvANDa2gqZTIbQ0FCrvw57Zt++fcjNzaWpZBSKmaGCTqGMgaOjI3bt2oXly5fj6NGj8PT0xPTp0yGTyeDv74/e3l4UFxfD399fL+ykrKwMjY2NWLFihQ1Xb180NTVh27ZtOH78OFxdXW29HAplQkEFnUIxkLVr12Lt2rUAgCVLlqCjowPAgHD/9NNPWLhwIf+1fX19uHjxIoKCgngPbwqQm5sLkUiE9PR0/mMMw+Ds2bN49913oVKprhizHgrF3qCCTqGMA12Rjo6Oxq233oqbb76Z/1hdXR0uXbrEn7FTBli0aBGKi4v1PnbXXXdhxowZeOKJJ6iYUygmQAWdQjGRoKAg/OMf/9D7WH19PXJzc/HKK6/YaFX2iZeXF2bNmqX3MQ8PDwQEBAz5OIVCMQ4q6BSKBVi1ahXy8vIwY8YMWy+FQqFcJVDrVwqFQqFQJgB0Dp1CoVAolAkAFXQKhUKhUCYAVNApFAqFQpkAUEGnUCgUCmUCQAWdQqFQKJQJABV0CoVCoVAmAFTQKRQKhUKZAFBBp1AoFAplAkAFnUKhUCiUCQAVdAqFQqFQJgBU0CkUCoVCmQBQQadQKBQKZQJABZ1CoVAolAkAFXQKhUKhUCYAVNApFAqFQpkAUEGnUCgUCmUCQAWdQqFQKJQJABV0CoVCoVAmAFTQKRQKhUKZAFBBp1AoFAplAkAFnUKhUCiUCQAVdAqFQqFQJgBU0CkUCoVCmQBQQadQKBQKZQJABZ1CoVAolAkAFXQKhUKhUCYAVNApFAqFQpkAUEGnUCgUCmUCQAWdQqFQKJQJABV0CoVCoVAmAFTQKRQKhUKZAFBBp1AoFAplAkAFnUKhUCiUCQAVdAqFQqFQJgBU0CkUCoVCmQBQQadQKBQKZQJABZ1CoVAolAkAFXQKhUKhUCYAVNApFAqFQpkAUEGnUCgUCmUCQAWdQqFQKJQJABV0CoVCoVAmAFTQKRQKhUKZAFBBp1AoFAplAkAFnUKhUCiUCQAVdAqFQqFQJgBU0CkUCoVCmQBQQadQKBQKZQLw/wFu9ZCCZ6TgeQAAAABJRU5ErkJggg==", "text/html": [ "\n", "
\n", "
\n", " Figure\n", "
\n", " \n", "
\n", " " ], "text/plain": [ "Canvas(toolbar=Toolbar(toolitems=[('Home', 'Reset original view', 'home', 'home'), ('Back', 'Back to previous …" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# plot\n", "%matplotlib ipympl\n", "plot_3d_lls(X_train[:,0], X_train[:, 1], y_train, lls_sol, \n", " \"Breast Cancer - Radius Mean vs. Area Mean vs. Perimeter Mean - LLS Mini-Batch GD\")\n", "print(\"w:\")\n", "print(w)" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "## Regularization\n", "---\n", "* A good way to **reduce** overfitting is to regularize (i.e. constrain) the model. \n", "* The fewer degrees of freedom it has, the harder it will be to overfit the data. \n", " * For example, in *Polynomial Fitting*, a simple regularization would be to reduce the number of polynomial degrees. \n", "* We will look at **Ridge & Lasso Regressions**." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Ridge Regression\n", "---\n", "* *Ridge Regression* (also called *Tikhonov Regularization*) is a regularized version of Linear Regression where the *regularization term* is equal to $\\lambda \\sum_{i=1}^n \\theta_i^2$ ($l_2$) and is added to the cost function.\n", "* This form of cost function **forces** the algorithm not only to fit the data but also **keep the model weights as small as possible**.\n", "* The regularization term is only added during **training**.\n", "* The **hyperparameter** $\\lambda$ controls how much to regularize the model.\n", " * **The Shrinking Effect**\n", " * When $\\lambda \\to 0$ we resort back to the (unconstrained) Least Squares solution.\n", " * As $\\lambda$ increases, then all the weights end up very close to zero and the result is a **flat line** going through the data's mean." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* **Ridge Regression Cost Function**: $$ J(\\theta) = MSE(\\theta) +\\lambda \\sum_i\\theta_i^2 $$\n", "* **In Vector Form**: $$ J(\\theta) = ||X \\theta - y ||_2^2 +\\lambda ||\\theta ||_2^2 = (X \\theta - y)^T(X \\theta - y) + \\lambda\\theta^T\\theta$$\n", "* **The Gradient**: $$\\nabla_{\\theta}J(\\theta) = 2(X^TX\\theta - X^Ty) + 2\\lambda\\theta $$\n", "* **The Closed-Form Solution**: $$ \\theta^{*} = (X^TX +\\lambda I)^{-1}X^Ty $$" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "w: [[0.88003189 0.1174982 ]]\n" ] } ], "source": [ "# ridge regression\n", "ridge_reg = Ridge(alpha=1, solver='cholesky', fit_intercept=False)\n", "ridge_reg.fit(X_train, y_train)\n", "w = ridge_reg.coef_\n", "lls_sol = X_train @ w.T\n", "print(\"w:\", w)" ] }, { "cell_type": "code", "execution_count": 12, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "5eede85057c6438b85e0fd8c763c0fe7", "version_major": 2, "version_minor": 0 }, "image/png": "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", "text/html": [ "\n", "
\n", "
\n", " Figure\n", "
\n", " \n", "
\n", " " ], "text/plain": [ "Canvas(toolbar=Toolbar(toolitems=[('Home', 'Reset original view', 'home', 'home'), ('Back', 'Back to previous …" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# plot\n", "plot_3d_lls(X_train[:,0], X_train[:, 1], y_train, lls_sol, \n", " \"Breast Cancer - Radius Mean vs. Area Mean vs. Perimeter Mean - LLS Mini-Batch GD\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", "\n", "Animation by Dustin Kenefake" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### LASSO Regression\n", "---\n", "* *LASSO (Least Absolute Shrinkage and Selection Operator) Regression* is just like Ridge, but instead of $l_2$, it uses $l_1$ norm of the weight vector.\n", "* * **LASSO Regression Cost Function**: $$ J(\\theta) = MSE(\\theta) +\\lambda \\sum_i|\\theta_i| $$\n", "* **In Vector Form**: $$ J(\\theta) = ||X \\theta - y ||_2^2 +\\lambda ||\\theta ||_1 $$" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* The LASSO cost function is **not differentiable** at $\\theta_i=0$, but Gradient Descent still works if we use the *sub-gradient vector*: \n", "\n", "$$ g(\\theta,J) = \\nabla_{\\theta} MSE(\\theta) + \\lambda \\begin{bmatrix}\n", " sign(\\theta_1) \\\\\n", " \\vdots \\\\\n", " sign(\\theta_n)\n", " \\end{bmatrix}, sign(\\theta_i)= \\begin{cases}\n", " -1 & \\quad \\text{if } \\theta_i < 0 \\\\\n", " 0 & \\quad \\text{if } \\theta_i = 0 \\\\\n", " +1 & \\quad \\text{if } \\theta_i > 0 \n", " \\end{cases}$$ " ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "* LASSO Regression tends to **eliminate the weights of the least important features** (i.e., set to them zero). It automatically performs **Feature Selection** and outputs a *sparse* model (few non-zero weights).\n", " * **Ridge** regression will tend to shrink the large weights while hardly shrinking the smaller weights at all. In **LASSO** regression, the shrinkage will be directly proportionate to the importance of the feature in the model.\n", " * Since $\\lambda$ is an arbitrarily selected constant, some feature weights can reach zero, meaning that these features will not be included in the model at all. And that is the built-in feature selection of LASSO regression.\n", " * In other words, ridge regression will try to find a good model with small-scale features possible while LASSO regression will try to find a model with as few features as possible.\n", "* $l_1$ regularization is much more **robust to outliers**." ] }, { "cell_type": "code", "execution_count": 13, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "w: [8.97848395e-01 1.71905813e-04]\n" ] } ], "source": [ "# lasso regression\n", "lasso_reg = Lasso(alpha=0.1, fit_intercept=False)\n", "lasso_reg.fit(X_train, y_train)\n", "w = lasso_reg.coef_\n", "lls_sol = X_train @ w.T\n", "print(\"w:\", w)" ] }, { "cell_type": "code", "execution_count": 14, "metadata": { "slideshow": { "slide_type": "subslide" } }, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "efab25ece6b34627933513d00c3103e2", "version_major": 2, "version_minor": 0 }, "image/png": "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", "text/html": [ "\n", "
\n", "
\n", " Figure\n", "
\n", " \n", "
\n", " " ], "text/plain": [ "Canvas(toolbar=Toolbar(toolitems=[('Home', 'Reset original view', 'home', 'home'), ('Back', 'Back to previous …" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# plot\n", "plot_3d_lls(X_train[:,0], X_train[:, 1], y_train, lls_sol, \n", " \"Breast Cancer - Radius Mean vs. Area Mean vs. Perimeter Mean - LLS Mini-Batch GD\")" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
\n", "\n", "* While LASSO is more robust to outliers and performs feature selection, Ridge provides a more balanced minimization of the weights." ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "slide" } }, "source": [ "### Other Regularizations\n", "---\n", "* **ElasticNet** - a combination between Ridge and Lasso (both $l_1$ and $l_2$ regularizations)\n", "* A more generalized regularizer: $$ \\frac{1}{2}\\sum_{i=1}^n (\\theta^Tx_i - y_i)^2 + \\lambda \\sum_{j=0}^M |\\theta_j|^q$$" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "
" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "subslide" } }, "source": [ "### Recommended Videos\n", "---\n", "#### Warning!\n", "* These videos do not replace the lectures and tutorials.\n", "* Please use these to get a better understanding of the material, and not as an alternative to the written material.\n", "\n", "#### Video By Subject\n", "\n", "* Supervised vs. Unsupervised Learning - Machine Learning - Supervised VS Unsupervised Learning\n", "* Cross-Validation - What is Cross Validation and its types?\n", " * K-Fold Cross Validation - K-Fold Cross Validation - Intro to Machine Learning\n", "* Linear Regression - Lecture 3 | Machine Learning (Stanford)\n", " * Machine Learning Lecture 13 \"Linear / Ridge Regression\" -Cornell CS4780\n", "* Bias/Variance Trade Off - Bias/Variance (C2W1L02)\n", " * Machine Learning Lecture 19 \"Bias Variance Decomposition\" -Cornell CS4780\n", "* Feature Scaling - Gradient descent in practice I: Feature Scaling\n", "* Regularization - Linear regression (6): Regularization" ] }, { "cell_type": "markdown", "metadata": { "slideshow": { "slide_type": "skip" } }, "source": [ "## Credits\n", "---\n", "* Icons made by Becris from www.flaticon.com\n", "* Icons from Icons8.com - https://icons8.com\n", "* Datasets from Kaggle - https://www.kaggle.com/\n", "* Feature Scaling for Machine Learning: Understanding the Difference Between Normalization vs. Standardization, Analytics Vidhya\n", "* Examples and code snippets were taken from \"Hands-On Machine Learning with Scikit-Learn and TensorFlow\"" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.18" } }, "nbformat": 4, "nbformat_minor": 4 }