Skip to content

Instantly share code, notes, and snippets.

@IvanYashchuk
Last active August 12, 2019 12:56
Show Gist options
  • Save IvanYashchuk/c211a84d8502c6fe7a7b8e458c18e11a to your computer and use it in GitHub Desktop.
Save IvanYashchuk/c211a84d8502c6fe7a7b8e458c18e11a to your computer and use it in GitHub Desktop.
ChainerX Gaussian Process Regression
Display the source blob
Display the rendered blob
Raw
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Implementation of Gaussian Process Regression with ChainerX\n",
"\n",
"The purpose of this notebook is to demonstrate the ChainerX linear albebra routines in the example of Gaussian Process Regression."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import chainerx\n",
"import chainerx as np\n",
"import matplotlib.pyplot as plt\n",
"%matplotlib inline"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"chainerx.set_default_device('cuda:0')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Unknown functions can be represented using [Gaussian Processes (GPs)](https://en.wikipedia.org/wiki/Gaussian_process). GPs can be used for regression and classification tasks. Here, we focus on a simple GP regression case. There is a whole book for details [Gaussian Processes for Machine Learning](http://www.gaussianprocess.org/gpml/).\n",
"\n",
"We are interested in learning a function $f(\\mathbf{x})$ from data $\\{ (x_i, y_i)\\; |\\; i=1,\\ldots,n\\}$, where $y_i\\in\\mathbf{R}$ is a noisy observation of $f(x_i)$."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"A Gaussian process is a [random process](https://en.wikipedia.org/wiki/Stochastic_process) where any point $\\mathbf{x} \\in \\mathbb{R}^d$ is assigned a random variable $f(\\mathbf{x})$ and where the joint distribution of a finite number of these variables $p(f(\\mathbf{x}_1),...,f(\\mathbf{x}_N))$ is itself Gaussian:\n",
"\n",
"$$p(\\mathbf{f} | \\mathbf{X}) = \\mathcal{N}(\\mathbf{f} | \\boldsymbol\\mu, \\mathbf{K})\\label{eq1}$$\n",
"\n",
"Here, $\\mathbf{f} = (f(x_1),...,f(x_N))$, $\\boldsymbol\\mu = (m(x_1),...,m(x_N))$ and $K_{ij} = \\kappa(x_i,x_j)$. $m$ is the mean function and usually considered to be zero function. $\\kappa$ is a positive definite *kernel function* or *covariance function*. Thus, a Gaussian process is a distribution over functions whose shape (i.e. smoothness) is defined by $\\mathbf{K}$. If points $x_i$ and $x_j$ are considered to be similar by the kernel the function values at these points, $f(x_i)$ and $f(x_j)$, can be expected to be similar too. The kernel function $\\kappa$ has free hyper-parameters $\\theta$.\n",
"\n",
"A GP prior $p(\\mathbf{f} | \\mathbf{X})$ can be converted into a GP posterior $p(\\mathbf{f} | \\mathbf{X},\\mathbf{y})$ after having observed some data $\\mathbf{y}$. The posterior can then be used to make predictions $\\mathbf{f}_*$ given new input $\\mathbf{X}_*$:\n",
"\n",
"$$p(\\mathbf{f}_* | \\mathbf{X}_*,\\mathbf{X},\\mathbf{y}) \n",
"= \\int{p(\\mathbf{f}_* | \\mathbf{X}_*,\\mathbf{f})p(\\mathbf{f} | \\mathbf{X},\\mathbf{y})}\\ d\\mathbf{f}\n",
"= \\mathcal{N}(\\mathbf{f}_* | \\boldsymbol{\\mu}_*, \\boldsymbol{\\Sigma}_*)$$\n",
"\n",
"So the posterior predictive distribution is also a Gaussian with mean $\\boldsymbol{\\mu}_*$ and $\\boldsymbol{\\Sigma}_*$. By definition of the GP, the joint distribution of observed data $\\mathbf{y}$ and predictions $\\mathbf{f}_*$ is\n",
"\n",
"$$\n",
"\\begin{pmatrix}\\mathbf{y} \\\\ \\mathbf{f}_*\\end{pmatrix} \\sim \\mathcal{N}\n",
"\\left(\\boldsymbol{0},\n",
"\\begin{pmatrix}\\mathbf{K}_y & \\mathbf{K}_* \\\\ \\mathbf{K}_*^T & \\mathbf{K}_{**}\\end{pmatrix}\n",
"\\right)\n",
"$$\n",
"\n",
"With $N$ training data and $N_*$ new input data, $\\mathbf{K}_y = \\kappa(\\mathbf{X},\\mathbf{X}) + \\sigma_y^2\\mathbf{I} = \\mathbf{K} + \\sigma_y^2\\mathbf{I}$ is $N \\times N$, $\\mathbf{K}_* = \\kappa(\\mathbf{X},\\mathbf{X}_*)$ is $N \\times N_*$ and $\\mathbf{K}_{**} = \\kappa(\\mathbf{X}_*,\\mathbf{X}_*)$ is $N_* \\times N_*$. $\\sigma_y^2$ is the noise term in the diagonal of $\\mathbf{K_y}$. It is the noise of target data $\\mathbf{y}$, so it is set to zero if training targets are noise-free and to a value greater than zero if observations are noisy. The mean is set to $\\boldsymbol{0}$. The sufficient statistics of the posterior predictive distribution, $\\boldsymbol{\\mu}_*$ and $\\boldsymbol{\\Sigma}_*$, can be computed with\n",
"\n",
"$$\\boldsymbol{\\mu_*} = \\mathbf{K}_*^T \\mathbf{K}_y^{-1} \\mathbf{y}$$\n",
"$$\\boldsymbol{\\Sigma_*} = \\mathbf{K}_{**} - \\mathbf{K}_*^T \\mathbf{K}_y^{-1} \\mathbf{K}_*$$\n",
"\n",
"Given this model, there are two basic problems to be solved. First, we need to *learn* the hyper-parameters $\\theta$ and $\\sigma_y^2$. Second, we need to *predict* mean and variance of $f(\\mathbf{x}_{*})$ at test input points $\\mathbf{X}_{*}$."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The squared exponential covariance function is given by\n",
"\n",
"$$k(\\mathbf{x}_i, \\mathbf{x}_j) = \\sigma_f^2 \\exp\\left(-\\frac{\\|\\mathbf{x}_i - \\mathbf{x}_j\\|^2_ 2}{2\\ell^2}\\right)$$\n",
"\n",
"where $\\sigma_f > 0$ and $\\ell > 0$ are hyperparameters of the kernel. There are many other kernels that can be used for Gaussian processes. This specific covariance function is perhaps the most common covariance function used in statistics and machine learning. It is also known as the radial basis function kernel, the gaussian kernel, or the exponeniated quadratic kernel."
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"def gaussian_rbf(x1, x2, l=1, sigma_f=1):\n",
" \"\"\"Returns the NxM kernel matrix between the two sets of input X1 and X2. \n",
" \n",
" Args:\n",
" X1: NxD matrix\n",
" X2: MxD matrix\n",
" alpha: scalar parameter\n",
" scale: scalar parameter\n",
" \n",
" Returns\n",
" NxM kernel matrix\n",
" \"\"\"\n",
" # distance between each rows\n",
" dist_matrix = np.sum(np.square(x1), axis=1).reshape(-1, 1) + np.sum(np.square(x2), axis=1) - 2 * np.dot(x1, x2.T)\n",
" return sigma_f**2 * np.exp(-1 / (2 * l**2) * dist_matrix)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"kernel = gaussian_rbf"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's visualize the entries of the kernel matrix."
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"<matplotlib.colorbar.Colorbar at 0x7f04781b4470>"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 2 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# create an Nx1 vector of equidistant points in [-3, 9]\n",
"N = 50\n",
"X = np.linspace(-3, 9, N).reshape(-1, 1)\n",
"\n",
"cov = kernel(X, X)\n",
"plt.pcolor(chainerx.to_numpy(cov))\n",
"plt.colorbar()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Prior\n",
"\n",
"Let's generate samples from a Gaussian process prior. Specifically, we will consider a zero-mean Gaussian process prior for functions of the form $f: \\mathbb{R} \\rightarrow \\mathbb{R}$ using the squared exponential kernel. That is,\n",
"\n",
"$$f(x) \\sim \\mathcal{GP}\\left(0 \\, , \\, k\\left(x, x'\\right)\\right).$$\n",
"\n",
"Let $f_n = f(x_n) \\in \\mathbb{R}$ be the value of the function $f$ evaluated at a point $x_n \\in \\mathbb{R}$. Furthermore, let $\\mathbf{f} \\in \\mathbb{R}^{N}$ be the vector of function values for each of the points of $\\mathbf{X}$ . \n",
"\n",
"The Gaussian process prior for $\\mathbf{f}$ becomes\n",
"\n",
"$$\\mathbf{f} \\sim \\mathcal{N}\\left(\\mathbf{0}, \\mathbf{K}\\right),$$\n",
"\n",
"where $\\mathbf{K}$ is the kernel matrix.\n",
"\n",
"To draw random functions from the GP we draw random samples from the corresponding multivariate normal. The common way of drawing such samples is using Cholesky decomposition, the algorithm is described in [wikipedia](https://en.wikipedia.org/wiki/Multivariate_normal_distribution#Drawing_values_from_the_distribution)."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"def generate_samples(mean, K, S):\n",
" \"\"\"Returns M samples from a Gaussian process with kernel matrix K.\n",
" \n",
" Args:\n",
" K: NxN kernel matrix\n",
" S: number of samples\n",
" \n",
" Returns:\n",
" SxS matrix of samples\n",
" \"\"\"\n",
"\n",
" z = np.random.normal(size = (mean.shape[0], S))\n",
" L = np.linalg.cholesky(K + 1e-6*np.eye(mean.shape[0])) # added `1e-6*eye(N)` to covariance matrix to fix non positive definitness\n",
" w = mean + L.dot(z)\n",
" return w"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Next, let's define a function for plotting a Gaussian process with specified mean and 95.45% confidence interval ($2\\sigma$)."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"def plot_gp(Xp, mu, Sigma, title=\"\", num_samples=0):\n",
" \n",
" mean, std = mu.ravel(), np.sqrt(np.diag(Sigma))\n",
"\n",
" \n",
" # plot distribution\n",
" plt.plot(Xp, mean, color='k', label='Mean')\n",
" plt.plot(Xp, mean + 2*std, color='k', linestyle='--')\n",
" plt.plot(Xp, mean - 2*std, color='k', linestyle='--')\n",
" # plt.fill_between does not work well with chainerx arrays so we are converting the input to numpy arrays\n",
" plt.fill_between(*map(chainerx.to_numpy, (Xp.ravel(), mean - 2*std, mean + 2*std)), alpha=0.1)\n",
" \n",
" # generate samples\n",
" if num_samples > 0:\n",
" fs = generate_samples(mu, Sigma, num_samples)\n",
" plt.plot(Xp, fs, color='b', alpha=.25)\n",
" \n",
" plt.title(title)\n",
" plt.legend()"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# Finite number of points\n",
"X = np.arange(-1, 7, 0.025).reshape(-1, 1).astype('float64')\n",
"\n",
"# Mean and covariance of the prior\n",
"mu = np.zeros(X.shape, dtype='float64')\n",
"cov = kernel(X, X)\n",
"\n",
"# Plot GP mean, confidence interval and samples \n",
"plot_gp(X, mu, cov, num_samples = 5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Prediction from noise-free training data"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's implement computation of $\\boldsymbol{\\mu_*}$ and $\\boldsymbol{\\Sigma_*}$ using previously defined formulas."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"def posterior_predictive(X, X_train, Y_train, l=1.0, sigma_f=1.0, sigma_y=1e-8):\n",
" '''Computes the suffifient statistics of the GP posterior predictive distribution \n",
" from m training data X_train and Y_train and n new inputs X.\n",
" \n",
" Args:\n",
" X: New input locations (N x D).\n",
" X_train: Training locations (M x D).\n",
" Y_train: Training targets (M x 1).\n",
" l: Kernel length parameter.\n",
" sigma_f: Kernel vertical variation parameter.\n",
" sigma_y: Noise parameter.\n",
" \n",
" Returns:\n",
" Posterior mean vector (N x D) and covariance matrix (N x N).\n",
" '''\n",
" K = kernel(X_train, X_train, l, sigma_f) + sigma_y**2 * np.eye(len(X_train))\n",
" K_s = kernel(X_train, X, l, sigma_f)\n",
" K_ss = kernel(X, X, l, sigma_f)\n",
" \n",
" z = np.linalg.solve(K, K_s).T\n",
"\n",
" mu_s = z.dot(Y_train)\n",
" cov_s = K_ss - z.dot(K_s)\n",
"\n",
" return mu_s, cov_s"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"and apply them to noise-free training data `X_train` and `Y_train`. The following example draws five samples from the posterior predictive and plots them along with the mean, confidence interval and training data. In a noise-free model, variance at the training points is zero and all random functions drawn from the posterior go through the trainig points. "
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"<matplotlib.legend.Legend at 0x7f0478005518>"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAD4CAYAAADvsV2wAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOydd3hURReH39n03kkjhRZ6BwFBQUQpiqACCgooShMElCIgIKJYsCMIgtKkKEgXROGjW+idJJAESA/pvW12vj+GKklI2SQo932efSC7c+/MZrPnzj3ld4SUEg0NDQ2N/z66ql6AhoaGhkbloBl8DQ0NjfsEzeBraGho3CdoBl9DQ0PjPkEz+BoaGhr3CaZVvYDicHV1lf7+/lW9DA0NDY1/DceOHUuQUroV9to9bfD9/f05evRoVS9DQ0ND41+DEOJKUa9pLh0NDQ2N+wTN4GtoaGjcJ5Tb4AshfIQQe4QQ54UQ54QQYwsZI4QQc4UQIUKI00KIFuWdV0NDQ0OjdBjDh68Hxkspjwsh7IBjQoidUsrzt4zpDtS59mgDLLj2r4aGhoZGJVHuHb6UMkZKefza/9OBQMD7H8N6ASuk4m/AUQjhWd65NTQ0NDRKjlF9+EIIf6A5cOgfL3kDEbf8HMmdF4Xr5xgmhDgqhDgaHx9vzOVpaGho3NcYzeALIWyB9cA4KWVaWc8jpVwkpWwlpWzl5lZoKqmGhoaGRhkwisEXQpihjP0qKeWGQoZEAT63/Fz92nMaGhoaGreQkJBQYec2RpaOAL4HAqWUnxcxbAsw6Fq2TlsgVUoZU965NTQ0NP5LHDt2jDp16rB06bIKOb8xsnTaAwOBM0KIk9eemwr4AkgpFwLbgR5ACJAFvGyEeTU0NDT+M5w9e5bHH38cO3t7OnbqVCFzlNvgSykPAuIuYyQwqrxzaWhoaPwXiY6Opnv37phbWPLzlh34+vpWyDxapa2GhobGPYCvfw1Wrt2AX40aFTbHPS2epqGhofFf5npPcWtHV9b/8hsqJFpxaDt8DQ0NjSpi0aJFdH/iSZJS0irc2INm8DU0NDSqhBMnTjB27Fjy8vKxsraulDk1g6+hoaFRyaSlpdGvXz+cXVz4etH36HSVY4o1H76GhoZGJTNp0iTCwsLYuP13XF0rT1FA2+FraGhoVCIpKSls3rKFYa+9Tpt27St1bm2Hr6GhoVGJ2Nk7sO+vY5hZWFT63NoOX0NDQ6OS2LlzJ0npWTg4OWFdSYHaW9EMvoaGhkYlcPToUbp27coXn31WZWvQDL6GhoZGBSOlZMyYMbi6VWPIsBFVtg7Nh6+hoaFRwaxevZq//vqLL+Z/i529fZWtQ9vha2hoaFQgGRkZTJo0iWbNW/DcgBerdC2awdfQ0NCoQGJjY3F2ceX9OZ9VWoFVUWguHQ0NDY0KxNe/JjsP/F1irRyDoeLWohl8DQ0NjQpi48aNNGnVDlsHxyLHpKZAXJwgLRXy88FgELjaw8MPGX89mktHQ0NDowIIDQ2lX79+zPlodqGvZ2XB+bOCs2d0pCSDnT14eYGvnwEfn0IPKTfaDl9DQ0OjApg5cyZmZmaMGvvmHa/FxUJoiMDEBPxrGPD0glvd+y42FbOm/+QOf+/evbz99ttVvQwNDY37lHPnzrFq1SpeHjoCdw/PG89LCVcuCUIu6nB0hOYtJd7Vbzf2FYlRphFCLBFCXBVCnC3i9U5CiFQhxMlrjxnGmLcodv1vNx988AG//vprRU6joaGhUSgzZszAxtaW0W+Mv/GcwQAXggWRkQIPT0n9hhJz88pdl7GuK8uAbncZc0BK2ezaY5aR5i2UyVOmEFC3HsOHDyc9Pb0ip9LQ0NC4jfz8fPL1BQwfNQZnZ5cbz4eGCBLiBf41DNSqLamEBld3YBSDL6XcDyQZ41zGwMLCgs/nLSAyMpKpU6dW9XI0NDTuI8zMzFi66icmTL7pVg6/IrgaJ/DxVS6cqqIyffjthBCnhBC/CiEaFjVICDFMCHFUCHE0Pj6+zJO1eqAtrwwfyfz58zly5EiZz6OhoaFRUqKjozl7Pgi9Qd7Iu09MhIhwQTV3ia+frNL1VZbBPw74SSmbAl8Dm4oaKKVcJKVsJaVs5eZWvk4wU6a/y+dff0OzZs3KdR4NDQ2NkjB79mxat2xOakoKANnZEHJBYGsrqVW7ao09VJLBl1KmSSkzrv1/O2AmhHCt6HltbG3pP/AlcgrAUJHlaxr3HKmpqezevZv169dX9VI07hPi4uJYsmQJz/R9HgdHR6SEC0Fql1+3vqy0TJziqJQlCCE8xLX7GyHEA9fmTayMuQE2b9lK48aNSbl21dX473Lq1Cl69+6Nq6srjz76KKNHv05aTj7ZeQWMGTOG9957j8zMzKpepsZ/kK+++oq8vLwbmTnRUZCRIahVR2JpWcWLu4ax0jLXAH8BdYUQkUKIV4QQI4QQ14Wf+wBnhRCngLnA81LKSru/8fD0IjAwkFmzKjQ5SKOK+frrr2nevDn79x9g2Guv89PGX9j06y6y8wpIy8knMiaOGTNm0KBBA7Zs2VLVy9X4D5GZmcmCBQvo0bMXNWvVJidHBWqdXSSuFe7LKDmiEu1uqWnVqpU8evRoqY/LLzCQlJl323MTxozix1UrOHv2LHXr1jXWEjXuEQwGSciVSBZ8M58Ro8fi6ORU6LhDf/3B5PHjCDx3lhkzZjBz5swSi1ppaBTF7t276d69O+t/2UHrNu0IPCdITVWFVWVpXetiY46pSdn240KIY1LKVoW+dr8Y/ISEeNo0bUD3bt1Yt26dsZaoUcWEhoYyc+a7fPL1AoTOpETH5OXlMXHcaH7ZtIHTp09Ts2bNCl6lxn8dg0ESfCUKJ2cXEhMFwYE6atQ04OVdtvNVlMG/B8IIlYOrqxsjRo/l559/JjAwsKqXo2EEoqKi6NKlC9u3bycyMrLEx5mbm/Pl/G/ZeeBvXL0qSKVK474hKyuLrPwCnF1cMRgEl8MENjYST6+qXtmd3FfiaSNGjaFN23aaS+c/QFZWFj169CAhMZGft/yKj69fqY4XQlCzVm2y8wpYOP9rcrMymT59egWtVuO/zNNPP42ljR2Llq0kJhpycwV1AgxVUkl7N+6bHT6Anb09Dz/yKDl6g5am+S9n1KhRnDlzhsXLVtGsRcsyn0dKydFjx5kxYwbbt2834go17gfOnTvH77//TsNGTcjPh8gIFagtRv6+SrmvDP51Pp4zh0ceeYR7OX6hUTRRUVFs2bKFNyZN4ZEuj5XrXEII5nzxNQ0bN2HgwIGEh4cbaZUa9wPz5s3D0tKSF18eQmS4wGAAP/97167clwbf0dGZ/fv3s23btqpeikYZqObhyZ4/jzD+LePoJFlZWbF4+Sry8vMZMGBAhd/96fV6fvvtN/bs2UNeniQxUdKmzZO8+OJIfv55G3l5eXc/iUaVk56ezsqVK+n1bF+srV2JiQF3D7C2ruqVFc19afD79n8BHz8/3n//fW2X/y9CSsnGjRtJyczF3dMLE5OSZeWUhJq1ajN7zuf8+eef/PHHH0Y7760YDAaWLVuGj09DunUby6TJm/hlZz67/8ghN68umzZH0rfvJ3h49GLOnFXk5xdUyDo0jMPq1avJyMhg8JBXiQgX6HTg43tv25P/ZNA2PByEBdjYFv66mZkZr4+bwKQ3XmfXrl089lj53AIalcOaNWt44YUX+Oa7pTzT93mjn79f/xdo2rwFbVoaX3spPDycfv36c+hQAgF1O/PysL507tIaJ2cDdnaCDdtmk5GhZ9+eAyz+Zh1vvbUYIfwYMqQDLi53P79G5fPcc89jMLWgXoPWnDyu2hNWtr59afnP5eHn58Pfhwyk5+bTuGnRDQZyc3Np26wBAXXqsHfv3vIvVqNCycjIoE6dOnh4ebNt1z6j7u7/ialOcDUijPr16xvtnF/P+57Jk1fy6vCJDBjUGR9fKOotSCn5ZcsefH0exaTADBeXbJo2tbknsz7uZ7Ly9KTn6AkLFcTFQotWZSuyKgwtD7+EmJlBw0aQlwfBgSqIUhgWFhZ8Me9bPvtqbuUuUKNMzJkzh9jYWN7/+NMKNfYAv/22gwYNGhglxpORkUFUbAF+dV7km8WrGPl6Z/xrFG3sQQWSe/bqTONmkvi0QDp3HsgPPxzlHt6b3XfMnj2b775fQl6e6k/rVg2jGfuK5D9n8AHs7KB2gCQtTRAWWvS2qNOjXfCtWVfz49/jRERE8Omnn/J0n360eqBthc/3UMdHqF0ngDfeeKNcAdTt27fj59eGH38KxNISOj3qjJNzyY/X6aBxU3scnVIYMWIGP/10WjP69wApKSnMnj2bo4cPExMtMBgEXt7/jg/mP2nwAdzcoHp1SVysIDam6HGRkRH07fccx44dq7zFaZSKmJgY6tZvwNsz36uU+czNzZn10SdcvHiRefPmlekcx44d49lnR+Po9CC1A9xp1KRsiomurm5s2LYYZ5dURoyYwYEDxfwxa1QKP/zwA9nZ2bww+FViY8DZRd7TmTm38p/z4cNNLR0pIei8IDkZGjeR2NnfOTYtNZUWDevQ88knWb16tRFWrWFscvILSM3Or/R5n3+6J2dPnyQsLAw7O7sSHxcVFUWrVl0pKKjPwqVf0a69c7m10C9eCKZrp4H4VH+QzZu/pG7de0Rv9z5DSkmjRo0wt7Ti+x8OcvmSjiZNDYXalvKg+fDLgBDKtWNhAcFBgvxCbIa9gwMvDh7C2rVrtaKbe5CNGzcSG1817ZLfmvYOubm5pbr7y8/P5+mnB5KS6s17H82k7YPlN/YAdQLqMu/bt/Hzt+bSJUi6ZzpI31/88ccfnD9/nhdfGkpMtMDBofCN5L3Kf9rggwri1q0nyc9XXeML49URrwGqgYHGvUNQUBB9+vThy88/rZL5m7dsxYnAUNo/9HCJj8nK0uPk8ghj3nyD7j1rFRucLS09evZi6ZoPsLE3IyiIQjcwGhWLlJLOXR6jw8N9yc0VeFe/dz0khfGfN/gAtnaqICIxQVBYX/TqPr70eqYPixYt0rpi3UN89NFHWFpaMvS10VW2Bls7OzJy8gkKCrrrWCkhMNic18ZOYvhrnSoka8PEBDJzzzB+/BSOH08z/gQaxdL2wfas+nkLSUk22NjIUgXh7wXuC4MP4F0d7Owkl0IFhSVevDbmDYYOH6ll7NwjREVFsXr1avoPHIyra/ma2ZeXd9+ZQatWrbh69WqRYzIyMnj00SH8cSiYWrVlkUV/xkCITA4fXsusWd+SkFBx82jczvnz54mMiSMpCbKz/n27e7iPDP51f75eD1cu3+naadSkKZNnzML+XpW5u8/4+uuvKSgoYNjI16t6KfTt/wLZ2dl88cUXRY6ZNGkWe/aEYWWVjlu1il1PsxYtGT6qN9u3r2XTprNF1ppoGJcRI0bQo+tjREcKLC0lVbwPKRP3jcEHJWrk7Q1X4wRpqXe+XmAwsGHzFg4ePFj5i9O4jZDQUJ54qjd+NWpU9VKoXSeAJ57qzYIFC0hLu9ON8tdfh1iwYC89evbgqaebVsqaxr81mWruGcyZM5fwcE1zp6IJCQnhwIEDPNFzEGlpKu/+31j5fF8ZfIDqvhILC0lYqLijiEVKyZtjX2fGjBlVsziNG3y/fDXzFy+t6mXcYPS48aSmpvLtt9/e9nxeXh4vvzwTZxdf3p39MmZmlbMeWzs7Zn0whYsX/+Lbb3doAdwKZtmyZeh0Oto+2B8zM0k196peUdkwisEXQiwRQlwVQpwt4nUhhJgrhAgRQpwWQrQwxrxlwcQE/GtKMjMFMdH/fM2Ewa8MY8+ePZw7d65qFnifo9fruRIeQa6+APN7SImqafMWPNTxEdb9/PNtcZ7585cRHJzFxClD8fErea6+Mej1bF+++HoiHTs+yuXLlTr1fUVBQQHLly/noU5PojNxx8OzeGmMf6LXQ0ICXI2D+HhIS6XK3HDG2uEvA7oV83p3oM61xzBggZHmLROuruDkJIkIvzOAO2DQS1hYWDB//vyqWdx9zoYNG6hdqyYnTxyv6qXcwdyFi9m+czfi2r28lPBAm8FMnj6dF1/qWOnrEULQf1A/qnmZEBUlycys9CXcFxw7dozIyEgefWwIOp3E0+vuwVqDAaIi4cwpweG/VVPzixd0XAjScea0jiOHBBcvCNIrOdHKaJW2Qgh/4BcpZaNCXvsW2CulXHPt52Cgk5Sy2Drx8lbaFkd2Npw8LnBxhYC6t/8OxowcyrbNG4mKisLBwaHU82uUnUceeYTQS5f4+8S5ChdJKwsCsDcXmJubERlp4OQ5Sc06BdjaQWICpCSrTYSUSgvHyQnsHCROThilAKsw9u89yFtv/MDSJV/QseO/LE+wGKSUNy6uXbo8xoULwdjY2FC3bl26dOlCnz598PDwqJS1HDoZREREdTy9zKhdp3ibmZoCIRcFOTkCW1uJoyM4OUvMLaCgAHJyIClRkJQIer1qiVij5u3SGxVVaVtZevjeQMQtP0dee+4Ogy+EGIa6C8DX17fCFmRlBV7eqgelu/vtPSiHDB3BXwcPEBISQsuWZe+XqlE6goKC2Lt3L2+/M6vSjL3BoL6gaWmCrExVzCTlzQeAtQ04OkrsHSAmIpiFXR9lcJNmfH88m0Zvz8PatgFpwcow2diolEwhIDcHYmIgKkqHEBJnF6XvZGtkz4+nlxuXL+3lm29+okmTkTg5Gff8lU1WVhYffvghO3fuYsfufeQVGGja6gFcPTzJSE/nxKmTbN68me2//sr2Suhal6c3YG7hj06nw9OzaF9MQYHKAIyJFlhZSRo1NhTa29bGBlxcJAUFEBMtiYwQnDwuqFdf4ljBn9091wBFSrkIWARqh1+Rc1X3kSTEQ1iooFmLm1H3Zi1acujkOdwd/iWKSP8RFi5ciJmZGc8PHFzhc6WmQFycIDlJ7bKEkFhZqR18coqeixfCiImKITY6hawsPXn5AhsLMz449j7zk1Kw2vs/WmNG2vKxnG6xAz9/XaESuVJCcrIkNVkQFweJCTqcnCS16hhPO71OQF2e6duBzZvXMnz4c3Tu/O/d5e/fv5/Bgwdz+fJlnnr6WZJS0rCxtWXS1Om3jbt4IRiDwUByZh7ZaUmcPXOaxx9/3OjrWbZsGT+tW8+rI1bg4GhdZH1Fbi6cPyvIyhJ4eUl8/eVd/fwmJlDdB9yqSc6fFQSeF9RvULFGv7IMfhTgc8vP1a89V6WYmIBfDUlwoI74q7dH3nUmJqRkZJGbmYanp2fVLfI+IT8/n5UrV/LEU71xq6BEdimVBk1UhCA9XWBmpnbdllYGYqJi2bR+Pwf2HeVicBD5+bnXjrJAp6uGEC50K4jGn0BsUWmQduRjGnwYx4NL8X7jlULT9IQAZ2dwdpb4+MHVOEn4FbWjq1lLGi1nf+yEiaxf25MlSzbQsuWr/Ns8kVJKFixYwNixY/H182fDtt95sMNDRY6vE1AXgLwCA+/Meo/vFn7DBx98wOTJk2+4gYzB0qVLiYzKQQhrPIrw3WdnK2Ofnw8NGhpKXX1rYQGNmkjOnVFGv159iYuNERZfCJVl8LcAo4UQPwJtgNS7+e8rC1dXiLJVAVxXN3nDzyqlpEP7tgTUrsWmTZuqdpH3AWZmZuz/6zDpmTkVcv7MDHUnl5amimb8axgo0MOunSfZuG4zRw8fQspkage4MPDlFtSt1x4b24Y4u1TDzs4CRyfw/f49bL7Nh1u+9+YFuWx7dwOrV8Yx7LU2DHzpEUxNC/e9mpoqN6KTsyTkguBCsI6UFEmt2rLc/v2AuvV4sldbNmz4iREj+tGhw79I0QtVqfz1vPl0evQxvlm8FPtSXLHenvk+SYmJTJ06lcTERD755BOjGP3w8HD279/PkGHfYG5Boa0mM9Lh/Dk1V6PGZXfXmZlBw8bK6AcFCqo5gHMF7PSNYvCFEGuAToCrECISeAcwA5BSLgS2Az2AECALeNkY8xoLP3/JubM6YmMkXt7qOSEEnbs8zsJ5XxETE6Pt8isYKSWu7l4Yu31rQQGEXxbExCiD6+tnwFAg2PnbOVYu/YGzZ/bg7GzgtTFP8sLgF3B2rk1YmCA7SwXTlM9dVWjH+7WkmqkllvnZN85vsLTGtMnDJAa5MmX8QeZ+vo/p7z7GM32L3p1aWakdXUQ4KlMsF+o1uLsL4G5MnDqVzp2jyMuzJT1dNQK615FSYjAYEOZWbN7xP2zt7Eodv7GysmLeoiU4ODry2WefYWFhwezZs8u9NiWXbkmrB57A3ePOoHtaqjL2ZmZQv2H5NfGvG/0zpwRnz0L7B0uX/lkS/tN6+KXh3BlBZia0bH3zixcWGsKDLRoze/Zspk6dWup1aJSMkydPMnHiJN7/5Ev8a9Y02nlTkpVCak6OwN1DYmYmOXcmiUXffMPff27GxTWbcRNGMPDlVzE3tyQ0RHA1TgXc/GtKnK/dmhcUqL+PlMQCvAbUo2luNDYYyBaC2Bo10O88xfkgwbofj7Nt82FSU6/Qpq0FC5eOxNOr+I3C1Ti4eEGHg4OkfsPyG329Hi6eMcfNVUfDhuU7V2Xw2WefceCPP5m3eBlm5axak1IycexoDuzdzalTJ8uVYSelpHHjxpiY1Wfm7JW0bH17zCUnB06fFJiaQcNGxovHgHIRWWKOh7umh19h+PlL8vMFUZE3bwVr1qrNgw89zHfffYdBEyypMJYuXcr+/ftwcDSOjpHBoAz9ubM6hA78/Q2kp0m+W7iN0cOf5cSx+bw17VkOnz7E0JGjMTW15PxZZeyr+0iatbhp7EGl2KWnC84HHuHhHH/2DhuFmDWLmXXr8liBATsnQZt2Osa+2Zqvvx3GY117cPQItG85m9Ur1hcryFfNHQLqGkhLU35gvb58712IAn5c+wXLlm0jK6t856potmzZwsSJEzFIjJKVJYTgo8+/Ytvu/ViWU71Or9fT+5l+dO0xFGeX24Px1wO0UkK9+sY19qDuAF1djXvO62gG/xq2duDiKomOul1nfOBLQ7h06RL79++vusX9h8nLy2PVqlV07fEkTs7lzy4pKFBdzmJjBJ6eKgf61MkUpk6cxOIFI2jZyoq9fx3kjYmTsbGxITdXFcekpSnD6+d/uz89IhwS4gXV3A3UqtWIcW8O4PG5n8C0aTSfPp2wy5fYu3sXOp1KAGjeypSRYzoxbdZrODnX583XD/DygPFkFlMV5VYNAupJ0tOVISkohzSOiYkJJ0/8wooVywkPL+fVowI5f/48AwYMoGmz5ny1YDE6IxUpmJqa4urqxtWUDKa+/XaxCqfFYWZmRv9Bb9H2wU54et68YBsMEByoai0aGMGNU9loBv8WfP0kBgNEht/c5Xd/shebf91Jx46VX0l5P7Bt2zYSExN57oWB5T5XXp4ymMnJguo+BtLT4bdtJxg/5hnOn/2eOV+8w9rN26hRq9aN8efOCHJz1Zf3nxkziYkQfkWHWzWJPl/gUc2Bjz4acUPy4dlnn8Xd3Z0lixbeOMbNTd3it2jpyYefjeTx7k+wY7sZXTqM4PKly0Wu3dUV6taXZGTAheA7dZ5Kw8jXX+Pq1TOsW7e/XBePiiIrK4t+/fphZW3DsjXrsK4AqxkWGsrnn33G4MGDS313XlBQwIaNGwkLy8Pa+vYancuX1N1enbr/rk5X19EM/i1YW6tb7NhYddsGYGlpSZsHO5BXoLl0KoLly5dTzd2DTp27lOs8mZlw6oQgIwOcXQ1ERwlWLlvDu9OfwtUtmR179jFoyKs3sjfy85Wxz8tTBvqfuc+ZGXAxWGBnJ3F1NfDBu18RG/XnbeJo5ubmDBs2jP/9voMrly7deN7BUQVlnZ1NeGX44wwfNZiY6Fp06TCBfXv2FfkeXFygRk1JUqIgOKjsRv+xbj3w9TNnzZqfKeMGt0K5ePEiV6/GM+/b7/Hw9KqQOerVb8DM2R+zY8cOvvvuu1Idu3fvXp59ZjD7dv99m4xC/FWIiRZ4e8tCM3b+DWgG/x9U95FIyW2+fL1ez+uvj2HhwoXFHHlvYjConWx2NmRlUa6dY0XQrn0HXn9jPKamZU8Yy8xQxttgUFWMsdF65n0xkx+WjaD3s53Ysecg9RveVPwwGJTbJydH+WD/uVPT6yEoUGBqqnbdm9YfZedvG0lLO3/H3MOHD+fzL77E5R9OVxsbaNxUYm8PXbo2ZOo7o7Cy7sALfb5i0/r1Rb4XTy/wr2EgMUEQXkjfhpJgYmLCsFGDCQw8wo4dZ8p0jookoH4j/j51nk6Plu8ifzdeenUY7R/qyMSJE4mMjCzxcStXrsTaOoA2D7a5cdeXmaniQg4OEr8a99iXqBRoWTqFEBoiiIuFFq1u6lv0fPwR0lKSCQwMNGphR0WRkqLuVOLibjfyOh3Y2oKXF1SrVnH6LiUlNSufHH3Z/Q7X86ALClQKW0JCKp99NJRzZ7YwZfq7jBk/8Y7P60KwIP6qIKCeAbdCmliEXBRcjVN51bm5kj5PjSct5W8uXfoLiyIidAkZuRQY7vwu6fXq4pKSArHRGcz5YCXhV37j/Y+78+qIoUW+r5CLgrhYQd36hjIF8DIzMhg9fBp9np5Enz41sb8H3A/JycksW76c5wa9ikk5LvCl4XJYGI882IquXbuycePGu47Pzs6mWrXqPNBuPLM/Hk+NWiol9/QpQYEemjaXVIaIa0Vp6Wg7/EKo7qNkFiIjbhqKFwYPITg4mAMHDlThyu5Obi6cPg0nTyopVi8vCAiAevXUw9v7WmAzCA4fVheGqmLz5i2kpGeU+fj0NDh3VpB/TawsOiqOWdN6EhaykyUrf2TshEl3GPuIcIi/KvD1K9zYJyVBXKzA21sF8jdvOEFw4GGmTRtapLHX6/WsWb6EXb/vuOM1U1OVo21vD57etsycPYR69Xsx7a1f+GLOJ0W+t5q1JHZ2qkCrLCqYNra2fLfiS/xq+hEdfffxlcG4ceOYOGECFy4EV9qc/jVr8vW33zNt5rslGr9161YyMmx5uFOXG5W1IRcFOdlKZPEeUuwuE5rBLwQLC3D3UDnS2ddqbJ56+lns7O1ZsmRJ1S6uGBIS4MgRSE2F2rWhfXuoU0cZfQ8P9ahVC1q3hiZNVNn/qVNw+eXTn3sAACAASURBVHLlu3pOnjxJ7969+HHVD2U6PjNT7exzsqHAAFGRYcyY8ijJSRdYu3kbPXr2uuOYhISbQVifQnT58vMh9KLAxkbi4yeJiZasXLYKL69cXnllSJFrMTEx4Zt5c/lizodFvA4NGklsbMDRyZypMwfSpNkAPp69my8/mVPoMTqdcieZmKiskLI0ODExgeSUELZs+avKG6Ts3r2bFStW8PobE6jfoHILBJ7s9TQ+NQMwGORde1bv3LkLZ5cmtH+4EVZWEB0FiQkCP39ZqBDavw3N4BfBP3f51tbW9Hq6D+vWrSM9Pb2KV3cnkZFw9qzK4W3VCqpXL95d4+ysxrm7K4N/6hSVmtGxbNkyzM3N6fVMn1Ifm5sLgeeU8JkEIsJP8/bETuh0mWzasYvWbdrdcUxOjjLmtraySHnbsFBlWOsEqGytiHBBr14P88knU4rc3YPK/x4xYgTHjhzm7OlThY4xNVVVlJZWYGtrwvi3+tGk6Ut89P6uIo2+hYWKMeTkqAByWS7KC76ZyocfziE8vOwuzvKSl5fH6NGj8fX3Z+yEt8p9vtxcFUC9cllwKVQ9YmOUe6+o31FOTi7P9u3Lp59+Wuy5p8/8ho8+W4h3dR1pqSorx8VV4l293Mu+J9AMfhGYm6tdfvzVmxk7AwYNpu/z/YvNqa4KLl+GkBCV2te8OZiZ6Tl8+DBfffUVixYtujHu2LFjt12sTExuunpSU9UFozLqy/Lz81m1ahWPd3+i1Ln3er0K0EaEg5mZIDb6OG9P6oxbNQe2/r6n0N2jlHAhSF24A+oVrluTnKTy7av7gI2tCtobDDqmTunHgAED7rquQYMGYWVlxfLvFxc5xswM6jeQmJqCi6uONyc/R5Omr/DR+1tYOO+rQo+xs4datSXJyWUL4r4yfBApKZdZs2ZflQXs586dS2BgIO9//BlWVlZlOoeU6rt46oTg6GEdF4J1REXC1asqThUaouPUSR1//yk4d0aQkHC78bewsCAzK5v33nuPxMTEIue5Ei7xru6KrS0EBwksLbmr/v2/Cc3gF4OXt/qgo69l7LRo9QBzvpyHu/u909AyMlIZfA8PcHKKZtq0qbi7u9OmTRvGjZvEF18uIS4ll+TMPAa/9BJubm707t37tkbtHh5Qty4kJ8O5cxXv3tm5cycJCQn0ff7uhvRWDAa1sw+5ILC2FqSnH2PKxMfw8a3Opl934uPrV+hx4ZdV7nTN2koCubDzXgpTkgrVfSR5ebB751n+OrAWS8uS+UKcnJx4/vnnWb/uR9ILaXR+HSsrZfTz88HDXfD6m8/RoNEIZr69jJ9Wryz0GHcPcPeQREYKUpJLtJwbPPLoY3j7mPLzz9tILuWxxqJV6zaMGD2Wx7v1KNPxcbFw7IgSmyswqCymps0MtGsvadNO0vZBScvWBgLqGfDwUHdzwYE6ThwXpN4So3p75ntkZGTwySeFx066d+/DtwvX4OGpel7r9VC3nrpA5+Up7Zz4eOUavBcz3krCPaeHfy9haQmubmoHUd1X7dAMBsnBv/6mlr8fXl4Vk0NcUq5eVTt7NzcVmB006GNWrdrNg+2H0K5DF+o3bIKTkzN//KGM2cgxizh6eAO/bFrO5s0P0aVLF+bPn09AQAAeHsrwXbgAgYFQvz6Fyv0ag927d+Pg6EinRx8r8TFSquya4CCBrR0YDMeZMLYrXl5e/LxlO25FdJVOToLISKWlU1iQFlQruuxsQYOGBnQ65cr5YdkykhK2MeudkrucRo4cydlz54iNjcGumLQYO3uoU1fJcvvVgOGj+jP3s0zGvfYhjo6OdO3x5B3H1KgpSUtTAcRmLZQRKgk6nY7BL/fig1m/cuBAKL161Srx+zEWTVq1oX7z1qU+Li9PueGSkgT29pKatQ04ORX+d2lpqR5ubhL/mpCYKAm/LDh7RoeXl0qlrFe/Ac/0fY65c+cybty427plhYWFsWPHaQa+9BRCXPfbG25sMpKS7pxUp1OVts4uEk8vbvtMCgogK1NdGDIzxY3/X++EJoR66HTqjtLcDGxs1YbE2A1ybkVLy7wLmZlw8rgOXz8DPr4QGxNN8/q1mT59Ou++W7LIf0WQmQnHjkFubgJubhkkZ1QnIjKF3Jx06jbwwtICLCwlpmaQkw0ZGWq3o9cLIJu///iBb7+ZzuLFi+nX96ZRi4iA0FDw9QUj6pjdRm5+AedDLuFVCsdo6EXBkcMCKyswMzvD2Nc64+Liysbtv+N5XeL0H+TlKReAqSk0aVa4MFlODpw4JnByVv7yrCxYuewMM6YM4KuvXmfMmDGlem8GgyQhI5eSfKsiI+DKZR2urgYO/ZXH3C/mERezkJ82fV+oFnxGukoPdHFVO8+SkpAQT9tmTzLm9U+YMeNxo2u/FMXhw4dZtuIHxk16u9Q6SSnJcPGC2mX7+SuDWtoNyK0dqKytJQH1JHGxITzUuhljx47ls88+uzH23XdnM3PmDlb9vBJbWw+kQV2YU1JU3wQPT7CzU20KpUEZ76xMQXo6pKYK9HqJra3E1laQlaU2ENcxMVEBe2sb9R6uu02lBH2++i7n54PBoI6xspZUdzOjcSPjp2VqBr8EBJ5TH+x1Jc3nn+7JpdCLhIWFGU0DpDQYDMrYnzsXwpQpr2JrX5cv5s3Fv4bExbXoYG1BASTEQ2yMICNDIGUqTZvZ4e5qwt8H9tKuXTtsbW25cAGio1UmjxHkbe4gLSef7LySR4gvX4ID+0ywtpI4Op3ltVcfxcHRkY3bf8e7uk+hx0ipZBbS0pSxtymioUTgeXUhbN5SiWAFBQpGD5tAwtUthIUFlcnnfDkqjqTUtCLXdivBQYLEBPDykvy+I5t5X35KRsYyfvl9Ew0aNb5j/PWLREBdQ6mapyQnZRMZ6kBAHR0V2Dn0BlJKOnbsSFDwBf46fgbbUmg1R0XC5Uu6G0a6qM+upCQlQcgFVZhXp67kwN519OjaBZ9rSqZSSmrXfgQLq6Z8MGcOV68K7B3A1lbi5a2M/T83C5mZysWTkiKIjVYVuOnpAnsHSc3aEi8vibWNKsCzsLj7xcpgUJuP1BRIThI4WJnRormWh18leFdXSppX49TPz73wIleuXGHfvqLL5CuSkBA4eTKEMWNGkZfvxtQZI2neUmnBFHf9MTFR/uAmzSQBdQ1YWtpz5rTgyPGr9OrViw4dOhAXF0ft2kpmIigIo6fzDRkyhKmTS56pERsL+/eaYG4ucfc8x+hhj2Frb8fPW3cUa1CjItWXsUbNog1GUpJqJl3dRxn79DTYv+cMZ09vZfLkN8tk7PV6Pe1aNeODd2eUaHztOipzJz5e0LWHFUOGjsXM9CkG9OlHbMydCfTe1cHeXl6TfS75upycrbCwKeDixbLXPZSGrVu3cuDAASZMfrtMxt7VTRZ7oS4Nzs7QrIVylwSd1/FA277YONzURjh69BhhYbm0aduNS2ECC0ulq9WilcrOuW7sMzPUHcPxY4KTx3WEherISAe/GtClq4GevQuoW1/pLhXoBc7Oys1UkjsTnU595zy9VApvkyblf9+FzlMxp/1vYe+gvmTRUSo1rtsTT2Fnb8+yZcsqfS3x8bB//2XGj38HCwt7lq/5iB4965XYpwvqD9CtmtrVOjlBSrIbH366lQsXLtKhQwciIq7QoIHKiAk2Yo1Mamoqq1evJvd62tNdSEqE//2mQ6eT+PoFMuylx7GytmLD1t/w9Ss8QAvKcIdfUel0HkXI0RsMcDlMYGV9s+nNlcsCK0tB9+5NGDZsWGnfHqDUGvv27cvWTRtISIi/63gTE+We0eshI13Qvac9L7z0FslJbXmhT18y/pECLITapULpUzU/nfMGI0e+SUpKxd7V6/V63nrrLWrXCWDAoJdKfFx01E1jH1C3/L0BbsXc/Jq+kYvkUpiO3387T8dOnYiMjMRgcOGRzoPx8HoQSytJ8xYGatVW898w8kcFJ0+ozCALc6hV20CrBwy0biOpWUu1R/XyhmbNJZ5ekqgoYRS5a2OjGfwS4u0jyckRJMSrDju9n+nLrl270FfiJ6rXq4Dql1/+hM4ki2U/zqJ1m8L91yXBxER1WvLwlPj4PcRnX+/hanwinTp1IjU1ipo1VUaCsSo1N23aRG5uLk/36XfXsWlp8Nt2HfoCqFHzAkMHP46ZmSk/b92BX40aRR6n16vgroVF8el0sTHKz+pfQ6VpJicpX+yTPZuwffuWMqcPAowYMYK8vDx+XLmiRONtbFTqZUqKwMJc8NTTnvTtP4Xz5zwZ+tLAO/7GLC3V+LQ0QVTJJWJo0y6AixcD+d//zpXm7ZSaJUuWEBQUxNsz3ytxU5PoKLgUpsPFVRn7ikgYMDFRcRoXV0lysjN/7g9jx+uvU23RD/R38MDVyZyHOqlAbFio4PDfgr//yuXUiXjSUuNwcIildp1EGjQy4OF5Z8N6UDv1mrUkdQIMpKaqmEtp7sQqGs2HX0KkhJPH1V9h85aShIR4XBwd8HCuPJGSoCA4eFBiYpGMm0ckrdsEGO3c12+lY6LPM+mNR5gxfTKTJr3F6dMqR79VK8qt/d2tWzcCg4I5dOp8sXpE2dmwdZOOjAyoVz+UEUM6I6Vk4/bfqV2n+PccFChISoTGTYqWr9Xr4fhRgY2NKoa6/tlu37KVKW91xNOjiHSeUtCxY0euhEfw14mzJY7zXAoVREcL6gQYCAsTfL/wMFs3TWHQkAZ8/PlXd/zOrvv/mzQtWS/V1JQUmgR0pVu3QaxfP97o7fOucz4wiKUrVjJ+8tsl0p2KiYawUB3OLpJ69SvG2N+KlBB8zoDNE42om3IZGyDP1Jpw79pMaz2Y0MsxRIRfJjUlhLy8GCARuOnbNDc3x83dnerVfajXoBENGzWmWYuWNGzc5LZGLqkp6u9Rp1OpuKXJvqkoLR0tLbOECKGqby8E60hKkri6ul2LuBsqJXCblgbz5x+hXqNGdGxnjVs14xl7UL5hC0sDOl195i86QdduHoCkfn3BkSPKtdOsWdlTNePj49m1axejxr5ZrBHQ62HXDh2pKVCvwRVGDe2i9Mm3/XZXYx8bo9Lp/GsYitUqjwwX5OercaAC2ceOnOXrL1/D32cyb775Zpne462MHDmS/v37c/L4UVq0eqBEx/jXVHr4YaGCxk0lz7/QhqSkiaxY8jZ+/l8wauzt66pVW5KeJrhwQdC0iCykW3FwdKT7kw/w+6+7uXJlJDVrVkz3Dm//WkyYMq1EY2Njbhr7uvUq3tiD+htuHP4rNlkxWFzLp7LUZ+J55Qy5EfPIqONLu/bueFfviLOLC46OTgghKCgoIDc3h6txccTGxhB++RLr165h+fequNHR0YkHOzxE58e60u3Jnri6utG4qSTwnOD8OUHDxsaJSZQHYzUx7wZ8BZgA30kpP/rH6y8BnwBR156aJ6UsnUj1PYCrG4RfkURFCJydJUcO/c340SPZtu0XatWquPxmKWHu3EPMm7eckaMb0Kf/qxUyj6srmJlKzp/z5PRpCLE8zQ9L5vPOOwu4eNGE2Fgoay93vV7P0OEjeabf80WOMRhgzy4d0dGCmrUjGfdaZ3Jzcln/yw7q1qtf7PkzM1TxlJPTTZ98YeTkQEwMVHOX2NiqOcOvCNavXYybm5JIMAZPP/00p8+ep5pP0e6nfyKEqgQ+dUJwIUjwUCcDqaldSUlJ4L0Zs6ju43ubFIWpqXJbnTur48pl5Uq4Gy+9+hyb1r/NypW7mDHjqbK8tSLJyspi5GuvMXT0G9QJqHvX8YmJqkLW2VkZ+8pIeEtNSeX7RT/iPn8BY/Nu97XYCsmKtwaQM/ntEp/PYDAQceUKR4/8zcF9+9i/bzfbf9nCpDde58EOD/Psc/15vNuzhIXacva0oH4DiX3ZW+2Wm3L/ioUQJsB8oDvQAOgvhGhQyNCfpJTNrj3+dcYe1BfS00v5TtNSwbu6DxcvXmDFipL5asvKoUPxfPTRJnz985g07cUKncvB8WY++s/rgli8+HvmzZuGoyOEhZU9a8fDw4N3P/yEevUL+9NQhnf/XkFYqKC6TwxTJzxCZkYmazdvu03LvjAKCiA4WOXb1w4ofpcYca2bmY+vMo6xMXDqxHlOHPuJSZMmGq37koWFBY0b1sdEV7otq4WF0vLJyhKEXxF0fxIGD3keP//RjB72Fof//vO28Y5O4O0tiYlW2kJ3o0279rz/8QRatnz0hjCgsViwYAErli8nKTHhrmOzslSqpK2tpG79ijf2IRfDeX3EbJrWe4E5szcSZF2bHGF52xhpZY2hWfNSnVen0+FXowbP9uvPF/MXcvRMMP87eIgxb04kOiqSN0YNp23zmqxdM57I8BDOnxNkVKEUlzF+zQ8AIVLKMCllHvAjcKdU4X8Edw8wM5NERQo8vbzp+MijLF++vMKanOfmwujRP5Gbm8Di5aOwsrK8+0HlxMlZKTV27NybJ3rO4KOP5nDy5Fr0elWUVVri4uLYtXsPBcX8jk4eU7taD8843p3WidTUFH7a9AuNmjS96/kvhQmyswR1AoqXr83MVHosnp4q8KnXK3G89Wu/wc3NjJEjR5b+zRVDQUEBw196kc+LUNEsCidnqF5dEherjEPPXia8NPQVHBwGMui5IYSFhtw23tdfYmUtuRQm7qqFJIRg4Ms9Mbe0JDa2tO+oaDIyMvj444/p1LkLbdq1L3bs9SbgOl3R2kbGIiwklpGvfEzHNiPZsPZPHurYijXrP6XxtHUEOTYlU2eGAcg1syQx4AHSHuparvmEEDRs3ITJ02fyx7HTbPp1J127P8Ha1d8xalgLpk8ex4plR8lIr5rYqTF+1d5AxC0/R1577p88K4Q4LYT4WQhRZAK1EGKYEOKoEOJofPzd09oqGxMT8PCEpCSlU/7cAJWTv3fv3gqZ74svDnDs2AWGjmhC02bFuzWuk52t9EdiopX+++UwQXSU2s2WVPfNxUXpfw96ZQKNm77IiBEvkZp6htjY0mvor1ixgse7PEr45ct3vCYlnD6pcpvtHZL46L0uJCcn8uPGrTRt3uKu546/qvTrq/vc2abwn4RfEZiYqIwrgKgIQWZmPm5u6UyerJqaGxMTExOys7JY/v0i8kt5a+TrL2/k21tawVO9rRn8yhjy8rvS/5nnSbxlF63TQY0akuxscVsPh6KwsIAdv69l7ty1RtODmT9/PvHx8UycWrzvXkqVTqrXqz7C5UiGKpaY6BwmjvuOjm2HsWXjHnr2eoCDR79lxY8zMTVtTESEKYdm70Ws2YDuvffIWraGv2b/wvkgU6PVngghaPtgB+YtWsLJoDDeef9d4q/+xqxpE3ns4eH8tOpnCiq56XC5s3SEEH2AblLKV6/9PBBoI6UcfcsYFyBDSpkrhBgOPCel7Hy3c99LWTq3nT9fiTk5u4CPbxZN69agd69eLF++3KjzpKbCnDnRHDi4llXrX8HcvOgUt5wciI0WJCXdXtYNSvPjetk2gIWF0vZ2dJK4uhYfiI2LhWNH05j85vN0fsSLceNWIoTK2inpzqxFixZIoWPHnoN3vBYUKPjzoMDSIoX5Xz1KbOxlftq4lVYPtL3rebOy1MXCxkblWBf3PtJS4czpmxIZ2dkqM8fFFR5saY5OUCGdzLZu3cpTTz3FdytW82Svp0t1bG6uWqOlpSqWu3hBsHr5Fb77dhqNm15lw7aNt6WPXrwgiL+qOnXdzU/8Yt9hHDmURlDQWjw8ytfVIy0tjRo1atCsZWtW/7yp2LFXLgkiI0WpK4VLvJZUWLPyD+Z/tZircVd4rGsT3nl/LLUDlE5I4HnBwf2ClKQknnzCkU4PqTvmnNw81m/9AxubB7GzU9lbFRFAzsvL48dVPzP3851Ehl+mVp1sJk+bxBNP9b4t+eNe7ngVBdy6Y6/OzeAsAFLKRCnl9Wqb74CWRpi3yjAzUzryCfEghBWTp8/kqd69jT7PwYNg71yNxStGFGnsc3NVS8bjRwXR0UqNsWYtA81bGmjdxkC79kpVsHUbAy1bG6hdx4Ctnco7vxCk49gRlctdVC2Uuwc0bmLPzPeXMn7KIurUUYY2IqLw8f8kODiYEydOFJp7H34FDv0pMDVJYf5X3YiJucSqdRtLZOyv9501MVFugbt9OS9fElhY3AzoXg4TXL4UiqkIwUQnKqxtZY8ePfD19WXFktKHrSwsoFYdSUaGkkYOqCvp3dePvs9P5MQxW157dfRtrsQaNVXF8HUNmuIY/OqzpKbGs2bNnlKvqzAGvfwKk6ZOL3ZMfPwtQnZGNvZ6PfxxIIbnn32Hd6ZOw8Y2mrWbZ/LD2q9uGPvoKOU6NDeXbFj3HjPevqnW+ukncxjY71EsLK6Qmiq4FFYxfw/m5uYMenkAe//8nremzSIvrx5DB7/M4w8/yM7ffr1rg5byYgyDfwSoI4SoIYQwB54Httw6QAhxa27HU0CgEeatUryqqw8mJkowZOgIHut+p8Jhefjzzyu8M3M11TziC3VVSKlcNCeOKckHD0+l9VO/oRKasrZW1YXXNw3m5spv7e6hgrKt20jqNzRgaany748fFURGUKih8PGFRk09CI8w5VJEHKdP7yY8vOiLxK2sWbMGIcQdjU7ir8Lff+rIzU1lwddPEhNzgVXrNtH2wQ4l+v3c2nbubmJg8fGQni7w9VOpi0pSAZYunkL/Pk9U6G21iYkJw4YNY//e3Xf43kuCq+tNaeSkJJVzP2BQE3r0HM2vv1xlxpT3bow1NVUB39xc7mqwOnfpQjV3HWvW7Cx3NaiVjS2TZ8yiWYui93EZ6SpIa28vS5RNVBrir0o+nr2DF/uN4vzZ7Ux7txf7D23j4U4db4xJTrpeRyOQhiguhR3m6acfvvH6yy+/jImJCZs3zr8RBI+NMeoyb8PWTsfQEQ8xd8FSJkz5mbS0bAb2e4aejz/Cob/+qLB5y23wpZR6YDTwG8qQr5VSnhNCzBJCXM/7GiOEOCeEOAWMAV4q77xVjYWFkieIi7vWGu/SZaNl6xgM8Oabmzl96k8aNMq64/WcHNUEJDREh729KgSrWevuhu9WhFAaI42aSFq0MuDoqES5jh9VF5B/bjRq11E9Vqe8tYw333yR8+cDCQu7+zw7d+6kXYeH8PC8KSWdmgJ/HhQkxKfy/bf9iYs7z6p1m2jXvmTGPjKi5G3nDAalh29jo3aV1yUVzp07xOmTW5k8efJtxTIVwSuvvMLkqW8XK5lcHDVrSWxsJBeDBbm5Shfm1RGdaNt+MN8tPMTCectujLV3gOo+cDVOUFwIzMTEhH79u3DkyDGOHy9Fue4/WLp0KZu2/FLsmPx8VSRmZqYqu40VpM3Ohn174hnc/x2+/vw9GjctYN9faxk9btxtFb6ZmcqVk5Ii8PQ2cOTQLoSIpH//mynC3t7e9O7dm7WrVuLumY2Tk9LET63Ans+2dtCwkaDDw48zb/FxPvz0GyIjwunVrQvPPvM0WVl3fvfLi1F+9VLK7VLKACllLSnl7GvPzZBSbrn2/ylSyoZSyqZSykeklEHGmLeq8fKWFBQo+dWfVq9i8ODBXLp0qdznXbPmBIcOhdF/YF38a9we/05PU37rzEyoXcdAg0YSy3Im7lhZqUbbTZupHf/FCzpOnbi9ebZOp76sLw8bgotrG2bNeo2QkHSK6fUBwPbfdzF/0c0+wJmZytiHX0ln2eKXuXr1b1b/XHJjn5BwTVLYrWRt52KiISdHSSgIoW7rs7Ng7eqZ+Pn5MmRI0b1qjYWHhwcfvP8e1crox7je3xZUf1sTE2jWAsZNeIb6jZ5i5ts/8suWmw3UfXzVxTnsLgJrg4Y8T5NmAVy6VDbDkpyczLhx41i25Psix1wP0ublKc2gEiot3JW4WPh23h+88uJwzp/dwHsfv8Cm7RvvkN3IylIZQUmJAi9vMDM3cGD/dh57rOltevgAw4cPJzk5ie1bNxJQT4naBQdVrDSCvYOqwi3Qm9KsxRD2HTrLlOnvYmpqarQU4VvRtHTKgY2Nan4QGwNP9+kPwA8/lK0p93X0esmMGZuxd5BMn3V7R6i4WDhzWn3hGzeVuHsUcZIyYmunzhtQz0BenrqwREfd3O2bm0ObtnZMmPwOsbGZfPrpBwQHy2IzPQowuaFXn5urjH1QYBrLvx9KfMJe1qwvuRsnPU0ZDzs7SZ2Au7sF8vNV2qWzs8rgyclRefjBQfs5e3oP06dPx7y4PE4jIqVk145f2L9nd5mOt7JSNQaZmYIrlwTW1tC8hY5JU16ius/DjHzlaw7uPwncFFiTUrlRivp8fP38WLLyK1zd6pQpJ//zzz8nLS2t2Kra8CuC5GSlWlpc9XNJ0evhzKk8Jr3xDR+9/yY1aiWx+89NDB3x2h0V71lZ6k44L49rWvWS4PNBXI09zMCBd9azdO7cmVq1arF75++YmipDLKWSR69IySwHRzVXTg6Ehdgw4vVJ/LR2XYXMpRn8clL9mnSyubkf7R/qyIoVK8oVeJk373fCwlIYMbo9dnY30wRjYyDkog5HR2jaXJZb16YohFAdtJq1kDg6KkGr82dv7nJsbKHbE/V4YfAY9u//H2vXbi9UXM1gMPBwx46sXqXa9un18PefguNHEvl+0TBSUvbw44bNJTb2OTkqSGthoe5GSuIWiAwXFBSoFEdQWjVCQG7uWRo1asSgQYNKNLcxEEIwa8Y0Pp5d9qY5Li5KNz/6mn/ZwRFatrZk6juvYWdXl5f6f8AfBy4C14P3ktTU4gXW3KrB1YREDh8OL9VaEhMT+fLLL+nZ+5lCdftBbVAiIwTV3ItWLS0Naanwy+YIXh08lt+2f8nwUY+wbdfv1KxV+46x2dnK2INqZGJhCQJo1KQ+O/+3ht6FJFnodDr27dvHd0tVtp2VlYp3ZWernsgVGU91dFJGPzcXzlSg4Jpm8MuJnT04OCjp5L79BxIaGsoff5Qt6KJ2ES3pcGrIRAAAIABJREFU2qM9o8Y+ceP52JibJej1GpS8vV15MDdXhrV2HQPp6SrgFRujdvvV3GH4qGfpN2A4LVu149KlOytw//zzTw7s349Op4qBjh4W7N8by/eLRlOg/4sNv/x61wKd6+j1apdlMCi3UkncAllZSkLB3UPdiSUmqtoJH1/JhAmvc/z48RIrORoDIZRsw7Ejhzl7+lSZz+NfU97wL6enqc+idRtHZr4/ERMTe14dNJMDe6NufE6ubpLwK0VXd5qaFjD6ted4991vSye1/OmnZGZmMr4IGYK0VJU95uQkqVW7fJZSSgi/DHM/38nYkUNJS9nLynVzefeDjwq9Q7tu7KUE/xpKb8jaGnJyBf7+gi6dO2Jra1voXN7e3thYmN7IfnJwvNlEvqIyd67j6KRSagsK4NQpVUFubDSDbwS8qktycwVt2/XGwdGR48ePl+k8oaGQm+/EW28/hYWFCiTGxd6iN1IJJej/xN1D7fbt7NQ6ru/2a9UWjBj1IsLclsREPRcv3p6ys2bNGqysrOjavSenTwq2/xLO0kXjsbY+xpbffqNJCUvY8/PVlzc7W+22Snpnc+WyquT08VVfoEuhAp3IIixkDzbmJpVq7K8zePBgLC0tWf794jKf47rejoWF8i/n5aksqgfaevHO+9PJy81h9LBp7NuTQH6+Mlbm5nChiFRNExMTnuzVhv37D3HhQskLHWvUrMnQkaMKlcrIy1MS1ZaW5a+kzc2Fw39nM2HsHOZ9OYmWrU3Y89f/6PJ4t0LHX3fjGAzQsJEkIV5gaioxFMDZM4eY/+UEUlNTi53zpzWrebBFI7Kv+bncPW7KVxTSk8ao2Nqpnb6X151dtozBf9LgX71aspRBY+HsDNbWkuRkW46dC+G1UaPvftA/0OsNjBy5iNj4UGpf808nJCg3jpNT1Rj761haqkKUWrVv7vbjYtWX2cY2nwkTZzNx4jc3/MB6vZ5169bxWLceREba8cPSsyxbPJVq7ufZ8tvvJRLWAuXGOXNK9QitW//uGTnXSU1Rnay8qytjF3FFkJsr+N/Or+nbqweBgVWTFezk5MTzzz/P+nU/kn63aHcxmJqqi19+/s0mKP41JW0e9Gf6ex+SmhrBxHFTObAvjdwclaqZkw1B5wuXXnjp1T4UFGTz3XfbSryGfi8MZtaHn9zxvMGgAsv5+SpIW5670bRUWL/2IiNeHsmBfQt4a9pzrNuy5f/snXd4FFUXh9+76b1DAgESagi9CNKVKqDYwAaCyKeogCiKgIoiChbsFZCOCCqIoqgUEUWRToAQAqEEUiG9bbLJZu/3x01CQhLSNsUw7/PkSXZ2Zu7Zze6ZO+ee8ztFMr4Kk5io1p2kVM5eSnVXZ+8g0WcKdv++jB++/67MimpfX1/CL1xgy+ZNBdua+Uvc3ZV8RUJC5V9TeXB2gSZld8esFPXO4asGGEVDEDWBbxMleGUwOJBlNFW4lH7p0m388cdJrK1DsbZW2Sz5ecvmTGWrCt4+RWf7Z0IFrdtY0L6jF7/8EsIHH+xESvj999+Ji4ujd7+JLFr4F+vWLKJtuxh+2v7rdTtVFSYjXTn7nBz15fXwKPsYUP/v8+cEtrYqiycjXTVwsdBdZsXSt7jvvvsIDCxZwK0mePLJJ2nYoCHhF8qR03odHByvNk0Jzws1NG8h6dWnDbNfeZ/LMceYPeMZ9vyZjDFXOf2UFLXgey2B7QJp39GHb7/9nezs639hYmNj+fyLL0jTFw8yS6nuOlJTBS1bKzXSyhIbY2Lh/E3MnP4EkkNs/mUtz84sPYU24hKcOqnDzi6vh7GjWqC3tJRkGwRZWYn89cfXPPzww1iWcRUaMGAAbdq0Ye3Kq8Vy+QvhDg4qnl8eobq6SB1wI+bF0hK6dQdHR+WUgo/XTMcZTy+wtVXSyf97dCJ33HFHuY/NzTXx9ts78fJyYvKUQaq1YKEq0rrg7POxtVU9N5u3MJGaqhZSn3p6PH7+LiyYv5nT762k/Q8/8FqPkWz5LomtW1YzZJhk8y9b8GrQsFxjJCaqbCQhVJ1AReRkY2NArxc085cFDsjaGlavmIPRaGThwoWVfOXm4aabbuLMmTN07NS5yudq0FCpt0ZHq2wqIVQh2i0DO/Hiq18SG3uSubOfYu+eRLKyri74lpSfP3bCcC5diuTvv6/f0/Ltt99m2tSpREYUX+Q9f06lP/o3N+FVyR4yJhMcPpjEo+PmsWrZIgYPa8Afe/8qdXE/v+L60kUdXg0k7Tte7U+cP7vPyhIEHd5Ibq6RRx55pEwbhBA8/vjjHDqwn1Mngwu2W1qqO107OzVmdeboVxd1yJWYDzs79Y9p2cqEXg/Bx5WccXUihMrLT0sTeDZoxfbt24mMLF9By9q1v3PpUjbjJt6Mvb0FYafVRapNQMWKqWoKJROtsoWcXSAtzZoZzz3FT9k/03TWkzRasoTnDu7gtX9fYOKj3qxav6JcOcVGo6qgPXVSh62tShGtiJ5ZTo5KA3R1VRpB+Y2+E+L2sHnjOmbNmlWtfQvKgxACS0sLZG4OcVcuV/l8/s2v9mmNj1f/mzYBksHDOjF3/gri4i7w6osTObjvMplZ4OAgORdWfBH3vgfvYu36Vbi6BpQ6VnR0NF988QVjHhxbLDPmYrggNkaJ2F2vH8H1MBhg7cojjBszheDjP7Ng0SOs+vpr3NzdS9w/v3o2MQH8/E1F+uBGXBJYWEgy9QJ7exM//biEm266iXbt2pXLlgkTJmBtbc3Xa1cV2Z7v9G1tVTFXbUodV4Z66fDzaeitVr1BCWedDavefNoGDZV0cq/eY5FSljsnf9GiXbi6OjL5qQFEXBIkJqq85dpslFAe7OzUAlPbdiZ6JR+jj2U89iYDQkocZDb9rBN5b1jvMitZc3NVQVS+TISvr6Rj54pf7C5dVGmYfs0lVy5D3BWBbxMwGuO4qUcP5syZU4VXaz5MJhP9enZj9ozpVT5XvoN3dlaVuKkpalvLVpIhQ9vz6usrSUtN4cUX7uLQ/jPk5Kj3+1SIKLLO5eDgQGDHhiQmqUlSSbz55psYjUZmzCz6PsZEq/TLht6SZn6Vi6EmJhqZOX0Fs2bMxs09ml93rWLS40+UqHGUm6su5iEndVhYKLmJwkV4aamQlKTCOTk5gkaNDfTo0YMpU6aU2x4PDw8+/vhjxtz/YLHnrKzUXa6VFYScFOVWoK0L1GuHDyrW2aWbpHFj5QSOBwmSk6pnLAsLdYttZ9+Em3oOZvXq1WXm5F++bKKp3wAeHN8bsCLikspbLmVdqk7i7g5t9EFY5RSt3rHOyca4/3iJcrMmk5qhhZ0RHDoguHBexV87dZY08694GCstFWJjBD4+yuGdPydwcZE0aSp54P4x7N+3r0qNyc2JTqdjzJgx/PLzlkrp6xQ/n0pXtblm1unXXDJ4WAALFn2FtY03c+fcwf59+8nNhaQklepaOPXPzj6DF1+cw7vvFp+oREZGsnTpUh4YO75INWtc3NUWhZVNv9zz5xluH/IM365fzX0PtGXHnq2069Cx2H6F9aNiYwS+vpJOXYr3ilVNblTmnLu7pKG3DUu/XMqECRMqZNfkyZPpfXPPEp+zsVFOX6dTWUH/lZh+vXf4oByxX3NZMNs/Gaxm++bSvS6MTyOwsJAMHDyJ06dPs3///uvuf/SojlsGDmLqM0M4d1ZgZ29+canqJiM9nU//CiKDojP5HCs7whw7cWCfjgP7BMeOKt37fXsF//6jI+SkjsQE1fCjYycT7TtWbqHPZIKwMLVQ69tUzXR1OrC2Ps/alV9iY1F9apiVZdq0aVhbW7P4k4/Mcj4rK6Uvb2Gh1j8S8xxQk6Zw6yBf3npvBU18+/PGK/ey9cfV5OYqrf2Q4KtO393dgdTUcFat2obJVPQzGB8fT/uOnZj+/KyCbclJKkvI2bly/WgNBiOvzFnBg/c8R3xcGF8sf4aPl3xUYhZNYiIcPaL0o2xsoENHU4kTg/zZfW6uuhD6NMri2JHDWFdSajjkxDFWLVta4nP5oWNrawg5qSO8HA1oapsbwuHn4+yiskzyuwnlK0Sa859kaQne3tCh0zDmvfE+rVq1KnXfoKBwvvr6AL7NjCQmqFvsVq3KbkZdlzh0YB8Dbr6Tl/924IxbACZ7B6QQpGPFPulJ9pDe+PmbcHNXTsnBQUlL+zZRoaCbekpat6la2f2lcNXxqkVLlTaXni7wb2Fk5jP/Y8G8uVy+XPVYubnx9vZmwoQJfPP1WrPE8oEC3Xx7e5WCmb8469UA+vV3Y+G7H9NvwFOsXjGfpZ9NwWhM58hhwdHDV0OdYycM48KFSH755UCRcwe278gvv/9VkGWVmqIWLvN1mCp6R3b44DGG9HuSpZ9/S/9bm/P3obXcPbpov4DcXFWHcjxIresgISDQRIdOpYc7Iy4JMjIkoNYT9vy5lWG39uXPP/+smIF5fPftN7w86zni4q6U+Hx+VpBPI0lUlCD4uDB760hzckM5fFBX/Wb+ks5dTTg7KyGuo4fNu+Lu01hi72DP0BFP4epW8oITwLx5P7J+3dc4OsRx5bKgsS9m0RupCQwGA2/On88dQ59Er+/EU8++RPb2/cj16xHz53Pg2YUMzBnBjGcW4+JiolVrSWB7NRP0b6Five7u5W+iUhppqRAVJfD2kaSnq7h902Ymvlv/Mfv/3csnn3xCo0Z1Mz723HPPkZ2dzZYfvjfbOa2t1azTyUnNvvMlfp2coUcvW+YtmMOkJz7g4P5gFsy7jbgrRwk6ItizW93xPvTwKOzsLViy5KeCc27YsIHI2KupPTHREHxCZT8Ftq9Yrn1KcjLPT5/D7UOeJTY2mfc+eYqvN35EgwYNCu2jQn0H9wvOhukw5qoeD527Xj89Ny0V4uMExhxwdpb4NoEN69bi6+tLv379ym9kIR555BGMRiPff7uh1H10OpUS26aticxMdYGqarO+6konr3LHq+qkJjpepSSrBaDMTFUG7tfcPDo1Z8MEMdG5xEZ8Q/tAf2655ZYiz0dFJeLnN4+evVsw++UnsLJSM4W6lIJZGn//9ScvPDOH8+cs6dNvLPeMeYQWrTL58vMnWfD6PDp16oSU8Oyz3/DRR39w733deOPth7nOta9SmEwQdFQgTeDhKYmKVCqaOdlBjBjUn5EjR7Jp06Y6F84pTHBwMA2btcRk5q9hbq7KF09MVDPdwoupcXGw7ZdQPnjnPaIiDzNg4Fh6951CMz97hg43MXP6a2zZfJgLF74hPv4SXbp0YfpzL/DCS/M4d1Zw5bKKjbdqU35nn5WVxaplX/Lhu9+RkuzAyFHDmbdgPL5NVDWdwUDBQntmpsqw8fSChg3Lf/d38oQg9JRaA+vaTZKWFk3XwFbMnj2bBQsWVPQtLKBnz56kZejZ9c+BMj9LWVnqfU9LE3h4qrXDikzi0tMgJkZggxU3dTd/x6saUGWp27jkiZHFRKvG5EFH1MJfI9+qpUT6NlFho/mvraTHTc7FHP6iRVsxGgV33TOUnBwVf63rzj4qMoIF817h++/+oaF3b+a8Mo3Adt3xaQQH9q1m608/sPANJQ4mBCxceB+hoRHs2nGcA3frCWxvj2+T67dUrAj5oRx3d1OBs/fzz2Zg74fx8vJiyZIlddrZA7Rv357M7FziUzPMqtxpYaEWcs+dVRk0+gzwb6HSCb284IGxAXTt/hmffbiBzRu/4uC+w9zcZzLx8X244+6xONi0JmvjDo5/9jL32dvz0EPTOHFMhcuaNFWL4eV5a/V6Pd+tX8cHiz4jNsaZDp368MGnExg8LAB9hmp5mJ4OycnqZGqh3YS7R8WkBVJTICJCYGkJzfxUssbyJeswmUwVXqy9lokTJ/Lkk09y4lhQmZIg+enEEZfUnVBCvA4XF0mDhirF2N6h+OffaFQJDDHR6kJhYSFp1UxNaMztE274GX5hsrNVat+Vy3mqkQ2gsW/lGy1fOCdY9OYKftj0AlFRYQX620ajiUaNZmHv2ID3Pp6On78J32oqpTYHyUlJfPLBuyxbvBKTqSV33vsYEybehzHXFmtr9QEfM2oQ6akpnDhxooiTjYjI4d9/U7B2cMaEkpNu1brqAnBxcXDmlE59KSzUDD9/4fDwv3/i5uxMr169qjZIDbFs2TJemz+fP/cdwaEUUa+qEBOtWjwCNGqkmrjnv/9pqbB71zkWf7qWwwfPYGfflZ49erLs4us0iTqERXYmOVY2pAb25sBbP9MiQEcpafFFuHDuHGtWLuPrNetISXGiVevbuO+hRxg4uAM6nSA1FYxGgRBKd97TE7waVP67djxIJQW0bSfp2l1NnoYO6I2Hmyu7dlVOkjqfpKQkAgMDmbfwHe66d0y5jzMa1RpEbIwgK0u9/5aW6jVa20BONuQYIVOvnrOzl3h7Sxo0hIYu1dPT9oaf4RfG2lrlLzf2hego5fivXFaNrn19K55B0riJZOCQIXz/XVPWrVvHc889B8D+/bFI6cWAW3rh7Fy+Rh61QXx8HMsXf87yJUtJTXXllkEzGDt+PN1uakBsrMAk1SzycuxF9v+7lwULFhSbUTdpYkVqqieXLxtZumwR7TsPIVPfgTZtK1ZUVZj0NLUomRAPHp7g7aMym0JPneSmLp0YMWyYGV59zdG+fXsiIyJYsfQLps2Yafbz+zRSF9qL4aqBeEwMuHuAt7dKabzjrhYMuW0e234J4tMPv8Fh9yc04ADWqER9ixwDVmcO0Dv5N3LcR5Q4hsFgIPh4ELt27uSXLVs5FRKFEG3p2Pllet48lPadmuPtA8nJSnfKwxNc3Uy4uVVdJCx/AdnFVRJQSHNqx67d6FOqni/p5uZGZGQkSZlGcisQe7O0hMa+qiBTnyHR6yE1VZCpVyJv1lZgZwuenhIXF7UQXd03pNoM/zpkZ6uetTExkJurYvzejZROfHlvtS5dFIwdPR1LiyBOnvwXIQSbN0v+2ptD774mbrqZSs9qqovQUyGsXr6U9WvXkpXlQq8+DzP6/kfo0as5TZupTkpJSUrjxsUVPnl/EQtee4Vz587RvHnzYuczGuHPP9N49NEniIs7wEuvfUX7Dp1p5qd00ivyITcYVOPziEuCJs3UrL6hN6z4cjEvzZzBzp07GThwoBnfjZph5MiR/PvvPg4cP1XpVojlIT0NYmMF8XHqM63TqRmnrZ1aKDRkgcM7s+n164dFkmxzgU8b9GRL58E42NthZaUjK1NPaloiURGXiYyIJycnF3CmeYubad2mN+07dqRlGxf8m0u8vNQF3sERs8t7/71HcPqUjlsHmmieVwsgpcTT0abSs+SSSM/KIfpKAi6u5VTxqwIeDtoMv8axtlYZPY2bQEy0kkc9dVKHpaWaofj4lD3rb+wrGTR0EL9vTyIpKYmkJMn+A/a4uVvQqk3dcfapKSn8uvUn1q9dxb69B7Cy8qPfLbO4856HaBvYJK9jkZolJiYKWrQ0FahXunl4MG7cuBKdPagveJ8+Tnz++cdMm/YM81++i1deX0tubn9iYtTClodn2Y4gORl2/64j7oogINBEl65qhrpmxTJefP5Z7rjjDvr373/9k9RR5s2bR48ePVi25HOenTm72sZxdIKWThI/f0hJkaSnKjVSvV41CLGyAo+h/ZF/LIWsqyW3WTobTuhaEXQ4lmxDOsbcHGxtLLF3dKGBVyfaBDSmqV8ALVu1wsPTmcaNoUVriZNT9Samx0arArLmzU34NVfOPiY6igfuvoMvly6pdHZOSYwaeRsWVjZ89a35sqpqGrPM8IUQtwEfARbAMinlW9c8bwOsAboBCcD9Usrwss5b2zP8azGZlNOJj1P6Hbm5Ant7lV7o5q5W40uarUZFSiIvWtK3pxWPPfY5W7dasuab++l/i02138Jdj9iYaP78Yxe/bd3Crh3bMRisaNS4F4OHjueWwUPx93fBp5FqDwh5cfNQHQ29JS1bXf3c2FpZ4GJXtr58Zibs3p3Ms88+z+nTm3nptQX07TcRvV7NNN3cVSs6O3uwtVF3WNnZYDAI4q/AsWOqsKVnL3XrbmEhWfLZx8x7aTYjRoxg48aNdaaatjKMGjWKf/7Zy9HQc9jUpohSbi6e946C/ftAryfH0obktr35fspWbB10uLmB5OoH18JCZbY5OoGTk/o/1kSTnqws2PK9DoMB7r3fVJBd997bC1m08PVS7zgry5w5c3jnnXc4euosDc3Rwus6VNcMv8oOXwhhAZwBhgCRwEHgQSllSKF9ngI6SimfEEI8ANwtpby/rHNX1uEbcowFDr/w6yvt7zwbsbCwKHdWR/6CTHKSICUFpFTaHa5u+f0zr96+mkxw9IAJ9z3f8+uCVVzx7cKsffOwc6y5CispJZfCwzkWdITDB/eze9fvnD51CnDC0yuQvv3vpWevwQQEtsXbG7wbFV1AS02Bk8ECR0eV550f0jp54jjdO7fH2aF8uawGAxw4kMW8ee9w34N9uePe3mTq4coVdRE1GIq+/1Kqis6EBPBwhz79TQXpnfv2/s1dw4dw77338vXXX9dYf9rqIjQ0FH1WNo2bt641G86fO8s7b7zG5x99gE9QEMumvsbBXA/m7N9E+CVLEhMEdnaSJs1Urr+UKjOlpicuOTnw717BmVOCWwaZaJlX32g0GunRMYD27dqxbds2s44ZGhpK27ZtefWNN3ly2jNmPfe11GWH3wuYJ6Uclvd4DoCU8s1C+2zL2+dfIYQlEAt4yTIGr6zDd3BwQF+aAlQZWFlZYWllpX5bWmFlZYmVtTVWlle329raYm/vgJ29Pfb29tjYOmJh4YnAE4Q7lpb22NjYYmNri5OTJe6ucOdHz+N98TT2GDFa2ZBzUzfivt2MnaOj2VIHjUYjiQnxxF25QsSli1w4f47wCxc4F3aG4OPHSE5OARyxsXGnQ8dBtOs4gE6db6KZvz/u7uDqJvFqUHx2lpWlsiAsLVVGTn6zqOzsbDq3ac6oUXewcuXKCtgJZ86oRjXCIpdffl2Md2Mr7nvwIcCSTD1kGSi4EBgy1YJc6wBVxn7h3Dmat2yBs60Vv2//jeHDhxdrYP1fJjEjG32Woca7ckkpGX3HcIKPB3Hq1Cl8fHw4fz6ekDPONPE3qUXXJDh/Pj8dVukeVVd/5dLIyVFhnODjgkaNJf1vvToB+fXnLUwcez+bNm3innvuMfvYPXv2JF2fya5/DpS9cxWoyzH8xkBEoceRwLWKQwX7SCmNQogUwAOIN8P4xZg950XSMq9KARZ2qEX+LnRbapKqaYkxJ0f9NuaQk2PEmJNDdk52wfacnBwMWVno9RmkJCej12eQmaknU5+JXp+BwWAArABHwAlwYiSSyYTihIpnWuRkkb33X6b59mQrttjYWmBnZ4W9vQW2djY4OAjsHWyws7NAp9MhhCjyk5ubS1ZmJpmZerKyssnUZ5GclEpSUgoqqmZF/r/WycmHRr6tubnPc/j5B9KiZWua+vnh4GCJW56Td3YuvU9sfj9ZKYv3k921YxuJiQmMGVP+VDVQF5TAQJUiePq0iZ9/DCEk5CgfvPMzo+4eRf9bb8Pe3ovsbIGVlaRVGxOOTnp+/fkX1qxYxoF9ezl+/AQNAtowcuTIsgf8j/H8009y+Uo8K9Z9U6PjfvP1V/yz508WL16Mj48KWTRv7kl8vCQyIgNvH0tc3aBLV0l0lCTikqpbadwYfJvWjCRIVpZSqEyIAzeP4v0ivvziU5o2bcqoUaOqZfwJEyYwZcoUgo8fo33HTtUyRnVS5xZthRCPA48DNG3atFLnmD1nTrXE8MtDbm4umXo9mZl69HkXAZ/Fn+OwumhbPQckk27qiHPAMPQZOWRmZmMwGDBkZZKVZSA1NYv4K4a8EmuJlLLgbyEEtra2WFvb4OBojZu7Da0DnHFxdcHFxRUXFxe8GjbE27sRTs6O2Nmpgg9bW3ByVgVlDg5l39lJqaoGMzNVRs61M7nv1q+jQYMGDB06tFLvlasr9OhhxR9/fMr69TtZsmQ1n320jM8++pynZ0zj8SkPEnflFPfd9TBnQk9hMplo1qwZCxcuxN/fr1Jj/hdo2aIFK1es4Pcd2xg0pGZSTOPj43jtpdn06dOHxx57rMhzV64cZdzDr7N87Uz69u+KECrd0KvB1VTPuDho2kzdIVZXeCcjXamB5uYqTSZrG5XllY+UkienTMNaJ8vsalVZHnjgAVzdPWhZzjaddY16GdKprkXbyiJ++gXnRx7G1ng1zGRycCBlxRqyb1N5zSaTWqDMNYI+E0y56rGUhe9Oir9dQqg8Zp1OFSBZWqhMCwtL9Zy1deWr9c6fE8REC1q2MtHQu+hzSYmJdGztx9SpU3n//fcrN0AJhIaG8vPPPzN48GA6depEaGgoL7zwAl26dKF///7ceuutZerr/9cxGAx07tyZDL2e3fuOlNmD1RzMef4Zvlq1gqCgoGItIJOT0/D1vY+bbm7Fhu/fKXZsWqr6rKSnq/h+Y19VPGQuxy+lEkWLilSfbW9vycWLOlq0NHHt2qmbvTXWltUf3otLM2CqxpT2uhzSOQi0EkL4A1HAA8BD1+yzBZgA/AuMBnaV5ezrE/849cbZ6Etvi4vY5BrQY4ExsDvZhWZvOp2agQPXpHrWztsUHaVKvX19ZTFnD7Dt163k5OQwfvx4s44bEBBAQMDVrktt27blp59+us4R9Q8bGxuWLl1K//79WbTwdeYteKvsg6rIK6/OY+SwISX2+3V1dWLs2MF8+eVPhIWeo1VA0a5hTs5KByoxUbX4PBumIypKdR3z9lHrLpUlKVFJQ6SmCrwaKE2gUyfVhaXw5/LK5Vi+WrmcmTOm4+npWfkBy0F6ejofv/ceXXv0pndf86V91gRVvhRKKY3AVGAbcAr4Vkp5UggxXwiRH0hbDngIIc4CM4DqSzSuY6Qkw4pVwQzlQUJffYfzEybwAD68MWAmekPdnKn6O8CvAAAgAElEQVTGxcGF86qpRdNSOhg98NA49u8/QKdO/7045n+Bfv368dhjj7Fxw9ekplRff874+DhMuUaaNWrIvffeW+p+M2eOxcLCig/fL7mLmxDg4aEcf+s2Jqyt1Kz80AGluR8TTbn7T0ipsrKOHVVdrQwGaNVatTBMS4OMDFFMy2f1imW88+YbJCVVU3ejQlhbW7P4s09Z+eXiah/L3GiVttWIlHDkEIx/YA1Ojhe5cGEBJpOJQYNH0LXH04y4fRDtOlS8cUR1kpykFsWcnIqmX16LraUFLvY1m0Vyo5GWlkZSUjK2bl7VIpeblZXFPSOH4u7qwvbt28vMFrvnnnls3fonR059hafndXSK89Dr4UqsIDFJ6cUIoaptHZ3UGpKFBSDUxUKaID1NCallZKgal/zwkFcDdQcspWqCAmrhON/crKwsurdvw809e9TY3eDTTz/N0qVLOXb6Aq5ubmY/f10O6WiUQnQU/LnrIgnxF3nqSdWyTafT8ceu3wg9l01IKFy5XHLIpDZITytfU4v333kTfWoyH3/0YZ1Xo/wv4+TkhJOTE5nZOWz47nuG3z7KbO+3lJLZz03nyKGDbNy4sVznfeWVR+jdZxTZhvLpXNvbq05zfkBGuiQ+TpCWBvFxEBtT/MOl06n6lYYNwcnFhIdH0XWAy7HqwhEQaCqy/eu1q4iPu8KMGTPKZZc5mDBhAp988gk//fA9D0+cVGPjVhXN4VcTBoO6pd239zSWlsk8/XTRbAufhkZ27w7DwiIAF1dZEL+vLfR6NbO3slIz+9KSHAwGA8sXf86AAf01Z19D/PzjDzw67gFmzHqRF16ca5ZzLv38EzZ8tYZXXnnluqGcwnTu7IfJ5MflmGx8GpkqdGfq4AgOjldvUwwGicmkZu1SKsduZ1f6Qm9urvo+OTsXbYKSnZ3NZx++R+/evYtJkFcnXbt2JTAwkG/Xf/Wfcvj1p1qljnExXFXgDhx8GytXzsLTs6jg0vPPTueFGUPJyjJw9oyotg435cFgULn2oDoYXW+R7betP5GQEF8sdU+j+hg9ejQTJ07k/bcX8t2Gr6t8vu82fM2rL87i3nvv5dVXX63Qsb6+uXz+0RcsevPLKtlgY6McvL29SrG0t79+Vk9khCA7W+DnX/SLkpKcROcuXXnllVdqdAIihGDSpEl4enrm1d78N9AcfjWQnqY692SkCRzsLRgxong9wbhx40hLvULoqR9JSVGLWrVBTg6EBIuCJixlSdGsW72SZs2aMWTIkJoxUAMhBIsXL+bWW29l+pOPsfGb9VU63809ezBx4kTWrVtX4QrlBg0sSEq8wNLPfiI5qfoWkwtjMKjwqKdX8e5RDRo25Kcff2BYLUhiz5gxg+83/1C7ukcVRHP41cCF8wKk5POPl3P00OYSG0b079+fdu3asWnD+7i5mbgYLsjIqFk7c3OVrnxWFgS0VcqT1yP8/Hn+2r2LSZMm1ftc+LqGtbU1W7ZsYcCAAcycPoUrl2MrdLzJZOLH7zdiY6GjW8d2rFixotKO6vXXHyU9PZv3F62q1PEV5dJFNXNvdk3G2IF9e4m+dKFWQ4s2ljoiL13EZKpeVVBzoTl8MxMfr5ocREdFE3wilIYNL5e4nxCC6dOnc+J4EHFxf2NpqZxvTd0dZmerHqCpqYJWba4qYl4PnYUFk/73GI8++mj1G6hRDEdHR7Zu3crOnTvxa9IYKSUZ6ellHhd+/jz333U7kyc+zJ87f62yg7zllo4MGtSdtSt3EBeXUKVzlUVGumpC5ONDkXUuo9HIs1OeYNL4ccWEEGuSbdu20b1DAPv/3VtrNlQEzeGbEZNJ9Vm1sZH88tM+IIWnn76z1P3HjRuHh4cHO37bTGA7WSS8Up3k5Chnr9dDm7Ymylun4ufnx7Ivl9K4cePqNVCjVOzs7OjduzdOtlb8vGkDXQJb8u5bC4iJjiq2b9iZ07z0wgxu7d2doCOHWLx4sdk0Zt54YxKZ+lwWLVhllvOVhJSqgtfSUuLbtKhT3/DVGs6dDWPu3Lm1OsPv168fjo6OfPv1V7VmQ0XQHL4ZiYmGzEyBrV0u/+wJoVcvF/z8Su9faGdnx6FDh3h70fs4OELbQElWllpAzc2tHhszMuDEMRXGaRsoy+3sD+zbS8ixw9VjlEal6NG9GwNvvZV333yDLm1b0u+mzjzykBKyEwJemD6FNSuWMXr0aEJCQpg8ebLZnOPNN7dl7twn6N33oWq7K42JVnfL/s2LZo1lZGTwzsLX6dOnD3feWfqEqiZwcHBg9OjR/PTj95VW6K1JNIdvJnJyVCaBm5tk2y9BpKZe4Yknym615+fnh4OtJUajERdXaNVGkp4Op0+ZP3MnPl45+9xcaN9BFnSsKg+vvjiLxx595D8Tq7wR6NixI5s3byYkJIR3332XgDatsdIJGjjZ0MDJlrffepOIiAjWrFmDr6/5GyfPmXM/Lf2bFcTYzYlerzLd3N2VLk9hlnz2MVcux7Jo0aI6kRo8fvx40tPS2PZL3ZcA0Ry+mYi4pBypq5tEmny5/faOPPDA8HIdu+nbb+jevg2pKSl4ekKLlpKkJEGYmdI1pYTw86rvp709dOpSPNvhehw+uJ+jhw8xffr0eqU7X19o27Ytzz33HFu2bGHLli0FTrBfv340bNiwjKMrj50dONjFM+OpV9m1w3z68FLC2TMCCwto0ar4FyA3J5v777+fXr16mW3MqjBgwACaNm3Kdxuqlj1VE2iFV2ZAr4fYGGjoDbHRAk+PRnz22ZRyi0YFBAQQGxPN2lXLmTJ9Bg29ISfHxMVwHfoMaNO27HTJ0sjIgPNn1eKst4/Ev3npFbSl8eUXn+Hi4sKECRMqZ4RGvSUw0IHw8GDmzo6ld7+u2NpW3aVEXBKkpQlaB5hK/A4tXLCAGhDELDc6nY4NGzbg5m3+uyhzU4fetv8uFy+o2Yinp+TXrafIzDyHTwVaXnbt2pXBgwez+NOPyczMBMC3CQQEmsjOVt2moiLVonB5MZkg4pISoNLroXUbEy1aVtzZR0dF8tMP3/O///0PR8cyOrZr3HA4OzvwzjsTOXf2Am+/UfWisKRE5fC9Gki8vIo+9/dff3Jg7x6sLXV17k6zV69eNGlUvX1uzUHdetf+gyQnQWKioLGv5PJl2PD1z3y3YV6pHaRK4+WXXybuymXWf7W6YFu++qCzM4Rf0HH0sODKZZVSWRJSqt6zFy8olcJLF3V4eELX7kqAqjKEngrBzc2NqVOnVu4EGvWeCRPuYciQ1ixbvJF/9lys9HkyM+HMaYGDg6RFy2srapN55qnHmDVjOrnVldFQRXbt2MYzUybXappoWWgOvwpICeEXBLa2yqFu/+0ccZdP89hjFe8A1b9/f/r06cPnH71f5ANta6uEzNq1N6HTQdgZHQf36wg6IjgbJrgYLjh/TnDimGD/v4ITx3VERYGTE7TvYKJNQOntC8vDkKHDiIqKws/Pr/In0ajXCCFYteo1bGyS+HDRN1wpufTkumRmqlRhIVQI89q6vhdnPktMdDSrVq2qs0V/UZERbPhqDSeOBdW2KaWixfCrwJXLSpu7dYCJuCuw89fDWFtfZuLEijdPFkLw0UcfkZ0rS/xAu7pB566SjHRJcjKkJAuSElV2kE6n9EgaNARnFxMuLlTJyedz8cIF2rVpibW1JoOscX0aNWrEli3vYqQjZ8ME1tblK+aDPOG+YIHJpFppXrtetXnjN2z6dgOvvfYaPXr0ML/xZuL+++9n+vTpfLdhHR07d6ltc0pE08OvJLm5cOSQwMYG2neU/L4jm/+Nf5HbR1qyceOHlT6vlJK4dEO5s3Py9zN3dlpWVhY9OwVy+8gRLFu2zLwn16i3GI2wc3c6p0IvMWR482Jx+GtJSlRhHCGUs3e4Zpko4tJFbrm5Gx07duSvv/6qtl615mLMmDH8sftPgkLPYVWFWVd16eFrIZ1Kkq/e599ckhAPx45EYaG7wpNPVnx2X5i0tDSefvzRcqsiClE9TaO/WbeWy7ExPPTQtd0qNTRKx9IS1q99hpdmTea3n89x8oQgIYFiE5iMDDgdqjpa2diotaprnT1Ak6ZNWbBwIRs3bqzzzh5UTn5CfBx//L6jtk0pEW2GXwkMBjW7d/eANgGSo4cFJ08IbumTS9eutlUqBpFS0qNHD2IvX+bvQ8exrQWhfIPBQN/uHWnk48O///5bJ4pbNP47REVF0atXb9Iz3HnrvY14ejXB2lo1N8nNhawsMBgEFhYSHx/wbVo8Zp+YmEBsdDR9enSrkabk5iInJ4fhI0bw6BNTGTSk8gqe2gy/DnEx/Kp6X0oyXAw34ukBLVvaVdk5CiF46623iIyIYPXypeYwt8KsXbWciEuXmD9/vubsNSpM48aN2b59GzoRwcuz+pJrDMLFVU2UpARnF2jewkS3myTN/Is7+7grl7n39tt4+L67MRlrv1VpRbCysmLnjh0MGXpbbZtSIprDryD5WveNGqkMmuhowaZvf+XVlyZga2seUZFBgwYxZMgQPnr3HVKSk81yzvIipWTL9xu59dZbNc17jUoTEBDAX3/9hb2dDXOeH0fzFjl07irp0EnSuo3Ep1HJiQX/7PmLIf17E37+HKtWraqVO1xzkJ2ZztmwM7VtRjGq5PCFEO5CiB1CiLC83yWuywshcoUQQXk/W6oyZm1z4bzKQGjcRJKZCZGXTBzc/wf+/pZmbYTw9ttvk5SUyEfvvWO2c5YHIQQ7du5i3bp12uxeo0oEBgayb98+Nn+/CSd7G/R6PceDjpaYp56bm8vM6VMZM2o4zk6O7N27l8GDB9eC1ebhrpHDeeapx2vbjGJUdYY/G/hdStkK+D3vcUlkSik75/2YR5+1FsjXum/SVKn3xUQL/v3nFMlJp5g8+S6zjtWlSxdWr17N09OfNet5r0dCQjyZGek4O9jiU5FSYQ2NUvD29qZDhw4421qx9OP3GDqgN4P69mTu7Of56N13+GDRWwDYWVuRlZnOlClTOHz4MJ07d65ly6vGffeN4dCB/Zw7G1bbphShSou2QojTwC1SyhghhA+wW0rZpoT90qWUFa7Lr0uLtiYTBB1R6WOdu0pyc+HffwQLX1vMhXPriI3dUy2tzoy5JuJSMxE6XbXPuJ+cNIFjRw9zKiQE6/IKAWlolJOkpCQ2bNjAunXrOHr0KHq9HisrK+Lj43F2dkZKWW/uKqOjo2nSpAnTn3uBWS9XrG8w1N1F24ZSypi8v2OB0qT5bIUQh4QQ+4QQ150KCyEez9v3UFxcXBXNMx/5Wvd+/hIhlFjaxfAMjgft5KGHBlRbX8vkpETuvG0Q31Rzg4W9f+9h88ZveXjcOM3Za1QLbm5uPPnkk/z999+kp6eTmZmJwWDA2VlJt9YXZw+qEG3w4CF8t2FdnZIUL9PhCyF2CiGCS/gp0nlAqluF0m4XmuVdcR4CPhRCtChtPCnlUilldylld6+yqjZqiMJa927uKtMgOkogsOONhVOYPv2xahvb3d0dWxtrXpkzs8SuRuZAr9fzwjNTadasGbNmzaqWMTQ0CiOEwNa2ainMdZ0JE8YTGRFB0JGKRymqizIrGaSUpa6cCCEuCyF8CoV0rpRyjqi83+eFELuBLsC5yplc8+Rr3TfzV9ez+Di4HCvw9LTgjtuG4eZWfR9anU7HihUr6NixI889PYV13202+5dk4WtzORt2hp07d2JXWR1mDQ2NItx1110cPnqMxs1b17YpBVQ1pLMFyBdJnwD8eO0OQgg3IYRN3t+eQB8gpIrj1hgZGVe17h0c1LaoKEHY6Sh2/74KIVKq3YaWLVvyzjvvsGvHNj7/+AOznttoNHLhbBhTp05l0KBBZj23hsaNjL29PV07d0RXh+5iqurw3wKGCCHCgMF5jxFCdBdC5AuwtAUOCSGOAX8Ab0kp/zMOP/y80rpv2kzN7pOTIDZGcGD/dtatfb3GpFCnTJnCmDFj2PTNerJL00euBNZWVmz77Tfee+89s51TQ0NDkZKSwlP/m8CWzZtq2xSgimqZUsoEoNi0UEp5CPhf3t97gQ5VGae2SEiA5GRB8xamgiKRyAhBXFwOB/Zt4p57+uHmVk5JwCoihGD58uVICQZhjamKF5rMzEzmznqeN15/DS8nXywstIVaDQ1z4+TkxKED+4mLi2fU3ffWtjlapW1pmExqdm9vL/HOS0lPTVG596dDDqLPOMWkSY/WqE1OTk44Ozthp8tl9ozpREdFVuo8BoOBxx8Zx7o1Kwk+fszMVmpoaOSj0+kYP348f+3eVW1JFxWyp7YNqKtERUJWllLDzA/BRUYIEuLh4P7NNGliycCBA2vFtksXw9n07XruGTmswuXbBoOB/41/iB2//cLnn3/OyJEjq8lKDQ0NgIcffhgpJRu/qf0m55rDLwGDAaIiBe4eV5s4pKWqbB1rGxMe7ik8/vikWuurGRgYyI4dO8hIT+P2wbew+/ed5TouOiqSO4cNZMdvv/DFF1/wxBNPVLOlGhoaLVu2pE+fPmzc8HWttz/UHH4JXAwXSAn+za/+cyIuCeLjwLeJ4Kety3nppZdq0ULo2bMn+/fvp1EjHx645w4+XPR2qfvmf8hcnJ3R6QSbN2/WnL2GRg0ybdo0ho8cadaEi8pQ9zsK1DCpKUoN07eJJF+oLz0NwsMF9g4mrG0isbNuUScKRvz9/Tl06BALFy7k9lEjsbbQERwSws5tv9LQ25v0tHSCjx/jxLEg/vjrLxo08uLQwYN1wnYNjRuJ+++/n7vvHUOSXnP4dQYplRqmjY3Et0nR2X1CPGTqD/HkpH7s2LGjzij52dnZ8frrrxc83vHT90UeOzk5MXToUPSpybg62teGiRoaGoCFkOzasY2evfrg4FhhaTGzoIV0CnHlMqSnC5r5XW3KkJEB588JHBzg7z+/xMPDg759+9auoddh/vz5JCQkEBoayqVLl0hOTmbjxo00atSotk3T0LihOXjwIA+NvosfNn1XazZoDj8Po1HF7p2dJV4Nrm6/eEEQFwfOzvHs2vk1EyZMqPNNGdzd3WnTpg1NmjSptYVlDQ2NovTs2ZP27duzduWysneuJjRvkEfERUFOjiiyUKvXQ/AJdREIOrIWo9HI44/XvaYGGhoadR8hBJMnTybo6BGOBx2tFRs0h49y7DEx0NBb4uh0dXtIsCAlWdChk2TL5lXccssttGlTTO5fQ0NDo1yMGzcOOzs71q5cXivja4u2FNfLAUhLgxPHBI0am2jREn7d8TtZaTXbX1ZDQ6N+4erqyv3338+/+/ZiMplqPORa7xy+lBAWBpb24OJa9v6JiZCUJPBvbqJw3499e3UYc6FXH4lOB00beWNpoS18amhoVI33338fS1t79Dk1X4RV70I6ubmQkgyhpwR6/fX3zdfLsbOX+BTy5WfD4MI5Qbv2EqMxltG330bQ0SPVa7iGhsYNgZubGw421hhzcmq88rbeOXxLS2jfAXQ6CA0RXK+wLTpKtS3097+ql3M5Fg7ut8DVTdLtJsmGr9by954/cXJyKv1EGhoaGhXg4MEDdA1sxdHDB2t03Hrn8AFsbSGgrcRggODjgoyM4vtkZysxNHcP1bYQVN/a48d0mHIlPXtJrKxMfLVmhbZYq6GhYVbatm1LZqae5Uu+qNFx66XDB3ByhvYdJLm5avE1IaHo8xfOK70cP39JTo4qrjp/TkdWJrRsrSpt/9z1O5fCw5k8eXLtvAgNDY16ibOzMxMnTmTL5k1cjo2psXHrrcMH5fQ7dpbY2UFoiI5TJwVJiXA5BuLjBN4+kJQIRw4JYmPAwUHi7q6ydSwsYO3KZXh6enL33XfX9kvR0NCoZ0ybNg2j0ciaGkzRrHdZOtdiYwMdOkkiI1Rv2vh4HWfPgKUVCECiw9VV0tRPEnZGLeA29FbHDh46jGFDBmFjY1Orr0FDQ6P+0bJlS0aMGMHq5V/y9IyZNeJn6r3DB7WA27SZxLcJBB8DD0+ll+PqCq5uJlxcVcOTTL2gbTsT+amxTz4xGVsri9o1XkNDo94yf/58LscnYW1dMy1GqxTSEUKMEUKcFEKYhBDdr7PfbUKI00KIs0KI2VUZsypk6iE9AwLbSzp1kTTzl7i4qoYnEZcE7u4qpJOZmclXq5aTm51VW6ZqaGjcAHTt2pVhgweiqyHJ8qrG8IOBe4C/SttBCGEBfAYMBwKBB4UQgVUct8JICefOCiwtoZlf0dzX8At5C7h5Ojo/bPqO56dP5cCBAzVtpoaGxg1GUlIir8yZyeGD+6t9rCo5fCnlKSnl6TJ26wGclVKel1JmAxuAO6sybmWIjYG0NCWOZmV1dXtyklrAbewLdnaqO9SKJV/Qvn17brnllpo2U0ND4wbDxsaGDevWsvTzT6t9rJrI0mkMRBR6HJm3rUSEEI8LIQ4JIQ7FxcWZxQCDAS5dFLi6FpU+Npng/HmBnd3VhicH9u3lxPEgpk2bpnWG0tDQqHYcHR2ZNGkSP/+4mciIS9U6VpkOXwixUwgRXMJPtczSpZRLpZTdpZTdvby8zHLOC+cFJhM0b1k0lJO/UOvXXBYs1C5f8gWurq6MHTvWLGNraGholMX06dMRQrD080+qdZwyHb6UcrCUsn0JPz+Wc4wooEmhx75522qExERIiFc9au3srm5PS1ULtZ5eaqEWwGg0Enf5MpMmTcLBwaGmTNTQ0LjBadq0KQ8++CDrVq8kKTGx2sapibTMg0ArIYQ/ytE/ADxUA+OSm6tE0OztJY19i24POyOwtoYWhWb9lpaW7Nq9GwtMNWHeDUdOTg6RkZFkZWnZT3UFW1tbfH19sSq8sKVRK7zwwgsgdBgM1ff9qJLDF0LcDXwCeAFbhRBBUsphQohGwDIp5QgppVEIMRXYBlgAK6SUJ6tseTmIuCjIyhJ06Hg1tx5UiCczU9C+gwnLvHdAr9djyNTT0N83z0wNcxMZGYmTkxN+fn7a+kgdQEpJQkICkZGR+Pv717Y5Nzzt27dnzepVxKUZqm2MqmbpbJZS+kopbaSUDaWUw/K2R0spRxTa7xcpZWspZQsp5YKqGl0e0lIhOlp1sXJ2ubo9Ph4uxwp8fWURvfz1X62mS2ArLly4UBPm3ZBkZWXh4eGhOfs6ghACDw8P7Y6rjnHmZBDbt22rlnPXSy2d/JCNjY0SR8vHYIBzYQJHR0mTQt2tjEYjSz77mC5dumgznWpGc/Z1C+3/UfeYN/dljh8/Vi3nrpfSCmfPQlaWUsvMD9lICWfPqAKrVm1kkRDP1i0/cCk8nA/ff792DNbQ0NDIY/v27dV2Ia53M/ycHFVM1diXIqGcSxcFycmq8Mre/up2KSVffPwBrVq1YtSoUTVvsEaNIoRg3LhxBY+NRiNeXl7cfvvttWiVhsZVqvOuq97N8K2soFt3SM68GrKJiVbNThp6X1XCzCf0VAjHgo7yxRdfYGGhLdbWdxwcHAgODiYzMxM7Ozt27NhB48al1gFqaNQr6p3DB9XmMD9kczkWzp/T4e4hi6Rg5tM2sB3BJ0Pw92tWw1be2DzzzDMEBQWZ9ZydO3fmww8/LHO/ESNGsHXrVkaPHs369et58MEH2bNnDwAZGRlMmzaN4OBgcnJymDdvHnfeeSfh4eE8/PDDZOS1T/v000/p3bs3u3fvZt68eXh6ehIcHEy3bt346quvtNi4Rp2k3oV0CnPlMpwN0+HmJmkTcLVvbT5GoxFbSwsC2wZgV7gqS6Ne88ADD7BhwwaysrI4fvw4PXv2LHhuwYIFDBw4kAMHDvDHH38wc+ZMMjIyaNCgATt27ODIkSN88803PP300wXHHD16lA8//JCQkBDOnz/PP//8UxsvS0OjTOrlDB8gLg7Ohin9nIDAoou0+Tw67gGa+jbiy6VLa97AG5zyzMSri44dOxIeHs769esZMWJEkee2b9/Oli1bePfddwGVSnrp0iUaNWrE1KlTCQoKwsLCgjNnzhQc06NHD3x9VWVf586dCQ8Pp2/fvjX3gjQ0ykm9dPhxcRB2WuDsTKnO/njQUbb/upU33nij5g3UqHVGjRrF888/z+7du0ko1PBYSsmmTZuKNa2fN28eDRs25NixY5hMJmxtbQueK9ypyMLCAqPRWP0vQEOjEtS7kE5ODpwOBUdHaNtO9aYtiQ8WvYmrqytTp06tWQM16gSPPvoor776Kh06dCiyfdiwYXzyySdIqdZ7jh49CkBKSgo+Pj7odDrWrl1Lbm5ujdusoVFV6p3Dt7KCDh1VV6vSnH1I8Al+/fknpk+fjouLS8k7adRrfH19i8Th85k7dy45OTl07NiRdu3aMXfuXACeeuopVq9eTadOnQgNDdXE9TT+k4j8mUxdpHv37vLQoUMVPi4n10RiRnapzz/+yDj+3LWTCxcu4ObmVhUTNSrAqVOnaNu2bW2boXEN2v+lfiGEOCylLLHlbL2M4ZfF+x98zPkzJzVnr6GhcUNxQzn8/LuZZr7etGjWqJat0dDQ0KhZ6l0M/3rs/n0ntw8eQHRUZG2boqGhoVHj3DAzfJPJxMLXXiEtNRlvb++yD9DQ0NCoZ9wwDn/zxm84cTyINWvWYG1tXdvmaGhoaNQ4N0RIJyM9nddfeYlu3bppzck1NDRuWG4Ih79q+VJiY2L4+OOP0ZVUdqtRN8nNhZ9/htdfV7+rWOyUkJBA586d6dy5M97e3jRu3LjgcXZ26Wm8hZk4cSKnT5++7j6fffYZ69atq5KtJbFz507uuuuu6+5z5MgRfvvtN7OPrVE/uCFCOk88NZXundrTu3fv2jZFo7zk5sKwYbB/P2RkgIMD9OwJ27ZRakVdGXh4eBQodM6bNw9HR0eef/75IvtIKZFSljoxWLlyZZnjTJkypVL2mYMjR44QHBzMbbfdVms2aNRd6v10Nzs7G08XR625yX+NX39Vzj49XbUrS09Xj3/91exDnT17lsHcxjUAAAlOSURBVMDAQMaOHUu7du2IiYnh8ccfp3v37rRr14758+cX7Nu3b1+CgoIwGo24uroye/ZsOnXqRK9evbhy5QoAL7/8coE4XN++fZk9ezY9evSgTZs27N27F1AyzPfeey+BgYGMHj2a7t27lygXvXXrVtq0aUPXrl358ccfC7bv27ePXr160aVLF/r06UNYWBiZmZnMnz+fdevW0blzZzZu3Fjifho3LlVy+EKIMUKIk0IIkxCixMquvP3ChRAnhBBBQoiKl85Wkj927qBX53acCQ2pqSE1zMXRo2pmX5iMDDCzhn4+oaGhPPvss4SEhNC4cWPeeustDh06xLFjx9ixYwchIcU/QykpKQwYMIBjx47Rq1cvVqxYUeK5pZQcOHCARYsWFVw8PvnkE7y9vQkJCWHu3LkFmj2F0ev1TJ48mV9++YXDhw8THR1d8Fzbtm3Zs2cPR48eZe7cubz88svY2dnxyiuvMHbsWIKCghg9enSJ+2ncuFQ1pBMM3AMsKce+t0op46s4XrnJSE/nhWen4uTkSKtWrWpqWA1z0aWLCuOkp1/d5uAAnTtXy3AtWrSge/erc5b169ezfPlyjEYj0dHRhISEEBgYWOQYOzs7hg8fDkC3bt0Kmqhcyz333FOwT3h4OAB///03s2bNAqBTp060a9eu2HEhISG0bt2aFi1aADB27FjWrFkDQHJyMuPHj+fcuXPXfV3l3U/jxqBKM3wp5Skp5fVXsGqJdxbOJ+LSJb788ssi8rUa/xGGD1cxe0dHEEL97tlTba8GCouhhYWF8dFHH7Fr1y6OHz/ObbfdRlZWVrFjCqf3Xk8WOf/zZ07p5Jdeeolhw4YRHBzMDz/8UKJ9FdlP48agpmL4EtguhDgshHj8ejsKIR4XQhwSQhyKi4ur1GAHDxzgyy8+44knntAaUfxXsbBQC7Tr18P8+ep3FRZsK0JqaipOTk44OzsTExPDtm3bzD5Gnz59+PbbbwE4ceJEiSGjwMBAwsLCuHDhAlJK1q9fX/BcSkpKQS/eVatWFWx3cnIiLS2tzP00bkzKdPhCiJ1CiOASfu6swDh9pZRdgeHAFCFE/9J2lFIulVJ2l1J29/LyqsAQV1m37iu8vb156623KnW8Rh3BwgJuvx1efln9rqEm8127diUwMJCAgADGjx9Pnz59zD7GtGnTiIqKIjAwkNdee43AwMBiUt329vYsXryY4cOH0717d3x8fAqemzVrFjNnzqRr164UVrwdOHAgx44do0uXLmzcuLHU/TRuTMwijyyE2A08L6Usc0FWCDEPSJdSvlvWvpWVR5ZSEhkZSZMmTSp8rEb1ocnwXsVoNKqeyra2hIWFMXToUMLCwrC0rPlMae3/Ur+oVXlkIYQDoJNSpuX9PRSYX8ZhVR1Tc/YadZr09HQGDRqE0WhESsmSJUtqxdlr3FhU6RMmhLgb+ATwArYKIYKklMOEEI2AZVLKEUBDYLMQIn+8r6WUWimgxg2Nq6srhw8frm0zNG4wquTwpZSbgc0lbI8GRuT9fR7oVJVxNOoPUkryLv4adQAtrn9jUe8rbTXqDra2tiQkJGhOpo4gpSQhIQFbW9vaNkWjhtCChho1hq+vL5GRkVQ23VbD/Nja2uLr61vbZmjUEJrD16gxrKys8Pf3r20zNDRuWLSQjoaGhsYNgubwNTQ0NG4QNIevoaGhcYNglkrb6kIIEQdcrOThnkCNqXNWAM2uiqHZVXHqqm2aXRWjsnY1k1KWqEtTpx1+VRBCHCqtvLg20eyqGJpdFaeu2qbZVTGqwy4tpKOhoaFxg6A5fA0NDY0bhPrs8JfWtgGloNlVMTS7Kk5dtU2zq2KY3a56G8PX0NDQ0ChKfZ7ha2hoaGgUQnP4GhoaGjcI9drhCyHGCCFOCiFMQohaTbsSQtwmhDgthDgrhJhdm7YURgixQghxRQgRXNu2FEYI0UQI8YcQIiTvfzi9tm0CEELYCiEOCCGO5dn1Wm3bVBghhIUQ4qgQ4ufatiUfIUS4EOKEECJICFHxFnbViBDCVQixUQgRKoQ4JYToVQdsapP3XuX/pAohnjHLuetzDF8I0RYwAUsoZwvGarLDAjgDDAEigYPAg1LK4p2ra5i8/sLpwBopZfvaticfIYQP4COlPCKEcAIOA3fV9nsmlJi/g5QyXQhhBfwNTJdS7qtNu/IRQswAugPOUsrba9seUA4f6C6lrHPFTUKI1cAeKeUyIYQ1YC+lTK5tu/LJ8x1RQE8pZWWLUAuo1zN8KeUpKeXp2rYD6AGclVKel1JmAxuAijSBrzaklH8BibVtx7VIKWOklEfy/k4DTgGNa9cqkIr0vIdWeT91YtYkhPAFRgLLatuW/wJCCBegP7AcQEqZXZecfR6DgHPmcPZQzx1+HaIxEFHocSR1wHn9VxBC+AFdgP21a4kiL2wSBFwBdkgp64RdwIfAC6i72rqEBLYLIQ4LIR6vbWMK4Q/EASvzwmDL8vpu1yUeANab62T/eYcvhNgphAgu4adOzKA1qoYQwhHYBDwjpUytbXsApJS5UsrOgC/QQwhR66EwIcTtwBUpZV1slNtXStkVGA5MyQsj1gUsga7AF1LKLkAGUJfW16yBUcB35jrnf74BipRycG3bUA6igCaFHvvmbdO4Dnkx8k3AOinl97Vtz7XI/7drxypxBWEUx/+nU4RgEyRgkRSyrW2wUxLEIpAugVjZJg/gC6TMG9gpggiCsJBFUJBAIIFAAkF7NxBT5QGEY3EnuqRemLt7z6/Z5RbLgb1838x8Y/+VdAasA7WH3ivAC0kbwAzwQNKu7TeVc2H7V/n8I+mI5ojzvG4qoNlpD0d2aIe0qODTNMhvtq/H9YMTv8KfEF+BJUlPStd+BRxXztRqZTi6A1zY/lA7zz+SHkqaL99naQbxl3VTge1t24u2H9O8X6dtKPaS5srQnXJc8pz6zREA27+BK0m98mgNqH6RYsRrxnicA1Ne8CW9lDQEngJ9SYMaOWzfAG+BAc3w8cD2zxpZ/idpH/gM9CQNJW3VzlSsAJvA6sj1tI3aoYBHwJmkHzSN/MR2a65AttAC8EnSd+AL0Lf9sXKmUe+AvfJ/LgPvK+cB7prjM2CsO9upvpYZERH3pnqFHxER91LwIyI6IgU/IqIjUvAjIjoiBT8ioiNS8CMiOiIFPyKiI24BfjdSgmSoiiIAAAAASUVORK5CYII=\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# Noise free training data\n",
"X_train = np.linspace(0,6,5).reshape(-1,1)\n",
"Y_train = np.sin(0.07*X_train**3)\n",
"\n",
"# Compute mean and covariance of the posterior predictive distribution\n",
"mu_s, cov_s = posterior_predictive(X, X_train, Y_train)\n",
"\n",
"plot_gp(X, mu_s, cov_s, num_samples = 5)\n",
"plt.plot(X_train, Y_train, 'ro', label='Training data', markersize=5)\n",
"plt.legend()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Prediction from noisy training data\n",
"\n",
"Now let's add some noise to the model, then training points are only approximated and the variance at the training points is non-zero."
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[<matplotlib.lines.Line2D at 0x7f047801ba58>]"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# Noisy training data\n",
"noise = 0.3\n",
"X_train = np.linspace(0,6,100).reshape(-1,1)\n",
"Y_train = np.sin(0.07*X_train**3) + noise * np.random.normal(size=X_train.shape)\n",
"\n",
"# Compute mean and covariance of the posterior predictive distribution\n",
"mu_s, cov_s = posterior_predictive(X, X_train, Y_train, sigma_y=noise)\n",
"\n",
"plot_gp(X, mu_s, cov_s, num_samples = 10)\n",
"plt.plot(X_train, Y_train, 'rx')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Effect of kernel parameters and noise parameter\n",
"\n",
"The following example shows the effect of kernel parameters $l$ and $\\sigma_f$ as well as the noise parameter $\\sigma_y$. Higher $l$ values lead to smoother functions and therefore to coarser approximations of the training data. Lower $l$ values make functions more wiggly with wide confidence intervals between training data points. $\\sigma_f$ controls the vertical variation of functions drawn from the GP. This can be seen by the wide confidence intervals outside the training data region in the right figure of the second row. $\\sigma_y$ represents the amount of noise in the training data. Higher $\\sigma_y$ values make more coarse approximations which avoids overfitting to noisy data."
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 864x360 with 6 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"params = [\n",
" (0.3, 1.0, 0.2),\n",
" (3.0, 1.0, 0.2),\n",
" (1.0, 0.3, 0.2),\n",
" (1.0, 3.0, 0.2),\n",
" (1.0, 1.0, 0.05),\n",
" (1.0, 1.0, 1.5),\n",
"]\n",
"\n",
"plt.figure(figsize=(12, 5))\n",
"\n",
"for i, (l, sigma_f, sigma_y) in enumerate(params):\n",
" mu_s, cov_s = posterior_predictive(X, X_train, Y_train, l=l, \n",
" sigma_f=sigma_f, \n",
" sigma_y=sigma_y)\n",
" plt.subplot(3, 2, i + 1)\n",
" plt.subplots_adjust(top=2)\n",
" plt.title(f'l = {l}, sigma_f = {sigma_f}, sigma_y = {sigma_y}')\n",
" plot_gp(X, mu_s, cov_s)\n",
" plt.plot(X_train, Y_train, 'rx')\n",
" plt.legend()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can define the learning criteria and estimate optimal values for these parameters.\n",
"Optimal values can be estimated by maximizing the marginal log-likelihood which is given by\n",
"\n",
"$$\n",
"\\log p(\\mathbf{y} | \\mathbf{X}) = \n",
"\\log \\mathcal{N}(\\mathbf{y} | \\boldsymbol{0},\\mathbf{K}_y) =\n",
"-\\frac{1}{2} \\mathbf{y}^T \\mathbf{K}_y^{-1} \\mathbf{y} \n",
"-\\frac{1}{2} \\log \\begin{vmatrix}\\mathbf{K}_y\\end{vmatrix} \n",
"-\\frac{N}{2} \\log(2\\pi)\n",
"$$\n",
"\n",
"Instead of the maximization problem we turn it into the minimization problem. We will minimize the negative marginal log-likelihood w.r.t. parameters $l$ and $\\sigma_f$, $\\sigma_y$ is set to the known noise level of the data. If the noise level is unknown, $\\sigma_y$ can be estimated as well along with the other parameters.\n",
"The minimization problem is solved using gradient-based algorithm."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We avoid computing the determinant of $\\begin{vmatrix}\\mathbf{K}_y\\end{vmatrix}$ by using the fact that the sum of the logarithms of the diagonal entries of the Cholesky decomposition of a matrix `S` equals half of the logarithm of determinant of matrix `S`. i.e. `0.5 * log(det(S))` and `sum(log(diag(chol(S))))` are equal up to numerical precision."
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[<matplotlib.lines.Line2D at 0x7f04716afbe0>]"
]
},
"execution_count": 19,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAD6CAYAAACiefy7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOyddVgUXRuH71lYQlGxu7sVwUJBLOwWREWxu1873s/u7i4sbGwMBFvBwO7uAkFq2T3fHyAvKCixsKhzX9dcsBPn/GZm95kz5zzPcyQhBDIyMjIyfz4KXQuQkZGRkUkeZIMvIyMj85cgG3wZGRmZvwTZ4MvIyMj8JcgGX0ZGRuYvQTb4MjIyMn8JiTb4kiTlliTJXZKkW5Ik3ZQkaUAM+0iSJC2QJOmBJEk+kiSZJbZeGRkZGZn4oa+FMsKAIUKIy5IkpQG8JUk6KoS4FWWf+kDhiKUSsDTi70/JlCmTyJcvnxYkysjIyPwdeHt7fxBCZI5pW6INvhDiNfA64n9/SZJuAzmBqAa/KbBBhEd5nZckyVSSpOwRx8ZKvnz58PLySqxEGRkZmb8GSZKexrZNq334kiTlA8oDF77blBN4HuXzi4h1MjIyMjLJhNYMviRJJsBOYKAQ4ksiyukuSZKXJEle79+/15Y8GRkZmb8erRh8SZKUhBv7TUKIXTHs8hLIHeVzroh1PyCEWCGEMBdCmGfOHGM3lIyMjIxMAtCGl44ErAZuCyHmxLKbK9AhwlunMuD3q/57GRkZGRntog0vHUvAEbguSdLViHWjgDwAQohlwEGgAfAACAQ6aaFeGRkZGZl4oA0vndOA9It9BNAnsXXJyMjIyCQcOdJWRkZG5i9BG106MjIyMn89QUFBGBsbA7Br1y5u3LiBkZERWbNmJW/evOTJk4f8+fMTPuypG2SDLyMjI5NAQkND2blzJ8uWLePSpUs8f/MefX0D9h04xLo1q6LtmzZtWt59+ISBvoIlS5bw9u1bihcvTrFixShatCipUqVKcr2ywZeRkZFJAKdPn6Z79+7cvn2bvPny06v/ID5/+UqatHpMmT2fSTPnEhIczLu3b3nx/BkBAf74BqkAOOR2jEP7XdFoNABIkkTt2rVxc3MDQK1Wo6enp3XNssGXkZGRiSdXr17FysqKXLlzs26zC3XrN0Sh+G9IVKFQoFAo0DcxIb+JCfkLFox2/OqNWwgODubxo4fcv3uHB/fvYWJiwnv/EPQUEsZ6AmPZ4MvIyMjonoLFSrJ+y3Ysq1uT2sQkQWUYGRlRvERJipcoGblOIwQatSCNoVJbUqMhe+nIyMjIxJExY8Zw6NgJAkPV1K3fMMHG/lck1cDuH2nwfX19+eeff3jy5ImupcjIyPwhLF++nMmTJ3Ng/wFdS0kwf6TBP+nhyfz58ylYsCBdunTh5csY0/bIyMjIxImLFy/St29fbGrXZdS/E3QtJ8H8kQa/YaNGXPS5TbdefXB2dqZw4cJMnz4dtVqta2kyMjK/GSqVii5dupAla1aWrV6fJN4zycUfafABcuTMxfgpMzjt5YNNrTocPX5CpwEPMjIyvyebN2/mxo0bTJ01j3SmprqWkyj+eC+dPHnzstp5K0FBQXwKVOH/8S13bt2kfv36upYmIyPzG2Dv0A7jdJmwrllL11ISzR/bwo+KJEmkSpUKtUYwZuy/NGzYkKVLl+palszvyowZ4O4efZ27e/h6mT+KkJAQAlWaP8LYw19i8KMyacZs6tRrQO/evZk6daqu5cj8jlhYgJ3df0bf3T38s4WFbnXJaJWHDx+SK1cu3I4c0bUUrfHXGfxUqVKxeuMWWtq1YdSoUcyZE9ucLTIysWBjAy4u4UZ+3Ljwvy4u4etl/hgmTpxIQEAAJUqV0rUUrfHXGXwApVLJgmWraNS0OatWrSY4OFjXkmR+N2xsoFcvmDgx/K9s7P8oHj9+zMaNG+nYpRtZs2XXtRyt8ccP2saGnp4ei1euJejrVyT9pAljlvmDcXeHpUth7NjwvzY2stH/g1i+fDmSJNGjT39dS9Eqf2UL/xuGhoaYZsjAm09fGDBgAK9fy9PsysSBb332Li4wYcJ/3TvfD+TK/JaEhoayevVqbOs3JEfOXLqWo1W0YvAlSVojSdI7SZJuxLK9hiRJfpIkXY1YxmmjXm3x7OlTVq1ahZ2dnRycJfNrLl2K3mf/rU//0iXd6pLRCgYGBrjsdmXoqLG6lqJ1pPDpZhNZiCRZAQHABiHEDyMckiTVAP4RQjSKT7nm5ubCy8sr3npUag2fvobG65gd27bQt3tnpk2bxvDhw+Ndp4yMzJ+BEIL3ASFowTQmmIypDdDXS1h7XJIkbyGEeUzbtNLCF0J4Ap+0UZauaGnXhsbNWjB27Fh8fHx0LUdGRkYH3L17l67duvP82TNdS0kSkrMPv4okSdckSTokSVLJ2HaSJKm7JElekiR5vX//PtnESZLEtDnzSZ8hIz179ky2emVkZFIOmzdvZu2a1SiVf6YjR3J56VwG8gohAiRJagDsAQrHtKMQYgWwAsK7dJJJHwAZM2Zi5Xpn8ufLk5zVysjIpACEEGzevBnL6tZ/lCtmVJKlhS+E+CKECIj4/yCglCQpU3LUHV8qVbEkS/bchIapZf98GZm/CG9vbx48eECL1va6lpJkJIvBlyQpmxSRqlKSpIoR9X5MjroTghCC5s1b0LVrV11LkZGRSSa2bNmCUqmkQeOmupaSZGjLLXMLcA4oKknSC0mSukiS1FOSpG+d4a2AG5IkXQMWAG2ENtyDkghJkihUpCibNm3i8uXLupYjIyOTDJikSUPrNu0wTZ9e11KSDK24ZSYVyemW+T1f/PyoXK4kZmblOXr0aKLKkpGRSfkEhobhHxymaxlACnfL/BNJmy4dA4cO59ixYxz5g7LlyURBTnMsE8Hbt28JCkkZxj4pkQ3+T+jYpTu58+Zl9uzZupYikxTIaY5lIqhXrx6Obe10LSPJ+WuTp8UFQ0ND1jpvo0TRGD1IZX53oqY57tUrPAmanOb4r+PZs2dcvXqVxi1a61pKkiO38H9BqTJl0TNKhVqtJiWPd8gkEDnN8V/P/v37AbBtEK/ML78lssGPA48ePKRkqdIcO3ZM11JktM33aY7ljJd/Ha6urhQoWIhChYvoWkqSIxv8OJAjVy6+fPnC9OnTdS1FRpvIaY7/eoKCgjh58iS1bevpWkqyIBv8OGBoaEiXHr04fvw4V69e1bUcGW0hpzn+69HT02Pjlm04ODrpWkqyIPvhxxHfz58xK1mYVi1bsn79eq2WLSMjozs+fQ1FpdboWkY0ZD98HWOaPj0O7TuwZcsWXr58qWs5MjIyWmDx4iX4XL+uaxnJhmzw40GPPgNYvHwVWbJk0bUUGRmZRPL27Vv69u3D0UMHdS0l2ZD98ONBnrx5yZM3L5JCT9dSZGRkEsnx48cBsLKpqWMlyYfcwo8narWayVOnsXnzZl1LkZGRSQRubm6kT5+B0mXL6VpKsiG38OOJQqFg144dCE0YDg4ORGR9lpGR+c1wd3enanUr9PT+njd2uYUfTyRJomOXbly/fp2zZ8/qWo5MMvF51CjOT53K4sWLcXZ2Dl8pJ1r7bXn79i1v376lSrXqupaSrMgt/ATQrJUd/xszgmXLlmFpaalrOTJJyJw5c1i7di2ZbtzABRgJBFepSsvM2TBq78CO1q2xePKEfPny6VipTHzImjUrz958ICAoRNdSkhW5hZ8AUqdOTes2bXFxceHDhw+6lvPnoeO0xVeuXEEIQUiYmkdPnpHGND01p87g/vRZHE2fgQMVK2PYzoEXC5fRaeNGSpQowerVq5NFm4z2kPSUpE6dWtcykhXZ4CeQDp27YVXDhs+fP+tayp+HDtIWq9Vqdu3aRbVq1TAzM2P3gcP4BqoYM3Equw+40b13Pwr27ENwtx5kmD+HwC7dMGjQkFMXr2BRqQpdu3ZlzJgxcoK934RWrVqzZdNGXctIdrQ1xeEaSZLeSZJ0I5btkiRJCyRJeiBJko8kSWbaqFeXFCteAufteyhQsJCupfx5RE1bPG7cf/lukiCTZVBQELNnz6ZIkSK0bNmS5y9fMmn6LMqWDw9UjDoor/T0INXqlQQMG0mq1StRenqQI2cutuxypX3HzkyePJnJkydrXaOMdnn9+jU7d+7g08e/7+1cW33464BFwIZYttcHCkcslYClEX9/e+4+eISeCKNo0aK6lvJnETVt8dixWjX2T58+5dmzZ1SvXh2NpGDGjJnkzZ+fkf9OpEHjpjF6bSg9PTB1ao/vOmdUVtaEVreO/IyVNTPmLUShUJA9Ry6t6ZRJGjw9PQGoXLWajpUkP1ox+EIIT0mS8v1kl6bAhoiJy89LkmQqSVJ2IcRrbdSvK9RqNTWqW1LNsiq7du3StZw/i+/TFtvYxN3oz5gR3v0Tsf/Xr1+5vmAB7w4cYPjHj9y5c4fcefJy6fptQMLjwmXSZ8gQY1Gp5s1GZWaO8rJXpLFXenpEflZe9kJlZY1CoWDGvIVIhOdy0ldIsstuCsXT05PUJiZ/lf/9N5KrDz8n8DzK5xcR635r9PT0sHNoh6urq5xfR5skNm2xhQWa1q0RJ04ghGBx69YUHDWKxRcvki1HLiZMncHmnXv51t0em7EHUJmZY+rUHpWZeaSxj/o5cOCQaPsLYMXqtTRq1Ejuz0+heHp6MjtbdozPnom2XunpQap5f/Z0pilu0FaSpO6SJHlJkuT1/v17Xcv5JY5OXVCr1axZs0bXUv4c4pu2OIpXz/379+m+ZQvDP39G07AhgSNGM/DsOW5OnMKKF+/Yunsf3Xv3o3CRonFqgausrPFd54ypU3tST54QrVsnNoKCgjl48CBbt26N96nLJC1hYWEUKFiI1DVrY+rUHqWnB0C0B/mfjNbSI0d06ewXQpSKYdty4KQQYkvE57tAjV916aSk9Mg/w65pQ548esCjR4/+qqi9FIO7O5rWrVlaowYD9uyhlkLBdj09QuvUJdM+VwKGjeTr6HGJqiL15AmYzJgap7LUajUNalrx/t0b7t69i4mJSaLqltEu/sEqAkPVkUY+sEs3Uq1e+csHeXLyu6dHdgU6RHjrVAb8fvf++6i0d+rMq1evuHbtmq6l/JWEWlrSwdAQu5072VGqDAdM0sDY8WQ4ezaaR01Cick752fo6ekxacYsXr16xZQpUxJc72+PjuMpYiIkJASVOryRq7KyJrBLN0xmTCWwS7cUY+yTEm25ZW4BzgFFJUl6IUlSF0mSekqS1DNil4PAI+ABsBLorY16Uwr1Gjbm6p0HmJn99t6mvyUaSQ+rydPwc+xIs2tXUNWth8mcmfiuc+br6HGRXTIJMfpRvXPiU5ZFpSo0b2XH/Pnz8fX1Teip/d7oIJ7iV9SoUYOuTo5A/B/kfwJaMfhCCAchRHYhhFIIkUsIsVoIsUwIsSxiuxBC9BFCFBRClBZCxL+fJgVjYGBApsxZCQlTywN1yURoaCgdO3ZkzfqN+AWpaJUpCwUOHiBg2EgM9+0hYPDQyBabysqa4MZNMdrpEq2MuAzSffPGCapqyeOHDzkWFobnwMH4H3WLvNep5s3+wVgoPT2YmyMXO/bsx9TUVItn/huRjPEUcSE4OBhvb29y5MyV4Af5706KG7T9XQkJCaFOrdpy4E0yEBgYSMOGDdmwYQOPHj/58ce7bRcmc2ZG+/EGt2yN0T7XeA3ShYSEsClXLhyXL6Fw7qxUMSuFffNG2IwdRb4FczAvXZS5M6fhX7xkjAOAaerUxaxSlRQ3fZ420Wg0qNXq2HeIGk/Rq5fOjD2Ep8xQqVRUsKgYzc3Wz9eXEQf30SVNGlwG9OHIwf0605jUyHPaapHWTRrw7MkjHj58KA/eJhGBgYE0adKEEydOMGfRMhzad4j0lY/aB/vNVz6q22RcB+nevX3D+jWr2LBmFe/fvSVb9uzYNmhMObMK5MtfAID7d+9waL8r7sePki17drb1HUi1OTNRlSqN8oo3vptcUFlZo1ar2eTYhkoIrPb/WYYkKCgIOzs77Bza0ahpC9yPHubEMTcWLlz4nwfUt26cXr3C4yl02MKfM2cOQ4YMwefeY7JkzQbAhw/vaVK3Jk+fPMbKphaPHz7gyeNH9O4/iLETJussliKpBm3lbJlapL1TZ3p0cuTo0aPUq1dP13L+OFQqFU2bNuXEiRPMX7oSO4d2AD/4wkN4N873xjzqIF3AsJE/bL914zrLlyxk9/ZthIaGUtu2Pt169qF6DRsUiug/viqW1ejQuSvely4woFd3ao4bxem69ah86ADC2DhyP6Mzp+nmdoTu6dJRNSwMff0/5yfXs2dPDhw4QP2mLQhVa7jg5c3ixYvJkycPw4YNix5P8S1wTofdOufOnSN33ryRxl4IQQ8nR169fMGOfYepYlkNlUrFmOH/sGTBXDJkzEjfGL5bvzNyl44Wqd+oCZkyZWbFihW6lvJHoqenR3nzSpHGPra+89j65WMapFOpVLgdOoB9s0bUtKyI6+6dtOvYiTPePji77MK6Zq0fjH1UKlhU4rD7afqVKUvBQwe41LAxQqnEtJ19pN/+hWEj2f7hA/v/oBa+q6srGzZsYNCwkbS0cwBg0NARNGzSjLFjx/Lo0aP4x1MkMS1btqL/oKGRn7dv3cyZUx5MnDaLKpbhaRaUSiXTZs+jcbMWTJ3wL14Xz+tEa1Ihd+lomQljR7F88QKeP39O9uzZdaLhTyM4OJjnz5+TOWdeQsL+6w//Pr/N95+jEq2fv3IVHq9YRvmJ/9JeaYCr/xeyZstGl+69cezU5aeRtzGh9PQgnVM7BubIyfI7tzkzfTbmo4cjBQURMGwkfsNHUbFMMUqVLMmRI0e0ck10iVqtpnTp0qg1guNnLqJUKiO3vX71EssKZWjSpAlbtmzRocofCVap8QtSAeFvi5XLlyRz5iwcPO75w0Pd/8sXalQ1x8QkDUc9z2FgYJCsWn93P/y/hg6dujJ5+sy/Ls92UhEaGoqdnR1Vq1ry7sOnaNt+iIK1bxHNOwf+a/H7Hz3C1tb2tF26iOL5c2I9ejit1Gpa5M3LRpddeN24R/8hQ+Nt7CHck8dv3Sb6uB4iS9aszJk+BY2+khDrGqRavRLjs2do17Ezbm5uPHjwINHXRNccP36c27dvM3TU2GjGHiB7jpx06tYTFxcXnj17piOFP/LkyRNu3LoV6VnlunsHL58/Z/DwUTG+waVJm5apM+dy9/Yt1q/+c97Y5RZ+EiABmUwMUSjk5FmJISwsjDZt2rBz506mz1lAxy7dYtzvWxRskH1bDI+54bvOmWDLatxfvJDyk8bTNV1atr97B0DuPHmpWceWmrXrUM2qBqm1HAX7aPkSCg0bwoLq1vTdfzjyzeLR7HlMPe3Jv2NHkydPHq3WmdwIITh16QqFixSL0Vi+fvWSe7dv0qppo592hyUnQ4cOZcGCBTx4+R4DAwPq16xOQEAAHue9f6qxZaN63L1zmwtXb2r9u/Iz5Bb+b0RwSAjLV67i/Pk/q/8vUcQz6lKtVuPo6MjOnTuZOG1mrMY+ar+84TE3Pvbui6F9C7blyUaJsSNpI0FAxUpMmz2fc5dvcLNLN+Y0a4Ftg0aRP+A0A/qSZkCfH8tNQCKtUkFB7GrnyKRTHpw66R75FpL16ROmzJ5Pzly5411mSiNYpaFosRKxGsrsOXJiXasughTS4Jkxg7CjRylZugwGBgbcv3eXdN5eLM6X/5cPpJHj/seH9+9YtWxJMolNWmSDn0SMHT2KWbNm6VpGyiGeUZfz589n69atjBk/iW69+sa4T9R++S8jRrPNrg3SpPHsDAxkUEAAz5s0Y8mTV6zd5IJT1+7kL1gwMvtlVJ95o13bMdq1UyuJtAIHDqHenAXkzZef0cOHoFKpIrNqajQajrt7cPv27XiXm1IYMWIEnTo5/TLA0M/Xl3+GDo3MPa9LNBUqMNrHB/sI75wr06fiAhRs5/jLY80rVqZu/YYsWTAX3z9gdjvZ4CcBhoaG2LVtz549e+S0yd+IZ9RlW6euLFm19qducd+CZ+7mzEmDWla0XbqILfkL0NbIiIBhIyl35jRpL12MdkyM2S83ueC7aVu0dcGNm/5YXxxb/UZGRkyYOoN7d26zduWyyPVBQUE0b9qIefPm/bKMlEhYWBhr164lICDgl/7phkZGrF2zOkXM9Xs/Vy5aC0E/z5OknjSetru2M6W8GambNo/T8cNGjcXPz5eNa1clsdKkRzb4SYRT1x4IIViy5M94FdQKv4i6fPXqFR07duTpq3dI+ga0aN3mp8UFDhzC9vfvqGNVhWdPnuA6dAR9/Pz4sn3PT8PlY0qa9f264JatE5U+t279hljb1GLerBl8/foVgNSpU1O/YWO2b99OaKjuxpgSiru7O+/evaOl3c/vC4Q/9OrUa8D+/fsJCwtLBnWx4+3tzUngXSt7TGZOY7FGQ3ZHpzgfX6pMWaxq1GT1iqW/5X2Limzwk4i8+fJh26Ahy5cvJygoSNdyUgbfz2IVpU/fw8MDMzMzduzYwVUfn18WJYRg9vQp9OzcgZKlSnPs9HlqmJhEc8n81ppXXo4+8B+TP/7364BY8+A/f/YUlUr1U32SJPHPyNF8+vgB53X/tXJb2LXh8+fPHDp0KM6XLaWwf/9+jI2NsaldN3Ldz2IhbOs35NOnT5w9eza5pUbD1taWI2P+Jfe+PZyobk0voEU8vbF69hvAm9evcd29M2lEJhOywU9CuvToQ568eeVuHYh1FqvQI0cYNWoUtWrVIk3adBw8ceqXc41qNBqGDujLzCkTae3Qjh37DpMzV24CBw6JObo2hvQK0ZJmtbPDtJ39D4m0EILHDRphMmMqFyqYI9nYYKivwKJ0MQrmyETrJg3Yu2sHGk3MuXIsKlXBsro1SxbMJTg4GACrGjVJnz7DbzclphCCffv2YWlVA+MokcQqM3PSODpwefoUnj97Gu1NyKZWHQwMDHB1ddWhcshw7Rq1ly7Gb90mRmvUjCxYiAJDBsYrUZpNrToUKVac5YsW/NYJEmWDn4RYVrfi0InTFCxY8MeNyZkrPCXkJY8l6vLg+PGETZ3K/2xqcejEKYoVLwHE3l+u0WgY3K8XzuvXMGDIMBYsXRmvoJioSbMgIpNmi9YEt2gZue5r5Sq4OnXiRCdH0m5cxyRJopTnSdJfPEM6YyUrV66kX79+NLx+jS2dHGlY25pHDx/EqHvgP8N5++YNW503hG9XKqllW48zZ878PoYjwsulZWs72rT7L7Ww8dxZjDp8gLq+nzGfMpEjZUtg2KZl5PU1SZOGuvUbJo9rZizfcc20aXjOnsOtydP4YlGRy5cuYtCgUYxvfj9DkiR69O7HdZ+rnD19SsvikxEhRIpdKlSoIBJCaJhavPELSjHL89fvxO3bt6OLPHFCiEyZwv/G9FmbJGddv0Cj0QhPT09x79498TVEJc56+4jj/5sk1BkziY/7Dos3fkHi477D0T5/W159/ioc2ncUgBg8fFSS3a9mLVqLGiA+6umJ3f37i0+fPsV4zdTHjomgNGmEfZYs4pzXtRh1v/YNFOXMKojCRYqK176B4o1fkLj56Ln45B+Y7Nc+wUScu//BI9Huz8LW9gIQHbt0F/fadxACxFEQrkNHRLueIW5HhZg+PVk0fv8df7J2rQDEgmWrxPa9BwUgnLfvTtD34snbzyJjpsyiTr0GSW4zVGHqBF8KwEvEYlPlFn4SI4TAtk4tnJy+c2VLzlzhKSAvuUqlYvPmzVSqVAkrKyv+N3Ey/sFhFChUmJKDhvxy3ljjubNYY9eCLc7rGTx8FENHjtHapNP37t5h5D8Deff6JWmNlAwfOphlnTtjeuQIzebPJ3369DHmgVHUqoXR3r1s0WiotGc7pk7teThrbjTdqefPYaxVDe7fu8uZU+Euitlu3sRg7txE6042bGy4PX48xo5tI+/PxRGj6L/DBYf2HZnTtDkFDx3Eb8gwKikUWM2cRtD+8G4cpacH+m0dUCf15ECxfMc9I7LWlilXnjOnPNDT06NyFcsEVWFkZESnbj04evggDx/c16b65CO2J0FKWP6UFv70OQsEII4dO/aj2LFjhYDwv0lNctYVhQULFogcOXIIQBQsVFhMnTVPPHz14Yfr5D9spBAg/IeN/KFlP7luffEOxGL7tuK1b2CsbwFxXS7fui8mTpspypYzE4AwMDAQLi4uCTvBiOt63tZWpE2XTly4eiuyno/7DouwDBlFo9QmonGzFpG6dw76R9SqVVtoNBrtXuwk4PPnz0JPT08crlw1/P4MHSEsKlUW2XPkEC9ddke7D1cXLxOfQXw1MBD+w0YKdcZMolOBgqJfv37JI/a77/jgwYOFkZGRePHRX1hUqizMzC0S9Vu+fv+JMDAwEJ279/x7W/iSJNWTJOmuJEkPJEkaEcN2J0mS3kuSdDVi6aqNen8X7Ns5kjVbdsaPHx+9lf8TrxWtk5x1AU+fPkUIQWiYhvsPn1CoSDGct+/m1KWrdOrW44dcQ7FNNyeEYPSwwYx2O8S21vb0POaGyZSJsSZJi42goCCePX0KQODXr1QqW4KxI4aipwjPk/7s2TNat24d/xONcl3NL13CKiyMHp0dI933VFbW+K3fxGZ1GBX37iZth7b4rnPmaYGCHD9+7LeYB9nDw4PqajU1bt0kYNhI9JcvIfWF8wwaOpI0t29G9ocrPT3I1r4ji+vYcl6lwmTGVFSlSvEobz7ck/j7BsT4Hffx8aFY8RKEBAdzxdsLy+qJm7c2c5asNGtlx9ZNG/H7HaeujO1JENcF0AMeAgUAA+AaUOK7fZyARfEt+09p4b/xCxIz5i4UgNi5c2e4yD+0D//JkyfC0dFRKBQKsX3PvsgW+vfX48v4ST/02ftNnh65Xp0xk/joekj06NtfAKJXv4HitW9grG8BMS2Xrt8R/06aKqxtagkjIyNRpaqlCFGphUajEZs3bxa3bt1K3MnGcF2D06YVNUD06Ns/mpZnXXsIAeJYNavIliIgJk6cmPiLnsQsbtVKvAPxZpereOMXJAaZmYsPkiTe7t4X7U3mW0v/yqLwVn6Ivr5Qp00nVjo6CQhuVo4AACAASURBVEC8e/cu6UTG8h1vlTGjcGjfUWzZ6SoAcdXR6Ye3wo/7Dosv4yfF+bd81POcAMT/Jk/7K1v4FYEHQohHQohQYCvwY5jiX07bDk4UL1nqv1Dz5MwVngx1ffnyhWHDhlG0aFFcXFzo3rsfhYuFe9zE5KURNcWB8rIXAYOHYjJnZuTMVb5rN3J65jTSLlrAjEZNGDdxCganPEm1eiVB9m1JtWj+T93qhg/qj0XpYowfM5IP797Qs2dP/h03FgN9BZIk4eDgQPHixRN30jFcV8M9e+hXqRIrlyzihk94613p6UHO3TtZkys35c+dQelxksxZslKmbPnfIl2y6uxZJpQuA7Xq8OjhA+Ze9mKHQ3uMff57O4mMYG5nR9mhg9BT6NExYyZ8nbfi6LqbGsDJkyeTTmQs3/H1/QcwZsIkzp05hb6+PpmbNE1UQB1A6bLlqFzVktXLl/58eseUSGxPgrguQCtgVZTPjnzXmie8hf8a8AF2ALnjUvaf1MJ/4xckHrx4J/yCQhN0TikZjUYjKlSoICRJEnZt2wuvG3fjdD2+tQq/9fV+a3m99g0UfQcOEYCYYls/svUf09+orbUdrofE4zcfhX+wSuzZs1dMnjxZPHr0KNmvx+fPn0WRIkXEmo1borV8Z85bJGqACElnKj7uOyz6Dx4q9PT0hK+vb7JrjCu+vr5CoVCIISNGizd+QWLgP8OFQqEQV+88jPGeBlvbhI9n1G8oALH/qLt4t+eAGK1Uij59+iSv+OnTxdfDbuKNX5CoZlVDlC1vJj7uOyy+OnWO8XsXn2XVhs0CEGs3bfvrWvhxYR+QTwhRBjgKrI9tR0mSukuS5CVJktf79++TSV7yYJImDUGharwvX005fbeJ8NEPDQ1FM306IUePM2TkWPa5nWDB0pXkf/w4Tt4zMaU4CAsLY3DfXiyaN5sOnbvhtHUHvuucMZk8npDadTGZMxPfdc4E9e0f2Xfs+/kzPTo50qpJfbatW4mJoT5NmzZh1KhR5N++PdljEExNTbl16xatW7eK5vffqGlzziiVLLapifKyF7YNGtKsRUv8/PySTEtiSZ06NW4nPLBv64gQgl3bt1Hd2oZs2XP8sK/S0wPljesEDBuJ+YXz1FUq2emyFY1NTfTGjadu3box1JB0eAYFoW7WFOnEMa5e8aZNthzheZJa2kV+7/ydOsd5HCgq9Ro2Jmfu3KxctjgJlCcd2jD4L4GoOV9zRayLRAjxUQgREvFxFVAhtsKEECuEEOZCCPPMmTNrQV7KQq1W09quFa1bt8bf31/XcuKdxfIbN2/epGLFijjff4hBWwdsDY0wr1g5Xq/I3w/Uao4eoYujA1uc1zNkxGimz5mPQqEIfzD0HYDxts2oSpWOPF5lZY1nVUvGWpSl+J5djB8/ngEDBmjl/BKLnp4eRnrgkj8/gVWqApA+QwZq1rFlxoVz+PcbSAWLSixbsyFl5sePaAjo6+tT3rwiefLm5cmq5dg/fUJLe4cfdv8+gtlv/SZcJAn/3TsRQtC7/yBsGzRM1lNY/egRTqmMSd/JkaH+/vQ/dRLfdc4ASIsXMN3AANWcmZweNzreZevr69O5W0/OnvLk1o3r2paeZGjD4F8CCkuSlF+SJAOgDRAtllqSpKhz/TUBft/8sIlET0+POQuX8fDhQ9q0aRMZcq8z4umjr5k+nZ19+1KhQgVevHyFVMeWgMFDMbVvEasPfUx8byBuT5mOoX1LQg/uZ8qsuQwdOSYyI2PUB4Pyijem7exRenpwcN9e5tevw9KPH+mwcCHjxo3DyMgoUeenTY4fP07XDu3YtX1b5Lrmrex48/o1586cBkCtETx4+Ci691ZKIOJB6dKrF6dOeaL09KDU2FFcViiwrf+j4Y4pgvlYr77k//CBGz7X0Gg03Lh5i7dv3ybbKfj4+PDZzJzLVasxDngfMfduuo7t6JY2HSdq1GBCyZJYzZ/Dg6WL4l1+uw6dME6VipVLf6NWfmx9PfFZgAbAPcK9dUZHrJsANIn4fypwk3APHnegWFzK/dP68KMuM+ctEoCoV6+eCAxMAVGXv/LRnz5dvNu2TfxToYJ4B+Ifi0ri0cbNkf2hgfZt4+w988bvPy+d176BYu7i5SJN2rSisUkaccWxY4z9/NE8etKmFeq06cSLrt2Fr4GB8NuzJ/HnlwRoNBpRtmxZUahwEfHyU4B44xckHr3+KIxTpRIdu3QXb/yCxPylKwUg7t69m2y64krggQPiHYjDlasKdcZMomO+fKJqdasf7uXDVx+Ex4XL4snbz9HW33j4TEiSJIaMGC187j0WgJgzZ06yaA8NDRUGBgZiXotW4ouRkZhqYCDCMmYUX526iE/7D4uPHz+KwMBA4efnJ9pmzy4W5cufoN9xxy7dhYGBgbh299Hf04cvhDgohCgihCgohJgcsW6cEMI14v+RQoiSQoiyQggbIcQdbdT7O+PYqQtzFi7lyJEjLF6s4xZCbD76Ufv3LSxI1707etev86pKVab7XCVfFyeMdu0kYPBQDI+5xcl75huBA4dwK1s22rZsyqA+PShZqjTjzl0i+6Jl0faLqeV4evho1GZm5Fy1gnTDh5O26S+cwpI5BuEbkiQxevRoHty/x/69uwFIlSoVi/Pmw393eNK1ipWqAHB7yZLkzW0UB84bG7MUsD1/ljet7Vj/5Am169aLts+2zc6ULZof60pmmJcsgtvhg5HbMmXKTPkK5ni6HydL1mzkzpMn2WaBu3fvHlVDQ+nudpjBufOwv1IV/NZtQrl3FwqNmgwZMmBsbEzatGkZdfQozS/GfypVgF59BxAWFsaKJfF/Q9AFcmoFHdK2gxPb9x6kXecehKk1fPz4MflFxJLFEnd3sLBA07o17uPGEWhZnZBho5iqUlEkTz4kIZBCQ1GZmWMyZ2ak0Q8Y/W+MOeij8uzpU4YN6k+NyhW4dPE8k2fOYdcBN3LnyfvDvt9nwFwweyb/jh6O2tsrbgb8Z+eXDLRo0YKiRYuybNH8yHWm9Ruy7NMnVO0dKPLiOa0yZcJm6dLwbpTkTmr3E15t2kQv4EO/gZhuXE8NoE69BtH2qWBWnnq2tqxbt45cuXLSuZ0958+ejtxuWd2aK95efA0IoHwFCy5cuJAs2t+8eUONVKm4MXEK6588ppxZBVRW1gzNm4+NfXtF27dkyZJkzZg+QfXkK1CAJs1bsn7Nyt8iEEs2+DqmmnUNFPpKHr98R3kzM7p37568+fN/4qN/TK2mrZ4epSZOxHfAAFLPmUmwnQPG2zaDnh5B9m0xOHkcVemyMXrPfM8Nn2v07upElfIl2bJxHY6dunL+6k26dO8Vp4yKC2bP5OyEcew1NES5a1fcDHhyxjvEgJ6eHr179ybA/0vkFHnF+w3EQaEg89HDpLdvwTo/PzoaG4d3jybDgHKccHen0caNDMqZC/WkqYwrXpIdCgUlXr8Gwme/UuopsKxYge3bt9OxY0fc3d0pUaIE165cjizGsnq459XF8+cwM7fg6dOnvHnzJsnl16pVi95vPvK+TFlUKhXlK5jz4cN7Fl334XX79j/sf+b0KWytq/L+XfzHGPoOHEKAvz/rVq3QhvQkRTb4KYRUJiY0b2XPypUrqVKlCk8j0gAkOcOG/TCA+aVCBXo8fEidOnXwTmeKb/sO5Fi2hJDadTE8tB9hbIxQKglq34Fg+7YYnjxOSO260bpdvuWgV6vVHNrvSotGttSuXpkjhw7QtVcfLly7xbTZ88iUKW6eWAvnzmLKhHF0L1eO1Pv3o1e7dviGXxnwGM4PG5vw9clEr169uHHzFqbpw1uR6TNkINiyOmvSpEUKCsIAifJ+fqhbtUr2pHaxcukSA3PkQGVdg8DAQBbf8GFtw8Yor3gD0LWDAyOH9I92iKmpKVeuXGHokP/mH7CoXAWlUsmZUx5UsKgIkCyt/PA+ayIfPuXMKnDQdS9qtRo7O7sf9s+YMSPXrl5h764d8a6rVJmy1Kxjy/LFC/D/8iXR2pMS2eCnEPT19Rn17wSct+/m8ZMnVK1aFZ84zPwUiZZy3oeFhVGpUiVWrVpF7/6DODV1JgUPHSTIvi1GLlsIrWLJZ5fd+G5ywbSdHYaHDhBk3xbDfXuideP4f/nCdfsWDC5ekE7t7Hn6+BFjxk/i1oq1zMychRw5c8VZ06OHD5g28X+0adOG1pcu/Wfsv5HMBjxORLkfSqUSA309FMePopw5DYAexYph9/EDL7r1QE9PwTggrGvXlGHsAYYNY/Gd+0ydNY+L584SHBxMTqfOBA4cwmXvSxw+sJ88uXP/cJhCocBIqeCMpztfv34lderUlK9gzplTHpQuW55NLjupXr16ksuvU6cOq1cu49b162TIkJGcuXLjdvggBQoUoEyZMj/sX7JkSUqUKMGRgwcSVN+wUWP59Okjy5csTKz0JEU2+CmM2nXrsffwcQQwanQ8/IMT6W/+5s0bhBCoNBL9Bg9jn9sJJtaxJUuPLviucyasRAn8J03DwCtqS1oiuEVLvqxYje+2XZg6tUftdpiFc2dRsUxxJh8+xNKPH9k/YhQXrt1mkJk5ufr1ilcYO0CJokXx9PRkw4YN6Ovrx+tYnfHd/fiwfTtSy6Ycev8OpacHbXa4YAec0AgwMEQYG2O0YkWyjS3EhTAhSJ06NefOnEJPT4+6l71RenqwYvFC0qVLR//+/WNsVFy6dImWjRuwY+tmILxbx+fqFVShodS2bUCGeE4vGF8+f/7M8ePHCfz6lVs3r1O8ZClUKhXnTntia2sb6wTsDRo04PzZ03wNCIh3neXMKtCwcVOWLZrPp086GIuLI7LBT0Zimv8zzYC+pBnQJ9q6Mh8+cKlNO+YuXklomAbN9Om/br0n0N88NDSUadOmUbBgQdZu3MKXYBUt7R2oYFEpmodM4MAh0frnlZe98N20Df/54R5GIdWqs71NW1Y6tWfy/8ZiZlGR4SdOIXbvp/7KFaSbPiXeGS63OG/g6MF9pEulpGrVqiiVyjgdl2C0OTPYd/cjU+/ejCxQgNlel1Be9sJ/wxZy5sxFkw1r8d20jTsLluJToiQiGQeUf8b06dMZ1L8fAOfOnKJseTOkylVJ26Et/rt30rlzZ9J4ecXYqLCwsKB8+fJsXBs+l69ldWvUajXnz53h7t07zJk7L0njDq5fDw+EKl6iJLdv3aREqVLo6+uz98Bh+vXrF+txDRo0QKVSccojYdd/2Jh/+RoQwILZMxN0fHIgG/xkJGrCMAgPKDLatR2jXTt/SOZkbFOLtKamvP7ox+DNmwlq0uTXrXcbG+jVCyZODP/7E2MvhMDFxYXixYszcuRIatSsTRmz6AHQP5sjNuq2m9d9aFzXhjaLF3C4bDn2uZ1g0/bdlK9g/kP6hG9pdKMS00QmmzeuZ3DfnmzbtCH5gpK0HZX73f0o2rMnV7y9uNG4KSora5rnzo098MWiIsfUYZQ9d5aHU6cm24Dyz9i1axe3b98iMDCQK95eVLGsjsrKmg1Nm7FFo2F0SEisjQpJkujYsSM3rl/j7p3bVKhYCQMDA86e8uTsKU+GDB7Es2fPkkz7t67QNGnTERQYSPGSpVEoFFSzrPLThHmWlpbY2duTJm26BNVbtFhx7Ns5snr5Eu7fu5ugMpIa2eAnI5EZBaPO7LTJBd9N22Kd7UkjBDez5aBBQACBjRohxo6NvfUeD39ze3t77O3tMTROxZadrqzZtC1Gt8ifno9KxfTJE6hrXZUnjx+zcPlqdh88ivW5s9EeYFEzXAp95U+zFQohWDhnJoP79qRu3bq4uLjE+gqudbQdlfvd/eiYJw+SJLF7hwsAIf8Mx02l4vyZ05ETt7uFhup8PCIkJIQrV65gVsGCy5cuolKpqFItvN+93D/DedG4MRmXLPlpo6JNmzbo6emxy2UrGZcvwSl/Aa54X6Js+fCZr56uW5dk7qc+Pj5kyJCRd2/DvYFKlCzFhjWrOO1x8qfHGRgYsG3rVqpZWSW47tH/m4hxqlSMHjo45UVPgzzjlS6WmHK6/yzP+/MPX0RLuzZiPAgBImzUqB9POg45769cuSICAgJEUGiYWLfZRcxbsiIyAjS+y7nLN0T5CuYCEPbtHMWdJ69+iI6NLcPlt78xZcns0aefAISDg4MICQlJ0P1PNNqIyo3lfgyMiLx97RsoHr3+KAwNDcXOKlXFR9dDIlv2HMLBweG//ZN6HthYuHTpkgDEyvWbxODho4RCoRD3n78N/54ecgs/r7FjfzmnQu3atYWZuUV4vnkjI2FrYCAevHwvauvpiQBj4ySbU/lonTpiim39/7S/eCfqGxuLXZUrx+n4Ww8exzgjW1yXyTPnCECsWOf8Z0baysSdmGZ2im22p8hjlEqWt+/AICMjJgCBs2f/2Hr/ib/5uXPnaNSoEeXLl2fuoqX4Bamo17Axbdo5ohcx52dcEUKweeN6alWvxOOHD1m5fhPzl6yIdDmE/95kYstwKYWpfsiSCeFdAYZKfQYMGICzszMGBgbxvr6JRltRubHcjwGWlqxYvRYIj7qtVMWSPS9eYNrJkc4FCnLmzJlkS/AWIzNm8GrTJgDKlDPj3JlTdMpfkKxrVnJ1xlT07VrGOYhtw4YNuB33QGVlzcl+A9kYGoo0ahjbgP+VKJFkHknWw4Yz3OsSSg93ChQsxNe9u1gfFESqGjV+eayXlxclCuXH/Zhbguvv2LkbZcqWZ+Q/gyLfMlIMsT0JUsLyp7Xwf5YX5vvZn6Lm6I66bvnajcJn8TKhyZRJfNyxQzx//jzW67BixQphZhY+Z2uGDBnFiDH/E3efvk6w/tuPX4qGTZoJQFhWtxaXb92P95tM1PP51sI/NXWm2Od2XASrwkRYWJju5nlNypnBpk+PLCdEpY68Dkfq1gvPpbNmgwhInVpMAKHOkCHJWr+/5MQJEWhiInoWKy6evP0s6iqV4ouRkfi477CYnyu36Fm0aLRziXwTieWNJCg0TLzxCxIXrt2OfEPdVaacyJo1a5Lc57CwMBEQHCo+7jssPigUYkvRYuJrahNRA8S9e/d+eXxISIgwMjISPfr0S9Rv3ePCZWFkZCRsatcVr30D5Rb+30hMeWGCW7QmuEXLaOu+j1SNelzTFq3I0r4jn9c6c3TaNHLnzk2ZMmVwcHCgW7du9O3bF41GEKxSs27DRkJUYUyZNZdL1+8wcOhw0pmaJkj7mVOe1LS0wO3QAcZOmMx214M/9aWP7a0lapbM5736MKtyZYqOHIr78KEY6uuhp6eXfH3235OUUblRBoSvX7vCzu5dMHVqT5bG4XmADgUHE9K1B2MBRZ8+uvPHt7HBcM9eFr9/T+jg/jirVJwePJTXJUow8OULsjk4RB/ctrH573MMbyQL581h+OABFHr6hN6SxI5SpWn84hkPVq9Jkvu8d+9ecmXPileaNCzWaGhz9w6HCxTEJ0MGChUq9MvjDQwMKF++PFcveydKR9FixRk3aSrux9xYvnhBosrSJr+JU/OfgX+/gRw5uJ/9XZ24dfMGIcHB5MyVm2pWNWj79g1ZsmYDwo1+VO+Yb1GrUVFZWZN3xTrG7NuDx4njXLzkhb+/P9lz5OCdfzCSJLHGeRvpTE0T9cNSqVTMnDKRhXNnUaBgIQ4c20GZcuV/ekxUo66ysia0unXkZ+VlL94uX8WiSxdZ1N4e/y9fMG7enKlmZgnWqDViGiy1sdGO8Y0yICwsLGh96BA3lqwgq0M7MvxvLL47tmF67RqBw0eSaulS7dUbT4QQhFpZEdalG4VnTGUC0KxjJ9yPHUUIQaNGjaBChf+6c3r1Cu/6imVw+/nz57zZuI50e3Yy0qIiuz9/pub6zZg6tU+SqGIfHx+++PlhdOYUvQCfFq2wcd1D9woV4vw7qFixIitWrCAsLCxRcR+duvbg9El3JowdRYFChan7XR4iXSC38JOJ82dPY1PVgs7t2+B50p28+fJTtrwZX/z8mDbpf1QqW4JJ/44hwN8/Rn/9mFwX8xcsSN+BQ9juepCzl69z/f4T3DzORn6xTdOnT5Sxf/zwIY3r2rBgzkzaOjrh5nnul8YeYn6T+WbsAwcOYb+/P1MmjMPaygofHx/67dqFwZgxCdb52xDhpml+6BBLgW1v3qBQKOhRtBh9T7rju3Yjq/MXZHblysma4C0qly5dwj5TRpTLlrCtaDH6KBTkunePo4cPki1bNsqXLx/tXH7lAmxjY0OZ0FDODx2Bvm0DHty/x/vSZVhWszYXFmk/w6SPjw8O2XNgPmUSdkDo+Eno793P5AcP4nw9K1asSFBQEPfuJG7aDkmSWLhiDaXKlKNn5w6JfmvQBrLBT2KEECxbNJ+WjeoRFBjIinXOXLv7iPVbtrNszQbcPM9y7vINGjZpxuL5c6hrXZV76UwTPdFyYlCr1axcuoia1SpGDszOXriE1KlTx+n47/33b924Tu+tm5in1CedsRKndvacOXMGV1dXSpUqlVSnkfKIMiDcT1+f91vDZ1+qnd6UVhoND/Lk5cnjRww7eJDAdet04o//evNmVn35wuM58xkS+JW5ltUwdWpP9ls3adiw4X9J7uI4uG1tbc0sSWLvly+YmYd3+Vy97M2K+/cYlQTZJX18fKhtaspC6xp4p01Lrtx5kGrVQrF9e5yvZ82aNVm9Zh3Zc+RMtJ7UqVOzcdsOMmbMROumDaJlEtUFssFPQjQaDf/078P/Ro/Atn5DTpy5SJPmLX/wjMlfsCCLVqxh1wE3goICqTpsMK5OnWL1zU9K7t65TRPbmowdMZQqltU5cfYijZu1iHc5QghOHj+GfbNG1LSsyL7dO9GEBmOk1EOpDI+c/av4Lk3zwY4dmXzvHv67d6I/bgIngTOnPKhYqQoajYYzBgY68cdXnT1L1zRpCbasxsvnzzFs0Ajfdc7Md+zw37wN8Ug5nSFDBsqWLcuZUx7UuHCOGsBl70uULWeGt7c34sQJrfnjBwQE8PDhQ+40b8nuz58oUbIUW5w30Lt7V9RWVnG+ntmyZaNjxw6k11IKiKzZsrP38DGyZcuOQ4smCUrQpi1kg59EaDQaBvfrxaYNaxn4z3BWbdxCmrRpf3pMFctqHDt9AfNKlWk+eyYeJUv94LqYVAQGBjJjykTqVK/M44cPWbxyDZu27yZnrh8TZMWFkUMG0qZFY+7ducXUqVN58eIF//77r5ZV/0Z8NyBcbtAgepiaEuR+giJFi5E5S9bIjJIKhSLcPVMHTFap+GJRkSve4U4DFSwqhn/3hg/D0NAwxnP51eB2q1atyJs3H4rKVdmppwfH3ChTrjzl/fzQtGqlNffT0NBQBv8zlOrWNty6eYPiJUvjceIYpzw94u1+/PjRQ/bt2akVXQA5cuZi96GjlCxdhh6dHBk2sB8BOpjTWhIpMRosAnNzc+HlFf+ZaFRqDZ++hiaBorghhGDEkIGsX72CISNGM3Tkr/unU82bjcosPBVBSEgIK1s2ZcgpD17myEmxkGD81m1KEqMvhGDfnl2MHzuSl8+f07yVHROmzSRz5izxLuvcmdMUKlSIfHly4eN9kXv37uHg4PCfoZCJRAhBmFrDp0AVAD06OXLh3Bmu3H5InepVyJY1M0ePHk1WTcHBwaRJk4Y+AwZHzOK0kPsv3jGoTw8K5M3NrFmzElx2QEgYX0PCWNK0Af09PfBz7Eia9Wu5PmYMNSdO1No5+AWquPfwERXLFGPG3IWsWraYYkWLsHfv3niVM27cOKZMmcK952/j3JUZF1QqFdMnjWfRvNlkyZqNEWP/R+s2bX/IE5UxtQH6eglrj0uS5C2EiLH/VystfEmS6kmSdFeSpAeSJI2IYbuhJEnbIrZfkCQpnzbqTaksmjeb9atX0HfgkDgZe4ieZ8fkwnnGX7uCgdKAvq9eMqZwUdJ1bBenqQN/RdQB4Vs3rtOikS2bndozSKVi1wE3lq5eH29j/8XPj749utC8QR3Wr1iMiaE+VatWxcnJSTb2sSBJEkp9PfSkcONvWd2aN69f8/jRQ2rUqk0qLRqZuBIUFETf/gOwqV0Hb6+LlCpTFj09PdwOHSAkJCRRZSv1JAIDA9GrW4/FGg0F1q/FOU0aHubJoyX18OzZM/z8/bl9Mzx5WuGiRXhw/16M6ZB/RcWKFVGr1dzwuao1fRAeRDlm/CQOHvcgd548DO7bk4plijF35jQePXyg1bpiItEGX5IkPWAxUB8oAThIklTiu926AJ+FEIWAucD0xNabUtmzw4XJ/xtL81Z2jPp3QpyPi5ZnZ1b45Qne5UqtaTOZfuEcfbNkQXXaM9H6VGbmpOngwKrWzaldvTKZr13jQOrUdFy+mqrV4p+n/NyZ09SqVpHd27cxduxYJkyI+zn/7Vy5coUKpYpx8fxZKluG59K5cO4sY8ZPYotL8vfzpk+fnv9NnIp5xcpcu3IZM3MLfK5e5uvXr1hbJ+7tsmmjhnRuZ08thSLcXbJlawYYGNItDr7xccXR0ZFWTRty6+YNACRJD41GQ9myZeNdlkVEN1PU2bu0iZl5Rfa5ubPRZRdFihZn+qTxVDUrTXWLcvzTvw8bNqxHo9FovV5ttPArAg+EEI+EEKHAVuD7WaWbAusj/t8B1JJ0Fl2TdJw7c5r+vbpR2bIa85asiNO0fVH5llnS0MOdwJ59UFlZ061XX1asc2bt40dY7tzO7Vs3E6zv86dPjD16hEYBAXR3O8ze8mbsVOoTvHUn6ho1412ey5ZNtGhYF0MDA86cOcOECRMwNjZOsL6/jfz58/Py5QtOHj9GocJFSJ8+AxfPnQVAFSaSPfnWkydP8Av4yv27dwgKDKR8BQvOngpvZFglIqEYQL58+Uh94RyVZ0zDHnAuWgzfdc5aSwcthMDHx4fiJUtz68Z18uUvgFqtpkSJkgky+FmzZiVr1qzcvHE90dpiQ6FQUMe2Ptv27Mfrxl0mTZ9FVYUD1gAAIABJREFU7rz5cN2zk3/HjYu3/YhTnVooIyfwPMrnFxHrYtxHCBEG+AEZYypMkqTukiR5SZLk9f79ey3ISx7u3b1Dp7Z25M2Xn7XO2xLUlRFbdGrjZi1w2XsAf39/GtSsjvO6NfF6+r9//47pk8ZTuVxJli6cR9qWrQnt0ZtG3l4Ed+0R77EBIQQKSaJJg3r079+fK1euUKlSJe3mk/8LMDU1pVKlSrgfP4pCocCichUuXTgHQIumjejUqVOy6mnWrBlO7ey5dvUKAGXLm3H2zGmKFy9OlizxH9OJSqVKlSjx9Su3Jk/jRdFiXLtyGQ+FAidjY94fPJho7c+fP8fX15cSJUtx++YNSpQshWV1K675+FC4cOEElVm2bFluXU86gx+VXLnz0LVnHzbv2MOdJ684dTppBu1TnJeOEGKFEMJcCGGeOXPc5jvVNS9fPMehRWOUBgY4b9+TIHeuqNGpX0ePi+ze+Wb0K1Wx5Nipc5hXrMw/A/rQsLY1pz1Oxmr4hRBc9rrIsEH9sShVlHmzpmNZ3Yrjpy+wvF0Hsu9wiTVZW2xoNBpWLV+CY+vmmBrpkTd3TubNm4eJiUn4DtrOJ/8XYGtry7Url/n48QMWlSrz4P49Pnx4j5GREadOnUrayqM8oIOCgrhx4wYt0qcnx7rVpDYxoWChwphbWGjlwVO5cmVmAieEoGy58ly74o2pqSkbnj/naPlfB/P9imvXrgFQqEhRHj18QPFSpZEgwQOfAEuWLGGna8KmPEwMCoWCXLniPgVovMrWQhkvgai+e7ki1sW4jyRJ+kA6IMnmAWvRvBkzpkzk6ePHSVVFJJ8/fcKhZVO+fPnClp17yZsvX4LK+Vl06jeyZM3Gtj37Wbh8Na9evqBVk/pULleSEUMGsmrZYpzXrWHJgrn07d4ZizLFaFDLmm2bNtDKvi2nLl1lzaZtlP306acPlth48/oVbVs2ZcywIRga6BMUFPjjTtrOJ/8XYGtrixCCUyfdqVi5CgBeF84zICSEPI8e8e7du/921vbbUpQHtI+PD9XVaroePsgxPz9KlymLQqFgzLh/GTp0aKKrKlq0KOnSpeOyV3hO/Ldv3pA2XTqMjIzw9k58BOqVK1eQJAk9PQWa/7d33mFRHV0cfmcpIohib1iwK72LWFAx1thi18QSNYkpxhTT1MQYE790NVFj1GiiYo29YsNKkyaKXaNYsUtv8/2xCzaQtssueN/n2Qf27uyd32Xh3OHMKZmZNG9hS1tPF3766ae835wLDRs2pHr1krHozC/aqKUTAjQWQtigNuyDgCFPjdkIDAeOAP2APVJHDsr4+HhSkpP55fvv+Pl/39LSuzUjXh9Lj159tN4PNSE+nlcH9uW/C+dZvnYjdg4F9xVmkVu9nMfdLVmhm/0HDaF7z95s3bSBs3N/w3LpEiYlJ2ePq1a9Bq7u7nz06SS69ehJ+QqPOvg878aSm2tn0/p/mfj+u6SkJDNv3jzGjh2be8mGx1PuJ08umLH//nu1EXr8PXv3quO7Da1JuZZwdXXlzbfeom79+rSwtcfU1JTgwCMM9e3EKv8dnJ4/n2qTJj2Z7KQtHrtBq9zcWAVc+HkW/7w3jtd8O3H79i2sq1UGit5aUqVSMW3aNCrXrEM5S0s+Bm6v9MPWzoHs0OsifNZ9+/bFqlpNLpw/D6j/Bk6djClSie2EhAR++vEnnD1a4aXZVC/x5FZGsyAPoBtwGjgHfKE59jXQU/O9GbAaOAsEAw3yc96ilEcOO3FGfjZ5qqxv00ACsnadOvLLb76Tpy9d10qp45gLV6Sru4dUqVRy4T9+hT5PxMlzctX6LfJgaGSeZVRzLK9cuYq8vXGbPHbmogyPOStPX7qer3KsD6Z+80QJ5qzzPZj6zRPHLt64K+vVt5Hu7u7y1KlTef/ws0oK56NBRq7v1UV5YgPnxgP1z9vds6V092wpL964K32NjOTDsmUL97MsCJqGLz+ULSv3HAqWgPz9z0Wy/+Chsl69elqd6k58ijx3JU52EEI+LGsu/9e9hyxXrpzM2LWrSNeYkZEpr99PkqPfHCfNLSzkstXrJCD37dtXaK1paWmyTJky8s13xmvFZhhCeWS917x/3kMb9fCv3ImXS/xWS6/WbSQgLcqVk2PHvSODo04W+sMIDD8um7WwlaampvKvZSsLfZ5PJ30ljYyMJCAB6eTiKrfuDsiX0X+6W1RBH7nePDTPV63fIi/euCOTUtPluXPnZGpqat4/eG0Y7KLcMEooGRkZ8mBgqDz13zU57r0J0tTUVF68cVeO/3CiPNGvnyxy963nofl533N3l8lly8rV4z+QgDwYGimH1qotF7doobWp0tPT5aGgEBlx8pxs2ryF/NDNQyZZWsp/7R1kRuXKhf6sHz58KFeuXiNPXIiVrdq0la7uHnLS1G8kIO/cuVMkza6urrKtT4dSY/ANbtNW2xgZGdG5Ww/WbdnJzoDDdOnWg0Xz59HSqQVjhg/laEhQvs8lpWT9mlV0aufFtStXWLZ6PV179Cy0tsGDBjBhwgT27NnDrFmzuB13k7fHjCQ9PT3X9zzdFPx5ETbPq7qZY3/dxUu55eDI+HFjGdC7Oyv+mo+ZiRENGjR4JhMwR7RRT74AjdhLC+Hh4bRu6cZu/+14erUiNTWVyPCjTPHpQLN9+4refSs3HnMTmX/7HSYmpnSbM5uuZctS+9RJfrl6BWMvL61Nd+fOHbw93Vm7agUOTs78fek/0t98mz7HolCNG1fozzo0NJSB/fsREXaUmOhoWmhCM+vUqUPFxzqxFQZHR0dORB8r9hBZXVHqDf7jODg58/uffxEcFcO49yawf+8euvv60LVDGxbOn8uN69dyfF9GRgYH9u2ld1df3nx9OE2aNmPXwUDa+BTuFzT28iXMTFQ42Dbnhx9+oH379rz77rscO3aMjRs2YGVhRnp6Ognx8c+8N692iI/zePZu1nsfr7r5xM1j1Gj8rl2ljYcza1Ys5/PPP+f9998v2IVNnPjsH2379gXzyWqrxWAJwsnJiQoVKnBo/37cPFsCEOe3DKsRw7jy6yxujx+fZzvBQqG5Qf/XoAH/3rvP9YWLycjM5CtTU2q8NYYBQK2hQ7U2XdWqValXrx5REeE4Ortge/MGZgv+IOmTz8icM6fQ1xYRoc6GrVatOnfv3qG5nR2Ozi4MHz68yJqdnJy4dSvO8FoVFpbclv6G8NB1i8NzV+LkN9//JJvb2mW7VRo2aix79e0nR785To4c84bs1qOnrFS5igRk9Ro15IyfZsrLtx4U+l+1izfuypq1asvx48c/9xrefvc92dzWToYce+R6yssNk/V43D+fNSZx4BCZYW6eY+vEhxM/kw/MzKQPSHd3dxkaGlqon3uReYF9+C+//LKsb9NAXr+fJBs1biLnN2kqzyz+RwJy1qxZ6kE6amw+a9YsCcjQ6NNyupGRlCB3tmotVSqVfPDggVbn6tu3r7Rp0FAe/v4neRPk9i+myPYdO8l37ewK/Vm/9tprsnqNmnLZmvUSkOu3+cu7CSla0RsQECDNzMzkhu27FJdOSceiXDlGvzGOvYdD2HsklMlfT6dBo8Yci4xg5fKl/LtqJWdOn8KnQ0f++OsfjoQfZ8Tosflzb+TCymX/cO3qFV5++eXnjuvZoztXYy/TtUPb7GScHFskvtwLs7VPRm5IYxOsBvbFZL+6gXSK70uUXbmclJd7Z7/XZH8A5YcP4b9fZ5E25Suuz5rN1nLlCPzuO1xdXQt9fUVCly0GDZz27dtz8cJ5rsRexqNlKz6Ji8OiV19q1qrNkSNHsgbpJFopNDSU6jVqotqzizEZGYT26Em749GsHz8eS0tLrc7l5ubGhfPnqHfzJoOEYEdaGg0bN2bRhQtk+PkV6rOOiIjA1t6BE5qsWJsGDUlLLVrtnyy8vb25dfc+nl7eWjmfvnmhDf7jNG9hy9vjP+CflWs5HHaMM5dvcOrSNQ6ERDBnwWJ69e2Hubl5oc6d5UvPzMxk7uxf8fT0pIMQz42pfumllzg9ejSdTYx5pUcX1qz0yw7dfLzzVfIr/THbtPEJt025n38g/osvsRoxjPJjX8dslR9JA4dQZtdOTPYHkJycTOisX+idksKk7VupaGFK4zGjKbtxIyotxEQXGm24hEooWbVqggMP497Si7t373Du7Blc3T0eGXwdERoaymt169L084nZXaKSlq/g5X/+0bpLLWsxcaStD1ebtyAyPAwHJxcSEhI4VatWgT/rlJQUTpw4gb2DIzHHo6ldpw4Be/dQs0pFzp4tejEyIyMjzEyLHpZqKCgGvxjI8qWf+X02Fy+c55uOHREDB+aZgVqte3f+TknhrWbNmPj+O8Sv//eZzle5bb4mvfNe9so+ecBgHsxfyN2//qbs4P5MtG9KZ/8d4OPDF59/9mjCLOOqlEgodhwcHNiwcRMdO3XBzcMTgJCgQNw8PLl48SLXr+vGhxwfH09MTAztLMoxu60PoZaWWFpasjMlhaQlS7T+35WXlxebtmzFycUVR2cXoiLCs9tmFiYBy9TUlLDIaF4d+Tonjkdja6de6RsbG1O/kEmQT/PHvLmMf2uMVs6lbxSDXwxkGWWnaV/yvzJl6Dh/fv4yUNu3R7V6Nb9cvcrpwYOx+XA89/76h2stnixGmlPkjsn+AMpsWv/Eyv6H4CC6xT/Ey8iIXbt2sXnzZlq0eLqwKUqJBD1gbGxMz5d7UMGqAg0bNcbKqiJHg4Nw12zi6mqVHx4ejpSSe+PeYdXNG9g7OnHoQAAD+vYq1Io7LywtLenapQvlLC1xcHLmVtxNzM3NsbCwoDC9L4QQ1K3fgKrVqnP29KnsWjp2dnZaS7S8dOkS/65eSVpamlbOp08Ug19MpLVth3z7XSampCAKEm7Yvj3irbeovXAhqrfeYv2t63g52zFn1i88fPAAeDZyp+xvs7AaMYzri5cxp5U3QRM/xWrkMCbYNmf00qW8FxtLx44dnzunUiKh+Dl//jyzfvqehPh4XNw9CA0Jws7BiT8WLsbbWzc+ZG9vb4LDonBz9+T4sSgcnVyIDA+nTJky2Nra6mTOY1GRLPxjDk7OLgBER0Uyeep0evfuXeBzzZkzh5UrV3D6ZAzp6ek0a2HLiehjhaqBnxv29vakpaVx9sxprZ1TXygGv5gw2R9A+b8XFzzc8Kkwxc4mxrTy8uLryZ/jYtuYBf16U3Zw/+zaOEcmforZV5NYWL06TUa9ykfj32bZpQuoVq2i+qX/GDp0aP7Krr6A8fD65ty5c0yf+iVHQ4Jx8/Dk9MkYUpKT6d1vYJGrVeaGSqXCpnETrsReJiUlBUdnZ6Iiw3F0dCxScMLz2L59O19M/JDa1nUxMjIiKiKcEWPeoHUhOrrNmDGDrZs3cixKXTyttnUd4uJuatXgZ53r5IlorZ1TXygGvxgw2R+A6YC+rB04KM+mz0+QQ7Po2hMmsP2TT7j0zjt84u5O6oEAxtesgbFvBywOHyDolx/pnJbGzdjLDB0yhEOHDjF79mwlHr4E0LJlS1QqFcFBR3Bz99RUPA0hNvYyv/3+O6mp2m/b+fnnn7Pb3z+7JLK9ozNREeG4ueXYIU8rZG3cnjl1kqaajdv09HQOHjrElStP113MnStXrnD58mVc3T2IjorAolw5GjRqxHf/+55OnTppTW/Tpk0xNjYm5njhe1EYCorBLwaMj4Yw1NiYjQ/uqw/kN9zwOWGKdfr25YvISKZs2sTX+/dT/shByr06hF7TprHo/Hk+v3ePP/74g1atWuVe6Cw3crjRaD3pR+EZLC0tcXR0JPjIYZxd3VCpVISGBHE0JJh333knO8FIW9y/f5/vvvuOiLCjRIaHYVm+PFJK4h8+1Globta5I8KPql1IEeHE3bxBh3ZtWbFiRb7PExgYCICbhyfRUZHY2tlTuXIV3p/wgVbdUaampvj6dsKsFDT3UQx+MRDo04GNDx/w0ksvPTqYnxX388IUNcbfaPBgasyZk22gG44ejY2NTcGN/OO8wPHw+ubLsmWxDA6krLk5zVrYEhoUiI+UfIz2N27DwtTt+9TRMmHYOzrRoGEjQsIi6Nmz8CVD8qJSpUrY2NhoQjKduXP7Funp6TRq3IQ9e/bk+zyBgYGYmprS3Nae49HHsHNwJDoqkhvXrmpd87ZtW/nwk8/yHmjgKAa/GNi/dzcAvr6+2j2xrvzsL3A8vL4p37EjS5KTubt6BW7unlgEHaHBRxO4WKWK1g1+VlRMc1s7TkQfw9HJBSMjIxwd7KlSpYpW53oaNzc3TsWcwNFFvXEbFRGOd1sf9u/fn+9omGvXruHo7MLVK7EkxMdj5+DE+HFjefONsTrRbKKDloPFTcm/ghLAvj27cXR0pEaNGto9seJnL3W0+uILzLdso8UXnzHuxnX+Skgg+utvSGnTjsOHD2t1rpCQEOrUrUvczRuaDVsX/vh9Fju2Fb3lYF7MnTuXA0FHaWFrj7GxMZHhYbRu2474+Ph8h2cuWvw3azfv4Fik2tXVvIUtp0/GFKqHbV6EhobiZNuU4EDtfgbFjWLwdYyUkopWVvTq9XRf9yKi+NlLJWXKlKFsl84kvj4Gz62bmQvszsjE3bMlly9f5saNG1qb68bNm7i6exIZrnbtODg688N337B9+3atzZEblStXpqyZKZXm/c6rdeoRGR5GqzbqRunnFizIV5JfWkYmpqamREdFYmJigpGREWlpaVqN0MmiWrVqXPrvIieiS3akjmLwdYwQghWr1zB16lTtnljxs5da9kyeRPovP/Lw408ZJwSJG9czcMirXLp2k+rVq2ttnh3+u5n9x0Iiw8OoUMGKjMwMnW/YZpGZmcmnH33A7vv3+fVqLOVDgqlUqTKBP/7C0I0b80zy+/bbbxnY7xUyMzOJjoqkabMWnDl9EkAnBr9OnTpUqFCBmBMlO1JHMfg6Ji0tDZMiNFLOFcXPXjrZuxfvWbPok5bGyVdH8D83D8bt3U3lyAgqL/hTqyUv0jIyMTExITI8LHvjFigWg69Sqdi5fRsLzp5h66gxLIx/SMbED3H/bjqZvfs8+4anrnP9+vXE3YpDCMGxqEjsHB05ER2NqakpTZs21bpeIQT29vbEHD+m9XMXJ0WyREKISkIIfyHEGc3XHLsNCCEyhBARmsfGosxZ0ni5kw9vvVE66nAoFAMhIcT+9BP7gKMhgZh06cYrGRmkHzzAzrv3iO/eXSslL6ZNm8aIV4eRnJxMzIno7Lo2ZmZmOZfb0AGurq5ERYRRccAg5gK158/l3rDX+OP+fVL79Mn1Oq9fv077kBBGNWjE1Sux3L4Vh52DI2/bORAxZIjOEsbs7e2JOXG8RDdDKerS81Ngt5SyMbBb8zwnkqSUTpqH7uK9DIz4hw+JioygVq1a+paSP5Siafpn4kRsRo3CwsKCkKBAXD082QfscndnnxD0TklB9u9f5JIXW7Zs4erVK8QcjyYtLU0d7RIbi5OTk9Zq0OSFm5sbly9douF/F3kL2ObZkgpLl7Bt21Z+8vDItbTHhg0bCAZGbt3ElSV/AfCSsQlNP/2I5q+9pjO9nTt35pX+A0hKStLZHLqmqAa/F7BE8/0SoODFMEoxYaEhZGZm0rp1Cel4rxRNMwiMjY1xd3fnaHAwzi6u6gSs4CBatW7L7sxMznfuXKRQ3OTkZMLCwp7YsHV0duGvpX4FioMvKm5ubvgA1u+N47OGjZhR1pz7i5exLC2Nw4cPkz5mTI7XuW7dOv5r0JD4v/3wmfUL3xob4/zNV/w7ZChXmjTRmd5evXox67c5hS6TbggU1eBXl1Jm9QW8DuS2o2QmhAgVQgQKIV6Ym0Jw0BGEELRs2VLfUvKHUjTNYOjWrRv16tfH3MKCFrb2HA0Ows2zJR1VKmqsW1ekUNyjR4+SlpaGu2dLIsPDqFS5CtZ16mJspKJsMWaTuri44FuhAv5vvk1K67ZEhoeR2qYtYZO+pEdCAum//ZbjdbZp146RY94grZ0PKypW4rP0dE77vkT/2TM5efKkTjWrkDy4f1+nc+iSPA2+EGKXECI6h8cTcYaa1lq5ObfqSSndgCHAr0KIhs+Zb6zm5hAaFxdXkGsxOEICj+Dg4ECFChX0LSX/KEXTDIKPP/6YpX4rEELg5tmSsKMhlA8JYrVKxSc2NkUKxc1K4HL18FSXN3B2Yee2LQwfMpDbt2/r4nJypEKFCky8fQePL6bg5OLK/fv3OH/uLPaOTvRTqfg0l+t8e/yHjB33LnKXP4OvXSWoWTPqbd6ID+Ds7KxTN6SrsyOffPCeTs5dHORp8KWUvlJKuxweG4AbQoiaAJqvN3M5xxXN1/PAPsD5OfPNl1K6SSndqlatWohLMhz6DRhY8Ebg+kZJ5jIYTIxUpKam4ubhSfzDh9zfsZ1VAwdzslYtMjMzCx2KW716dbq93BMLi3KcijmBk7MLhw4EsHPH9mJfnBirBAJo6a12ex7aH4BpeBhbRo7mpp29umCc5jpTDh5k8eIlJKakYbI/gIqjXmUy4HjlCisbNmKNSkWlxYt16oZs3KgRJ46X3Fj8orp0NgJZreGHAxueHiCEqCiEKKP5vgrgDZwo4rwlgtdGjGDEiBH6lpF/lGQug+Llbl0ZM3wobu7qDlj/Nm5C7znzWb9526MS14UIxR08ZCiLlq7k+LFIMjMzsyN0inPDNovNmzfj6dSCClZW1KhZk0MHAkh8/0Ne+ulXZs5b+Cjipn17vkpMZOTIERw/FoVJWCjL+/bjV+DanPn0PnGcqFq11K5IHboh7e3tOXfmtE4qlxYHRTX4M4BOQogzgK/mOUIINyHEAs2Y5kCoECIS2AvMkFKWeoN/9sxprl76r2SFcCnJXAZF7dq1OBocRN369alcpSqhIUEApKVnkpGRUahzPnz4kIeJ6iiTrA1bOwdHjkVF6qV5vaWlJRcvXOBYZASt2/pw6MB+pJQIIciUkiMh4YwePZpJkyYxY8YMBgwZhoOTM4nvf8iauDjq1bdBduzE7xkZtI+NhQ8+0Kkb0t7envT0dM6ePqWzOXRJkQy+lPK2lLKjlLKxxvVzR3M8VEo5WvP9YSmlvZTSUfN1oTaEGzo/zZiObwefolWtLG6UZC6DwsvLi1u34rh08SJuHp4cDVYb/AnvvVPoDlizZ8/GukY1Hj54QER4GNWq1yA+Pp6E+Hjc9RCN5aIpnqYurdCOW3E3OXUyJvv1ffsDWL58OdOnT6dbj5788OtvgLpkSWhIMO6eLbE6GsrnlSqT+OGHOndDZmXxxpTQZihKpq2OiAg7ioeHh75lKJRgvLy8AAgJDsTN3YNzZ89w585tatW2JigoiAeTJhU4b2LPnj1MtbSkUkQ4YSHBuLi5kZgQz9gmTegcFaXLy8mR8uXL07hJE6Iiwmjd1geAg/v3Zb/++ti3CI85R3jMWRYtW0mZMmUAOHP6FDdvXKdflapYjRhGxooVmP/4o87dkE2aNGHKl1/Rwk775RuKA8Xg64C7d+5w4fw5xeArFAlbW1ssLS0JDQ7CzVNt/I+GBOPbuQsAB1NTC5Q3kZyczKFDhxAtvSg/fAh1z53F1d0Tz8Qk5t25Q/Xu3Yvlup7G3c2NyPAw6tarR9169Tm8P+CJ160qVqRmrdpPHMsqOe5lYsLyV/oz+5im5IGO3ZAmJiZMmTKF5i100+9X1ygGXwdk+Ub18S+yQunByMiIKVOm0MG3E47OLhgbGxMaFEgLO3tq1qrNwnPnCpQ3ceTIEZKTk6k2eBh73n6PVcCrp05iNWIYrFyptxDcHj160OmlLqSlpeHdph2HDx7Ic48iYO8e6ts0oMzUb5hx5BC7/Hc+elHHbsgH9+8RfOSQzs6vSxSDrwPCw9T1vHXZF1ThxeCjjz6ix8s9MTc3p4WdAyHBgQgh6NipM/7+/qR6e+c7b8Lf3x8jIyNatmrNloQE/hCCpn7L+Dkxgdl6LPs7ePBgZs2Zi4mJCe19O3Hv3l2CA3Nv9pKUlMShAwG06+BLcnIyp2JOZO8FFAd//vknPbv4cu/u3WKbU1soBl8HDB72Gps2bSpZCVcKBklmZibnz5zk+rWreHl7ExYSTEpKCgOHDmPSlC/J2LUr33kTAwcO4vtfZlO+QgUydm5nnJERF0e8zvCkJFposc5+YTBCcu/uXTp26kyZMmXYuml9rmP3791NYkICXbu/zLHIcNLT04t1cWVvbw9QIkslKwZfB9SuXZsePXroW4ZCKeD+/fu4OzuyctlSWrVuS3JyMuFHQ3D39OLNFvaUHT4833kTLbZuZYRNA1R79/Bl9DEWde1OSPny/Av46DnJrq23F+PHjcWiXDnadfBl66YN6uQywPzXnzB5zK+/ZdMGelhY0DkijJAgdSPzVq1aFZvWLIN/ogSWSlYMvpaJi7vJovlzuXpV+42UFV48KlasSLNmzQgNCcLTyxshBIcPHgAg9eABdoweTWa7durBz9mw9Pf351BqOlYjhpGyYD4DgKbNmtN5/lw2mZujWr1ar/kWzZs3z9776tOvP1diYzkYsA+ANBc3rEYMw2R/AAkJCSRsWMey9HQy3T15cP8+dnZ2Wm0Mkxe1a9emYsWKnDyurPBfeIKPHOaTDydw+fJlfUtRKA18/z0j69cnLCSYClZW2Nk7krRpA+a//sSaRo3oMmMGBw4ceDQ+lw3LTz/9lAkb1nFv8VIq7d1Fe6DLgj/4sE5dEj09UXXsqNd8Czc3N65fu8r1a1fp2qMXFStW4p/F6pSdtLbtuLd4KVYjhnFz1Gv8lZDAsanfkNa2HZ9N+YqoYg4nzW6GUgJj8RWDr2Uiwo5ibGysk0bKCi8g7u68e/Ag9rdvceH8OUbY2DA5KpJEewde6toDCwsL/v777+ee4uTJk4SFhdF3wCDS2rZjU526TAFSRr9BzaGv8poOa8jnl6wS4kcOHcDMzIyBQ19y5pfeAAAbG0lEQVRl66YNnD93FlAb/YRRo2m5fSurqlSh4ZtvA2BqpNJLcuOMGTOY8eOvxT5vUVEMvpYJPxqKo6MjZmZm+paiUBpo357rM2eyClBN+oxxe3YxAAgsWxYLCwv69BuIn58fd58TMbJw4UKMjIzo1bcfJgH76HjqJGta2GK+aAGftPQ0iHpPzs7OVKhQgYMBal/92+MnUMbMjOlfTUZKicn+AIzm/sbXwKjkZEwP7Gf9mlX4tvPWamP3/OLl5YWra/FFBmkLxeBrkczMTCIjwpT4ewWtUm/ECG7174/n1s3Ej3ydgMf8+MNfH0NSUhJ//fXXk2/SdC9LSEhgwYIFdO/Zm/prVlNhQB/6S8mVd8Zz4eeZmA0dbBDF8YyMjJg5cyYDhwwFoGq16oz/YCJbNq7n3zdHYz5kAP0yMtjk4UnS8tVYjRjGnTWrOHP6NFWqVCl2vYmJiaxdtYKTMSWrLJhi8LXI5UuXiH/4UDH4ClpFFRBAs717SZj4GZWWLWWkTQP271N3prJ3dMK7TTuOHAl88k2a7mU3VqygjJkZE1t6UW76VHZ18GUf0LptO77as5u+qanI4OBiv6acGD58OG3atMl+/u4HH9H7lf5cXLGcrg8fcLJmLRb8vZy0dj7cW7wU07CjtGzZEiMjo2LXmpGRweiRw9m+eVOxz10UircWaimnXv36XI+7g3kZ5ceqoCU05RJuzJrFvHMXGDNnPjNHDKX3xQvcu3sXq4oVWbx8FVUqWT35Pk3Ejs2AAZx8dQTlv5/BvZX/8vOc2dg0aIh1nbocDQ2mjpcX4pNP9HNtT5GZmUngoYMYm5lj7+iESqVi7sIlBL4+luZ3btOufUcsypUD4IadHR/fvMFXhSwiV1QsLS2xsbEpcaGZygpfy1S0Kk85zS+lgkKR0ZSsjqxUiamTv+CImRnHpn6DS2Ym+/bsAsCyfHlS0jM5dfYc4eHhgLqa5OL//iNx5OtY/fIjia+PIamVN0cOHaB1Wx8ePnjAyRPHizV+PT+8OmQg8+f+lv1cCIGXd2u6vdwr29gDHNi3FyklnTp10odMQF0582QJS75SDL4WmfTJh/yzZLG+ZSiUJjQlqz091U1QQkOCqDf6DRZWqsyunduzh2VmZvJK3760a9eOKVOm0LVrV5aMHIlq7hziJ36G+cI/if1rAQ8fPKB1Ox/CQkOQUmZX5DQEVCoV7du35/CBgDz7SFSuXIX+/fvr1X3q4ODA2TOnSUxM1JuGgqIYfC2RmprK34sWEBMTk/dgBYUCYmVlha2tLSFBgRgZGdHetxN7/HdmFxlTqVQsXfUvTi5uTJs2DVVAAJstLEhavoqEL6Zwb/FSbKd8gQ/Qqk1bQkOCEEJk30gMhU6dOnElNjbHlfPjGbdtfdqzatUqjA8c0Fn/2rxwdXUlMzOTmBLk1lEMvpaIOR5NamqqsmGroDNat25NSNARMjIy8O3chTu3bxGhKdQHULNWbVZv3Mqp/66x4tMvSFmxlrR2PoA6jv0Da2t61apN1arV6NNvAH///Y/B1Xvq2bMnQgi2bHqmW2p2xm38+n+5E3c9z3LQusbX15ezF/7Dxa3klEEvksEXQvQXQhwXQmQKIXKtXiSE6CKEOCWEOCuE+LQocxoq4UfVf3iKwVfQFT4+PqQkJ3P+3Fl8OnRCpVLhv33bM+MqWFmRNOEj0tq2yz524/o1/jx9mhujxgDQuHFjhg0bWmza80v16tXx9vbmYMCzoaJZGbfV3hzNmsYNyOzfX6f9a/PCwsKCenWsS1RXu6Ku8KOBvsD+3AYIIYyA34GuQAtgsBCiRRHnNTjCw0KpWrUq9evX17cUhVJK7969uXbzFo2bNKVipUp4ebdh0/p/89U3ecfWLQB07tadq1diWbd6Bffu3dO15ELh5+fHpm07c3wtpXUbFpma8kVmJqpx4/Rm7LPYsnkTU78oOWvYova0jZFS5tXN1wM4K6U8L6VMBVYAvYoyryFibGxMx44dS9TdXqFkYWZmhpXlo0iV3q/059zZMxyLjMjzvevWrqJho8Y0a2HLvt27GDtqhF4yVPODtbU15c1zzlSPmfkLQ+7fJ7pvX533r80PkZGRzPt9Fgnx8XrVkV+Kw4dfG3i8klis5lipYvbv8/Dz89O3DIVSzqZNGxnYuzvp6en06NUHExMT1q7ye6aEMIDJ/gDMf/2JK7GXCTx0kL4DBiGEIDjoCJUrV6ZJkyZ6uoq82bJpA+293Lhz53b2MZP9AbhMn8obVlY08fPTef/a/ODq6oqUkmNRed90DYE8Db4QYpcQIjqHh05W6UKIsUKIUCFEaFxcnC6m0AnGRsrKXkH3JCUlEbB3D9FRkVSsVIku3V9m5bKlJNjaZ5cQBrVxtBoxjDQXN1b7LUdKSd9+A5FScjBgLz4+Pgb932jjxo05GXOC2T//mH0sce8uBgIuH36IqampzvvX5gdXV1fgUVtTQydPgy+l9JVS2uXweHYbPWeuAHUee26tOZbbfPOllG5SSreqVavmcwr9Mnf2r7Rp6UFKSoq+pSiUctppat8fPqjeNhv++lju3bvLihvXs0sIW0z/GqsRw7i3eCmJXq1Ysmg+7dp3xKZhQ/67cIHYy5fp0KGDPi8jT+zs7BgxYgTz58zGf8c2pJSYfjWN307E8MEHHzwaqOP+tXlRo0YNrK2tS4/B1wIhQGMhhI0QwhQYBGwshnmLjdCgQBIS4ilTpoy+pSiUcmrWrEnTpk05oGkO4t2mLS3s7Jn98w8ktfIm8fUxlPv+OxJfH0Na23asX7uaa1ev8vqb4wB14hZAx44d9XUJ+WbmzJnY2dnx6oC+vDFiGBXNTWncqCHm5ub6lvYELVu2JCUlWd8y8kVRwzL7CCFiAS9gixBih+Z4LSHEVgApZTrwDrADiAFWSSlLVj5yHoSHheLhUXJicRVKNl26dOHIwf0kJiYihODjzyZx/txZDkz+HPOFf2Zn1spd/sz45ivsHZzwfakLAP0GDubc+QsG7b/PwtLSkoCAAL7//nvcXZ0xMTLMtKGVK1ey1G+VvmXkiyJV+ZJSrgPW5XD8KtDtsedbga1FmctQuX7tKlevXDG4jEWF0kvv3r05efoMd27fwty8Ll26v8w79g74zv2NU/MXUWnAIFLbtMOif28aJyfz9m9/oFKpjaVKCBrY1Ner/oJQoUIFPv74Y33LeC4qlQoTIxUp6Zn6lpInhnnLLEFkJVwpK3yF4sLHx4eNGzdhXacuoC4wNrHjS4w0N8f3q0msWenHx5vW0zM5mXe8WtHGRx2rfupkDGNHDOX06dP6lF/qkFLySu+X+XHGdH1LyRPF4BeRSpUrM2DQIJycnPQtReEFwtRYRdzN69lJV6ZTpzFhxx6MjIx5Z+woFs2fR8M33sJz06MCa/t2+7P+37VKNzYtI4Tgwf377N+7W99S8kQx+EXE08ubFcuXK39ECsXK2rVrsW9s80THJVt7BwIjjrN970HCY84y/fufMTExyX7df/s2mjdvTt26dfUhuVTTqlUrIsPDDD5STzH4RSAzM5Ob168adDyzQukkq6zxru1Pbo0ZGRnh5OJKzVpP5jbeu3uXI4cO0KtXqUtyNwi8vb1JSUkhKsKwwzMVg18Ezp45jUPThixbtkzfUhReMGrVqoWbmxubN6zP1/jd/tvJyMgwXIOv6cH7BHv36q30cUHJaiQTEhSYx0j9ohj8IhAarI5pdnEped3rFUo+gwYNIjIijPPnzj53nPmvP1Hn9Gna+fg8Ci4wNGOq6cGbbfT37oUePcD4qUBCQ9OtoXr16rz66mvUqm2tbynPRTH4RSA48DCVK1emWbNm+pai8AIycOBAhBCsXbXiuePSXNzovmgB+6ZMUYdn6rmOfI5klUkYMACmTFF/nTYNvvvuyZuAoel+jL//XsIr/QfoW8ZzUbptF4GQwCO0atVK8eEr6AVra2tWrFiBo0fr546LqVkTxyV/U27AAHjrLXWVST3Wkc+V9u3V+qZNg8mT4YMPwNlZbeQNWfdjJCU8JDE5DauKFfUtJUeUFX4huXUrjnNnz+Dt7a1vKQovMAMGDKB2zWq5vp6ZmcmIwf3p/9usR8b0rbcM02ju3as26pMnPyp9/PhNwFB1a7h37x4NrGuyZOGf+paSK4rBLyRlzcqycPHf9O3bV99SFF5w9u/2552xo3JshLLbfwdnz5zmA2fnZ42pIZHlrlm1Cr7++pF75+efDVv3Y2T1HQ7YZ7jx+IrBLyQW5coxbOhQGjdurG8pCi84165dZc1KP7ZtfrImYUZGBt9OncLgGjXw/fPPZ42pIRnPkJAn3TXt28Nnn6n9+Yas+yk6depESOAREhIS9C0lRxSDX0g2r/+XC+fO6FuGggKvvfYadnZ2fPbRBO7euZN9fN5vM4k5Hs0nHTognjameq4j/wwTJz7rrklPh02bDFv3U3Tq1Im0tDSCjhzSt5QcEfnph6kv3NzcZGhoaIHfl5aRyZ2EVB0oUpOUlESTOtWZMGEC//vf/3Q2j4JCfgkPD8fT05PmLez47c9FNGrchFf798aqQnlWr16tBBYUE0lJSVSsWJERY95g6vTC24bKFqYYF7I6qBDiqJTSLafXlCidQnA0JJi0tDRat35+dISCQnHh7OzMhg0b6NevH0EHA3B1tGP7tq2kpqYqxr4YKVu2LIsXL6ZBcwd9S8kRxeAXgoMBezEyMsruPqSgYAh07dqVS5cuUaZMGcxN1X/axk8nLinonEGDBnE/MY3k9Ax9S3kG5behEBwI2Iu7uzvly5fXtxQFhSeoXLmyviUoAFs3rScxVd1o3pBQNm0LSGJiIsciIwy+J6iCgoL+mPv7bH7637f6lvEMygq/gJibm3Pu0hXKqAx3s1tBQUG/vPLKK7z//vucP3eWBg0b6VtONkXtadtfCHFcCJEphMhxV1gz7qIQ4pgQIkIIUfCwGwOjWuVKVK1aVd8yFBQUDJSshMzN65/pAKtXiurSiQb6AvvzMba9lNIpt3ChksKH745j3drV+pahoKBgwNSpUwdvb29W+S3NMQNaXxTJ4EspY6SUp7QlxtC5cf0ay/7+i/Pnz+tbioKCgoEzduxYMjMzuXH9mr6lZFNcm7YS2CmEOCqEGPu8gUKIsUKIUCFEaFxcXDHJyx87t6m7C/Xo0UPPShQUFAydIUOGEHPyJDVq1tK3lGzyNPhCiF1CiOgcHgVpndNaSukCdAXeFkK0zW2glHK+lNJNSulmaH7yHdu2YGNjg52dnb6lKCjkTAnvHFWaMDY2xtTYiLSUZIOprZOnwZdS+kop7XJ4bMjvJFLKK5qvN4F1gEfhJeuHhPh4DuzbQ8+ePZXMRQXDJafOUQbcNKS0ExcXh0PTBvz15zx9SwGKwaUjhLAQQlhmfQ+8hHqzt0RxKy4Oz5Ze9OljWIkUCgpPkFPnKANvGlKaqVq1Km5ubsybPdMgVvlFDcvsI4SIBbyALUKIHZrjtYQQWzXDqgMHhRCRQDCwRUq5vSjz6oN6Njbs2r1bKaegYPiUoKYhLwJTp07l1q04g1jlK9Uy80FqaioPH9yjRYO6WjungoLOyHLjlJC2gC8C3bp1IygomOCoGMpZWuY5XlfVMpXSCvlg145tODZtSGFuPgoKxUpunaMMuGnIi8DUqVO5c+c2O7dv0asOpbRCPlix7B+qVq2Kk5OTvqUoKDyfnDpHZTUNUVb5esPd3Z2wsDDqN7ElNSNTbzoUg58HN29cZ/fO7Xz00UdKqVkDIi0tjdjYWJKTk/Utpdgxi43F2sQEEx+fRwf37lUb9YkTn31D+/aKsTcAnJ2dSc/IJDAsirr1bShTpkyxa1AsWB6sWelHRkYGI0eO1LcUhceIjY3F0tKS+vXrv1BhslJKbpcrR+yWLdhIqTbkj7txFAya2MuX6Njak559XmH2HwuL/XdX8eHnweoVy/Hy8qJp06b6lqLwGMnJyVSuXPmFMvYAQggqW1uT7OmphF6WQOrXr8+UKVNYs9KPb6dOKfb5lRV+Hqz5dz0J9+/kPVCh2HnRjH0WQggwM3sUejl5smLsSxCff/45ly9fZvYvP5KSksyUad8Vm7tYWeHngpQSpKRpowa4K1mKCjkghGDYsGHZz9PT06latWrx1FpKTlaHXE6erP6qROGUGIQQ/P7777z33nvMn/Mbq/yWFdvcisHPhb27/RnQu7tBVbpTMCwsLCyIjo4mKSkJAH9/f2rXrq37iR88gLg4JfSyBGNkZMTMmTPZtm0bb415HWOVIPxoKHFxN3U6r2LwcyAtLY1pU77g8n8XqVKlir7lKBgw3bp1Y8sWdWy1n58fgwcPzn4tISGBUaNG4eHhgbOzMxs2qMtPXbx4kTZt2uDi4oKLiwuHDx8GYN++ffj4+NCvXz+aNWvG0KFDc66lnpgIVavmHHqpUKLo0qULZqbGlFVl8NrAvrg0b8SY4UM5c+aMTuZTfPg58Mfvs4g5Hs26deswNTXVtxyFPHj//feJiIjQ6jmdnJz49ddf8xw3aNAgvv76a3r06EFUVBSjRo3iwIEDAEyfPp0OHTqwaNEi7t27h4eHB76+vlSrVg1/f3/MzMw4c+YMgwcPzk7qCw8P5/jx49SqVQtvb28OHTpE69atn5y0Rg24e/fJY0roZYnG3NycgIAA5s2bx4oVK7AwL6uTeZQV/lNEhB3lf99MpU+fPvTu3VvfchQMHAcHBy5evIifnx/dunV74rWdO3cyY8YMnJyc8PHxITk5mUuXLpGWlsaYMWOwt7enf//+nDhxIvs9Hh4eWFtbo1KpcHJy4uLFi8V8RQr6onnz5sycOZNr165Rt65uyrgoK/yn+PG7b6hRowYLFizQtxSFfJKflbgu6dmzJx999BH79u3j9u3b2cellKxdu/aZkN6vvvqK6tWrExkZSWZmJmZmZtmvPZ6MY2RkRHp6uu4vQMGgUKl0tw4vlSv8e/fuMfPH7wkLDc73H0xGRgYqIVjht5ydO3dSqVIlHatUKC2MGjWKL7/8Ent7+yeOd+7cmdmzZ2f74cPDwwG4f/8+NWvWRKVS8c8//5CRkVHsmhVeTEqlwY8IO8p3076kW8d2NLepzfDB/Vkw7/dndsCllMScOM7nH09gYK/uVDAzonKlijRr1kxPyhVKItbW1rz33nvPHJ88eTJpaWk4ODhga2vL5MmTARg3bhxLlizB0dGRkydPYmFhUdySFV5QSmV5ZFB3mtmzZw+7d+9m9+7dnD9/niOh4TRp1oI//5jH7F9+4MGDBzx88ABjY2NGjhzJL7/8ovzxlRBiYmJo3ry5vmXojRf9+hVy53nlkUutD79q1aoMHDiQgQMHAupQuHr16iGEwK5pQ17q1AkLCwscHR3p0aMHNWrU0LNiBQUFBd1Sag3+09SvXz/7++7du9O9e3f9iVFQUFDQA0VtcfiDEOKkECJKCLFOCGGVy7guQohTQoizQohPizKngoKCgkLhKOqmrT9gJ6V0AE4Dnz09QAhhBPwOdAVaAIOFEC2KOK+CQs5ZqC8AL+p1KxSdIhl8KeVOKWVW3GMgYJ3DMA/grJTyvJQyFVgB9CrKvAoKZmZm3L59+4UzflJKbt++/UTsvoJCftGmD38UsDKH47WBy489jwU8tTivwguItbU1sbGxxMXF6VtKsWNmZoa1dU5rKwWF55OnwRdC7AJyCmH5Qkq5QTPmCyAdKHKdTyHEWGAsoLP0YoWSj4mJCTY2NvqWoaBQosjT4EspfZ/3uhBiBNAD6Chz/v/6ClDnsefWmmO5zTcfmA/qOPy89CkoKCgo5I+iRul0ASYCPaWUibkMCwEaCyFshBCmwCBgY1HmVVBQUFAoOEWN0vkNsAT8hRARQoh5AEKIWkKIrQCaTd13gB1ADLBKSnm8iPMqKCgoKBQQgy6tIISIA/4r5NurALe0KEdbKLoKhqKr4BiqNkVXwSisrnpSyqo5vWDQBr8oCCFCc6snoU8UXQVD0VVwDFWboqtg6EJXqayWqaCgoKDwLIrBV1BQUHhBKM0Gf76+BeSCoqtgKLoKjqFqU3QVDK3rKrU+fAUFBQWFJynNK3wFBQUFhcco1QZfCNFfCHFcCJEphNDrLryhlogWQiwSQtwUQkTrW8vjCCHqCCH2CiFOaD7D8frWBCCEMBNCBAshIjW6pupb0+MIIYyEEOFCiM361pKFEOKiEOKYJlencC3sdIQQwkoIsUZT5j1GCOFlAJqaan5WWY8HQoj3tXLu0uzSEUI0BzKBP4CPpJR6+WXTlIg+DXRCXTwuBBgspTyhDz2PI4RoC8QDf0sp7fStJwshRE2gppQyTAhhCRwFeuv7ZyaEEICFlDJeCGECHATGSykD9akrCyHEB4AbUF5K2UPfekBt8AE3KaXBxboLIZYAB6SUCzSVAMyllPf0rSsLje24AnhKKQubk5RNqV7hSyljpJSn9K0DAy4RLaXcD9zRt46nkVJek1KGab5/iDpLu7Z+VYFUE695aqJ5GMSqSQhhDXQHFuhbS0lACFEBaAssBJBSphqSsdfQETinDWMPpdzgGxA5lYjWu/EqKQgh6gPOQJB+lajRuE0igJuAv5TSIHQBv6KubZWpbyFPIYGdQoijmmq4hoINEAf8pXGDLRBCWOhb1FMMAvy0dbISb/CFELuEENE5PAxiBa1QNIQQ5YC1wPtSygf61gMgpcyQUjqhrvzqIYTQuytMCNEDuCmlPKpvLTnQWkrpgrrr3dsaN6IhYAy4AHOllM5AAmBI+2umQE9gtbbOWeKbmOdVvtlAKFCJaAU1Gh/5WmCZlPJffet5GinlPSHEXqALoO9Nb2+gpxCiG2AGlBdCLJVSDtOzLqSUVzRfbwoh1qF2ce7XrypA/Z927GP/oa3BgAw+6htkmJTyhrZOWOJX+CUEpUR0AdFsji4EYqSUP+tbTxZCiKpCCCvN92VRb8Sf1K8qkFJ+JqW0llLWR/37tccQjL0QwkKz6Y7GXfIS+r85AiClvA5cFkI01RzqCOg9kOIxBqNFdw6UcoMvhOgjhIgFvIAtQogd+tBhyCWihRB+wBGgqRAiVgjxur41afAGXgU6PBae1k3fooCawF4hRBTqG7m/lNJgQiANkOrAQSFEJBAMbJFSbtezpsd5F1im+TydgG/1rAfIvjl2ArT6n22pDstUUFBQUHhEqV7hKygoKCg8QjH4CgoKCi8IisFXUFBQeEFQDL6CgoLCC4Ji8BUUFBReEBSDr6CgoPCCoBh8BQUFhRcExeArKCgovCD8H38RA6pByDjYAAAAAElFTkSuQmCC\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"from scipy.optimize import minimize\n",
"\n",
"def nll_fn(X_train, Y_train, noise):\n",
" '''Returns a function that computes the negative marginal log-\n",
" likelihood for training data X_train and Y_train and given \n",
" noise level.\n",
" \n",
" Args:\n",
" X_train: training locations (M x D).\n",
" Y_train: training targets (M x 1).\n",
" noise: known noise level of Y_train. \n",
" \n",
" Returns:\n",
" Minimization objective.\n",
" '''\n",
" def step(theta):\n",
" K = kernel(X_train, X_train, l=theta[0], sigma_f=theta[1]) + \\\n",
" noise**2 * np.eye(len(X_train))\n",
" return np.sum(np.log(np.diag(np.linalg.cholesky(K)))) + \\\n",
" 0.5 * Y_train.T.dot(np.linalg.inv(K)).dot(Y_train) + \\\n",
" 0.5 * len(X_train) * np.log(np.array(2*np.pi, dtype='float64'))\n",
" return step\n",
"\n",
"def objective_and_grad(params):\n",
" \"\"\"Returns both value and gradient. Suitable for use\n",
" in scipy.optimize\"\"\"\n",
" theta = np.random.normal(size=params.size).require_grad()\n",
" fun = nll_fn(X_train, Y_train, noise)\n",
" out = fun(theta)\n",
" out.backward()\n",
" return chainerx.to_numpy(out), chainerx.to_numpy(theta.grad)\n",
"\n",
"init_params = 0.1*np.random.normal(size=2)\n",
"\n",
"# Minimize the negative log-likelihood w.r.t. parameters l and sigma_f.\n",
"res = minimize(objective_and_grad, init_params, method='L-BFGS-B', jac=True, options={'gtol': 1e-9, 'ftol': 0})\n",
"\n",
"# Store the optimization results in global variables so that we can\n",
"# compare it later with the results from other implementations.\n",
"l_opt, sigma_f_opt = res.x\n",
"\n",
"# Compute the prosterior predictive statistics with optimized kernel parameters and plot the results\n",
"mu_s, cov_s = posterior_predictive(X, X_train, Y_train, l=l_opt, sigma_f=sigma_f_opt, sigma_y=noise)\n",
"plot_gp(X, mu_s, cov_s)\n",
"plt.plot(X_train, Y_train, 'rx')"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
" fun: array([[66.3643472]])\n",
" hess_inv: <2x2 LbfgsInvHessProduct with dtype=float64>\n",
" jac: array([ 76.02868139, 160.05121638])\n",
" message: b'CONVERGENCE: REL_REDUCTION_OF_F_<=_FACTR*EPSMCH'\n",
" nfev: 8\n",
" nit: 3\n",
" status: 0\n",
" success: True\n",
" x: array([0.40666028, 0.90399203])"
]
},
"execution_count": 20,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"res"
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"<matplotlib.legend.Legend at 0x7f0471976400>"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 864x576 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.figure(figsize=(12, 8))\n",
"plt.plot(X_train, Y_train, 'rx')\n",
"plt.plot(X, np.sin(0.07*X**3), label='true')\n",
"plot_gp(X, mu_s, cov_s)\n",
"plt.legend()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## References\n",
"\n",
"\\[1\\] Carl Edward Rasmussen and Christopher K. I. Williams. [Gaussian Processes for Machine Learning](http://www.gaussianprocess.org/gpml/). \n",
"\\[1\\] Martin Krasser. [Gaussian Processes](http://krasserm.github.io/2018/03/19/gaussian-processes/). "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"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": "3.7.3"
}
},
"nbformat": 4,
"nbformat_minor": 2
}

Implementation of Gaussian Process Regression with ChainerX

The purpose of this notebook is to demonstrate the ChainerX linear albebra routines in the example of Gaussian Process Regression.

import chainerx
import chainerx as np
import matplotlib.pyplot as plt
%matplotlib inline
chainerx.set_default_device('cuda:0')

Unknown functions can be represented using Gaussian Processes (GPs). GPs can be used for regression and classification tasks. Here, we focus on a simple GP regression case. There is a whole book for details Gaussian Processes for Machine Learning.

We are interested in learning a function f(\mathbf{x}) from data \{ (x_i, y_i)\; |\; i=1,\ldots,n\}, where y_i\in\mathbf{R} is a noisy observation of f(x_i).

A Gaussian process is a random process where any point \mathbf{x} \in \mathbb{R}^d is assigned a random variable f(\mathbf{x}) and where the joint distribution of a finite number of these variables p(f(\mathbf{x}_1),...,f(\mathbf{x}_N)) is itself Gaussian:

p(\mathbf{f} | \mathbf{X}) = \mathcal{N}(\mathbf{f} | \boldsymbol\mu, \mathbf{K})\label{eq1}

Here, \mathbf{f} = (f(x_1),...,f(x_N)), \boldsymbol\mu = (m(x_1),...,m(x_N)) and K_{ij} = \kappa(x_i,x_j). m is the mean function and usually considered to be zero function. \kappa is a positive definite kernel function or covariance function. Thus, a Gaussian process is a distribution over functions whose shape (i.e. smoothness) is defined by \mathbf{K}. If points x_i and x_j are considered to be similar by the kernel the function values at these points, f(x_i) and f(x_j), can be expected to be similar too. The kernel function \kappa has free hyper-parameters \theta.

A GP prior p(\mathbf{f} | \mathbf{X}) can be converted into a GP posterior p(\mathbf{f} | \mathbf{X},\mathbf{y}) after having observed some data \mathbf{y}. The posterior can then be used to make predictions \mathbf{f}_* given new input \mathbf{X}_*:

p(\mathbf{f}_* | \mathbf{X}_*,\mathbf{X},\mathbf{y})
= \int{p(\mathbf{f}_* | \mathbf{X}_*,\mathbf{f})p(\mathbf{f} | \mathbf{X},\mathbf{y})}\ d\mathbf{f}
= \mathcal{N}(\mathbf{f}_* | \boldsymbol{\mu}_*, \boldsymbol{\Sigma}_*)

So the posterior predictive distribution is also a Gaussian with mean \boldsymbol{\mu}_* and \boldsymbol{\Sigma}_*. By definition of the GP, the joint distribution of observed data \mathbf{y} and predictions \mathbf{f}_* is

\begin{pmatrix}\mathbf{y} \\ \mathbf{f}_*\end{pmatrix} \sim \mathcal{N}
\left(\boldsymbol{0},
\begin{pmatrix}\mathbf{K}_y & \mathbf{K}_* \\ \mathbf{K}_*^T & \mathbf{K}_{**}\end{pmatrix}
\right)

With N training data and N_* new input data, \mathbf{K}_y = \kappa(\mathbf{X},\mathbf{X}) + \sigma_y^2\mathbf{I} = \mathbf{K} + \sigma_y^2\mathbf{I} is N \times N, \mathbf{K}_* = \kappa(\mathbf{X},\mathbf{X}_*) is N \times N_* and \mathbf{K}_{**} = \kappa(\mathbf{X}_*,\mathbf{X}_*) is N_* \times N_*. \sigma_y^2 is the noise term in the diagonal of \mathbf{K_y}. It is the noise of target data \mathbf{y}, so it is set to zero if training targets are noise-free and to a value greater than zero if observations are noisy. The mean is set to \boldsymbol{0}. The sufficient statistics of the posterior predictive distribution, \boldsymbol{\mu}_* and \boldsymbol{\Sigma}_*, can be computed with

\boldsymbol{\mu_*} = \mathbf{K}_*^T \mathbf{K}_y^{-1} \mathbf{y}
\boldsymbol{\Sigma_*} = \mathbf{K}_{**} - \mathbf{K}_*^T \mathbf{K}_y^{-1} \mathbf{K}_*

Given this model, there are two basic problems to be solved. First, we need to learn the hyper-parameters \theta and \sigma_y^2. Second, we need to predict mean and variance of f(\mathbf{x}_{*}) at test input points \mathbf{X}_{*}.

The squared exponential covariance function is given by

k(\mathbf{x}_i, \mathbf{x}_j) = \sigma_f^2 \exp\left(-\frac{\|\mathbf{x}_i - \mathbf{x}_j\|^2_ 2}{2\ell^2}\right)

where \sigma_f > 0 and \ell > 0 are hyperparameters of the kernel. There are many other kernels that can be used for Gaussian processes. This specific covariance function is perhaps the most common covariance function used in statistics and machine learning. It is also known as the radial basis function kernel, the gaussian kernel, or the exponeniated quadratic kernel.

def gaussian_rbf(x1, x2, l=1, sigma_f=1):
    """Returns the NxM kernel matrix between the two sets of input X1 and X2.

    Args:
        X1: NxD matrix
        X2: MxD matrix
        alpha: scalar parameter
        scale: scalar parameter

    Returns
        NxM kernel matrix
    """
    # distance between each rows
    dist_matrix = np.sum(np.square(x1), axis=1).reshape(-1, 1) + np.sum(np.square(x2), axis=1) - 2 * np.dot(x1, x2.T)
    return sigma_f**2 * np.exp(-1 / (2 * l**2) * dist_matrix)
kernel = gaussian_rbf

Let’s visualize the entries of the kernel matrix.

# create an Nx1 vector of equidistant points in [-3, 9]
N = 50
X = np.linspace(-3, 9, N).reshape(-1, 1)

cov = kernel(X, X)
plt.pcolor(chainerx.to_numpy(cov))
plt.colorbar()
<matplotlib.colorbar.Colorbar at 0x7f04781b4470>
output_9_1.png

Prior

Let’s generate samples from a Gaussian process prior. Specifically, we will consider a zero-mean Gaussian process prior for functions of the form f: \mathbb{R} \rightarrow \mathbb{R} using the squared exponential kernel. That is,

f(x) \sim \mathcal{GP}\left(0 \, , \, k\left(x, x'\right)\right).

Let f_n = f(x_n) \in \mathbb{R} be the value of the function f evaluated at a point x_n \in \mathbb{R}. Furthermore, let \mathbf{f} \in \mathbb{R}^{N} be the vector of function values for each of the points of \mathbf{X} .

The Gaussian process prior for \mathbf{f} becomes

\mathbf{f} \sim \mathcal{N}\left(\mathbf{0}, \mathbf{K}\right),

where \mathbf{K} is the kernel matrix.

To draw random functions from the GP we draw random samples from the corresponding multivariate normal. The common way of drawing such samples is using Cholesky decomposition, the algorithm is described in wikipedia.

def generate_samples(mean, K, S):
    """Returns M samples from a Gaussian process with kernel matrix K.

    Args:
        K: NxN kernel matrix
        S: number of samples

    Returns:
        SxS matrix of samples
    """

    z = np.random.normal(size = (mean.shape[0], S))
    L = np.linalg.cholesky(K + 1e-6*np.eye(mean.shape[0])) # added `1e-6*eye(N)` to covariance matrix to fix non positive definitness
    w = mean + L.dot(z)
    return w

Next, let’s define a function for plotting a Gaussian process with specified mean and 95.45% confidence interval (2\sigma).

def plot_gp(Xp, mu, Sigma, title="", num_samples=0):

    mean, std = mu.ravel(), np.sqrt(np.diag(Sigma))


    # plot distribution
    plt.plot(Xp, mean, color='k', label='Mean')
    plt.plot(Xp, mean + 2*std, color='k', linestyle='--')
    plt.plot(Xp, mean - 2*std, color='k', linestyle='--')
    # plt.fill_between does not work well with chainerx arrays so we are converting the input to numpy arrays
    plt.fill_between(*map(chainerx.to_numpy, (Xp.ravel(), mean - 2*std, mean + 2*std)), alpha=0.1)

    # generate samples
    if num_samples > 0:
        fs = generate_samples(mu, Sigma, num_samples)
        plt.plot(Xp, fs, color='b', alpha=.25)

    plt.title(title)
    plt.legend()
# Finite number of points
X = np.arange(-1, 7, 0.025).reshape(-1, 1).astype('float64')

# Mean and covariance of the prior
mu = np.zeros(X.shape, dtype='float64')
cov = kernel(X, X)

# Plot GP mean, confidence interval and samples
plot_gp(X, mu, cov, num_samples = 5)
output_14_0.png

Prediction from noise-free training data

Let’s implement computation of \boldsymbol{\mu_*} and \boldsymbol{\Sigma_*} using previously defined formulas.

def posterior_predictive(X, X_train, Y_train, l=1.0, sigma_f=1.0, sigma_y=1e-8):
    '''Computes the suffifient statistics of the GP posterior predictive distribution
    from m training data X_train and Y_train and n new inputs X.

    Args:
        X: New input locations (N x D).
        X_train: Training locations (M x D).
        Y_train: Training targets (M x 1).
        l: Kernel length parameter.
        sigma_f: Kernel vertical variation parameter.
        sigma_y: Noise parameter.

    Returns:
        Posterior mean vector (N x D) and covariance matrix (N x N).
    '''
    K = kernel(X_train, X_train, l, sigma_f) + sigma_y**2 * np.eye(len(X_train))
    K_s = kernel(X_train, X, l, sigma_f)
    K_ss = kernel(X, X, l, sigma_f)

    z = np.linalg.solve(K, K_s).T

    mu_s = z.dot(Y_train)
    cov_s = K_ss - z.dot(K_s)

    return mu_s, cov_s

and apply them to noise-free training data X_train and Y_train. The following example draws five samples from the posterior predictive and plots them along with the mean, confidence interval and training data. In a noise-free model, variance at the training points is zero and all random functions drawn from the posterior go through the trainig points.

# Noise free training data
X_train = np.linspace(0,6,5).reshape(-1,1)
Y_train = np.sin(0.07*X_train**3)

# Compute mean and covariance of the posterior predictive distribution
mu_s, cov_s = posterior_predictive(X, X_train, Y_train)

plot_gp(X, mu_s, cov_s, num_samples = 5)
plt.plot(X_train, Y_train, 'ro', label='Training data', markersize=5)
plt.legend()
<matplotlib.legend.Legend at 0x7f0478005518>
output_19_1.png

Prediction from noisy training data

Now let’s add some noise to the model, then training points are only approximated and the variance at the training points is non-zero.

# Noisy training data
noise = 0.3
X_train = np.linspace(0,6,100).reshape(-1,1)
Y_train = np.sin(0.07*X_train**3)  + noise * np.random.normal(size=X_train.shape)

# Compute mean and covariance of the posterior predictive distribution
mu_s, cov_s = posterior_predictive(X, X_train, Y_train, sigma_y=noise)

plot_gp(X, mu_s, cov_s, num_samples = 10)
plt.plot(X_train, Y_train, 'rx')
[<matplotlib.lines.Line2D at 0x7f047801ba58>]
output_21_1.png

Effect of kernel parameters and noise parameter

The following example shows the effect of kernel parameters l and \sigma_f as well as the noise parameter \sigma_y. Higher l values lead to smoother functions and therefore to coarser approximations of the training data. Lower l values make functions more wiggly with wide confidence intervals between training data points. \sigma_f controls the vertical variation of functions drawn from the GP. This can be seen by the wide confidence intervals outside the training data region in the right figure of the second row. \sigma_y represents the amount of noise in the training data. Higher \sigma_y values make more coarse approximations which avoids overfitting to noisy data.

params = [
    (0.3, 1.0, 0.2),
    (3.0, 1.0, 0.2),
    (1.0, 0.3, 0.2),
    (1.0, 3.0, 0.2),
    (1.0, 1.0, 0.05),
    (1.0, 1.0, 1.5),
]

plt.figure(figsize=(12, 5))

for i, (l, sigma_f, sigma_y) in enumerate(params):
    mu_s, cov_s = posterior_predictive(X, X_train, Y_train, l=l,
                                       sigma_f=sigma_f,
                                       sigma_y=sigma_y)
    plt.subplot(3, 2, i + 1)
    plt.subplots_adjust(top=2)
    plt.title(f'l = {l}, sigma_f = {sigma_f}, sigma_y = {sigma_y}')
    plot_gp(X, mu_s, cov_s)
    plt.plot(X_train, Y_train, 'rx')
    plt.legend()
output_23_0.png

We can define the learning criteria and estimate optimal values for these parameters. Optimal values can be estimated by maximizing the marginal log-likelihood which is given by

\log p(\mathbf{y} | \mathbf{X}) =
\log \mathcal{N}(\mathbf{y} | \boldsymbol{0},\mathbf{K}_y) =
-\frac{1}{2} \mathbf{y}^T \mathbf{K}_y^{-1} \mathbf{y}
-\frac{1}{2} \log \begin{vmatrix}\mathbf{K}_y\end{vmatrix}
-\frac{N}{2} \log(2\pi)

Instead of the maximization problem we turn it into the minimization problem. We will minimize the negative marginal log-likelihood w.r.t. parameters l and \sigma_f, \sigma_y is set to the known noise level of the data. If the noise level is unknown, \sigma_y can be estimated as well along with the other parameters. The minimization problem is solved using gradient-based algorithm.

We avoid computing the determinant of \begin{vmatrix}\mathbf{K}_y\end{vmatrix} by using the fact that the sum of the logarithms of the diagonal entries of the Cholesky decomposition of a matrix S equals half of the logarithm of determinant of matrix S. i.e. 0.5 * log(det(S)) and sum(log(diag(chol(S)))) are equal up to numerical precision.

from scipy.optimize import minimize

def nll_fn(X_train, Y_train, noise):
    '''Returns a function that computes the negative marginal log-
    likelihood for training data X_train and Y_train and given
    noise level.

    Args:
        X_train: training locations (M x D).
        Y_train: training targets (M x 1).
        noise: known noise level of Y_train.

    Returns:
        Minimization objective.
    '''
    def step(theta):
        K = kernel(X_train, X_train, l=theta[0], sigma_f=theta[1]) + \
            noise**2 * np.eye(len(X_train))
        return np.sum(np.log(np.diag(np.linalg.cholesky(K)))) + \
               0.5 * Y_train.T.dot(np.linalg.inv(K)).dot(Y_train) + \
               0.5 * len(X_train) * np.log(np.array(2*np.pi, dtype='float64'))
    return step

def objective_and_grad(params):
    """Returns both value and gradient. Suitable for use
    in scipy.optimize"""
    theta = np.random.normal(size=params.size).require_grad()
    fun = nll_fn(X_train, Y_train, noise)
    out = fun(theta)
    out.backward()
    return chainerx.to_numpy(out), chainerx.to_numpy(theta.grad)

init_params = 0.1*np.random.normal(size=2)

# Minimize the negative log-likelihood w.r.t. parameters l and sigma_f.
res = minimize(objective_and_grad, init_params, method='L-BFGS-B', jac=True, options={'gtol': 1e-9, 'ftol': 0})

# Store the optimization results in global variables so that we can
# compare it later with the results from other implementations.
l_opt, sigma_f_opt = res.x

# Compute the prosterior predictive statistics with optimized kernel parameters and plot the results
mu_s, cov_s = posterior_predictive(X, X_train, Y_train, l=l_opt, sigma_f=sigma_f_opt, sigma_y=noise)
plot_gp(X, mu_s, cov_s)
plt.plot(X_train, Y_train, 'rx')
[<matplotlib.lines.Line2D at 0x7f04716afbe0>]
output_26_1.png
res
     fun: array([[66.3643472]])
hess_inv: <2x2 LbfgsInvHessProduct with dtype=float64>
     jac: array([ 76.02868139, 160.05121638])
 message: b'CONVERGENCE: REL_REDUCTION_OF_F_<=_FACTR*EPSMCH'
    nfev: 8
     nit: 3
  status: 0
 success: True
       x: array([0.40666028, 0.90399203])
plt.figure(figsize=(12, 8))
plt.plot(X_train, Y_train, 'rx')
plt.plot(X, np.sin(0.07*X**3), label='true')
plot_gp(X, mu_s, cov_s)
plt.legend()
<matplotlib.legend.Legend at 0x7f0471976400>
output_28_1.png

References

[1] Carl Edward Rasmussen and Christopher K. I. Williams. Gaussian Processes for Machine Learning.
[1] Martin Krasser. Gaussian Processes.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment