{"cells":[{"metadata":{},"cell_type":"markdown","source":"# CNN Architecture\n\n![image.png](attachment:image.png)\n\nIn this notebook, we introduce various components of convolutional neural networks and how they can be used to assemble effective CNN classifiers. For more information about training CNN models, please see the following notebooks where we use ensemble CNN models to classify [digits](https://nbviewer.jupyter.org/github/cschupbach/deep_learning_cnn/blob/master/digits_ensemble.ipynb) and [fashion](https://nbviewer.jupyter.org/github/cschupbach/deep_learning_cnn/blob/master/fashion_ensemble.ipynb) items, achieving testing accuracies of 99.67% and 93.55%, respectively. Note that these examples utilize the MNIST datasets of the Keras distribution, where the `digits` dataset is composed of 60000 training samples and 10000 testing samples.","attachments":{"image.png":{"image/png":"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"}},"execution_count":null},{"metadata":{},"cell_type":"markdown","source":"**Import modules**","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"import numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport matplotlib.gridspec as pltgs\nimport torch\nimport keras\nfrom keras.layers import Conv2D\nfrom keras.layers import MaxPooling2D, AveragePooling2D\nfrom keras.layers import Dense, Flatten\nfrom keras.models import Sequential\nfrom keras import backend as K\nfrom keras.utils import to_categorical","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Create class `Plot()` for notebook network plots:","execution_count":null},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"class Plot():\n    def __init__(self, width=10, height=4, cmap='magma', size=12, bot=0.1):\n        self.width = width\n        self.height = height\n        self.cmap = cmap\n        self.size = size\n        self.fig = plt.figure(figsize=(self.width, self.height))\n        self.bot = bot\n\n    def _network_dict(self, network_id):\n        c = 'Convolution\\n+ '\n        dictionary = {\n            0: [c + 'ReLU', r'Max Pooling $(2 \\times 2)$'],\n            1: [c + 'Tanh', r'Max Pooling $(2 \\times 2)$'],\n            2: [c + 'ReLU', r'Average Pooling $(2 \\times 2)$'],\n            3: [c + 'Tanh', r'Average Pooling $(2 \\times 2)$'],\n            4: [c + 'Tanh', r'Max Pooling $(2 \\times 2)$', c + 'ReLU'],\n            5: [c + 'ReLU', r'Max Pooling $(2 \\times 2)$', c + 'ReLU',\n                c + 'ReLU', r'Max Pooling $(2 \\times 2)$']\n        }\n        return dictionary[network_id]\n\n    def _no_ticks(self, ax):\n        ax.set_xticks([])\n        ax.set_yticks([])\n        return None\n\n    def _get_lims(self, ax):\n        return np.array(ax.get_xlim()), np.array(ax.get_ylim())\n\n    def _shape_label(self, x, xlim, ylim):\n        _shape = tuple([x.shape[i] for i in [1, 2, 0]])\n        plt.text(np.mean(xlim), np.max(ylim)*1.05, _shape,\n                 ha='center', va='top', size=self.size)\n        return None\n\n    def _level_locs(self, levels):\n        x1 = [1 / (levels * 4)]\n        x1 += [(i + 0.8) / levels for i in range(1, levels-1)]\n        x2 = [(i + 0.7) / levels for i in range(1, levels-1)]\n        x2 += [(levels - 0.25) / levels]\n        return np.array(x1), np.array(x2)\n\n\n    def _network_desc(self, levels, network_id):\n        x1, x2 = self._level_locs(levels)\n        labels = self._network_dict(network_id)\n        gs = pltgs.GridSpec(1, 1, left=0, right=1, top=self.bot, bottom=0)\n        ax = self.fig.add_subplot(gs[0,0])\n        for i in range(len(x1)):\n            ax.plot([x1[i], x1[i]], [0.8, 0.7], c='k')\n            ax.plot([x1[i], x2[i]], [0.7, 0.7], c='k')\n            ax.plot([x2[i], x2[i]], [0.8, 0.7], c='k')\n            ax.text((x1[i] + x2[i]) / 2, 0.55, labels[i],\n                     ha='center', va='top', size=self.size)\n        ax.set_ylim(0, 1)\n        ax.set_xlim(0, 1)\n        ax.axis('off')\n        return None\n\n    def _conv_desc(self, levels, activation):\n        x1, x2 = self._level_locs(levels)\n        label = 'Convolution'\n        if len(activation) > 0:\n            label += f'\\n+ {activation}'\n        gs = pltgs.GridSpec(1, 1, left=0, right=1, top=1, bottom=0)\n        ax = self.fig.add_subplot(gs[0,0])\n        for i in range(len(x1)):\n            ax.plot([0.27, 0.45], [0.50, 0.50], c='k')\n            ax.plot([0.45, 0.44], [0.50, 0.52], c='k')\n            ax.plot([0.45, 0.44], [0.50, 0.48], c='k')\n            ax.text((0.27 + 0.45) / 2, 0.47, label, ha='center', va='top',\n                    size=self.size)\n        ax.set_ylim(0, 1)\n        ax.set_xlim(0, 1)\n        ax.axis('off')\n        return None\n\n    def _plot_input(self, x_input, levels):\n        gs = pltgs.GridSpec(1, 2, left=0, right=1/levels, top=1,\n                            bottom=self.bot)\n        ax = self.fig.add_subplot(gs[0,0])\n        ax.imshow(x_input[0,0], cmap=self.cmap, aspect='equal')\n        xlim, ylim = self._get_lims(ax)\n        plt.text(np.mean(xlim), -np.max(ylim)*0.05, 'Input',\n                 ha='center', va='bottom', size=self.size)\n        self._shape_label(x_input[0], xlim, ylim)\n        self._no_ticks(ax)\n        return None\n\n    def _network_layout(self, x_input, x_list):\n        input_size = x_input[0,0].shape[0]\n        xrng = np.arange(len(x_list))\n        x = [x_list[i][0,:,:,:] for i in xrng]\n        layers = [x[i].shape[0] for i in xrng]\n        size_ratio = [x[i].shape[1] / input_size for i in xrng]\n        ws = [(1 / ((layers[i] - 1) * size_ratio[i])) - 1 for i in xrng]\n        hs = [(1 / ((layers[i] - 1) * size_ratio[i])) - 1 for i in xrng]\n        return x, xrng, layers, ws, hs\n\n    def _plot_network(self, x_input, x_list, levels):\n        x, xrng, layers, ws, hs = self._network_layout(x_input, x_list)\n        for i in xrng:\n            gs = pltgs.GridSpec(layers[i], layers[i], left=(i+1)/levels,\n                                right=(i+2)/levels, top=1, bottom=self.bot,\n                                wspace=ws[i], hspace=hs[i])\n            for j in range(layers[i]):\n                ax = self.fig.add_subplot(gs[j,j])\n                ax.imshow(x[i][j], cmap=self.cmap, aspect='equal')\n                self._no_ticks(ax)\n            xlim, ylim = self._get_lims(ax)\n            self._shape_label(x[i], xlim, ylim)\n        return None\n\n    def network(self, x_input, x_list, activation='', network_id=0,\n                channels='first'):\n        if channels == 'last':\n            x_input = np.transpose(x_input, [0, 3, 1, 2])\n            x_list = [np.transpose(x, [0, 3, 1, 2]) for x in x_list]\n        levels = len(x_list) + 1\n        if levels > 2:\n            self._network_desc(levels, network_id)\n        else:\n            self.bot = 0\n            self._conv_desc(levels, activation)\n        self._plot_input(x_input, levels)\n        self._plot_network(x_input, x_list, levels)\n        return None","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Define additional helper/plotting functions:","execution_count":null},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"def normalize(x):\n    return x.astype('float32') / 255\n\n\ndef plot_samples(x_train, y_train):\n    class_dict = np.arange(10)\n    sample_list = [x_train[y_train[:,i].astype(bool)][:12] for i in range(10)]\n    samples = np.concatenate(sample_list)\n    gs = pltgs.GridSpec(10, 12, hspace=-0.025, wspace=-0.025)\n    fig = plt.figure(figsize=(10, 8.5))\n    yloc = np.linspace(0.95, 0.05, 10)\n    for i in range(120):\n        ax = fig.add_subplot(gs[i//12, i%12])\n        ax.imshow(samples[i,:,:,0], cmap='magma')\n        ax.set_xticks([])\n        ax.set_yticks([])\n    return None\n\n\ndef display_conv(x, conv, activation):\n    x1 = K.eval(conv(x))\n    Plot().network(K.eval(x), [x1], activation=activation, channels='last')\n    return None\n\n\ndef display_conv_pool(x, conv_list, network_id, height=5):\n    x_list = [conv_list[0](x)]\n    for i in range(1, len(conv_list)):\n        x_list += [conv_list[i](x_list[i-1])]\n    x_list = [K.eval(x_list[i]) for i in range(len(conv_list))]\n    Plot(height=height).network(K.eval(x), x_list, network_id=network_id,\n                                channels='last')\n    return None\n\n\ndef display_fc(model, x):\n    label = model.get_layer(index=-1).name\n    fc = K.eval(model(x))\n    plt.figure(figsize=(10, 0.5))\n    plt.imshow(fc, cmap='magma', aspect='auto')\n    plt.gca().set_xticks([])\n    plt.gca().set_yticks([])\n    xloc = np.array(plt.gca().get_xlim())\n    plt.text(xloc[0], -0.6, label, ha='left', va='bottom', size=12)\n    plt.text(np.mean(xloc), 0.7, fc.shape, ha='center', va='top', size=12)\n    return None","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"**Read in data**:","execution_count":null},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"train = pd.read_csv('../input/digit-recognizer/train.csv').to_numpy()\ntest = pd.read_csv('../input/digit-recognizer/test.csv').to_numpy()\n\nx_train = normalize(train[:,1:]).reshape(-1, 28, 28, 1)\nx_test = normalize(test).reshape(-1, 28, 28, 1)\ny_train = to_categorical(train[:,0])\n\nprint(f'  Train data   shape = {x_train.shape}')\nprint(f'   Test data   shape = {x_test.shape}')\nprint(f'Train labels   shape = {y_train.shape}')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"The first 12 samples from each class are shown below.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"plot_samples(x_train, y_train)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Feature Learning\n\nConvolutional neural networks (CNN) classify high-dimensional input data by extracting feature maps in an iterative process known as feature learning.\n\n$$\\mathsf{input} \\ \\ \\longrightarrow \\ \\ \\mathsf{feature \\ learning} \\ \\ \\longrightarrow \\ \\ \\mathsf{classification} \\\\ $$\n\nThe feature learning segment of the network is a set of feature maps or layers primarily constructed using convolutional kernels convolved with the input layers in a sequence.\n\n$$\\mathsf{input} \\ \\ \\circledast \\ \\ \\text{conv}_{1} \\ \\ \\circledast \\ \\cdots \\ \\circledast \\ \\text{conv}_{N} \\ \\ \\longrightarrow \\ \\mathsf{classification}$$","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"## Convolutional Layers\n### Input/Output\n\nHere we discuss the role of convolutional kernels/filters in CNN architecture. Given a convolutional kernel function $\\Omega$, the kernel of $\\omega$ is convolved with an input layer $x_\\text{in}$, such that,\n\n$$x_\\text{out} = \\Omega(x_\\text{in})\\tag{1} \\\\ $$\n\nwhere $x_\\text{out}$ is the output layer. The shape of the output layer $x_\\text{out}$ is dependent on the shape of the input layer $x_\\text{in} \\in \\mathbb{R}^{H_\\text{in} \\times W_\\text{in} \\times D_\\text{in}}$ and the hyperparameters of the kernel function $\\Omega$: the number of kernels $K$, kernel size $F$, the stride $S$, and the amount of padding $P$ used on the input border.\n\n$$x_\\text{in} \\in \\mathbb{R}^{H_\\text{in} \\times W_\\text{in} \\times D_\\text{in}} \\quad\\overset{\\Omega}{\\longrightarrow}\\quad x_\\text{out} \\in \\mathbb{R}^{H_\\text{out} \\times W_\\text{out} \\times D_\\text{out}}$$\n\nwhere\n\n$$H_\\text{out} = \\frac{H_\\text{in} - F + 2P}{S} + 1\\tag{2} \\\\ $$\n\n$$W_\\text{out} = \\frac{W_\\text{in} - F + 2P}{S} + 1\\tag{3} \\\\ $$\n\n$$D_\\text{out} = K\\tag{4}$$\n\n\n#### Example 1\n\nLet the input layer $x_\\text{in}$ be a tensor of size ${5 \\times 5 \\times 1}$ and the hyperparameters of $\\Omega$ be $K=2$, $F=3$, $S=1$ and $P=0$, such that,\n\n$$H_\\text{out} = \\frac{H_\\text{in} - F + 2P}{S} + 1 = \\frac{5 - 3 + 2(0)}{1} + 1 = 3 \\\\ $$\n\n$$W_\\text{out} = \\frac{W_\\text{in} - F + 2P}{S} + 1 = \\frac{5 - 3 + 2(0)}{1} + 1 = 3 \\\\ $$\n\n$$D_\\text{out} = K = 2 \\\\ $$\n\nTherefore, the output $x_\\text{out} \\in \\mathbb{R}^{3 \\times 3 \\times 2}$ is a feature map with 2 layers and each layer is a $3 \\times 3$ matrix. For additional information about specific kernel functions and convolution, feel free to check out my GitHub [image processing](https://github.com/cschupbach/image_processing) repository.","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"### PyTorch\nExample 2 briefly introduces the [torch.nn](https://pytorch.org/docs/stable/nn.html) module as an introduction to the PyTorch library. While the remainder of the notebook uses the Keras library, both PyTorch and Keras offer similar tools for constructing convolutional neural networks.\n\nNote that PyTorch documentation specifies use of an input tensor in *channels first* format. This means the default input tensor for 2-dimensional convolution is expected to have the shape $N \\times C_\\text{in} \\times H_\\text{in} \\times W_\\text{in}$, where $N$ is the batch size, $C_\\text{in}$ is the number of input channels/layers, $H_\\text{in}$ is the input height or the number of rows, and $W_\\text{in}$ is the width or the number of columns. Alternatively, Keras 2-dimensional convolution expects an input tensor of shape $N \\times H_\\text{in} \\times W_\\text{in} \\times C_\\text{in}$, which is known as *channels last* format.","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"#### Example 2\n\nLet's use the 40th training image as our sample input `x`.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"x = x_train[39:40,:,:,:1]","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Currently `x` is a NumPy array of shape $1 \\times 28 \\times 28 \\times 1$. For convolution using PyTorch, we need `x` to be a PyTorch tensor in *channels first* form (currently in *channels last*). We solve this requirement below.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"x = np.transpose(x, [0, 3, 1, 2])\nx = torch.Tensor(x)\nprint(f'x = {x.shape}')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"Below, we use the `torch.nn` function `Conv2d` to create two 2-dimensional convolution objects, `conv1` and `conv2`. The first convolution object `conv1` uses 3 kernel filters ($K=3$) and no zero padding ($P=0$) of the input border, while `conv2` uses 6 kernel filters ($K=6$) and adds one pixel of zero padding ($P=1$) to the input border. Both convolution objects have kernel size $F=3$, stride $S=1$, and one input channel, given that the input tensor is of shape/size $1 \\times 1 \\times 28 \\times 28$.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"conv1 = torch.nn.Conv2d(in_channels=1, out_channels=3, kernel_size=3,\n                        stride=1, padding=0)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"conv2 = torch.nn.Conv2d(in_channels=1, out_channels=6, kernel_size=3,\n                        stride=1, padding=1)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"The kernels of each convolution object convolve the input `x` as follows:","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"x1 = conv1(x).detach().numpy()\nx2 = conv2(x).detach().numpy()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"We apply the `network` function of object `Plot` to display the results of each convolution as follows:","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"Plot().network(x, [x1], channels='first')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"Plot().network(x, [x2], channels='first')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Activation Functions","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"Most CNN architectures also apply an activation function to convolutional layers. We use activation functions to increase non-linearity in the network. In theory, the output layers of convolution with activation have a thresholding effect similar to that of action potentials in the biological firing of neurons.\n\n$$\\text{input} \\ \\ \\circledast \\ \\ \\text{conv}_{1} \\ + \\ \\text{activation}_1 \\ \\circledast \\ \\cdots \\ \\circledast \\ \\text{conv}_{N} \\ + \\ \\text{activation}_N \\longrightarrow \\ \\text{classification}$$","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"### Rectified Linear Unit  (ReLU)\n$$f(x) = \\begin{cases} 0 & \\text{for } \\ x \\leq 0 \\\\\nx & \\text{for } \\ x \\gt 0 \\end{cases}$$","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"Here we provide an example of convolution with ReLU activation using Keras. Because the Keras library uses *channels last* format (by default) for 2-dimensional convolution, we revert tensor `x` back into an array, transform to *channels last* format, and express the array as a Keras tensor. ","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"x = x.detach().numpy()\nx = np.transpose(x, [0, 2, 3, 1])\nx = K.constant(x)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"We create two 2-dimensional convolution objects with ReLU activation, `relu1` and `relu2`, using the Keras function `Conv2D` below. The first convolution object `relu1` uses 3 kernel filters ($K=3$) and no zero padding ($P=0$) of the input border, while `relu2` uses 6 kernel filters ($K=6$) and adds one pixel of zero padding ($P=1$) to the input border. Both convolution objects have kernel size $F=3$ and stride $S=1$ (default). Note that the zero padding parameter is either `'valid'` (default), where $H_\\text{out}$ and $W_\\text{out}$ are according to equations (2) and (3) with $P=0$, or `'same'`, where $H_\\text{out} = H_\\text{in}$ and $W_\\text{out}=W_\\text{in}$.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"relu1 = Conv2D(filters=3, kernel_size=3, strides=1, padding='valid',\n               activation='relu')\nrelu2 = Conv2D(filters=6, kernel_size=3, strides=1, padding='same',\n               activation='relu')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"We use the `display_conv` function to display the resulting ReLU-activated convolutional layers as follows:","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, relu1, activation='ReLU')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, relu2, activation='ReLU')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Softmax\n$$P(y=j| \\mathbf{x}) = \\dfrac{e^{\\mathbf{x}^\\mathsf{T}\\mathbf{w}_j}}{\\sum_{k = 1}^{k} e^{\\mathbf{x}^\\mathsf{T}\\mathbf{w}_k}}$$","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"softmax = Conv2D(filters=3, kernel_size=3, activation='softmax')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, softmax, activation='Softmax')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Sigmoid\n$$f(x) = \\sigma(x) = \\dfrac{1}{1 + e^{-x}}$$","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"sigmoid = Conv2D(filters=3, kernel_size=3, activation='sigmoid')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, sigmoid, activation='Sigmoid')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Hyperbolic Tangent (tanh)\n$$f(x) = \\tanh(x) = \\dfrac{e^x - e^{-x}}{e^x + e^{-x}}$$","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"tanh = Conv2D(filters=3, kernel_size=3, activation='tanh')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, tanh, activation='Tanh')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Exponential Linear Unit (ELU)\n$$f(\\alpha, x) = \\begin{cases} \\alpha (e^x - 1) & \\text{for } \\ x \\leq 0 \\\\\nx & \\text{for } \\ x \\gt 0 \\end{cases}$$","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"elu = Conv2D(filters=3, kernel_size=3, activation='elu')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, elu, activation='ELU')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Softplus\n$$f(x) = \\ln(1 + e^x)$$","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"softplus = Conv2D(filters=3, kernel_size=3, activation='softplus')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, softplus, activation='Softplus')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Softsign\n$$f(x) = \\dfrac{x}{1 + |x|}$$","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"softsign = Conv2D(filters=3, kernel_size=3, activation='softsign')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv(x, softsign, activation='Softsign')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Others\n\n- Leaky ReLU\n\n- PReLU\n\n- RReLU\n\n- SELU\n\n- GELU","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"## Pooling\n\nPooling methods are often used between certain convolutions for dimensionality reduction. The most common method for pooling input layers is *max* pooling using a $2 \\times 2$ kernel with non-overlapping strides. In certain cases with very large input, pooling with a $3 \\times 3$ kernel may be effective. Typically, *max* pooling produces better results than the less common *average* pooling.\n\n### Max Pooling","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"relu = Conv2D(filters=6, kernel_size=3, padding='same', activation='relu')\ntanh = Conv2D(filters=6, kernel_size=3, padding='same', activation='tanh')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"maxpool = MaxPooling2D(pool_size=(2, 2))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"We use the `display_conv_pool` function to display the resulting layers as follows:","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv_pool(x, [relu, maxpool], network_id=0)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv_pool(x, [tanh, maxpool], network_id=1)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Average Pooling","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"avgpool = AveragePooling2D(pool_size=(2, 2))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv_pool(x, [relu, avgpool], network_id=2)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv_pool(x, [tanh, avgpool], network_id=3)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Stacking Layers","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"relu = Conv2D(filters=6, kernel_size=3, padding='same', activation='relu')\ntanh = Conv2D(filters=6, kernel_size=3, padding='same', activation='tanh')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv_pool(x, [tanh, maxpool, relu], network_id=4)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"maxpool = MaxPooling2D(pool_size=(2, 2))\nrelu1 = Conv2D(filters=8, kernel_size=3, padding='same', activation='relu')\nrelu2 = Conv2D(filters=16, kernel_size=3, padding='same', activation='relu')\nrelu3 = Conv2D(filters=16, kernel_size=3, activation='relu')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"display_conv_pool(x, [relu1, maxpool, relu2, relu3, maxpool],\n                  network_id=5)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"A common CNN feature learning architecture is similar to that shown above, where feature maps are constructed with ReLU-activated convolutional layers and intermittent max pooling. It's also fairly common to normalize the batch of input layers prior to convolution in the feature learning segment.","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"# Classification\n\nIn the classification segment of the network, we flatten the final feature learning layer to form fully connected (FC) layers.\n\n## Fully Connected (FC) Layers\n\nHere, we introduce FC layers by building a network of feature learning layers using Keras [Sequential](https://keras.io/api/models/sequential/) class. We initialize a sequential object and add feature learning layers to the model as follows:","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"model = Sequential()\nmodel.add(Conv2D(8, 3, padding='same', activation='relu',\n                 input_shape=(28, 28, 1), name='conv_1'))\nmodel.add(Conv2D(8, 3, padding='same', activation='relu', name='conv_2'))\nmodel.add(MaxPooling2D(pool_size=(2, 2), name='maxpool_1'))\nmodel.add(Conv2D(16, 3, activation='relu', name='conv_3'))\nmodel.add(Conv2D(16, 3, activation='relu', name='conv_4'))\nmodel.add(Conv2D(16, 3, activation='relu', name='conv_5'))\nmodel.add(MaxPooling2D(pool_size=(2, 2), name='maxpool_2'))","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"![image.png](attachment:image.png)\n\n**Figure 1.** Example of fully connected layers where the flattened input layer is fed into a sequence of dense, fully connected layers. Neurons of each fully connected layer are linked to all activations in the prior layer.","attachments":{"image.png":{"image/png":"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"}},"execution_count":null},{"metadata":{},"cell_type":"markdown","source":"### Flatten\nThe output layers of the feature learning network above has the shape $4 \\times 4 \\times 16$. To begin the classification segment of the CNN, we flatten the output layers forming input array of length 256 similar to that of the simplified neural network shown in Figure 1 [[1](https://www.researchgate.net/publication/331525817_Temporal_Convolutional_Neural_Network_for_the_Classification_of_Satellite_Image_Time_Series)] and displayed the fully connected layer using the `display_fc` function.","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"model.add(Flatten(name='flatten_1'))\ndisplay_fc(model, x)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Dense\nWe sequentially add to the model two ReLU-activated fully connected layers of size 128 and a Softmax-activated output layer with a length equal to the number of classes (in our case 10).","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"model.add(Dense(128, activation='relu', name='dense_1'))\ndisplay_fc(model, x)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model.add(Dense(128, activation='relu', name='dense_2'))\ndisplay_fc(model, x)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model.add(Dense(10, activation='softmax', name='dense_3'))\ndisplay_fc(model, x)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Network Summary","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"The network can be summarized as follows:","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"model.summary()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Additional Layers and Functions\n\nMany additional layers and layer functions may be added to fine tune the base architecture of the CNN shown above based on the data being classified. While a few common additions are discussed below, we highly suggest reviewing documentation for the [Keras Layers API](https://keras.io/api/layers/).\n\n## Batch Normalization\nWe use batch normalization to scale values of inputs to mean zero and unit variance. This helps account for a problem called covariate shift, where the distribution of data varies across training and testing batch inputs. For some datasets, batch normalization may increase efficiency by reducing the number of epochs required to effectively train the CNN. However, batch normalization may not be suitable in some instances due to an increase in the amount of time required to complete an epoch. Results typically depend on the data and where batch normalizations are added in the model. For a datasets like `digits`, where the distribution of pixel values of is likely similar across large batches of input data, batch normalization may not be worth the increase in computational complexity.\n\n## Dropout Layers\nA dropout layer is a regularization method where a proportion of input values are randomly set to zero. Doing this reduces the amount of overfitting while training and results in a fit more robust to potential outliers. We typically apply dropout layers to inputs of the fully connected layers during the classification segment of the network. \n\n## L1/L2 Regularization\nRegularization techniques like L1 and L2 regularization can be applied to layers of the network to reduce/prevent overfitting. The regularizers are added as penalty terms to the weight parameters of the loss function to reduce model complexity.","execution_count":null},{"metadata":{},"cell_type":"markdown","source":"# Citations\n\n1. C. Pelletier, G. Webb, and F. Petitjean, \"Temporal Convolutional Neural Network for the Classification of Satellite Image Time Series,\" *Remote Sensing*, vol. 11, no. 5, p. 523, 2019.","execution_count":null}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}