{"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_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"![clusters of digits - small.png](attachment:d03637b1-b9b3-49b1-b097-db0c83c975a7.png)\n\nsource: [Hackernoon](https://hackernoon.com/latent-space-visualization-deep-learning-bits-2-bd09a46920df)","metadata":{},"attachments":{"d03637b1-b9b3-49b1-b097-db0c83c975a7.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"# 1. Introduction","metadata":{}},{"cell_type":"markdown","source":"This notebook was inspired by my latest readings on autoencoders and variational autoencoders (VAE). All the references can be found throughout this notebook and are aggregated in the last section.\n\nIn one sentence, an autoencoder is a type of neural networks that is trained to copy its input to its output. So if you are wondering \"why on earth a neural network that acts as an identity function could be of any interest?\", I hope that you will learn something interesting in this notebook. ","metadata":{}},{"cell_type":"code","source":"# Import libraries\nimport os, warnings\nimport numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nimport tensorflow as tf\nfrom tensorflow import keras\nfrom keras import layers\nfrom keras import backend as K\nfrom keras.models import load_model\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn.decomposition import PCA\n\n# Reproducability\n# https://keras.io/getting_started/faq/#how-can-i-obtain-reproducible-results-using-keras-during-development.\ndef set_seed(seed=42):\n    np.random.seed(seed)\n    tf.random.set_seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    os.environ['TF_DETERMINISTIC_OPS'] = '1'\nset_seed()\n\n# Hide info, warnings and errors\nwarnings.filterwarnings('ignore')\nos.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'  # filters out all TF messages","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_kg_hide-input":true,"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:37:44.06406Z","iopub.execute_input":"2022-01-28T17:37:44.064662Z","iopub.status.idle":"2022-01-28T17:37:50.614721Z","shell.execute_reply.started":"2022-01-28T17:37:44.064517Z","shell.execute_reply":"2022-01-28T17:37:50.613987Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Autoencoders can be applied to any type of input data. In this notebook, I'll use an image dataset, the famous MNIST dataset. So let's start loading the data and normalizing all image pixels between 0 and 1:","metadata":{}},{"cell_type":"code","source":"x_train = pd.read_csv('/kaggle/input/digit-recognizer/train.csv')\nx_test = pd.read_csv('/kaggle/input/digit-recognizer/test.csv')\ny_train = x_train.pop('label')\nx_train = x_train.astype('float32') / 255.\nx_test = x_test.astype('float32') / 255.","metadata":{"execution":{"iopub.status.busy":"2022-01-28T17:37:50.616206Z","iopub.execute_input":"2022-01-28T17:37:50.616523Z","iopub.status.idle":"2022-01-28T17:37:55.518225Z","shell.execute_reply.started":"2022-01-28T17:37:50.616492Z","shell.execute_reply":"2022-01-28T17:37:55.517526Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 2. Autoencoder","metadata":{}},{"cell_type":"markdown","source":"An autoencoder is a special type of neural network that is trained to copy its input to its output. For example, given an image of a handwritten digit, an autoencoder first encodes the image into a lower dimensional latent representation, then decodes the latent representation back to an image. An autoencoder learns to compress the data while minimizing the reconstruction error.","metadata":{}},{"cell_type":"markdown","source":"![autoencoder_small.png](attachment:163d9f12-deec-4a2d-b9b6-e37a3a0a78f4.png)","metadata":{},"attachments":{"163d9f12-deec-4a2d-b9b6-e37a3a0a78f4.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"The loss function in this case is \"a distance\" between input and output and is called a reconstruction loss. E.g. when the input is an image, a pixel-by-pixel comparison using RMSE or binary entropy can be used.  ","metadata":{}},{"cell_type":"markdown","source":"Below, the most simple possible autoencoder where the encoder and the decoder are composed of a single Dense layer. Let's train it and then use it to reconstruct the images of the test dataset:","metadata":{}},{"cell_type":"code","source":"# MNIST images are 28x28X1 represented as a 784 dimensional vector\ninputs = keras.Input(shape=(784,)) \n\n# Create encoder, the latent space has 32-dimensions\nencoded = layers.Dense(32, activation='relu')(inputs) \nencoder = keras.Model(inputs, encoded)\n\n# Create decoder\nlatent_inputs = keras.Input(shape=(32,))\noutputs = layers.Dense(784, activation='relu')(latent_inputs)\ndecoder = keras.Model(latent_inputs, outputs)\n\n# Instantiate our autoencoder\noutputs = decoder(encoder(inputs))\nautoencoder = keras.Model(inputs, outputs, name='simplest_autoencoder')\n\n# Compile our autoencoder with Adam optimizer and a pixel-by-pixel loss (using binary crossentropy)\nautoencoder.compile(optimizer='adam', loss='binary_crossentropy')\n\n# Train our autoencoder. We can use the images from the test set to evaluate the model loss on unseen data\nautoencoder.fit(x=x_train, y=x_train, epochs=100, batch_size=256, validation_data=(x_test, x_test))","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:37:55.519294Z","iopub.execute_input":"2022-01-28T17:37:55.519544Z","iopub.status.idle":"2022-01-28T17:40:21.857296Z","shell.execute_reply.started":"2022-01-28T17:37:55.519506Z","shell.execute_reply":"2022-01-28T17:40:21.85669Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Use the trained autoencoder to reconstruct the images from the test dataset\nx_test_reconstructed = autoencoder.predict(x_test)\n\n# Pick randomly and display 10 images and its recosntructed version \npos = 1\nfig=plt.figure(figsize=(20,5))\nfor i in (np.random.randint(0, x_test.shape[0], size=10)):\n    fig.add_subplot(2, 10, pos)\n    plt.imshow(np.array(x_test.iloc[i]).reshape(28,28))\n    plt.axis('off')\n    fig.add_subplot(2, 10, pos+10)\n    plt.imshow(x_test_reconstructed[i].reshape(28,28))\n    plt.axis('off')\n    pos = pos + 1\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-01-28T17:40:21.859476Z","iopub.execute_input":"2022-01-28T17:40:21.860192Z","iopub.status.idle":"2022-01-28T17:40:23.93299Z","shell.execute_reply.started":"2022-01-28T17:40:21.860159Z","shell.execute_reply":"2022-01-28T17:40:23.932205Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Autoencoders are lossy which means that the decompressed outputs will be degraded compared to the original inputs. The complexity of your model and the size of the latent reprensentation will allow you to manage the loss / complexity trade-off. Here with a very simple model and a 32-dimensional latent representation, this is not so bad at all!  ","metadata":{}},{"cell_type":"markdown","source":"**How does an autoencoder work?**","metadata":{}},{"cell_type":"markdown","source":"The answer is simple : it creates clusters in the latent space of inputs that are structurally similar. For instance, with the MNIST dataset, we can expect :\n* to find one cluster for the hand-written 1, another cluster for the hand-written 2, ... \n* it is likely that the 1's cluster is closer to the 7's cluster than the 5's cluster because a 1 look more like a 7 than a 5...  ","metadata":{}},{"cell_type":"markdown","source":"It's not straightforward to illustrate with the example above because the latent representation is 32-dimensional. Let's use a dimension reduction technique e.g. PCA to display the latent representation in a 2-d space. ","metadata":{}},{"cell_type":"code","source":"# Use our trained encoder to create the points in the latent space\nx_train_encoded = encoder.predict(x_train)\n\n# Scale the inputs before applying PCA\nx_train_enc_scaled = StandardScaler().fit_transform(x_train_encoded)\n\n# Create principal components\npca = PCA()\nx_train_enc_pca = pca.fit_transform(x_train_enc_scaled)\n\n# Convert to dataframe\ncomponent_names = [f\"PC{i+1}\" for i in range(x_train_enc_pca.shape[1])]\nx_train_enc_pca = pd.DataFrame(x_train_enc_pca, columns=component_names)\n\n# Add the labels\nx_train_enc_pca['label'] = y_train\n\n# Plot the points of the latent space, considering only PC1 and PC2\nplt.figure(figsize=(12, 10))\nsns.scatterplot(data=x_train_enc_pca, x='PC1', y='PC2', hue='label', palette=\"deep\")\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-01-28T17:40:23.934139Z","iopub.execute_input":"2022-01-28T17:40:23.934385Z","iopub.status.idle":"2022-01-28T17:40:26.525198Z","shell.execute_reply.started":"2022-01-28T17:40:23.934348Z","shell.execute_reply":"2022-01-28T17:40:26.524328Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"We can see the clusters in the above plot but it is a bit messy: the clusters are not clearly separated but this plot allows to understand the general idea. \n\n*Note : why the clusters are not clearly separated? it's because 2 principal components do not represent well enough the points in the 32-dimensional latent space (only 20% of the variance is explained by the 2 first Principal Components). I guess the clusters are much well separated in the 32-dimensional latent space but it's pretty hard to visualize*🤔","metadata":{}},{"cell_type":"code","source":"def plot_variance(pca, width=8, dpi=100):\n    # Create figure\n    fig, axs = plt.subplots(1, 2)\n    n = pca.n_components_\n    grid = np.arange(1, n + 1)\n    # Explained variance\n    evr = pca.explained_variance_ratio_\n    axs[0].bar(grid, evr)\n    axs[0].set(\n        xlabel=\"Component\", title=\"% Explained Variance\", ylim=(0.0, 1.0)\n    )\n    # Cumulative Variance\n    cv = np.cumsum(evr)\n    axs[1].plot(np.r_[0, grid], np.r_[0, cv], \"o-\")\n    axs[1].set(\n        xlabel=\"Component\", title=\"% Cumulative Variance\", ylim=(0.0, 1.0)\n    )\n    # Set up figure\n    fig.set(figwidth=8, dpi=100)\n    return axs\n\nplot_variance(pca)\nplt.show()","metadata":{"_kg_hide-input":true,"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:40:26.526672Z","iopub.execute_input":"2022-01-28T17:40:26.527079Z","iopub.status.idle":"2022-01-28T17:40:26.831409Z","shell.execute_reply.started":"2022-01-28T17:40:26.527042Z","shell.execute_reply":"2022-01-28T17:40:26.830831Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 3. Variational autoencoder (VAE)","metadata":{}},{"cell_type":"markdown","source":"Variational Autoencoder is a type of autoencoder with added constraints on the encoded representations being learnt. Instead of letting the neural network learn an arbitrary function, it is learning the parameters of a probability distribution modeling your input data. The implementation below was proposed by F. Chollet [here](https://blog.keras.io/building-autoencoders-in-keras.html) with some minor adaptation. \n\nOur VAE is an end-to-end autoencoder mapping inputs to reconstructions with :\n* an encoder mapping inputs to the latent space (here a 2D latent space)\n* a decoder that can take points on the latent space and output the corresponding reconstructed images.\n\nAutoencoders use a pixel-by-pixel loss function. Here, our VAE loss combines 2 different loss functions:\n* A reconstruction loss (pixel-by-pixel) function\n* A KL (a.k.a. Kullback-Leibler) loss function  \n\nIn this implementation, it is a weighted average of these 2 losses : vae_loss = r * reconstruction_loss + (1-r) * kl_loss","metadata":{}},{"cell_type":"code","source":"def build_model(r=0.5):\n    \n    ##########################################################################################\n    # First, here's our encoder network, mapping inputs to our latent distribution parameters:\n    ##########################################################################################\n    original_dim = 28 * 28\n    intermediate_dim = 64\n    latent_dim = 2\n\n    inputs = keras.Input(shape=(original_dim,))\n    h = layers.Dense(intermediate_dim, activation='relu')(inputs)\n    z_mean = layers.Dense(latent_dim)(h)\n    z_log_sigma = layers.Dense(latent_dim)(h)\n\n    ##################################################################################\n    # We can use these parameters to sample new similar points from the latent space:\n    ##################################################################################\n\n    def sampling(args):\n        z_mean, z_log_sigma = args\n        epsilon = K.random_normal(shape=(K.shape(z_mean)[0], latent_dim), mean=0., stddev=0.1)\n        return z_mean + K.exp(z_log_sigma) * epsilon\n\n    z = layers.Lambda(sampling, name='encoder_sampling')([z_mean, z_log_sigma])\n\n    ################################################################################\n    # Finally, we can map these sampled latent points back to reconstructed inputs:\n    ################################################################################\n\n    # Create encoder\n    encoder = keras.Model(inputs, [z_mean, z_log_sigma, z], name='encoder')\n\n    # Create decoder\n    latent_inputs = keras.Input(shape=(latent_dim,))\n    x = layers.Dense(intermediate_dim, activation='relu')(latent_inputs)\n    outputs = layers.Dense(original_dim, activation='sigmoid')(x)\n    decoder = keras.Model(latent_inputs, outputs, name='decoder')\n\n    # instantiate VAE model\n    outputs = decoder(encoder(inputs)[2])\n    vae = keras.Model(inputs, outputs, name='vae')\n    \n    #We train the model using the end-to-end model, with a custom loss function: the sum of a reconstruction term, and the KL divergence regularization term.\n    reconstruction_loss = keras.losses.binary_crossentropy(inputs, outputs)\n    reconstruction_loss *= original_dim\n    kl_loss = 1 + z_log_sigma - K.square(z_mean) - K.exp(z_log_sigma)\n    kl_loss = K.sum(kl_loss, axis=-1)\n    kl_loss *= -0.5\n    vae_loss = r * reconstruction_loss + (1-r) * kl_loss\n    vae.add_loss(vae_loss)\n    vae.compile(optimizer='adam')\n    \n    return encoder, decoder, vae","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:40:26.832388Z","iopub.execute_input":"2022-01-28T17:40:26.832674Z","iopub.status.idle":"2022-01-28T17:40:26.842026Z","shell.execute_reply.started":"2022-01-28T17:40:26.832648Z","shell.execute_reply":"2022-01-28T17:40:26.841556Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"![vae.png](attachment:82abae00-6262-4b7b-98ed-743b13db7152.png)","metadata":{},"attachments":{"82abae00-6262-4b7b-98ed-743b13db7152.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"It's time to train our VAE:","metadata":{}},{"cell_type":"code","source":"encoder, decoder, vae = build_model(r=0.5)\n\nhistory = vae.fit(x_train, x_train,\n        epochs=100,\n        batch_size=32,\n        validation_data=(x_test, x_test))","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:40:26.842811Z","iopub.execute_input":"2022-01-28T17:40:26.843426Z","iopub.status.idle":"2022-01-28T17:55:10.093944Z","shell.execute_reply.started":"2022-01-28T17:40:26.843397Z","shell.execute_reply":"2022-01-28T17:55:10.093175Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"vae.save('vae.tf')\nencoder.save('vae_encoder.tf')\ndecoder.save('vae_decoder.tf')","metadata":{"_kg_hide-input":true,"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:55:10.095924Z","iopub.execute_input":"2022-01-28T17:55:10.096794Z","iopub.status.idle":"2022-01-28T17:55:13.054523Z","shell.execute_reply.started":"2022-01-28T17:55:10.096741Z","shell.execute_reply":"2022-01-28T17:55:13.053011Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"hist = pd.DataFrame(history.history)\nhist.plot()\nplt.show()\n\nfinal_loss = hist.iloc[99]['loss']\nfinal_val_loss = hist.iloc[99]['val_loss']\nprint(f'Final loss = {final_loss:.4f} and final val_loss = {final_val_loss:.4f}')","metadata":{"_kg_hide-input":true,"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:55:13.058812Z","iopub.execute_input":"2022-01-28T17:55:13.059118Z","iopub.status.idle":"2022-01-28T17:55:13.289809Z","shell.execute_reply.started":"2022-01-28T17:55:13.059077Z","shell.execute_reply":"2022-01-28T17:55:13.288992Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Now we can try our VAE with the test images. The output is not that bad when you consider the original image is a 784-dimensional vector and the latent input is a 2-dimensional vector : some reconstructed images display the same digit as in the original image and others display a different digit! And if you train the VAE with a larger r, that is to say giving more predominance to the reconstruction_loss over the kl_loss, the reconstruction will be even better.","metadata":{"execution":{"iopub.status.busy":"2022-01-20T14:15:20.283836Z","iopub.execute_input":"2022-01-20T14:15:20.284109Z","iopub.status.idle":"2022-01-20T14:15:22.085958Z","shell.execute_reply.started":"2022-01-20T14:15:20.284079Z","shell.execute_reply":"2022-01-20T14:15:22.085061Z"}}},{"cell_type":"code","source":"x_test_reconstructed = vae.predict(x_test)\n\npos = 1\nfig=plt.figure(figsize=(20,5))\nfor i in (np.random.randint(0, x_test.shape[0], size=10)):\n    fig.add_subplot(2, 10, pos)\n    plt.imshow(np.array(x_test.iloc[i]).reshape(28,28))\n    plt.title('img #' + str(i))\n    plt.axis('off')\n    fig.add_subplot(2, 10, pos+10)\n    plt.imshow(x_test_reconstructed[i].reshape(28,28))\n    plt.title('img #' + str(i))\n    plt.axis('off')\n    pos = pos + 1\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-01-28T17:55:13.291054Z","iopub.execute_input":"2022-01-28T17:55:13.291499Z","iopub.status.idle":"2022-01-28T17:55:15.580436Z","shell.execute_reply.started":"2022-01-28T17:55:13.291456Z","shell.execute_reply":"2022-01-28T17:55:15.579766Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Because our latent space is 2-dimensional, we can look at the neighborhoods of different classes on the latent 2D plane. Each of these colored clusters is a type of digit. Close clusters are digits that are structurally similar.","metadata":{}},{"cell_type":"code","source":"x_train_encoded = encoder.predict(x_train)\n\nz = pd.DataFrame([x_train_encoded[2][:,0],x_train_encoded[2][:,1],y_train]).T\nz.columns=['z_0', 'z_1', 'y_train']\n#z.head()\n\nplt.figure(figsize=(12, 10))\n#sns.scatterplot(x=x_train_encoded[2][:,0], y=x_train_encoded[2][:,1], hue=y_train, palette=\"deep\")\nsns.scatterplot(data=z, x='z_0', y='z_1', hue='y_train', palette=\"deep\")\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-01-28T17:55:15.581468Z","iopub.execute_input":"2022-01-28T17:55:15.58177Z","iopub.status.idle":"2022-01-28T17:55:19.859934Z","shell.execute_reply.started":"2022-01-28T17:55:15.581743Z","shell.execute_reply":"2022-01-28T17:55:19.85903Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Below, we can check the influence of the importance given to the reconstruction loss versus the KL loss. Here, I took \"extreme\" values for r for the sake of the demonstration:\n* when the predominance is given to the reconstruction_loss (r=0.999), **the clusters are clear and pretty-well separated**. On the other hand,the distribution of the latent vector spreads over a big range (z_0 in [-20,50] and z_1 in [0, 60])\n* when the predominance is given to the kl_loss (r=0.001), the clusters overlap more but **the distribution of the latent vector is a unit Gaussian distribution**.","metadata":{}},{"cell_type":"code","source":"encoder_big_r, decoder_big_r, vae_big_r = build_model(r=0.999)\nhistory_big_r = vae_big_r.fit(x_train, x_train,\n        epochs=50,\n        batch_size=32,\n        validation_data=(x_test, x_test))\nvae_big_r.save('vae_big_r.tf')\n\nencoder_small_r, decoder_small_r, vae_small_r = build_model(r=0.001)\nhistory_small_r = vae_small_r.fit(x_train, x_train,\n        epochs=50,\n        batch_size=32,\n        validation_data=(x_test, x_test))\nvae_small_r.save('vae_small_r.tf')\n\nx_train_encoded_big_r = encoder_big_r.predict(x_train)\nz_big_r = pd.DataFrame([x_train_encoded_big_r[2][:,0],x_train_encoded_big_r[2][:,1],y_train]).T\nz_big_r.columns=['z_0', 'z_1', 'y_train']\nx_train_encoded_small_r = encoder_small_r.predict(x_train)\nz_small_r = pd.DataFrame([x_train_encoded_small_r[2][:,0],x_train_encoded_small_r[2][:,1],y_train]).T\nz_small_r.columns=['z_0', 'z_1', 'y_train']","metadata":{"_kg_hide-input":true,"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-01-28T17:55:19.861325Z","iopub.execute_input":"2022-01-28T17:55:19.861797Z","iopub.status.idle":"2022-01-28T18:10:22.979877Z","shell.execute_reply.started":"2022-01-28T17:55:19.861756Z","shell.execute_reply":"2022-01-28T18:10:22.979021Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig=plt.figure(figsize=(20, 10))\nfig.add_subplot(1, 2, 1)\nsns.scatterplot(data=z_big_r, x='z_0', y='z_1', hue='y_train', palette=\"deep\")\nplt.title('r=0.999 -> reconstruction loss is predominant')\nfig.add_subplot(1, 2, 2)\nsns.scatterplot(data=z_small_r, x='z_0', y='z_1', hue='y_train', palette=\"deep\")\nplt.title('r=0.001 -> KL loss is predominant')\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-01-28T18:10:22.981074Z","iopub.execute_input":"2022-01-28T18:10:22.981296Z","iopub.status.idle":"2022-01-28T18:10:26.200566Z","shell.execute_reply.started":"2022-01-28T18:10:22.981268Z","shell.execute_reply":"2022-01-28T18:10:26.199942Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig=plt.figure(figsize=(20, 4))\nfig.suptitle('When predominance is given to the reconstruction loss, the encoder is trained without constraint on z distribution')\nfig.add_subplot(1, 2, 1)\nsns.histplot(data=z_big_r['z_0'], kde=True)\nfig.add_subplot(1, 2, 2)\nsns.histplot(data=z_big_r['z_1'], kde=True)\nplt.show()\n\nfig=plt.figure(figsize=(20, 4))\nfig.suptitle('When predominance is given to the KL loss, the encoder is trained with a constraint so z is a unit Gaussian distribution')\nfig.add_subplot(1, 2, 1)\nsns.histplot(data=z_small_r['z_0'], kde=True)\nfig.add_subplot(1, 2, 2)\nsns.histplot(data=z_small_r['z_1'], kde=True)\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-01-28T18:10:26.2015Z","iopub.execute_input":"2022-01-28T18:10:26.201856Z","iopub.status.idle":"2022-01-28T18:10:28.637233Z","shell.execute_reply.started":"2022-01-28T18:10:26.201823Z","shell.execute_reply":"2022-01-28T18:10:28.636229Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 4. Applications of autoencoders","metadata":{}},{"cell_type":"markdown","source":"In this section, I'm going to show 2 applications of autoencoders. There are, of course, many more possible applications of autoencoders and if you know them, please add them in the comments section below. ","metadata":{}},{"cell_type":"markdown","source":"## 4.1 Generative Models ","metadata":{"execution":{"iopub.status.busy":"2022-01-20T13:22:34.232704Z","iopub.execute_input":"2022-01-20T13:22:34.233004Z","iopub.status.idle":"2022-01-20T13:22:34.239179Z","shell.execute_reply.started":"2022-01-20T13:22:34.232973Z","shell.execute_reply":"2022-01-20T13:22:34.238168Z"}}},{"cell_type":"markdown","source":"An autoencoder is a \"generative model\". Once your autoencoder is trained, just sample points from the latent space, you can generate new input data samples. Below, you can see 10 new \"MNIST-like\" images generated by the decoder that we created in the VAE section. We also display the coordinates of the points in the 2d latent space that were used to generate these images:","metadata":{}},{"cell_type":"code","source":"# Take 10 points randomly in this latent space\nrandom_points = np.random.normal(loc=0.0, scale=1.0, size=(10,2))\n\n# Create \"new\" images for these 10 points\ndecoder = load_model('vae_decoder.tf')\nnew_images = decoder.predict(random_points)\n\n# Display new images\npos = 1\nfig=plt.figure(figsize=(20,5))\nfor i in range(0,10):\n    fig.add_subplot(1, 10, pos)\n    plt.imshow(new_images[i].reshape(28,28))\n    plt.title(np.around(random_points[i],2))\n    plt.axis('off')\n    pos = pos + 1\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-01-28T18:10:28.638473Z","iopub.execute_input":"2022-01-28T18:10:28.63878Z","iopub.status.idle":"2022-01-28T18:10:29.56544Z","shell.execute_reply.started":"2022-01-28T18:10:28.638749Z","shell.execute_reply":"2022-01-28T18:10:29.564549Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"*Note : in this [notebook](https://www.kaggle.com/cdeotte/dog-autoencoder) created by GM Chris Deotte, you can find another example of generation of images of dogs using encoders* ","metadata":{}},{"cell_type":"markdown","source":"## 4.2 Denoising Models ","metadata":{}},{"cell_type":"markdown","source":"You can use autoencoders to denoise inputs. To achieve this, the training process is slightly different : instead of training the autoencoder to copy its input to its output, the autoencoder is trained to map noisy inputs to clean outputs. E.g. train your autoencoder to map noisy digits images from the MNIST dataset to clean digits images. The implementation below is proposed by Santiago L. Valdarrama [here](https://keras.io/examples/vision/autoencoder/).","metadata":{}},{"cell_type":"markdown","source":"**Step 1 : create noisy inputs by adding a noise to x_train, x_test clean images**","metadata":{}},{"cell_type":"code","source":"# create noisy inputs\nnoise_factor = 0.4\nx_train_noisy = x_train + noise_factor * np.random.normal(loc=0.0, scale=1.0, size=x_train.shape) \nx_test_noisy = x_test + noise_factor * np.random.normal(loc=0.0, scale=1.0, size=x_test.shape) \n\nx_train_noisy = np.clip(x_train_noisy, 0., 1.)\nx_test_noisy = np.clip(x_test_noisy, 0., 1.)\n\n# format the inputs to the format expected by our convolutional autoencoder\nx_train = np.array(x_train).reshape(42000,28,28)\nx_test = np.array(x_test).reshape(28000,28,28)\nx_train_noisy = np.array(x_train_noisy).reshape(42000,28,28)\nx_test_noisy = np.array(x_test_noisy).reshape(28000,28,28)\n\n# display noisy inputs\npos = 1\nfig=plt.figure(figsize=(20,5))\nfor i in (np.random.randint(0, x_train_noisy.shape[0], size=10)):\n    fig.add_subplot(1, 10, pos)\n    plt.imshow(x_train_noisy[i])\n    plt.axis('off')\n    pos = pos + 1\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-01-28T18:42:24.843452Z","iopub.execute_input":"2022-01-28T18:42:24.844279Z","iopub.status.idle":"2022-01-28T18:42:28.13208Z","shell.execute_reply.started":"2022-01-28T18:42:24.844238Z","shell.execute_reply":"2022-01-28T18:42:28.131202Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Step 2 : create a convolutional autoencoder**","metadata":{}},{"cell_type":"code","source":"input = layers.Input(shape=(28, 28, 1))\n\n# Encoder\nx = layers.Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\")(input)\nx = layers.MaxPooling2D((2, 2), padding=\"same\")(x)\nx = layers.Conv2D(32, (3, 3), activation=\"relu\", padding=\"same\")(x)\nx = layers.MaxPooling2D((2, 2), padding=\"same\")(x)\n\n# Decoder\n# Note : UpSampling2D is just a simple scaling up of the image by using nearest neighbour or bilinear upsampling, \n# so nothing smart. Conv2DTranspose is a convolution operation whose kernel is learnt (just like normal conv2d operation) \n# while training your model. Using Conv2DTranspose will also upsample its input but the key difference is the model \n# should learn what is the best upsampling for the job.\nx = layers.Conv2DTranspose(32, (3, 3), strides=2, activation=\"relu\", padding=\"same\")(x)\nx = layers.Conv2DTranspose(32, (3, 3), strides=2, activation=\"relu\", padding=\"same\")(x)\nx = layers.Conv2D(1, (3, 3), activation=\"sigmoid\", padding=\"same\")(x)\n\n# Autoencoder\nautoencoder = keras.Model(input, x)\nautoencoder.compile(optimizer=\"adam\", loss=\"binary_crossentropy\")","metadata":{"execution":{"iopub.status.busy":"2022-01-28T18:42:36.131191Z","iopub.execute_input":"2022-01-28T18:42:36.131886Z","iopub.status.idle":"2022-01-28T18:42:36.203192Z","shell.execute_reply.started":"2022-01-28T18:42:36.131841Z","shell.execute_reply":"2022-01-28T18:42:36.2026Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Step 3 : train the autoencoder**","metadata":{}},{"cell_type":"code","source":"autoencoder.fit(\n    x=x_train_noisy,\n    y=x_train,\n    epochs=50,\n    batch_size=128,\n    shuffle=True,\n    validation_data=(x_test_noisy, x_test),\n)","metadata":{"execution":{"iopub.status.busy":"2022-01-28T18:42:38.495797Z","iopub.execute_input":"2022-01-28T18:42:38.496571Z","iopub.status.idle":"2022-01-28T19:04:57.523053Z","shell.execute_reply.started":"2022-01-28T18:42:38.496526Z","shell.execute_reply":"2022-01-28T19:04:57.522487Z"},"_kg_hide-output":true,"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Step 4 : denoise the images from the test set**","metadata":{}},{"cell_type":"markdown","source":"We can use our model to remove the noise that was added to the test images in step 1. and display the output of the autoencoder. As you can see, there is little difference between the original image and the output of the autoencoder, quite impressive! ","metadata":{}},{"cell_type":"code","source":"x_test_denoised = autoencoder.predict(x_test_noisy)\n\npos = 1\nfig=plt.figure(figsize=(20,2))\nfig.suptitle('Original images')\nfor i in range(0,10):\n    fig.add_subplot(1, 10, pos)\n    plt.imshow(x_test[i].reshape(28,28))\n    plt.axis('off')\n    pos = pos + 1\nplt.show()\n\npos = 1\nfig=plt.figure(figsize=(20,2))\nfig.suptitle('Images denoised by the autoencoder')\nfor i in range(0,10):\n    fig.add_subplot(1, 10, pos)\n    plt.imshow(x_test_denoised[i].reshape(28,28))\n    plt.axis('off')\n    pos = pos + 1\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-01-28T19:07:46.214139Z","iopub.execute_input":"2022-01-28T19:07:46.214877Z","iopub.status.idle":"2022-01-28T19:07:53.585048Z","shell.execute_reply.started":"2022-01-28T19:07:46.214835Z","shell.execute_reply":"2022-01-28T19:07:53.584186Z"},"_kg_hide-input":true,"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"*Note: there was a [competition on Kaggle](https://www.kaggle.com/c/denoising-dirty-documents) a few years ago where the participants were requested to remove noise from printed text. Many solutions implemented autoencoders.*","metadata":{}},{"cell_type":"markdown","source":"# 5. References","metadata":{}},{"cell_type":"markdown","source":"* [Tutorial on autoencoders and VAE (video in French)](https://www.youtube.com/watch?v=lTy_AAjSTD8)\n* [Auto-Encoding Variational Bayes](https://arxiv.org/abs/1312.6114)\n* [How to Calculate the KL Divergence for Machine Learning](https://machinelearningmastery.com/divergence-between-probability-distributions/)\n* [Dog autoencoder](https://www.kaggle.com/cdeotte/dog-autoencoder)\n* [Building autoencoders in keras](https://blog.keras.io/building-autoencoders-in-keras.html)\n* [Understanding latent space in machine learning](https://towardsdatascience.com/understanding-latent-space-in-machine-learning-de5a7c687d8d)","metadata":{}}]}