Representations_for_AI/Representations_for_AI.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Learning better data representations for AI with Scikit-Learn and Keras\n",
    "\n",
    "### James McDermott, Lecturer in Computer Science, NUI Galway\n",
    "\n",
    "### james.mcdermott@nuigalway.ie,   http://jmmcd.net, `@bleepbeepbzzz`\n",
    "\n",
    "<img src=img/NUI_Galway_logo.png width=400 alt=\"NUI Galway logo\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Introduction\n",
    "\n",
    "### This notebook/presentation is about learning representations (embeddings) for data using classic machine learning methods, and also using a modern method (a Siamese neural network with triplet loss). It is part of the Atlantec 2019 AI Tools and Techniques session.\n",
    "\n",
    "### We won't assume any prior knowledge, but will assume willingness to ignore unnecessary details :) Some knowledge of Python, Numpy, and neural networks is helpful."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# About me\n",
    "\n",
    "James McDermott is a Lecturer in Computer Science in the National University of Ireland, Galway, where he teaches courses in the area of artificial intelligence. He holds a BSc in Computer Science with Mathematics from the National University of Ireland, Galway, and a PhD in evolutionary computation and computer music from the University of Limerick. He has also worked on supercomputing in Compaq/Hewlett-Packard. His post-doctoral work was in evolutionary design and genetic programming in University College Dublin and Massachusetts Institute of Technology. His research interests are in program synthesis, evolutionary computing, artificial intelligence, and computational music and design. He has chaired the EuroGP and EvoMUSART international conferences and is a member of the Genetic Programming and Evolvable Machines journal editorial board. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "\n",
    "## Installation/setup\n",
    "\n",
    "### (Recommended): visit https://tinyurl.com/atlantec-jmmcd\n",
    "\n",
    "\n",
    "(Redirect to: https://colab.research.google.com/github/jmmcd/ML-snippets/blob/master/Representations_for_AI/Representations_for_AI.ipynb)\n",
    "\n",
    "* (Advanced alternative): clone from: https://github.com/jmmcd/ML-snippets and run locally.\n",
    "\n",
    "### I will present from these slides, and you can type Shift-Enter on each \"cell\" in your notebook to execute it.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Import all the usual libraries and tell Matplotlib to show plots in the notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "from sklearn.base import TransformerMixin\n",
    "from sklearn.manifold import MDS, TSNE\n",
    "from sklearn.decomposition import PCA\n",
    "from keras.models import Model, Sequential\n",
    "from keras.optimizers import Adam\n",
    "from keras.layers import Dense, Dropout, Flatten, Convolution2D, MaxPooling2D, Input, Lambda\n",
    "from keras.utils import np_utils, plot_model\n",
    "from keras.datasets import mnist\n",
    "from keras import backend as K"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "\n",
    "## Load and re-shape data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Tensorflow provides the MNIST dataset\n",
    "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
    "# Each digit is of shape 28x28 with just 1 channel (greyscale, not RGB)\n",
    "X_train = X_train.reshape(X_train.shape[0], 28, 28, 1)\n",
    "X_test = X_test.reshape(X_test.shape[0], 28, 28, 1)\n",
    "X_train = X_train.astype('float32') # convert to float\n",
    "X_test = X_test.astype('float32')\n",
    "X_train /= 255 # scale to the range [0, 1]\n",
    "X_test /= 255"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## AI in the past, present and future (approximately)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* Around 45 years ago, statistical learning (think: regression) was already successful on some tasks. But in the AI world the focus was mostly on \"good old-fashioned AI\", that is symbolic approaches driven by rules, planning, world-modelling.\n",
    "\n",
    "* Around 35 years ago, neural/connectionist approaches got a lot of hype, and succeeded on some tasks, but failed to deliver on some more ambitious goals, leading to an \"AI Winter\". Example: Multi-Layer Perceptron."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* Around 25 years ago, other statistical learning approaches came along and neural approaches went out of fashion. We started to see that a good general approach to data problems was to see how to map the data from one space into another which better reflects the true \"geometry\" of the data -- ie make an *embedding* of the data -- and to base our algorithms on *distances* between data points in the new space. Example: Support Vector Machines.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* Around 10 years ago, we finally understood how to make *deep* (many-layer) neural networks succeed, especially on vision tasks. Example: AlexNet, a Deep Convolutional Neural Network. It turned out that what they were doing internally and implicitly was something a lot like *learning good embeddings from the data*. Each layer in a deep neural network maps the data into a new space, slightly rearranging it.\n",
    "\n",
    "<img src=\"img/NeuralNetwork.png\" width=600 alt=\"Image of a small feedforward neural network\"> <font size=1>[source Wikimedia]</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* Now, we are seeing several approaches which do the same task deliberately. Examples: Word2Vec, Generative Adversarial Networks and Variational Auto-encoders. \n",
    "\n",
    "* We are also seeing neural networks taking on tasks such as time-series modelling, speech recognition, game-playing. Examples: Long Short-Term Memory and Reinforcement Learning.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "* In the future, I think that we'll come full circle, and we'll see symbolic approaches layered on top of these learned embeddings. For example, the system will learn an embedding using techniques like what we'll see today, and then produce an explicit symbolic representation of the implicitly-learned embedding. That will enable reasoning, explanation (and human-style reasoning is a lot like explaining), and robustness -- properties that are lacking in current neural approaches."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Our goal today is \n",
    "### (1) to develop an intuition for embeddings, and so understand a core underlying mechanism of AI \n",
    "### (2) learn some of the APIs of Scikit-learn for machine learning and Keras for deep learning."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## What is an embedding?\n",
    "\n",
    "### A mapping from an original data space to a new space, usually of lower dimension. In a *good* embedding the geometry of the new space reflects some property of the data.\n",
    "\n",
    "<img src=\"img/word_embedding.png\" width=500 alt=\"Example of a word embedding\"> <font size=1>[source mathworks.com]</font>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Regression: predict a real number, eg house price\n",
    "\n",
    "## Classification: predict a label, eg 0/1 (cat/dog)\n",
    "\n",
    "## *Embedding*: output a real vector, interpreted as a location"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "<img src=\"img/king-queen.png\" width=\"600\"> <font size=1>[source https://medium.com/scaleabout/a-gentle-introduction-to-doc2vec-db3e8c0cce5e]</font>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "# Singular - Plural\n",
    "\n",
    "<img src=\"img/plural-relation.png\" width=600> <font size=1>[source https://www.kdnuggets.com/2016/05/amazing-power-word-vectors.html]</font>\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Let's look at some sample images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x7200 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(nrows=1, ncols=5, figsize=(20, 100))\n",
    "for i, ax in enumerate(axs.flat):\n",
    "    ax.matshow(X_train[i].reshape(28, 28))\n",
    "    ax.axis('off')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Visualisation: how are the digits distributed in the embedding?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_viz(xy, ims):\n",
    "    \"\"\"Generate an image of our embedding. xy is an array of x-y coordinates of the data\n",
    "    points, and ims are the data points themselves (the original images).\"\"\"\n",
    "    canvas_size = 1500\n",
    "    imsize = ims[0].shape\n",
    "    canvas = np.zeros((canvas_size+imsize[0], canvas_size+imsize[1]), dtype=float)\n",
    "    pts = loc2pix(xy, canvas_size) # scale to fit canvas\n",
    "    for p, im in zip(pts, ims):       \n",
    "        # paint the image of the digit onto the canvas\n",
    "        canvas[p[0]:p[0]+imsize[0], p[1]:p[1]+imsize[1]] = im[:, :]\n",
    "\n",
    "    # 20x20 inches: we have to make it large to avoid subsampling\n",
    "    fig, ax = plt.subplots(figsize=(20, 20)) \n",
    "    ax.matshow(canvas)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "def loc2pix(x, size):\n",
    "    \"\"\"Helper function for make_viz. All values in x are in some range [min, max], and \n",
    "    we want it in [0, size].\"\"\"\n",
    "    x = (x - x.min()) / (x.max() - x.min()) # now in [0, 1]\n",
    "    x *= size\n",
    "    x = x.astype(int)\n",
    "    assert np.all(x >= 0) and np.all(x <= size)\n",
    "    return x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Classic algorithms for embedding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Multi-dimensional scaling (MDS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_samples_train = n_samples_test = 200\n",
    "\n",
    "mds = MDS(n_components=2) # fit_transform()\n",
    "x = mds.fit_transform(X_train[:n_samples_train].reshape(n_samples_train, 28*28))\n",
    "make_viz(x, X_train[:n_samples_test].squeeze())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Principal components analysis (PCA)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pca = PCA(n_components=2) # fit() and then transform()\n",
    "pca.fit(X_train[:n_samples_train].reshape(n_samples_train, 28*28))\n",
    "x = pca.transform(X_test[:n_samples_test].reshape(n_samples_test, 28*28))\n",
    "make_viz(x, X_test[:n_samples_test].squeeze())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Results are good!\n",
    "\n",
    "* We already have a useful embedding. Other algorithms (clustering, anomaly detection, classification, visualisation, interactive search...) can be expected to work well using that embedding.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Notice that the `scikit-learn` API is consistent.\n",
    "### Very easy to swap out one algorithm for another.\n",
    "\n",
    "```python\n",
    "mds = MDS()\n",
    "mds.fit_transform(X)\n",
    "tsne = TNSE() # t-SNE, another algorithm doing a similar job\n",
    "tsne.fit_transform(X)\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "\n",
    "<img src=\"img/Data_Science_VD.png\" width=600 alt=\"Data Science Venn Diagram by Drew Conway\">\n",
    "<font size=1>[source http://drewconway.com/zia/2013/3/26/the-data-science-venn-diagram</font>\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Next, we are going to train a deep convolutional neural network to do the same task.\n",
    "\n",
    "* We will use Keras\n",
    "* We will mimic the Scikit-learn API"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Convolutional: suitable for images. Each neuron \"looks at\" each patch of the input in succession and outputs a new \"image\"\n",
    "\n",
    "### Siamese network: we run our convolutional network multiple times on different images, and then...\n",
    "\n",
    "### Triplet loss: ... we ask it to \"push\" similar images closer together, and push distinct images farther apart."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "<img src=img/siamese.png alt=\"Siamese neural network\" width=\"600\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Triplet-loss and model source: https://github.com/Ao-Lee/Shares/blob/master/triplet%20inputs.py\n",
    "\n",
    "Adapted by James McDermott:\n",
    " - changed to MNIST task\n",
    " - changed to a small, simple network (not ResNet)\n",
    " - added generation of proper a, p, n triplets\n",
    " - increased alpha to try to prevent the embedding from winning trivially\n",
    " - removed the L2 normalisation (pushing all digits to surface of hypersphere)\n",
    " - added a 2D layout visualisation of our embedding"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Next up is the complete code for the `SiameseConvEmbedding` neural network. We won't read through this in full but will pick out highlights after."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "class SiameseConvEmbedding(TransformerMixin): # inherit from sklearn TransformerMixin\n",
    "    def __init__(self, n_components=2, \n",
    "                 learning_rate=0.001, batch_size=100, \n",
    "                 epochs=5, alpha=0.4):\n",
    "        self.n_components = n_components\n",
    "        self.batch_size = batch_size\n",
    "        self.learning_rate = learning_rate\n",
    "        self.epochs = epochs\n",
    "        self.alpha = alpha\n",
    "        \n",
    "    def __repr__(self):\n",
    "        return f\"\"\"SiameseConvEmbedding(n_components={self.n_components}, \n",
    "learning_rate={self.learning_rate}, \n",
    "batch_size={self.batch_size},\n",
    "epochs={self.epochs},\n",
    "alpha={self.alpha})\"\"\"\n",
    "        \n",
    "    def get_image(self, label, X, y):\n",
    "        \"\"\"Choose an image from our training or test data with the\n",
    "        given label.\"\"\"\n",
    "        idx = np.random.randint(len(y))\n",
    "        while y[idx] != label:\n",
    "            # keep searching randomly\n",
    "            idx = np.random.randint(len(y))\n",
    "        return X[idx] \n",
    "        \n",
    "    def get_triplet(self, X, y):\n",
    "        \"\"\"Choose a triplet (anchor, positive, negative) of images\n",
    "        such that anchor and positive have the same label and\n",
    "        anchor and negative have different labels.\"\"\"\n",
    "        n = a = np.random.choice(self.labels)\n",
    "        while n == a:\n",
    "            # keep searching randomly\n",
    "            n = np.random.choice(self.labels)\n",
    "        A = self.get_image(a, X, y)\n",
    "        P = self.get_image(a, X, y) # get two images of same label for A and P\n",
    "        N = self.get_image(n, X, y) # get image of a different label for N\n",
    "        return A, P, N\n",
    "\n",
    "    def generate_triplets(self, X, y):\n",
    "        \"\"\"Generate an un-ending stream (using yield) of triplets for\n",
    "        training.\"\"\"\n",
    "        while True:\n",
    "            A = np.zeros((self.batch_size, *self.imsize))\n",
    "            P = np.zeros((self.batch_size, *self.imsize))\n",
    "            N = np.zeros((self.batch_size, *self.imsize))\n",
    "\n",
    "            for i in range(self.batch_size):\n",
    "                A[i], P[i], N[i] = self.get_triplet(X, y)\n",
    "\n",
    "            # a \"dummy\" label which will come in to our identity loss\n",
    "            # function below as y_true. We'll ignore it.\n",
    "            label = np.ones(self.batch_size) \n",
    "            yield [A, P, N], label\n",
    "        \n",
    "    ############## Loss ###########################\n",
    "\n",
    "    def identity_loss(self, y_true, y_pred):\n",
    "        \"\"\"This loss function just takes the mean of y_pred. Because of the\n",
    "        way we wire the network (see complete_model below), y_pred is the\n",
    "        output of the triplet loss, so minimising it is what we want to\n",
    "        do.\"\"\"\n",
    "        return K.mean(y_pred)\n",
    "\n",
    "    def triplet_loss(self, x):\n",
    "        \"\"\"The triplet loss is ||A - P|| - ||A - N|| + alpha, where ||.||\n",
    "        is the Euclidean norm. Notice that this is not a loss function in the\n",
    "        format expected by Keras, ie f(y_true, y_pred).\"\"\"\n",
    "    \n",
    "        anchor, positive, negative = x\n",
    "        # axis=1 because here we are summing along the dimensions of the embedding\n",
    "        A_P = K.sqrt(K.sum(K.square(anchor - positive), axis=1))\n",
    "        A_N = K.sqrt(K.sum(K.square(anchor - negative), axis=1))\n",
    "        loss = K.maximum(A_P - A_N + self.alpha, 0)\n",
    "        return loss # this is a vector, 1 value per triplet\n",
    "\n",
    "    \n",
    "    ############## Model ###########################\n",
    "\n",
    "    def embedding_model(self):\n",
    "        \"\"\"A tiny vision model. Its input is an image and output is an embedding.  We\n",
    "        don't compile or add a loss here since this will become a component in\n",
    "        the complete model below.\"\"\"\n",
    "        model = Sequential()\n",
    "        model.add(Convolution2D(32, (3, 3), activation='relu', input_shape=self.imsize))\n",
    "        model.add(Convolution2D(32, (3, 3), activation='relu'))\n",
    "        model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "        model.add(Dropout(rate=0.25))\n",
    "        model.add(Flatten())\n",
    "        model.add(Dense(16, activation='relu'))\n",
    "        model.add(Dense(self.n_components, activation='tanh'))\n",
    "        return model\n",
    "\n",
    "\n",
    "    def complete_model(self):\n",
    "        \"\"\"This part of the model is quite tricky. Rather than a Sequential\n",
    "        model, we declare a Model and say which are its inputs and\n",
    "        outputs, and we declare how the outputs are calculated from the\n",
    "        inputs. In particular, there are no layers in this model, other\n",
    "        than the layers in the embedding model discussed above.\n",
    "\n",
    "        A further complication is that our triplet loss can't be\n",
    "        calculated as a function of y_true and y_predicted as\n",
    "        usual. Instead we calculate the triplet loss as an extra Lambda\n",
    "        layer. Then the Model's loss is set to be equal to the triplet\n",
    "        loss via the identity function.\"\"\"\n",
    "\n",
    "        input_1 = Input(self.imsize)\n",
    "        input_2 = Input(self.imsize)\n",
    "        input_3 = Input(self.imsize)\n",
    "\n",
    "        # call the base_model three times to get A, P, N\n",
    "        A = self.base_model(input_1)\n",
    "        P = self.base_model(input_2)\n",
    "        N = self.base_model(input_3)\n",
    "\n",
    "        loss = Lambda(self.triplet_loss)([A, P, N]) \n",
    "        model = Model(inputs=[input_1, input_2, input_3], outputs=loss)\n",
    "        model.compile(loss=self.identity_loss, optimizer=Adam(self.learning_rate))\n",
    "        return model\n",
    "\n",
    "    def fit(self, X_train, y_train):\n",
    "    \n",
    "        self.imsize = X_train[0].shape\n",
    "        self.labels = np.array(sorted(set(y_train))) # all the possible labels, sorted\n",
    "        \n",
    "        ############## Main ###############################\n",
    "\n",
    "        # create the data generators\n",
    "        train_generator = self.generate_triplets(X_train, y_train)\n",
    "\n",
    "        # instantiate the model and take a look\n",
    "        self.base_model = self.embedding_model()\n",
    "        model = self.complete_model()\n",
    "        print(self.base_model.summary())\n",
    "\n",
    "        # fit\n",
    "        model.fit_generator(train_generator, \n",
    "                                 epochs=self.epochs,\n",
    "                                 steps_per_epoch=10,\n",
    "                                 validation_steps=0 # no validation\n",
    "                                 )\n",
    "        return self\n",
    "    \n",
    "    def transform(self, X, y=None):\n",
    "        return self.base_model.predict(X)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# `get_triplet`: (anchor, positive, negative)\n",
    "\n",
    "```python\n",
    "def get_triplet(self, X, y):\n",
    "    n = a = np.random.choice(self.labels)\n",
    "    while n == a:\n",
    "        # keep searching randomly\n",
    "        n = np.random.choice(self.labels)\n",
    "    A = self.get_image(a, X, y)\n",
    "    P = self.get_image(a, X, y) # two images of same label\n",
    "    N = self.get_image(n, X, y) # image of a different label\n",
    "    return A, P, N\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Small convolutional network\n",
    "```python\n",
    "    def embedding_model(self):\n",
    "        model = Sequential()\n",
    "        model.add(Convolution2D(32, (3, 3), activation='relu',\n",
    "                                input_shape=self.imsize))\n",
    "        model.add(Convolution2D(32, (3, 3), activation='relu'))\n",
    "        model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "        model.add(Dropout(rate=0.25))\n",
    "        model.add(Flatten())\n",
    "        model.add(Dense(16, activation='relu'))\n",
    "        model.add(Dense(self.n_components, activation='tanh'))\n",
    "        return model\n",
    "```        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Triplet loss: push anchor and positive closer, push anchor and negative farther apart\n",
    "```python\n",
    "    def triplet_loss(self, x):\n",
    "        \"\"\"The triplet loss is ||A - P|| - ||A - N|| + alpha\"\"\"    \n",
    "        anchor, positive, negative = x\n",
    "        A_P = K.sqrt(K.sum(K.square(anchor - positive), axis=1))\n",
    "        A_N = K.sqrt(K.sum(K.square(anchor - negative), axis=1))\n",
    "        loss = K.maximum(A_P - A_N + self.alpha, 0)\n",
    "        return loss # this is a vector, 1 value per triplet\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Mimic the `scikit-learn` API\n",
    "```python\n",
    "def fit(self, X_train, y_train):\n",
    "    self.base_model = self.embedding_model()\n",
    "    model = self.complete_model()\n",
    "    model.fit_generator(train_generator, epochs=self.epochs,\n",
    "                        steps_per_epoch=10, validation_steps=0)\n",
    "    return self\n",
    "    \n",
    "def transform(self, X, y=None):\n",
    "    return self.base_model.predict(X)\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Train the network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_37 (Conv2D)           (None, 26, 26, 32)        320       \n",
      "_________________________________________________________________\n",
      "conv2d_38 (Conv2D)           (None, 24, 24, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_19 (MaxPooling (None, 12, 12, 32)        0         \n",
      "_________________________________________________________________\n",
      "dropout_19 (Dropout)         (None, 12, 12, 32)        0         \n",
      "_________________________________________________________________\n",
      "flatten_19 (Flatten)         (None, 4608)              0         \n",
      "_________________________________________________________________\n",
      "dense_37 (Dense)             (None, 16)                73744     \n",
      "_________________________________________________________________\n",
      "dense_38 (Dense)             (None, 2)                 34        \n",
      "=================================================================\n",
      "Total params: 83,346\n",
      "Trainable params: 83,346\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "Epoch 1/40\n",
      "10/10 [==============================] - 5s 510ms/step - loss: 0.3250\n",
      "Epoch 2/40\n",
      "10/10 [==============================] - 3s 349ms/step - loss: 0.1945\n",
      "Epoch 3/40\n",
      "10/10 [==============================] - 3s 331ms/step - loss: 0.1711\n",
      "Epoch 4/40\n",
      "10/10 [==============================] - 3s 304ms/step - loss: 0.1346\n",
      "Epoch 5/40\n",
      "10/10 [==============================] - 3s 320ms/step - loss: 0.1472\n",
      "Epoch 6/40\n",
      "10/10 [==============================] - 3s 305ms/step - loss: 0.1293\n",
      "Epoch 7/40\n",
      "10/10 [==============================] - 3s 304ms/step - loss: 0.1317\n",
      "Epoch 8/40\n",
      "10/10 [==============================] - 3s 305ms/step - loss: 0.1182\n",
      "Epoch 9/40\n",
      "10/10 [==============================] - 3s 307ms/step - loss: 0.1146\n",
      "Epoch 10/40\n",
      "10/10 [==============================] - 3s 304ms/step - loss: 0.0938\n",
      "Epoch 11/40\n",
      "10/10 [==============================] - 3s 305ms/step - loss: 0.1007\n",
      "Epoch 12/40\n",
      "10/10 [==============================] - 3s 305ms/step - loss: 0.0852\n",
      "Epoch 13/40\n",
      "10/10 [==============================] - 3s 313ms/step - loss: 0.0881\n",
      "Epoch 14/40\n",
      "10/10 [==============================] - 3s 310ms/step - loss: 0.0864\n",
      "Epoch 15/40\n",
      "10/10 [==============================] - 3s 315ms/step - loss: 0.0763\n",
      "Epoch 16/40\n",
      "10/10 [==============================] - 3s 316ms/step - loss: 0.0755\n",
      "Epoch 17/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0850\n",
      "Epoch 18/40\n",
      "10/10 [==============================] - 3s 315ms/step - loss: 0.0695\n",
      "Epoch 19/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0751\n",
      "Epoch 20/40\n",
      "10/10 [==============================] - 3s 318ms/step - loss: 0.0694\n",
      "Epoch 21/40\n",
      "10/10 [==============================] - 3s 316ms/step - loss: 0.0680\n",
      "Epoch 22/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0613\n",
      "Epoch 23/40\n",
      "10/10 [==============================] - 3s 320ms/step - loss: 0.0656\n",
      "Epoch 24/40\n",
      "10/10 [==============================] - 3s 315ms/step - loss: 0.0690\n",
      "Epoch 25/40\n",
      "10/10 [==============================] - 3s 316ms/step - loss: 0.0625\n",
      "Epoch 26/40\n",
      "10/10 [==============================] - 3s 315ms/step - loss: 0.0504\n",
      "Epoch 27/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0637\n",
      "Epoch 28/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0692\n",
      "Epoch 29/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0565\n",
      "Epoch 30/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0510\n",
      "Epoch 31/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0542\n",
      "Epoch 32/40\n",
      "10/10 [==============================] - 3s 318ms/step - loss: 0.0475\n",
      "Epoch 33/40\n",
      "10/10 [==============================] - 3s 319ms/step - loss: 0.0500\n",
      "Epoch 34/40\n",
      "10/10 [==============================] - 3s 315ms/step - loss: 0.0492\n",
      "Epoch 35/40\n",
      "10/10 [==============================] - 3s 318ms/step - loss: 0.0554\n",
      "Epoch 36/40\n",
      "10/10 [==============================] - 3s 315ms/step - loss: 0.0377\n",
      "Epoch 37/40\n",
      "10/10 [==============================] - 3s 317ms/step - loss: 0.0456\n",
      "Epoch 38/40\n",
      "10/10 [==============================] - 3s 315ms/step - loss: 0.0441\n",
      "Epoch 39/40\n",
      "10/10 [==============================] - 3s 318ms/step - loss: 0.0552\n",
      "Epoch 40/40\n",
      "10/10 [==============================] - 3s 316ms/step - loss: 0.0572\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "SiameseConvEmbedding(n_components=2, \n",
       "learning_rate=0.001, \n",
       "batch_size=100,\n",
       "epochs=40,\n",
       "alpha=0.4)"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nn = SiameseConvEmbedding(n_components=2, batch_size=100, epochs=40)\n",
    "nn.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Visualise results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = nn.transform(X_test[:n_samples_test])\n",
    "make_viz(x, X_test[:n_samples_test].squeeze())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Let's hack:\n",
    "* Add/remove layers in the `embedding_model` code\n",
    "* Run for more `epochs` and larger `batch_size` (more training $\\implies$ better embedding?)\n",
    "* Try different values of `alpha` $\\in [0.2, 1.0]$ and `learning_rate`\n",
    "* We may observe the training fails to converge with large `learning_rate`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Notes\n",
    "\n",
    "### By the end, we have achieved a very good embedding with the NN\n",
    "### Notice that the NN is *supervised*, in contrast to PCA and MDS\n",
    "### We have produced a 2D embedding because that is nice for visualisation, but higher-dimensional embeddings are common in practice."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Conclusions\n",
    "\n",
    "### We have gained intuition for embeddings and seen classic embedding algorithms\n",
    "\n",
    "### We have seen the `scikit-learn` API and mimicked it in a modern neural embedding model\n",
    "\n",
    "### We have seen some Keras for creating neural networks\n",
    "\n",
    "### We have used Jupyter Notebooks / Google Colab as a medium for coding, experiments and documentation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# We'll teach lots more in the NUI Galway MSc in Artificial Intelligence (part-time/online and full-time/classroom) starting September!\n",
    "\n",
    "## http://www.it.nuigalway.ie/engineering-informatics/information-technology/artificialintelligence/courses/\n",
    "\n",
    "<img src=img/NUI_Galway_logo.png width=400 alt=\"NUI Galway logo\">"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}