iris_classification_ttn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pennylane as qml\n",
    "from pennylane import numpy as np\n",
    "import qiskit\n",
    "from qiskit import *\n",
    "import math\n",
    "from sklearn import datasets\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from pennylane.optimize import NesterovMomentumOptimizer\n",
    "from sklearn.utils import shuffle\n",
    "import random\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import pandas as pdfrom \n",
    "from IPython.display import Image\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#the circuit diagram for the TTN\n",
    "Image(\"ttn.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#initialise pennylane QNode with 4 qubits\n",
    "dev = qml.device(\"default.qubit\", wires=4,shots = 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load data for training and validation\n",
    "iris = datasets.load_iris()\n",
    "idx = iris.target != 2\n",
    "features = iris.data[idx].astype(np.float32)\n",
    "labels = iris.target[idx].astype(np.float32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#split data into training and validation sets\n",
    "np.random.seed(0)\n",
    "num_data = len(features)\n",
    "num_train = int(0.75 * num_data)\n",
    "index = np.random.permutation(range(num_data))\n",
    "features_train = features[index[:num_train]]\n",
    "labels_train = labels[index[:num_train]]\n",
    "features_val = features[index[num_train:]]\n",
    "labels_val = labels[index[num_train:]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "#fuction for encoding features into the amplitudes of the 4 qubits,\n",
    "#the features(b[i]) are scaled to lie between (0,pi/2) and encoded \n",
    "#as cos(b[i])|0 > + sin(b[i])|1 >\n",
    "def statepreparation(b):\n",
    "    for i in range(4):\n",
    "        qml.RY(np.cos(np.pi*b[i]/max(b)), wires=i)\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#paramatrised TTN circuit with 3 layers,returns <Z> of 3rd Qubit. This is used as the label\n",
    "#for the classifiaction\n",
    "@qml.qnode(dev)\n",
    "def circuit(params,a = None):\n",
    "    statepreparation(a)\n",
    "    qml.RY(params[0],wires = 0)\n",
    "    qml.RY(params[1],wires = 1)\n",
    "    qml.RY(params[2],wires = 2)\n",
    "    qml.RY(params[3],wires = 3)\n",
    "    qml.CNOT(wires = [1,0])\n",
    "    qml.CNOT(wires = [3,2])\n",
    "    qml.RY(params[4],wires = 1)\n",
    "    qml.RY(params[5],wires = 2)\n",
    "    qml.CNOT(wires = [1,2])\n",
    "    qml.RY(params[6],wires = 2)\n",
    "    return qml.expval(qml.PauliZ(2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def square_loss(labels, predictions):\n",
    "    loss = 0\n",
    "    for l, p in zip(labels, predictions):\n",
    "        loss = loss + (l - p) ** 2\n",
    "\n",
    "    loss = loss / len(labels)\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def accuracy(labels, predictions):\n",
    "\n",
    "    loss = 0\n",
    "    for l, p in zip(labels, predictions):\n",
    "        if abs(l - p) < 1e-5:\n",
    "            loss = loss + 1\n",
    "    loss = loss / len(labels)\n",
    "\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cost(params,features,labels):\n",
    "    predictions = [circuit(params, a=f) for f in features]\n",
    "    return square_loss(labels, predictions)\n",
    "   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iter:     1 | Cost: 0.3377316 | Acc train: 0.5200000 | Acc validation: 0.4800000 \n",
      "Iter:     2 | Cost: 0.3235450 | Acc train: 0.5333333 | Acc validation: 0.4800000 \n",
      "Iter:     3 | Cost: 0.3050162 | Acc train: 0.5333333 | Acc validation: 0.4800000 \n",
      "Iter:     4 | Cost: 0.2844399 | Acc train: 0.5333333 | Acc validation: 0.4800000 \n",
      "Iter:     5 | Cost: 0.2603515 | Acc train: 0.5466667 | Acc validation: 0.5200000 \n",
      "Iter:     6 | Cost: 0.2355350 | Acc train: 0.5600000 | Acc validation: 0.5200000 \n",
      "Iter:     7 | Cost: 0.2116550 | Acc train: 0.5733333 | Acc validation: 0.5200000 \n",
      "Iter:     8 | Cost: 0.1895039 | Acc train: 0.6000000 | Acc validation: 0.5200000 \n",
      "Iter:     9 | Cost: 0.1694259 | Acc train: 0.6266667 | Acc validation: 0.5600000 \n",
      "Iter:    10 | Cost: 0.1505679 | Acc train: 0.6933333 | Acc validation: 0.6800000 \n",
      "Iter:    11 | Cost: 0.1337834 | Acc train: 0.7333333 | Acc validation: 0.7200000 \n",
      "Iter:    12 | Cost: 0.1190910 | Acc train: 0.8000000 | Acc validation: 0.7600000 \n",
      "Iter:    13 | Cost: 0.1056740 | Acc train: 0.9333333 | Acc validation: 0.8800000 \n",
      "Iter:    14 | Cost: 0.0950432 | Acc train: 0.9733333 | Acc validation: 0.9200000 \n",
      "Iter:    15 | Cost: 0.0858332 | Acc train: 0.9733333 | Acc validation: 1.0000000 \n",
      "Iter:    16 | Cost: 0.0778131 | Acc train: 0.9866667 | Acc validation: 1.0000000 \n",
      "Iter:    17 | Cost: 0.0714136 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    18 | Cost: 0.0656995 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    19 | Cost: 0.0609027 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    20 | Cost: 0.0567063 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    21 | Cost: 0.0526718 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    22 | Cost: 0.0487950 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    23 | Cost: 0.0449816 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    24 | Cost: 0.0411717 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    25 | Cost: 0.0374990 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    26 | Cost: 0.0340785 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    27 | Cost: 0.0309939 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    28 | Cost: 0.0281871 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    29 | Cost: 0.0255703 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    30 | Cost: 0.0233093 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    31 | Cost: 0.0212975 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    32 | Cost: 0.0196395 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    33 | Cost: 0.0182377 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    34 | Cost: 0.0170330 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    35 | Cost: 0.0160590 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    36 | Cost: 0.0152720 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    37 | Cost: 0.0146320 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    38 | Cost: 0.0141187 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    39 | Cost: 0.0137198 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    40 | Cost: 0.0133796 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    41 | Cost: 0.0130449 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    42 | Cost: 0.0127267 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    43 | Cost: 0.0124088 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    44 | Cost: 0.0120854 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    45 | Cost: 0.0117558 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    46 | Cost: 0.0114353 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    47 | Cost: 0.0111266 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    48 | Cost: 0.0108355 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    49 | Cost: 0.0105550 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    50 | Cost: 0.0102815 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    51 | Cost: 0.0100331 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    52 | Cost: 0.0098474 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    53 | Cost: 0.0096587 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    54 | Cost: 0.0094842 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    55 | Cost: 0.0093197 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    56 | Cost: 0.0091699 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    57 | Cost: 0.0090372 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    58 | Cost: 0.0089099 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    59 | Cost: 0.0087844 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Iter:    60 | Cost: 0.0086673 | Acc train: 1.0000000 | Acc validation: 1.0000000 \n",
      "Optimized rotation angles: [ 0.71084077  0.02889237  0.09987318 -0.34253389  0.17154279  0.41787331\n",
      "  0.25096693]\n"
     ]
    }
   ],
   "source": [
    "# initialise the optimizer\n",
    "opt = NesterovMomentumOptimizer(0.01)\n",
    "\n",
    "#batch_sie for stochastic Gradient Descent\n",
    "batch_size = 5\n",
    "\n",
    "# set the number of steps\n",
    "steps = 60\n",
    "# set the initial parameter values\n",
    "init_params = np.random.rand(7)\n",
    "params = init_params\n",
    "cost_list = np.zeros(steps)\n",
    "acc_train_list = np.zeros(steps)\n",
    "acc_val_list = np.zeros(steps)\n",
    "\n",
    "for it in range(steps):\n",
    "    \n",
    "     # Update the weights by one optimizer step\n",
    "    batch_index = np.random.randint(0, num_train, (batch_size,))\n",
    "    features_train_batch = features_train[batch_index]\n",
    "    labels_train_batch = labels_train[batch_index]\n",
    "    params = opt.step(lambda v: cost(v, features_train_batch, labels_train_batch), params)\n",
    "\n",
    "    # Compute predictions on train and validation set\n",
    "    \n",
    "    \n",
    "    predictions_train = np.array([(circuit(params, a=f)) for f in features_train])\n",
    "\n",
    "    #thresholding the predicted labels\n",
    "    predictions_train[predictions_train > 0.5] = 1\n",
    "    predictions_train[predictions_train < 0.5] = 0\n",
    "    predictions_val = np.array([(circuit(params, a=f)) for f in features_val])\n",
    "    predictions_val[predictions_val > 0.5] = 1\n",
    "    predictions_val[predictions_val < 0.5] = 0\n",
    "    \n",
    "    \n",
    "    # Compute accuracy on train and validation set\n",
    "    acc_train_list[it] = accuracy(labels_train, predictions_train)\n",
    "    acc_val_list[it] = accuracy(labels_val, predictions_val)\n",
    "    \n",
    "    cost_list[it] = cost(params, features, labels)\n",
    "    print(\n",
    "        \"Iter: {:5d} | Cost: {:0.7f} | Acc train: {:0.7f} | Acc validation: {:0.7f} \"\n",
    "        \"\".format(it + 1,cost_list[it] , acc_train_list[it], acc_val_list[it])\n",
    "    )\n",
    "\n",
    "print(\"Optimized rotation angles: {}\".format(params))\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'accuracy')"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot of cost function vs time steps\n",
    "plt.plot(range(steps),cost_list,color ='b',linewidth=3)\n",
    "plt.xlabel('number of steps')\n",
    "plt.xlabel('number of steps')\n",
    "plt.ylabel('accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'accuracy')"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot of accuracy vs time steps\n",
    "df=pd.DataFrame({'number of steps': range(steps),  \n",
    "                 'training accuracy': acc_train_list,\n",
    "                 'test accuracy': acc_val_list })\n",
    "\n",
    "plt.plot( 'number of steps', 'training accuracy', data=df,color ='c',linewidth=3)\n",
    "plt.plot('number of steps','test accuracy', data = df,color ='b',linewidth=3)\n",
    "plt.legend()\n",
    "plt.xlabel('number of steps')\n",
    "plt.ylabel('accuracy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "References:\n",
    "\n",
    "https://arxiv.org/abs/1804.03680"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}