tutorials/04_embedding_networks.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Embedding nets for observations\n",
    "\n",
    "!!! note\n",
    "    You can find the original version of this notebook at [tutorials/04_embedding_networks.ipynb](https://github.com/sbi-dev/sbi/blob/main/tutorials/04_embedding_networks.ipynb) in the `sbi` repository.\n",
    "\n",
    "## Introduction\n",
    "\n",
    "When engaging in simulation-based inference, the selection of appropriate summary statistics for observations holds significant importance. These statistics serve to succinctly describe the data generated by simulators, often leveraging domain-specific knowledge. However, in certain scenarios, particularly when dealing with highly complex raw data where domain knowledge may be limited or non-existent, it becomes necessary to learn directly from the data the appropriate summary statistics to employ. `sbi` offers functionality to learn summary statistics from simulation outputs with an **embedding neural network**, referred to as `embedding_net`. \n",
    "\n",
    "When an embedding network is used, the posterior approximation for a given observation $x_o$ can be denoted as $q_\\phi\\big(\\theta \\mid f_\\lambda(x_o)\\big)$ where $\\phi$ are the parameters of the conditional density estimator and $\\lambda$ the parameters of the embedding neural network. Note that the simulation outputs pass through the `embedding_net` before reaching the density estimator and that $\\phi$ and $\\lambda$ are **jointly learned** during training. `sbi` provides pre-configured embedding networks (MLP, CNN, and permutation-invariant networks) or allows to pass custom-written embedding networks.\n",
    "\n",
    "It is worth noting that only `NPE` and `NRE` methods can use an `embedding_net` to learn summary statistics from simulation outputs. `NLE` does not offer such functionality because the simulation outputs are also the output of the neural density estimator. \n",
    "\n",
    "## Main syntax"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```Python\n",
    "# import required modules\n",
    "from sbi.neural_nets import posterior_nn\n",
    "\n",
    "# import the different choices of pre-configured embedding networks\n",
    "from sbi.neural_nets.embedding_nets import (\n",
    "    FCEmbedding,\n",
    "    CNNEmbedding,\n",
    "    PermutationInvariantEmbedding\n",
    ")\n",
    "\n",
    "# choose which type of pre-configured embedding net to use (e.g. CNN)\n",
    "embedding_net = CNNEmbedding(input_shape=(32, 32))\n",
    "\n",
    "# instantiate the conditional neural density estimator\n",
    "neural_posterior = posterior_nn(model=\"maf\", embedding_net=embedding_net)\n",
    "\n",
    "# setup the inference procedure with NPE\n",
    "inferer = NPE(prior=prior, density_estimator=neural_posterior)\n",
    "\n",
    "# train the density estimator\n",
    "density_estimator = inference.append_simulations(theta, x).train()\n",
    "\n",
    "# build the posterior\n",
    "posterior = inference.build_posterior(density_estimator)\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inferring parameters from images\n",
    "\n",
    "In the example that follows, we consider a simple setup where the data points generated by the simulator model are high-dimensional (32x32 grayscale images) and we use a convolutional neural network as summary statistics extractor.\n",
    "\n",
    "First of all, we import all the packages required for running the tutorial"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "from sbi import analysis, utils\n",
    "from sbi.inference import NPE, simulate_for_sbi\n",
    "from sbi.utils.user_input_checks import (\n",
    "    check_sbi_inputs,\n",
    "    process_prior,\n",
    "    process_simulator,\n",
    ")\n",
    "\n",
    "seed = 0\n",
    "torch.manual_seed(seed);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The simulator model\n",
    "\n",
    "The simulator model that we consider has two parameters: $r$ and $\\theta$. On each run, it generates 100 two-dimensional points centered around $(r \\cos(\\theta), r \\sin(\\theta))$ and perturbed by a Gaussian noise with variance 0.01. Instead of simply outputting the $(x,y)$ coordinates of each data point, the model generates a grayscale image of the scattered points with dimensions 32 by 32. This image is further perturbed by an uniform noise with values betweeen 0 and 0.2. The code below defines such model.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def simulator_model(parameter, return_points=False):\n",
    "    \"\"\"Simulator model with two-dimensional input parameter and 1024-D output\n",
    "\n",
    "    This simulator serves as a basic example for using a neural net for learning\n",
    "    summary features. It has only two input parameters but generates\n",
    "    high-dimensional output vectors. The data is generated as follows:\n",
    "        (-) Input:  parameter = [r, phi] (1) Generate 100 two-dimensional\n",
    "        points centered around (r cos(phi),r sin(phi))\n",
    "            and perturbed by a Gaussian noise with variance 0.01\n",
    "        (2) Create a grayscale image I of the scattered points with dimensions\n",
    "            32 by 32\n",
    "        (3) Perturb I with an uniform noise with values betweeen 0 and 0.2\n",
    "        (-) Output: I\n",
    "\n",
    "    Parameters\n",
    "    ----------\n",
    "    parameter : array-like, shape (2)\n",
    "        The two input parameters of the model, ordered as [r, phi]\n",
    "    return_points : bool (default: False)\n",
    "        Whether the simulator should return the coordinates of the simulated\n",
    "        data points as well\n",
    "\n",
    "    Returns\n",
    "    -------\n",
    "    I: torch tensor, shape (1, 1024)\n",
    "        Output flattened image\n",
    "    (optional) points: array-like, shape (100, 2)\n",
    "        Coordinates of the 2D simulated data points\n",
    "\n",
    "    \"\"\"\n",
    "    r = parameter[0]\n",
    "    phi = parameter[1]\n",
    "\n",
    "    sigma_points = 0.10\n",
    "    npoints = 100\n",
    "    points = []\n",
    "    for _ in range(npoints):\n",
    "        x = r * torch.cos(phi) + sigma_points * torch.randn(1)\n",
    "        y = r * torch.sin(phi) + sigma_points * torch.randn(1)\n",
    "        points.append([x, y])\n",
    "    points = torch.as_tensor(points)\n",
    "\n",
    "    nx = 32\n",
    "    ny = 32\n",
    "    sigma_image = 0.20\n",
    "    im = torch.zeros(nx, ny)\n",
    "    for point in points:\n",
    "        pi = int((point[0] - (-1)) / ((+1) - (-1)) * nx)\n",
    "        pj = int((point[1] - (-1)) / ((+1) - (-1)) * ny)\n",
    "        if (pi < nx) and (pj < ny):\n",
    "            im[pi, pj] = 1\n",
    "    im = im + sigma_image * torch.rand(nx, ny)\n",
    "    im = im.T\n",
    "    im = im.reshape(1, -1)\n",
    "\n",
    "    if return_points:\n",
    "        return im, points\n",
    "    else:\n",
    "        return im"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The figure below shows an example of the output of the simulator when $r = 0.70$ and $\\theta = \\pi/4$\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1115x561 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# simulate samples\n",
    "true_parameter = torch.tensor([0.70, torch.pi / 4])\n",
    "x_observed, x_points = simulator_model(true_parameter, return_points=True)\n",
    "\n",
    "# plot the observation\n",
    "fig, ax = plt.subplots(\n",
    "    facecolor=\"white\", figsize=(11.15, 5.61), ncols=2, constrained_layout=True\n",
    ")\n",
    "circle = plt.Circle((0, 0), 1.0, color=\"k\", ls=\"--\", lw=0.8, fill=False)\n",
    "ax[0].add_artist(circle)\n",
    "ax[0].scatter(x_points[:, 0], x_points[:, 1], s=20)\n",
    "ax[0].set_xlabel(\"x\")\n",
    "ax[0].set_ylabel(\"y\")\n",
    "ax[0].set_xlim(-1, +1)\n",
    "ax[0].set_xticks([-1, 0.0, +1.0])\n",
    "ax[0].set_ylim(-1, +1)\n",
    "ax[0].set_yticks([-1, 0.0, +1.0])\n",
    "ax[0].set_title(r\"original simulated points with $r = 0.70$ and $\\phi = \\pi/4$\")\n",
    "ax[1].imshow(x_observed.view(32, 32), origin=\"lower\", cmap=\"gray\")\n",
    "ax[1].set_xticks([])\n",
    "ax[1].set_yticks([])\n",
    "ax[1].set_title(\"noisy observed data (gray image with 32 x 32 pixels)\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Choosing an `embedding_net`\n",
    "\n",
    "The outputs $x$ from the simulator are defined in a 1024 dimensional space (32 x 32 = 1024). To avoid having to setup a conditional neural density estimator to work directly on such high-dimensional vectors, one could use an `embedding_net` that would take the images as input and encode them into smaller vectors.\n",
    "\n",
    "`sbi` provides pre-configured embedding networks of the following types:\n",
    "\n",
    "- Fully-connected multi-layer perceptron\n",
    "- Convolutional neural network (1D and 2D convolutions)\n",
    "- Permutation-invariant neural network (for trial-based data, see [here](https://sbi-dev.github.io/sbi/latest/tutorials/12_iid_data_and_permutation_invariant_embeddings/))\n",
    "\n",
    "In the example considered here, the most appropriate `embedding_net` would be a CNN for two-dimensional images. We can setup it as per:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sbi.neural_nets.embedding_nets import CNNEmbedding\n",
    "\n",
    "embedding_net = CNNEmbedding(\n",
    "    input_shape=(32, 32),\n",
    "    in_channels=1,\n",
    "    out_channels_per_layer=[6],\n",
    "    num_conv_layers=1,\n",
    "    num_linear_layers=1,\n",
    "    output_dim=8,\n",
    "    kernel_size=5,\n",
    "    pool_kernel_size=8\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "!!! note\n",
    "    See [here](https://github.com/sbi-dev/sbi/blob/main/sbi/neural_nets/embedding_nets.py) for details on all hyperparametes for each available embedding net in `sbi`\n",
    "\n",
    "## The inference procedure\n",
    "\n",
    "With the `embedding_net` defined and instantiated, we can follow the usual workflow of an inference procedure in `sbi`. The `embedding_net` object appears as an input argument when instantiating the neural density estimator with `sbi.neural_nets.posterior_nn`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# set prior distribution for the parameters\n",
    "prior = utils.BoxUniform(\n",
    "    low=torch.tensor([0.0, 0.0]), high=torch.tensor([1.0, 2 * torch.pi])\n",
    ")\n",
    "\n",
    "# make a SBI-wrapper on the simulator object for compatibility\n",
    "prior, num_parameters, prior_returns_numpy = process_prior(prior)\n",
    "simulator_wrapper = process_simulator(simulator_model, prior, prior_returns_numpy)\n",
    "check_sbi_inputs(simulator_wrapper, prior)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sbi.neural_nets import posterior_nn\n",
    "\n",
    "# instantiate the neural density estimator\n",
    "neural_posterior = posterior_nn(model=\"maf\", embedding_net=embedding_net)\n",
    "\n",
    "# setup the inference procedure with NPE\n",
    "inferer = NPE(prior=prior, density_estimator=neural_posterior)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# run the inference procedure on one round and 10000 simulated data points\n",
    "theta, x = simulate_for_sbi(simulator_wrapper, prior, num_simulations=10_000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Neural network successfully converged after 96 epochs."
     ]
    }
   ],
   "source": [
    "density_estimator = inferer.append_simulations(theta, x).train(training_batch_size=256)\n",
    "posterior = inferer.build_posterior(density_estimator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualizing the results\n",
    "\n",
    "We now generate 50000 samples of the posterior distribution of $r$ and $\\theta$ when observing an input data point $x$ generated from the `simulator model` with $r = 0.70$ and $\\theta = \\pi/4$.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# generate posterior samples\n",
    "true_parameter = torch.tensor([0.50, torch.pi / 4])\n",
    "x_observed = simulator_model(true_parameter)\n",
    "samples = posterior.set_default_x(x_observed).sample((50000,))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The figure below shows the statistics of the generated samples.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# create the figure\n",
    "fig, ax = analysis.pairplot(\n",
    "    samples,\n",
    "    points=true_parameter,\n",
    "    labels=[\"r\", r\"$\\phi$\"],\n",
    "    limits=[[0, 1], [0, 2 * torch.pi]],\n",
    "    fig_kwargs=dict(\n",
    "        points_colors=\"r\",\n",
    "        points_offdiag={\"markersize\": 6},\n",
    "    ),\n",
    "    figsize=(5, 5),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Defining custom embedding networks\n",
    "\n",
    "It is also possible to define custom embedding networks and pass those to neural density estimator. For example, we could have defined our own architecture for the CNN as per"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "class SummaryNet(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        # 2D convolutional layer\n",
    "        self.conv1 = nn.Conv2d(in_channels=1, out_channels=6, kernel_size=5, padding=2)\n",
    "        # Maxpool layer that reduces 32x32 image to 4x4\n",
    "        self.pool = nn.MaxPool2d(kernel_size=8, stride=8)\n",
    "        # Fully connected layer taking as input the 6 flattened output arrays\n",
    "        # from the maxpooling layer\n",
    "        self.fc = nn.Linear(in_features=6 * 4 * 4, out_features=8)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = x.view(-1, 1, 32, 32)\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        x = x.view(-1, 6 * 4 * 4)\n",
    "        x = F.relu(self.fc(x))\n",
    "        return x\n",
    "\n",
    "# instantiate the custom embedding_net\n",
    "embedding_net_custom = SummaryNet()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}