Tutorial_Symmetry.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Tutorial: Incorporating Symmetry into Deep Dynamics Models\n",
        "(August 2023) Author: Rui Wang\n",
        "\n",
        "# [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/iaifi/summer-school-2023/blob/main/Tutorial_Symmetry.ipynb)\n",
        "\n",
        "This google colab tutorial is supporting material for the lectures by Robin Walters on *Incorporating Symmetry into Deep Dynamics Models* at [IAIFI Summer School](https://iaifi.org/phd-summer-school.html). We will 1) build regular group  convolution neural net layers; 2) define approximate equivariance and implement relaxed group convolution for translation group and C4 rotation group; 3) introduce basic ideas of steerable convolutions and [e2cnn](https://github.com/QUVA-Lab/e2cnn). 4) introduce several methods that can improve models' long-term dynmaics prediction accuracy.\n",
        "\n",
        "This tutorial is based on the following materials:\n",
        "\n",
        "* [Dr. Denis Boyda's tutorial](https://www.dropbox.com/sh/450ao3m5v4qt6o3/AABe1B2jEcyM5ZyoY_pmOStfa?dl=0) from IAIFI Summer School 2022.\n",
        "\n",
        "* [UvA Deep Learning Tutorials](https://uvadlc-notebooks.readthedocs.io/en/latest/index.html).\n",
        "\n",
        "* Prof. Erik Bekkers's course [Group Equivariant Deep Learning](https://www.youtube.com/@erikbekkers6398/videos).\n",
        "\n",
        "* Cohen et al. [Group Equivariant Convolutional Networks](https://arxiv.org/pdf/1602.07576.pdf).\n",
        "\n",
        "* Cohen et al. [Steerable CNNs](https://arxiv.org/abs/1612.08498).\n",
        "\n",
        "* [E2CNN](https://github.com/QUVA-Lab/e2cnn) & [ESCNN](https://github.com/QUVA-Lab/escnn) & [E3NN](https://e3nn.org/)\n",
        "\n",
        "* Wang and Walters et al. [Incorporating Symmetry into Deep Dynamics Models for Improved Generalization](https://arxiv.org/abs/2002.03061).\n",
        "\n",
        "* Wang and Walters et al. [Approximately Equivariant Networks for Imperfectly Symmetric Dynamics](https://proceedings.mlr.press/v162/wang22aa.html).\n"
      ],
      "metadata": {
        "id": "N2dSAwLACjSN"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 1. Group Equivariant Convolutional Networks"
      ],
      "metadata": {
        "id": "7YNFbmARCvTf"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.1 Group, Equivariance and Group Equivariant Convolutional Networks"
      ],
      "metadata": {
        "id": "p2xNgzMjGpeP"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.1.1 Definition of group\n",
        "\n",
        "A Group (or Abstract group) $G = (\\Omega, \\cdot)$ is a set with a binary operation, called group multiplication, that satisfies the following four axioms:\n",
        "\n",
        "1.   <ins>Closure</ins>: For any $g_1, g_2 \\in G$ the product $g_1 \\cdot g_2 = g_3 \\in G$.\n",
        "2.   <ins>Associativity</ins>: For any $g_1, g_2, g_3 \\in G$: ($g_1 \\cdot g_2) \\cdot g_3 = g_1 \\cdot (g_2 \\cdot g_3)$.\n",
        "3.   <ins>Identity</ins>: There exists an identity element $e$ s.t. $e \\cdot g = g \\cdot e = g$ for each $g \\in G$.\n",
        "4.   <ins>Inverse</ins>: For any $g \\in G$ there exists an inverse element $g^{-1} \\in G$, s.t. $g \\cdot g^{-1} = e$.\n",
        "\n",
        "E.g. Translation Group $(R^2, +)$ consists of all possible translations in $R^2$ and is equipped with group product and group inverse,\n",
        "\n",
        "$$g \\cdot g' = (x + x') ; \\;\\;\\ g^{-1} = -x $$\n",
        "\n",
        "with $g = (x), g' = (x')$ and $x, x' \\in \\mathbb{R}^2$.\n",
        "\n",
        "E.g. Roto-translation group $SE(2) = \\mathbb{R}^2 \\rtimes SO(2)$ consists of the coupled space of translation vectors and rotations and is equipped with group product and group inverse,\n",
        "$$g \\cdot g' = (x, R_{\\theta}) \\cdot (x', R_{\\theta'}) = (R_{\\theta}x' + x, R_{\\theta+\\theta'}) ; \\;\\;\\ g^{-1} = (-R_{\\theta}^{-1}x, -R_{\\theta}^{-1}) $$\n",
        "\n",
        "with $g = (x, R_{\\theta}), g' = (x', R_{\\theta'})$."
      ],
      "metadata": {
        "id": "DGxx9GLXY6_M"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.1.2 Equivariance and Invariance\n",
        "\n",
        "Formally, a function $f \\colon X \\to Y$ may be described as respecting the symmetry coming from a group $G$ using the notion of equivariance.  Assume a group representation $\\rho_{\\text{in}}$ of $G$ acts on $X$ and $\\rho_{\\text{out}}$ acts on $Y$. We say a function $f$ is <ins>$G$-equivariant</ins> if\n",
        "\n",
        "$$ f( \\rho_{\\text{in}}(g)(x)) = \\rho_{\\text{out}}(g) f(x) $$\n",
        "for all $x \\in X$ and $g \\in G$. The function $f$ is <ins>$G$-invariant</ins> if $f( \\rho_{\\text{out}}(g)(x)) = f(x)$ for all $x \\in X$ and $g \\in G$.  This is a special case of equivariance for the case $\\rho_{\\mathrm{out}}(g) = 1$.\n"
      ],
      "metadata": {
        "id": "EdQvrsc4aPzR"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![picture](https://drive.google.com/uc?export=view&id=1SVJOPGAzLo3D5BYdXT-uc5iGm1VLEAAE)"
      ],
      "metadata": {
        "id": "ccnwN7fpOYgo"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.1.3 Recap on CNN\n",
        "\n",
        "<ins>CNN is nothing else but group convolution with the translation group. </ins>\n",
        "\n",
        "At each layer of a regular convnet convolves or correlates stack of feature maps $f : \\mathbb{Z}^2 \\rightarrow \\mathbb{R}^K$ with a filter (or a set of filters) $\\psi : \\mathbb{Z}^2 \\rightarrow \\mathbb{R}^K$.\n",
        "\n",
        "\n",
        "$$[f * \\psi] (\\mathbf{x}) = \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2}f(\\mathbf{y})\\psi(\\mathbf{y}-\\mathbf{x}) = \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2}f(\\mathbf{y})T_{\\mathbf{x}}\\psi(\\mathbf{y}) $$\n",
        "\n",
        "A left regular representation is a representation that transforms the function $f$ by transforming their domains via inverse group actions."
      ],
      "metadata": {
        "id": "5d7OEDyP1e6s"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.1.4 Group Equivariant Convolution"
      ],
      "metadata": {
        "id": "cFQhHnEo--_2"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "<ins>G-CNN is a generalization of CNN for the case of an arbitrary group $G$.</ins>  \n",
        "\n",
        "By replacing the shift with a more general transformation from some group $G$, We can make models become equivariant to other groups.\n",
        "\n",
        "* The first layer of a G-CNN (Lifting Convolution): $$[f * \\psi] (g) =  \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2}f(\\mathbf{y})\\psi(g^{-1}\\mathbf{y})$$\n",
        "\n",
        "* For all layers after the first, both the input and the filters are functions on $G$:\n",
        "$$[f * \\psi] (g) =  \\sum_{h \\in G}f(h)\\psi(g^{-1}h)$$"
      ],
      "metadata": {
        "id": "9HVYP4R7_pJ7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "<ins>E.g.</ins> $G=\\mathbb{Z}^2 \\rtimes C_4$\n",
        "\n",
        "* <ins>Lifting Convolution</ins> maps signals $f_{in}: \\mathbb{Z}^2 → R$ to signals on arbitrary group G: $f_{out}: G→R$.  Group elements $g \\in G = \\mathbb{Z}^2 \\rtimes C_4$ can be parametrized using $\\mathbf{x} \\in R^2$ and $r \\in C_4$ such that $g=(\\mathbf{x}, r)$. For such groups lifting convolution has form\n",
        "$$\n",
        "\\begin{aligned}\n",
        "f_{out}(\\mathbf{x}, r) = (f_{in} * \\psi) (\\mathbf{x})\n",
        "    &=\n",
        "\\sum_{\\mathbf{y} \\in \\mathbb{Z}^2} f_{in}(\\mathbf{y}) L_{r}  L_{\\mathbf{x}} \\psi(\\mathbf{y}),\\\\\n",
        "    &=\n",
        "    \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2}  f_{in}(\\mathbf{y}) L_{r} \\psi(\\mathbf{y} -\\mathbf{x}),\\\\\n",
        "    &=\n",
        "    \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2}  f_{in}(\\mathbf{y})\\psi_{r}(\\mathbf{y} -\\mathbf{x}),\\\n",
        "\\end{aligned}\n",
        "$$\n",
        "<ins>So lifting convolution is the same as conventional convolutions with a stack of filter bank </ins>$\\{k_{r}(\\mathbf{x})\\}_{r \\in C_4}$."
      ],
      "metadata": {
        "id": "1uNB5aq6EVzH"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "* <ins>Group Convolution:</ins> Note that the input feature map, $f_{in}$, now has one additional group dimension defined over $C_4$, besides the usual spatial dimensions defined over $\\mathbb{R}^2$,\n",
        "\n",
        "$$\n",
        "\\begin{aligned}\n",
        "f_{out}(\\mathbf{x}, r) = (f_{in} * \\psi) (\\mathbf{x}, r)\n",
        "    &=\n",
        "\\sum_{\\mathbf{y} \\in \\mathbb{Z}^2} \\sum_{r' \\in C^4} f_{in}(\\mathbf{y} , r) L_{r} L_{\\mathbf{x}} \\psi(\\mathbf{y},r'),\\\\\n",
        "    &=\n",
        "    \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2} \\sum_{r' \\in C^4} f_{in}(\\mathbf{y} , r')  L_{r} \\psi(\\mathbf{y-x},r'),\\\\\n",
        "    &=\n",
        "    \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2} \\sum_{r' \\in C^4}  f_{in}(\\mathbf{y}, r') \\psi_r(\\mathbf{y-x},r')\\\\\n",
        "    &=\n",
        "    \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2} \\sum_{r' \\in C^4}  f_{in}(\\mathbf{y}, r') \\psi(r^{-1}(\\mathbf{y-x}),r^{-1}r')\\\\\n",
        "\\end{aligned}\n",
        "$$"
      ],
      "metadata": {
        "id": "aEAK-TdRWn65"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.2 C4 Cyclic Group"
      ],
      "metadata": {
        "id": "V31DXglsOlXQ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Let's first implement a function that can rotate images**"
      ],
      "metadata": {
        "id": "oRpZpbuvImdI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "from PIL import Image\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import torch\n",
        "import torchvision\n",
        "from torch import nn\n",
        "from torch import optim\n",
        "import torch.nn.functional as F"
      ],
      "metadata": {
        "id": "fMrerIAaMr5N"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def rot_img(x, theta):\n",
        "    \"\"\" Rotate 2D images\n",
        "    Args:\n",
        "        x : input images with shape [N, C, H, W]\n",
        "        theta: angle\n",
        "    Returns:\n",
        "        rotated images\n",
        "    \"\"\"\n",
        "    # Rotation Matrix (2 x 3)\n",
        "    rot_mat = torch.FloatTensor([[np.cos(theta), -np.sin(theta), 0],\n",
        "                                 [np.sin(theta), np.cos(theta), 0]])\n",
        "\n",
        "    # The affine transformation matrices should have the shape of N x 2 x 3\n",
        "    rot_mat = rot_mat.repeat(x.shape[0],1,1)\n",
        "\n",
        "    # Obtain transformed grid\n",
        "    # grid is the coordinates of pixels for rotated image\n",
        "    # F.affine_grid assumes the origin is in the middle\n",
        "    # and it rotates the positions of the coordinates\n",
        "    # r(f(x)) = f(r^-1 x)\n",
        "    grid = F.affine_grid(rot_mat, x.size(), align_corners=False).float()\n",
        "    x = F.grid_sample(x, grid)\n",
        "    return x.float()"
      ],
      "metadata": {
        "id": "gogWJEQMNHQI"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# C4 rotations: e, r, r^2, r^3\n",
        "x = torch.rand(1,1,5,5)\n",
        "fig=plt.figure(figsize=(20, 5))\n",
        "for i in range(1, 5):\n",
        "    fig.add_subplot(1, 4, i)\n",
        "    plt.imshow(rot_img(x, np.pi/2*i)[0,0])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        },
        "id": "aOHncymzNPo9",
        "outputId": "8442fc2b-6a38-4eb6-8b6a-f8ee454998de"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x500 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.3 Implementing the Lift Convolution"
      ],
      "metadata": {
        "id": "vDNYJxKIPn9b"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "$$f_{out}(\\mathbf{x}, r) = (f_{in} * \\psi) (\\mathbf{x})\n",
        "= \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2}  f_{in}(\\mathbf{y})\\psi_{r}(\\mathbf{y} -\\mathbf{x})$$"
      ],
      "metadata": {
        "id": "sWpduUdeCEYR"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**In lift convolution, the input convolves with a stack of rotated filters.**\n",
        "\n",
        "*The reason why we call it lifting is that we add a new axis $r$ in the output feature maps.*"
      ],
      "metadata": {
        "id": "d2IlLWUZMsyd"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![picture](https://drive.google.com/uc?export=view&id=17RprAkuSe43q_gjsi1al2RX18mZ0o00q)"
      ],
      "metadata": {
        "id": "Icc31YdfEcXn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class LiftingConvolution(nn.Module):\n",
        "    \"\"\"Lifting Convolution Layer for finite rotation group\n",
        "\n",
        "    Attributes:\n",
        "        in_channels: number of input channels\n",
        "        out_channels: number of output channels\n",
        "        kernel_size: kernel size\n",
        "        group_order: the order of rotation groups\n",
        "        activation: whether to use relu.\n",
        "    \"\"\"\n",
        "    def __init__(self,\n",
        "                 in_channels,\n",
        "                 out_channels,\n",
        "                 kernel_size,\n",
        "                 group_order,\n",
        "                 activation = True\n",
        "                 ):\n",
        "        super(LiftingConvolution, self).__init__()\n",
        "\n",
        "        self.in_channels = in_channels\n",
        "        self.out_channels = out_channels\n",
        "        self.kernel_size = kernel_size\n",
        "        self.group_order = group_order\n",
        "        self.activation = activation\n",
        "\n",
        "        # Initialize an unconstrained kernel.\n",
        "        self.weight = torch.nn.Parameter(torch.zeros(self.out_channels,\n",
        "                                                     self.in_channels,\n",
        "                                                     self.kernel_size,\n",
        "                                                     self.kernel_size))\n",
        "\n",
        "        # Initialize weights\n",
        "        stdv = np.sqrt(1/(self.in_channels*self.kernel_size*self.kernel_size))\n",
        "        self.weight.data.uniform_(-stdv, stdv)\n",
        "\n",
        "    def generate_filter_bank(self):\n",
        "        # Obtain a stack of rotated filters\n",
        "        # Rotate kernels by 0, 90, 180, and 270 degrees\n",
        "        # ==============================\n",
        "        filter_bank = torch.stack([rot_img(self.weight, -np.pi*2/self.group_order*i)\n",
        "                                   for i in range(self.group_order)])\n",
        "        # ==============================\n",
        "\n",
        "        # [#out, group_order, #in, k, k]\n",
        "        filter_bank = filter_bank.transpose(0,1)\n",
        "        return filter_bank\n",
        "\n",
        "\n",
        "    def forward(self, x):\n",
        "        # input shape: [bz, #in, h, w]\n",
        "        # output shape: [bz, #out, group order, h, w]\n",
        "\n",
        "        # generate filter bank given input group order\n",
        "        filter_bank = self.generate_filter_bank()\n",
        "\n",
        "        # concatenate the first two dims before convolution.\n",
        "        # ==============================\n",
        "        x = F.conv2d(\n",
        "            input=x,\n",
        "            weight=filter_bank.reshape(\n",
        "                self.out_channels * self.group_order,\n",
        "                self.in_channels,\n",
        "                self.kernel_size,\n",
        "                self.kernel_size\n",
        "            ),\n",
        "            padding = (self.kernel_size-1)//2\n",
        "        )\n",
        "        # ==============================\n",
        "\n",
        "        # reshape output signal to shape [bz, #out, group order, h, w].\n",
        "        # ==============================\n",
        "        x = x.view(\n",
        "            x.shape[0],\n",
        "            self.out_channels,\n",
        "            self.group_order,\n",
        "            x.shape[-1],\n",
        "            x.shape[-2]\n",
        "        )\n",
        "        # ==============================\n",
        "        if self.activation:\n",
        "            return F.relu(x)\n",
        "        return x"
      ],
      "metadata": {
        "id": "j9UENyEg0ucj"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lifting_layer = LiftingConvolution(in_channels = 1,\n",
        "                                   out_channels = 1,\n",
        "                                   kernel_size = 3,\n",
        "                                   group_order = 4,\n",
        "                                   activation = True)\n",
        "\n",
        "x = torch.rand(1, 1, 64, 64)\n",
        "out = lifting_layer(x)\n",
        "print(out.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QH5y_IPK0346",
        "outputId": "19d9269c-60d5-4fae-8e49-99c53abffdd5"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([1, 1, 4, 64, 64])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.4 Implementing Group Convolution Kernel"
      ],
      "metadata": {
        "id": "AQKnUvTTQhZx"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**After the first layer, both the filter and input are now functions on $G=\\mathbb{Z}^2 \\rtimes C_4$**\n",
        "\n",
        "*Be careful about this, rotation not only acts on the spatial dimension but also the new $r$ axis. A rotation means a planar rotation and periodic shft.*\n",
        "\n",
        "*A single channel from regular CNNs is expanded to four channels since we apply kernel with rotations four times.*"
      ],
      "metadata": {
        "id": "B__anHOmzvHD"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "$$f_{out}(\\mathbf{x}, r) = (f_{in} * \\psi) (\\mathbf{x}, r)\n",
        "    = \\sum_{\\mathbf{y} \\in \\mathbb{Z}^2} \\sum_{r' \\in C^4}  f_{in}(\\mathbf{y}, r') \\psi(r^{-1}(\\mathbf{y-x}),r^{-1}r')\n",
        "$$\n"
      ],
      "metadata": {
        "id": "Lj36BPj1CLaO"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![picture](https://drive.google.com/uc?export=view&id=1ADYVLN8veOHRgK_l50p7Vw3-Nvn2E0P9)\n"
      ],
      "metadata": {
        "id": "PZ_bF30ScLwX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class GroupConvolution(nn.Module):\n",
        "    \"\"\"Group Convolution Layer for finite rotation group\n",
        "\n",
        "    Attributes:\n",
        "        in_channels: number of input channels\n",
        "        out_channels: number of output channels\n",
        "        kernel_size: kernel size\n",
        "        group_order: the order of rotation groups\n",
        "        activation: whether to use relu.\n",
        "    \"\"\"\n",
        "    def __init__(self,\n",
        "                 in_channels,\n",
        "                 out_channels,\n",
        "                 kernel_size,\n",
        "                 group_order,\n",
        "                 activation = True,\n",
        "                 ):\n",
        "        super(GroupConvolution, self).__init__()\n",
        "        self.in_channels = in_channels\n",
        "        self.out_channels = out_channels\n",
        "        self.kernel_size = kernel_size\n",
        "        self.group_order = group_order\n",
        "        self.activation = activation\n",
        "\n",
        "        # Initialize an unconstrained kernel.\n",
        "        # the weights have an additional group order dimension.\n",
        "        self.weight = torch.nn.Parameter(torch.zeros(self.out_channels,\n",
        "                                                     self.in_channels,\n",
        "                                                     self.group_order, # this is different from the lifting convolution\n",
        "                                                     self.kernel_size,\n",
        "                                                     self.kernel_size))\n",
        "\n",
        "        stdv = np.sqrt(1/(self.in_channels*self.kernel_size*self.kernel_size))\n",
        "        self.weight.data.uniform_(-stdv, stdv)\n",
        "\n",
        "    def generate_filter_bank(self):\n",
        "        # Obtain a stack of rotated and cyclic shifted filters\n",
        "        filter_bank = []\n",
        "        weights = self.weight.reshape(self.out_channels*self.in_channels,\n",
        "                                     self.group_order,\n",
        "                                     self.kernel_size,\n",
        "                                     self.kernel_size)\n",
        "\n",
        "        for i in range(self.group_order):\n",
        "            # planar rotation\n",
        "            rotated_filter = rot_img(weights, -np.pi*2/self.group_order*i)\n",
        "\n",
        "            # cyclic shift\n",
        "            shifted_indices = torch.roll(torch.arange(0, self.group_order, 1), shifts = i)\n",
        "            shifted_rotated_filter = rotated_filter[:,shifted_indices]\n",
        "\n",
        "\n",
        "            filter_bank.append(shifted_rotated_filter.reshape(self.out_channels,\n",
        "                                                            self.in_channels,\n",
        "                                                            self.group_order,\n",
        "                                                            self.kernel_size,\n",
        "                                                            self.kernel_size))\n",
        "        # stack\n",
        "        # reshape output signal to shape [#out, g_order, #in, g_order, k, k].\n",
        "        filter_bank = torch.stack(filter_bank).transpose(0,1)\n",
        "        return filter_bank\n",
        "\n",
        "    def forward(self, x):\n",
        "        # input shape: [bz, in, group order, x, y]\n",
        "        # output shape: [bz, out, group order, x, y]\n",
        "\n",
        "        # Generate filter bank with shape [#out, g_order, #in, g_order, h, w]\n",
        "        filter_bank = self.generate_filter_bank()\n",
        "\n",
        "        # Reshape filter_bank to use F.conv2d\n",
        "        # [#out, g_order, #in, g_order, h, w] -> [#out*g_order, #in*g_order, h, w]\n",
        "        # ==============================\n",
        "        x = torch.nn.functional.conv2d(\n",
        "            input=x.reshape(\n",
        "                x.shape[0],\n",
        "                x.shape[1] * x.shape[2],\n",
        "                x.shape[3],\n",
        "                x.shape[4]\n",
        "                ),\n",
        "            weight=filter_bank.reshape(\n",
        "                self.out_channels * self.group_order,\n",
        "                self.in_channels * self.group_order,\n",
        "                self.kernel_size,\n",
        "                self.kernel_size\n",
        "            ),\n",
        "            padding = (self.kernel_size-1)//2\n",
        "        )\n",
        "\n",
        "        # Reshape signal back [bz, #out * g_order, h, w] -> [bz, out, g_order, h, w]\n",
        "        x = x.view(x.shape[0], self.out_channels, self.group_order, x.shape[-2], x.shape[-1])\n",
        "        # ========================\n",
        "\n",
        "        if self.activation:\n",
        "          return F.relu(x)\n",
        "        return x"
      ],
      "metadata": {
        "id": "dzTItCkechNT"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "groupconv_layer = GroupConvolution(in_channels = 1,\n",
        "                                   out_channels = 1,\n",
        "                                   kernel_size = 5,\n",
        "                                   group_order = 4,\n",
        "                                   activation = True)\n",
        "# The input has an additional dimension.\n",
        "x = torch.rand(1, 1, 4, 64, 64)\n",
        "out = groupconv_layer(x)\n",
        "print(out.shape)"
      ],
      "metadata": {
        "id": "YSUrd378cocX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fd7d5acf-1447-40be-b1c3-4455012413d9"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([1, 1, 4, 64, 64])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# C4 rotations: e, r, r^2, r^3\n",
        "filter_bank = groupconv_layer.generate_filter_bank()\n",
        "fig=plt.figure(figsize=(10, 5))\n",
        "i = 4\n",
        "for j in range(1, 5):\n",
        "    fig.add_subplot(1, 4, j)\n",
        "    plt.imshow(filter_bank[0,j-1,0,j-1].data.numpy())\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])"
      ],
      "metadata": {
        "id": "qZLLS4w3cpwb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 205
        },
        "outputId": "1cd4be0d-57c9-42eb-bda9-8da024820fcf"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x500 with 4 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.5 Test Equivariance"
      ],
      "metadata": {
        "id": "hLppDfHeQx8_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "$$ \\|f\\|_{EE} = sup_{x,g}\\|f(\\rho_X(g)x) - \\rho_Y(g)f(x)\\| = 0$$"
      ],
      "metadata": {
        "id": "GZA_6iEseHZA"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Let's build a group equivariant CNNs.**"
      ],
      "metadata": {
        "id": "UR6gszpjKT1Z"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class GroupEquivariantCNN(torch.nn.Module):\n",
        "\n",
        "    def __init__(self,\n",
        "                 in_channels,\n",
        "                 out_channels,\n",
        "                 kernel_size,\n",
        "                 hidden_dim,\n",
        "                 group_order,\n",
        "                 num_gconvs # number of group convolution layers.\n",
        "                 ):\n",
        "        super().__init__()\n",
        "\n",
        "        # First Layer\n",
        "        self.lifting_conv = LiftingConvolution(in_channels = in_channels,\n",
        "                                               out_channels = hidden_dim,\n",
        "                                               kernel_size = kernel_size,\n",
        "                                               group_order = group_order,\n",
        "                                               activation = True)\n",
        "        # Middle Layers\n",
        "        self.gconvs = []\n",
        "        for i in range(num_gconvs):\n",
        "            self.gconvs.append(GroupConvolution(in_channels = hidden_dim,\n",
        "                                                out_channels = hidden_dim,\n",
        "                                                kernel_size = kernel_size,\n",
        "                                                group_order = group_order,\n",
        "                                                activation = True))\n",
        "        self.gconvs = nn.Sequential(*self.gconvs)\n",
        "\n",
        "\n",
        "        # Final Layer\n",
        "        # To generate equivariant outputs\n",
        "        self.final_layer = GroupConvolution(in_channels = hidden_dim,\n",
        "                                            out_channels = out_channels,\n",
        "                                            kernel_size = kernel_size,\n",
        "                                            group_order = group_order,\n",
        "                                            activation = False)\n",
        "\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = self.lifting_conv(x)\n",
        "\n",
        "        out = self.gconvs(out)\n",
        "\n",
        "        # functions on (g,x,y) -> functions on (x,y)\n",
        "        out = torch.mean(self.final_layer(out), dim = 2)\n",
        "\n",
        "        # If we want to have a invariant classifer, we can average over the last three dimensions.\n",
        "        # out = F.sigmoid(torch.mean(self.final_layer(out), dim = (2,3,4)))\n",
        "        return out"
      ],
      "metadata": {
        "id": "skRK8RplcuTm"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "GCNN = GroupEquivariantCNN(in_channels = 1,\n",
        "                           out_channels = 1,\n",
        "                           kernel_size = 3,\n",
        "                           hidden_dim = 4,\n",
        "                           group_order = 4,\n",
        "                           num_gconvs = 1)"
      ],
      "metadata": {
        "id": "Fp6RplwdcwNS"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Test whether the model is equivariant under C4 transformations."
      ],
      "metadata": {
        "id": "U-n_Xhxr_i_6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Download mnist dataset.\n",
        "train_ds = torchvision.datasets.MNIST(root=\"data\", train=True, download=True)\n",
        "image = torch.from_numpy(train_ds.data[0].numpy()).unsqueeze(0).unsqueeze(0).float()\n",
        "image = (image - image.min())/(image.max() - image.min())\n",
        "\n",
        "\n",
        "## |rf(x) - f(rx)|\n",
        "# test whether the model is equivariant under C4 transformations.\n",
        "for i in range(4):\n",
        "    theta = np.pi/2*i\n",
        "    x = image\n",
        "    f_x = GCNN(x)\n",
        "    r_f_x = rot_img(f_x, theta).data.numpy()\n",
        "    r_x = rot_img(x, theta)\n",
        "    f_r_x = GCNN(r_x).data.numpy()\n",
        "\n",
        "    equiv_error = np.mean(np.abs(r_f_x - f_r_x))\n",
        "    print(\"Equivariance Error with Rotation {:0.0f} Degrees: {:0.5f}\".format(theta/np.pi/2*360, equiv_error))\n",
        "\n"
      ],
      "metadata": {
        "id": "xKuiLFe7NYsP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can also visualize $f(rx)$ and $r(f(x))$ and compare them."
      ],
      "metadata": {
        "id": "05Dfl0MO--cl"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "titles = [\"$x$\", \"$f(x)$\", \"$gx$\", \"$f(gx)$\", \"$g(f(x))$\"]\n",
        "vis_imgs = [x[0,0], f_x[0,0].data.numpy(), r_x[0,0].data.numpy(), f_r_x[0,0], r_f_x[0,0]]\n",
        "fig=plt.figure(figsize=(10, 5))\n",
        "idx = 0\n",
        "for i in range(6):\n",
        "    if i == 1:\n",
        "        continue\n",
        "    fig.add_subplot(2, 3, i+1)\n",
        "    plt.imshow(vis_imgs[idx])\n",
        "    plt.title(titles[idx], fontsize = 20)\n",
        "    plt.axis('off')\n",
        "    idx += 1\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 453
        },
        "id": "1UG7e0FggOgu",
        "outputId": "96e3e2b1-0292-41e7-cb6b-48b9681b3b66"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x500 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 2.  Approximately Equivariant Networks"
      ],
      "metadata": {
        "id": "YQBueEWbDLlg"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 2.1.1 Situations where perfectly equivariant models may not work"
      ],
      "metadata": {
        "id": "iUrJKGvtJUfU"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**1) If the image of 6 is perfectly rotated version of 9, Can GCNN distinguish them?**"
      ],
      "metadata": {
        "id": "lmDJuWxv3H7H"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_6, x_9 = np.zeros((16, 16)), np.zeros((16, 16))\n",
        "points_6 = [(5, 5),(5, 6),(5, 7),(5, 8),(5, 9),(5, 10),\n",
        "            (6, 5),(6, 10),(7, 5),(7, 10),(8, 5),(8, 6),\n",
        "            (8, 7),(8, 8),(8, 9),(8, 10),(9, 5),(10, 5),\n",
        "            (11, 5),(11, 6),(11, 7),(11, 8),(11, 9),(11, 10)]\n",
        "\n",
        "points_9 = [(4, 5),(4, 6),(4, 7),(4, 8),(4, 9),(4, 10),(5, 10),\n",
        "            (6, 10),(7, 5),(7, 6),(7, 7),(7, 8),(7, 9),(7, 10),\n",
        "            (8, 5),(8, 10),(9, 5),(9, 10),(10, 5),(10, 6),(10, 7),\n",
        "            (10, 8),(10, 9),(10, 10)]\n",
        "for x,y in points_6:\n",
        "    x_6[x,y] = 1.0\n",
        "for x,y in points_9:\n",
        "    x_9[x,y] = 1.0\n",
        "# plt.imshow(x_6, origin=\"lower\")"
      ],
      "metadata": {
        "id": "WxRSGrUBJT7L"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Let's train a GCNN classifier (modify the last layer the GroupEquivariantCNN in 1.5 first)**\n",
        "\n",
        "*You'll see the loss doesn't converge because the C4-GCNN can't distinguish these two digits.*"
      ],
      "metadata": {
        "id": "KFIBKyVxUJ09"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "GCNN = GroupEquivariantCNN(in_channels = 1,\n",
        "                           out_channels = 1,\n",
        "                           kernel_size = 3,\n",
        "                           hidden_dim = 4,\n",
        "                           group_order = 4,\n",
        "                           num_gconvs = 1)\n",
        "\n",
        "# image 6 and 9\n",
        "inps = torch.stack([torch.from_numpy(x_6), torch.from_numpy(x_9)]).unsqueeze(1).float()\n",
        "\n",
        "# labels\n",
        "tgts = torch.FloatTensor([0,1])\n",
        "\n",
        "loss_fun = nn.BCELoss()\n",
        "optimizer = optim.Adam(GCNN.parameters(), lr=0.001)\n",
        "\n",
        "\n",
        "for epoch in range(1000):\n",
        "    # forward + backward + optimize\n",
        "    outputs = GCNN(inps)\n",
        "    loss = loss_fun(outputs[:,0], tgts)\n",
        "    optimizer.zero_grad()\n",
        "    loss.backward()\n",
        "    optimizer.step()\n",
        "\n",
        "    # print statistics\n",
        "    if epoch % 100 == 0:\n",
        "        print(\"Epoch: {} | Loss: {:0.5f} \".format(epoch,  loss.item()))\n",
        ""
      ],
      "metadata": {
        "id": "XD0WZaiSNiL1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "2) Real-world dynamics data may not satisfy the strict equivariance.  \n",
        "\n",
        "*Wang and Walters et al. [Approximately Equivariant Networks for Imperfectly Symmetric Dynamics](https://proceedings.mlr.press/v162/wang22aa.html).*\n",
        "\n",
        "3) When the outputs have lower symmetry than input.\n",
        "\n",
        "*Smidt et al. [Finding Symmetry Breaking Order Parameters with Euclidean Neural Networks](https://arxiv.org/abs/2007.02005)*"
      ],
      "metadata": {
        "id": "po6ireyn3oN0"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![picture](https://drive.google.com/uc?export=view&id=1wtUfmwYxVxp_VcUfa3GRaRgxELS-Degg)"
      ],
      "metadata": {
        "id": "ANWBJVmVTc_t"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Other relevant papers:\n",
        "* Finzi et al. [Residual Pathway Priors for Soft Equivariance Constraints](https://arxiv.org/abs/2112.01388)\n",
        "* Elsayed et al. [Revisiting Spatial Invariance with Low-Rank Local Connectivity](https://arxiv.org/abs/2002.02959)"
      ],
      "metadata": {
        "id": "6kvBJtVULtaq"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 2.1.2 Approximate Equivariance\n",
        "Let $f \\colon X \\to Y$ be a function and $G$ be a group. Assume that $G$ acts on $X$ and $Y$ via representations $\\rho_{X}$ and $\\rho_{Y}$. We say $f$ is $\\epsilon$-approximately $G$-equivariant if for any $g \\in G$,\n",
        "\n",
        "$$ \\|f(\\rho_{X}(g)(x)) - \\rho_{Y}(g)f(x)\\| \\leq \\epsilon. $$\n",
        "\n",
        "Note that strictly equivariant functions are $\\epsilon = 0$ approximately equivariant."
      ],
      "metadata": {
        "id": "rvVCi9UmHoas"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 2.1.3 Relaxed Group Convolution\n",
        "\n",
        "**We relax the weight-sharing constraints in Equivariant Networks by introducing group element dependent parameters.**\n",
        "\n",
        "Group Convolution: $$[f * \\psi] (g) =  \\sum_{h \\in G}f(h)\\psi(g^{-1} h)$$\n",
        "The $G$-equivariance of group convolution results from the shared kernel $\\psi(g^{-1}h)$.  To relax this and consequently relax the $G$-equivariance, we replace the single kernel $\\psi$ with a set of kernels $\\lbrace \\psi_l \\rbrace_{l=1}^L$. We define the new kernel $\\psi$ as a linear combination of $\\psi_l$ with coefficients that vary with $h$ and thus introduce symmetry-breaking dependence on the specific pair $(g,h)$,\n",
        "\n",
        "$$ \\psi(g,h) = \\sum_{l=1}^L w_l(h) \\psi_l(g^{-1}h). $$\n",
        "We define relaxed group convolution by multiplication with $\\psi$ as such\n",
        "\n",
        "$$[f {\\Large\\star} \\psi](g) = \\sum_{h\\in G}f(h)\\psi(g, h)\n",
        "  = \\sum_{h\\in G}\\sum_{l=1}^L f(h) w_l(h) \\psi_l(g^{-1}h) $$"
      ],
      "metadata": {
        "id": "KNwaV2dyIG3U"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "By varying the number of kernels $L$, we can control the degree of relaxation of equivariance. The weights $w_i(h) \\in \\mathbb{R}$ and the kernels $\\psi_l(g^{-1}h) \\in \\mathbb{R}^{c_\\mathrm{out} \\times c_\\mathrm{in}}$ can be learnt from data.  Relaxed group convolution reduces to group convolution and is fully equivariant if and only if $g_1^{-1}h_1 = g_2^{-1} h_2$ implies $\\psi(g_1,h_1) = \\psi(g_2,h_2)$.  In particular, this occurs if $w_l(h_1) =  w_l(h_2)$ for all $h_1,h_2 \\in G$ and for all $l$."
      ],
      "metadata": {
        "id": "masbn5CZIc4w"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "e.g. To relax translation equivariance, instead of using a single filter, we use a linear combination of multiple filters where the coefficients can be varying across the input location."
      ],
      "metadata": {
        "id": "WPS02LvmLAbp"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![picture](https://drive.google.com/uc?export=view&id=1oNcmOFXuedlOkX-rQ8Pq4_fssEw4TOCm)\n"
      ],
      "metadata": {
        "id": "wHHYkI1tTX3B"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.2 Implementing Relaxed Translation Equivariant Layers"
      ],
      "metadata": {
        "id": "wiD0B1c-RjA3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class Relaxed_ConvBlock(nn.Module):\n",
        "    def __init__(self,\n",
        "                 in_channels,\n",
        "                 out_channels,\n",
        "                 kernel_size,\n",
        "                 h_size, # Image size\n",
        "                 w_size,\n",
        "                 num_filter_banks # L\n",
        "                ):\n",
        "\n",
        "        super(Relaxed_ConvBlock, self).__init__()\n",
        "\n",
        "        # Initialize multiple filters.\n",
        "        self.convs = nn.Sequential(*[nn.Conv2d(in_channels,\n",
        "                                               out_channels,\n",
        "                                               kernel_size,\n",
        "                                               padding=(kernel_size-1)//2)\n",
        "                                     for i in range(num_filter_banks)])\n",
        "        self.num_filter_banks = num_filter_banks\n",
        "\n",
        "        # For every location we initialize a set of coefficients for linear combinations.\n",
        "        self.combination_weights = nn.Parameter(torch.ones(h_size, w_size, num_filter_banks).float()/num_filter_banks)\n",
        "        self.activation = nn.ReLU()\n",
        "\n",
        "    def forward(self, x):\n",
        "        # Compute Convolution for each filter\n",
        "        outs = torch.stack([self.convs[i](x) for i in range(self.num_filter_banks)], dim  = 0)\n",
        "\n",
        "        # Compute the linear combination of kernels for each spatial location\n",
        "        out = torch.einsum(\"ijr, rboij -> boij\", self.combination_weights, outs)\n",
        "        return self.activation(out)"
      ],
      "metadata": {
        "id": "IWFSwo4pBe9U"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = Relaxed_ConvBlock(in_channels = 1,\n",
        "                          out_channels = 1,\n",
        "                          kernel_size = 3,\n",
        "                          h_size = 64,\n",
        "                          w_size = 64,\n",
        "                          num_filter_banks = 3)\n",
        "\n",
        "out = model(torch.randn(8, 1, 64, 64))\n",
        "out.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "imcqacHRBiAH",
        "outputId": "087b9e8e-07ec-442d-9b95-73ce38205a8c"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([8, 1, 64, 64])"
            ]
          },
          "metadata": {},
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.4 Implementing Relaxed Rotaion Equivariant Layers"
      ],
      "metadata": {
        "id": "oC2ZtqGCRzQ-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class Relaxed_LiftingConvolution(nn.Module):\n",
        "    def __init__(self,\n",
        "                 in_channels,\n",
        "                 out_channels,\n",
        "                 kernel_size,\n",
        "                 group_order,\n",
        "                 num_filter_banks,\n",
        "                 activation = True\n",
        "                 ):\n",
        "        super(Relaxed_LiftingConvolution, self).__init__()\n",
        "\n",
        "        self.num_filter_banks = num_filter_banks\n",
        "        self.in_channels = in_channels\n",
        "        self.out_channels = out_channels\n",
        "        self.kernel_size = kernel_size\n",
        "        self.group_order = group_order\n",
        "        self.activation = activation\n",
        "\n",
        "        self.combination_weights = nn.Parameter(torch.ones(num_filter_banks, group_order).float()/num_filter_banks)\n",
        "\n",
        "        # Initialize an unconstrained kernel.\n",
        "        self.weight = torch.nn.Parameter(torch.zeros(self.num_filter_banks, # Additional dimension\n",
        "                                                     self.out_channels,\n",
        "                                                     self.in_channels,\n",
        "                                                     self.kernel_size,\n",
        "                                                     self.kernel_size))\n",
        "        stdv = np.sqrt(1/(self.in_channels*self.kernel_size*self.kernel_size))\n",
        "        self.weight.data.uniform_(-stdv, stdv)\n",
        "\n",
        "        # If combination_weights are equal values, then the model is still equivariant\n",
        "        # self.combination_weights.data.uniform_(-stdv, stdv)\n",
        "\n",
        "    def generate_filter_bank(self):\n",
        "        \"\"\" Obtain a stack of rotated filters\"\"\"\n",
        "        weights = self.weight.reshape(self.num_filter_banks*self.out_channels,\n",
        "                                      self.in_channels,\n",
        "                                      self.kernel_size,\n",
        "                                      self.kernel_size)\n",
        "        filter_bank = torch.stack([rot_img(weights, -np.pi*2/self.group_order*i)\n",
        "                                   for i in range(self.group_order)])\n",
        "        filter_bank = filter_bank.transpose(0,1).reshape(self.num_filter_banks, # Additional dimension\n",
        "                                                         self.out_channels,\n",
        "                                                         self.group_order,\n",
        "                                                         self.in_channels,\n",
        "                                                         self.kernel_size,\n",
        "                                                         self.kernel_size)\n",
        "        return filter_bank\n",
        "\n",
        "\n",
        "    def forward(self, x):\n",
        "        # input shape: [bz, #in, h, w]\n",
        "        # output shape: [bz, #out, group order, h, w]\n",
        "\n",
        "        # generate filter bank given input group order\n",
        "        filter_bank = self.generate_filter_bank()\n",
        "\n",
        "        # for each rotation, we have a linear combination of multiple filters with different coefficients.\n",
        "        relaxed_conv_weights = torch.einsum(\"na, noa... -> oa...\", self.combination_weights, filter_bank)\n",
        "\n",
        "        # concatenate the first two dims before convolution.\n",
        "        # ==============================\n",
        "        x = F.conv2d(\n",
        "            input=x,\n",
        "            weight=relaxed_conv_weights.reshape(\n",
        "                self.out_channels * self.group_order,\n",
        "                self.in_channels,\n",
        "                self.kernel_size,\n",
        "                self.kernel_size\n",
        "            ),\n",
        "            padding = (self.kernel_size-1)//2\n",
        "        )\n",
        "        # ==============================\n",
        "\n",
        "        # reshape output signal to shape [bz, #out, group order, h, w].\n",
        "        # ==============================\n",
        "        x = x.view(\n",
        "            x.shape[0],\n",
        "            self.out_channels,\n",
        "            self.group_order,\n",
        "            x.shape[-1],\n",
        "            x.shape[-2]\n",
        "        )\n",
        "        # ==============================\n",
        "\n",
        "        if self.activation:\n",
        "            return F.relu(x)\n",
        "        return x"
      ],
      "metadata": {
        "id": "DTjuaL7tQslD"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class Relaxed_GroupConv(nn.Module):\n",
        "    def __init__(self,\n",
        "                 in_channels,\n",
        "                 out_channels,\n",
        "                 kernel_size,\n",
        "                 group_order,\n",
        "                 num_filter_banks,\n",
        "                 activation = True\n",
        "                ):\n",
        "\n",
        "        super(Relaxed_GroupConv, self).__init__()\n",
        "\n",
        "        self.num_filter_banks = num_filter_banks\n",
        "        self.in_channels = in_channels\n",
        "        self.out_channels = out_channels\n",
        "        self.kernel_size = kernel_size\n",
        "        self.group_order = group_order\n",
        "        self.activation = activation\n",
        "\n",
        "\n",
        "        ## Initialize weights\n",
        "        self.combination_weights = nn.Parameter(torch.ones(group_order, num_filter_banks).float()/num_filter_banks)\n",
        "        self.weight = nn.Parameter(torch.randn(self.num_filter_banks, ##additional dimension\n",
        "                                               self.out_channels,\n",
        "                                               self.in_channels,\n",
        "                                               self.group_order,\n",
        "                                               self.kernel_size,\n",
        "                                               self.kernel_size))\n",
        "\n",
        "        stdv = np.sqrt(1/(self.in_channels))\n",
        "        self.weight.data.uniform_(-stdv, stdv)\n",
        "\n",
        "        # If combination_weights are equal values, then the model is still equivariant\n",
        "        # self.combination_weights.data.uniform_(-stdv, stdv)\n",
        "\n",
        "\n",
        "    def generate_filter_bank(self):\n",
        "        \"\"\" Obtain a stack of rotated and cyclic shifted filters\"\"\"\n",
        "        filter_bank = []\n",
        "        weights = self.weight.reshape(self.num_filter_banks*self.out_channels*self.in_channels,\n",
        "                                      self.group_order,\n",
        "                                      self.kernel_size,\n",
        "                                      self.kernel_size)\n",
        "\n",
        "        for i in range(self.group_order):\n",
        "            # planar rotation\n",
        "            rotated_filter = rot_img(weights, -np.pi*2/self.group_order*i)\n",
        "\n",
        "            # cyclic shift\n",
        "            shifted_indices = torch.roll(torch.arange(0, self.group_order, 1), shifts = i)\n",
        "            shifted_rotated_filter = rotated_filter[:,shifted_indices]\n",
        "\n",
        "\n",
        "            filter_bank.append(shifted_rotated_filter.reshape(self.num_filter_banks,\n",
        "                                                              self.out_channels,\n",
        "                                                              self.in_channels,\n",
        "                                                              self.group_order,\n",
        "                                                              self.kernel_size,\n",
        "                                                              self.kernel_size))\n",
        "        # stack\n",
        "        filter_bank = torch.stack(filter_bank).permute(1,2,0,3,4,5,6)\n",
        "        return filter_bank\n",
        "\n",
        "    def forward(self, x):\n",
        "\n",
        "        filter_bank = self.generate_filter_bank()\n",
        "\n",
        "        relaxed_conv_weights = torch.einsum(\"na, aon... -> on...\", self.combination_weights, filter_bank)\n",
        "\n",
        "        x = torch.nn.functional.conv2d(\n",
        "            input=x.reshape(\n",
        "                x.shape[0],\n",
        "                x.shape[1] * x.shape[2],\n",
        "                x.shape[3],\n",
        "                x.shape[4]\n",
        "                ),\n",
        "            weight=relaxed_conv_weights.reshape(\n",
        "                self.out_channels * self.group_order,\n",
        "                self.in_channels * self.group_order,\n",
        "                self.kernel_size,\n",
        "                self.kernel_size\n",
        "            ),\n",
        "            padding = (self.kernel_size-1)//2\n",
        "        )\n",
        "\n",
        "                # Reshape signal back [bz, #out * g_order, h, w] -> [bz, out, g_order, h, w]\n",
        "        x = x.view(x.shape[0], self.out_channels, self.group_order, x.shape[-2], x.shape[-1])\n",
        "        # ========================\n",
        "\n",
        "        return x\n",
        "\n",
        "class RelaxedGroupEquivariantCNN(torch.nn.Module):\n",
        "\n",
        "    def __init__(self, in_channels, out_channels, kernel_size, hidden_dim, group_order, num_gconvs):\n",
        "        super().__init__()\n",
        "\n",
        "        # First Layer\n",
        "        self.lifting_conv = Relaxed_LiftingConvolution(in_channels = in_channels,\n",
        "                                                       out_channels = hidden_dim,\n",
        "                                                       kernel_size = kernel_size,\n",
        "                                                       group_order = group_order,\n",
        "                                                       num_filter_banks = 3,\n",
        "                                                       activation = True)\n",
        "        # Middle Layer\n",
        "        self.gconvs = []\n",
        "        for i in range(num_gconvs):\n",
        "            self.gconvs.append(Relaxed_GroupConv(in_channels = hidden_dim,\n",
        "                                                out_channels = hidden_dim,\n",
        "                                                kernel_size = kernel_size,\n",
        "                                                group_order = group_order,\n",
        "                                                num_filter_banks = 3,\n",
        "                                                activation = True))\n",
        "\n",
        "        self.gconvs = nn.Sequential(*self.gconvs)\n",
        "\n",
        "\n",
        "        # Final Layer\n",
        "        self.final_layer = Relaxed_GroupConv(in_channels = hidden_dim,\n",
        "                                            out_channels = out_channels,\n",
        "                                            kernel_size = kernel_size,\n",
        "                                            group_order = group_order,\n",
        "                                            num_filter_banks = 3,\n",
        "                                            activation = False)\n",
        "        self.sigmoid = nn.Sigmoid()\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = self.lifting_conv(x)\n",
        "\n",
        "        out = self.gconvs(out)\n",
        "\n",
        "        # Generate a image output\n",
        "        out = torch.mean(self.final_layer(out), dim = 2)\n",
        "        return out\n",
        "\n",
        "        # Generate a class probability\n",
        "        # out = torch.mean(self.final_layer(out), dim = (2,3,4))\n",
        "        # return self.sigmoid(out)\n",
        "\n",
        "Relaxed_GCNN = RelaxedGroupEquivariantCNN(in_channels = 1,\n",
        "                                          out_channels = 1,\n",
        "                                          kernel_size = 3,\n",
        "                                          hidden_dim = 4,\n",
        "                                          group_order = 4,\n",
        "                                          num_gconvs = 1)"
      ],
      "metadata": {
        "id": "VOhNGVpbBmuR"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Download a single sample from mnist.\n",
        "# train_ds = torchvision.datasets.MNIST(root=\"data\", train=True, download=True)\n",
        "# image = torch.from_numpy(train_ds.data[0].numpy()).unsqueeze(0).unsqueeze(0).float()\n",
        "# image = (image - image.min())/(image.max() - image.min())\n",
        "\n",
        "## |rf(x) - f(rx)|\n",
        "for i in range(4):\n",
        "    theta = np.pi/2*i\n",
        "    x = image\n",
        "    f_x = Relaxed_GCNN(x)\n",
        "    r_f_x = rot_img(f_x, theta).data.numpy()\n",
        "    r_x = rot_img(x, theta)\n",
        "    f_r_x = Relaxed_GCNN(r_x).data.numpy()\n",
        "\n",
        "    equiv_error = np.mean(np.abs(r_f_x - f_r_x))\n",
        "    print(\"Equivariance Error with Rotation {:0.0f} Degrees: {:0.4f}\".format(theta/np.pi/2*360, equiv_error))"
      ],
      "metadata": {
        "id": "lGSy5481Oi0p"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Exercise: let's train a Relaxed GCNN classifier to distinguish image 9 and 6. Does the loss converge?**"
      ],
      "metadata": {
        "id": "0a5H_QTdRRt3"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 3. Steerable CNNs"
      ],
      "metadata": {
        "id": "JY_rECDIC6qS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.1 Representation and Steerable Convolution"
      ],
      "metadata": {
        "id": "xHvlSvAr5XOd"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.1.1 Representation"
      ],
      "metadata": {
        "id": "zOnVcUDm541Y"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "A **linear group representation** $\\rho$ of a compact group $G$ on a vector space (called *representation space*) $\\mathbb{R}^d$ is a *group homomorphism* from $G$ to the general linear group $GL(\\mathbb{R}^d)$, i.e. it is a map $\\rho : G \\to \\mathbb{R}^{d \\times d}$ such that:\n",
        "$$\\rho(g_1 g_2) = \\rho(g_1) \\rho(g_2) \\quad \\forall g_1,g_2 \\in G \\ .$$\n",
        "\n",
        "In other words, $\\rho(g)$ is a $d \\times d$ *invertible* matrix.\n",
        "We refer to $d$ as the *size* of the representation."
      ],
      "metadata": {
        "id": "W1W8kZnu7hR2"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.1.2 Trivial Representation $\\rho_0$"
      ],
      "metadata": {
        "id": "lFSc3r6ASwh8"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "The simplest example of *group representation* is the **trivial representation** which maps every element to $1 \\in \\mathbb{R}$, i.e. $\\rho: g \\mapsto 1$."
      ],
      "metadata": {
        "id": "Wy1j_YqK7pwi"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.1.3 Two-dimensional Representaion $\\rho_1$"
      ],
      "metadata": {
        "id": "8bHl4ZhBSz1W"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we know cyclic group $C_4$ is a discretization of a group of all planar rotations by an arbitrary angle, $SO(2)$. Rotation of a vector $\\mathbf{x} \\in R^2$ by arbitrary angle $\\theta$ is defined by multiplication with a matrix $\\rho_1(\\theta)$:\n",
        "\n",
        "$$\n",
        "    \\rho_2(\\theta) = \\begin{bmatrix} \\cos(\\theta) & -\\sin(\\theta) \\\\ \\sin(\\theta) & \\cos(\\theta) \\end{bmatrix}.\n",
        "$$\n",
        "\n",
        "For $C_4$, we have only four different matrixes."
      ],
      "metadata": {
        "id": "Taf_SnbR7_I0"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.1.4 Regular Representation"
      ],
      "metadata": {
        "id": "-beoLhFCTJGb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "A commonly used representation in equivariant deep\n",
        "learning is the regular representation. The regular representation of a finite group $G$ acts on a vector\n",
        "space $\\mathbb{R}^{|G|}$ by permuting its axes.\n",
        "\n",
        "Specifically, associating each axis $e_g$ of $\\mathbb{R}^{|G|}$ to an element $g \\in G$, the representation of an element $g^* \\in G$ is a permutation matrix which maps $e_g$ to $e_{g^* g}$.\n",
        "\n",
        "For instance, the regular representation of the group C4 with elements $\\{p \\frac{\\pi}{2}|p = 0, . . . , 3\\}$ is instantiated by:"
      ],
      "metadata": {
        "id": "19mTP6uN-1G9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![picture](https://drive.google.com/uc?export=view&id=1cuAVAE6sFeppk5-bxHsymoo83URqetlv)"
      ],
      "metadata": {
        "id": "gdOOxSW3_El7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "*The Group Convolution feature maps can be considered as feature fields. For every position, we have a fiber and these signals are represented by regular representations.*"
      ],
      "metadata": {
        "id": "yqL-LBB3JroA"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.1.5 Induced Representation"
      ],
      "metadata": {
        "id": "gsdC4ZmicDdm"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "We call $f:\\mathbb{R}^d → \\mathbb{R}^{d_{\\rho}}$ a feature vector field, or simply a **feature field**, if its\n",
        "\n",
        "1) *codomain* transform via a representation $\\rho(h)$ of a group $H$\n",
        "\n",
        "2) *domain* transform via the action $g^{-1}$ of $G = (\\mathbb{R}^d, +) \\rtimes H$\n",
        "\n",
        "Representation $\\rho$ defines the type of the field, and together with the group action of $G = (\\mathbb{R}^d, +) \\rtimes H$ defines the ***induced representation***\n",
        "\n",
        "$$(Ind_H^G[\\rho](\\textbf{x}, h)f)(\\textbf{x}')= \\rho(h)f(h^{-1}(\\textbf{x}'-\\textbf{x}))$$"
      ],
      "metadata": {
        "id": "Mg4VP9rycLMg"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "*For example, consider an image as network input. A rotation of a gray-scale image is performed by moving each pixel to a new position. The invariance of the scalar pixel values under rotations is modeled by the **trivial representation** $\\rho:G\\to\\operatorname{GL}(1),\\ g\\mapsto(1)$ of $G$ and identifies them as **scalar fields**.*\n",
        "\n",
        "*Formally, a scalar field is a function $f: \\mathbb{R}^2 \\to \\mathbb{R}$ mapping to a feature vector with $c=1$ channels.\n",
        "A rotation by an angle $\\theta \\in C_4$ transforms this scalar field as*\n",
        "$$ \\big[R_{\\theta}\\, f\\big](x)\n",
        "   \\ :=\\ \\rho(\\theta)\\,f\\big(R^{-1} x\\big)\n",
        "   \\ =\\ 1\\cdot f\\big(R^{-1}x\\big)\n",
        "   \\ =\\ f\\big(R^{-1}x\\big),$$"
      ],
      "metadata": {
        "id": "Q-tB6PedobaA"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**But when rotating vector field, each vector is not only moved to a new position but additionally changes its\n",
        "orientation via the action $g \\in G$**"
      ],
      "metadata": {
        "id": "6ezAsJLYpalS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Download ocean current data from [here](https://drive.google.com/file/d/1F7-XUAOYxD2g1PNDmq-QrtTnHb9vr1ey/view?usp=sharing) or creat a random vector field**"
      ],
      "metadata": {
        "id": "A82MQzjcCFVe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# The file contains 10 steps of ocean current velocity fields from Atlantic Ocean\n",
        "ocean_sample = torch.load(\"ocean_atlantic_sample.pt\").numpy()\n",
        "ocean_sample.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Tkwt08GMBqIx",
        "outputId": "4f1b12f9-1d7f-41cf-dc48-8357837e0fc3"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10, 2, 253, 253)"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "ocean_sample = ocean_sample[0]\n",
        "fig=plt.figure(figsize=(10, 5))\n",
        "imgs = [np.sqrt(ocean_sample[0]**2 + ocean_sample[1]**2), ocean_sample[0], ocean_sample[1]]\n",
        "titles = [\"Velocity Norm Field\", \"Velocity along X direction\", \"Velocity along Y direction\"]\n",
        "for i in range(3):\n",
        "    fig.add_subplot(1, 3, i+1)\n",
        "    plt.imshow(imgs[i])\n",
        "    plt.title(titles[i])\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "vTqYm-lHKdmt",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        },
        "outputId": "ab80c5a4-ed32-4428-e55f-d450b726c8f4"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x500 with 3 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        " A function that helps visualize vector field"
      ],
      "metadata": {
        "id": "HFbYquC9QKcH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#  A function that helps visualize vector field\n",
        "def visualize_vector_field(vector_field, scale = 100):\n",
        "    # Input vector field must have shape (H, W, 2) or (2, H, W).\n",
        "    assert len(vector_field.shape) == 3\n",
        "    if vector_field.shape[0] == 2:\n",
        "        vector_field = vector_field.transpose(1,2,0)\n",
        "\n",
        "    H, W, _ = vector_field.shape\n",
        "    x_coords, y_coords = np.mgrid[0:H, 0:W]\n",
        "\n",
        "    # Extract the x and y components of the vector field\n",
        "    u = vector_field[:, :, 0]\n",
        "    v = vector_field[:, :, 1]\n",
        "\n",
        "    # Create a quiver plot with blue arrows, without axes and labels\n",
        "    plt.quiver(x_coords, y_coords, u, v, color='blue', scale=scale)\n",
        "    plt.axis('off')\n",
        "\n",
        "    # Show the plot\n",
        "    plt.show()\n",
        "\n",
        "# Let's just look at a small patch of it\n",
        "small_ocean_sample = ocean_sample[:, :16, :16]\n",
        "visualize_vector_field(small_ocean_sample, scale = 10)\n"
      ],
      "metadata": {
        "id": "0nLA-StsQHCv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "If we only rotate the positions of vectors"
      ],
      "metadata": {
        "id": "xFROIUxVQJFw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "rot_pos_sample = rot_img(torch.from_numpy(small_ocean_sample).unsqueeze(0).float(), np.pi/2).squeeze(0)\n",
        "visualize_vector_field(rot_pos_sample.data.numpy(), scale = 10)"
      ],
      "metadata": {
        "id": "RrW-esAQQRI1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "We also need to rotate the orientation of the vectors"
      ],
      "metadata": {
        "id": "YRLY25GfQTGe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def rot_vector(x, theta):\n",
        "    #x has the shape [c x 2 x h x w]\n",
        "    rho = torch.FloatTensor([[np.cos(theta), -np.sin(theta)],\n",
        "                             [np.sin(theta), np.cos(theta)]])\n",
        "    out = torch.einsum(\"ab, bc... -> ac...\",(rho, x.transpose(0,1))).transpose(0,1)\n",
        "    return out\n",
        "\n",
        "\n",
        "rot_direc_pos_sample = rot_vector(rot_pos_sample.unsqueeze(0), np.pi/2).squeeze(0)\n",
        "visualize_vector_field(rot_direc_pos_sample.data.numpy(), scale = 10)\n"
      ],
      "metadata": {
        "id": "pdJ3yka0QWLF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.1.6 Steerable Constraint\n",
        "\n",
        "Let $f$ be the input feature map $f \\colon \\mathbb{R}^2 \\to \\mathbb{R}^{c_\\mathrm{in}}$ with group $G \\subset O(2)$ acting on\n",
        "$\\mathbb{R}^2$ by matrix multiplication and on the channel space $\\mathbb{R}^{c_\\mathrm{in}}$ by $\\rho_{\\mathrm{in}}$.  Also fix an $G$-action $\\rho_{\\mathrm{out}}$ on $\\mathbb{R}^{c_\\mathrm{out}}$.  Let $\\phi \\colon \\mathbb{R}^2 \\to \\mathbb{R}^{c_\\mathrm{out} \\times c_\\mathrm{in}}$ be a kernel, then the standard 2D convolution $f *_{\\mathbb{R}^2} \\phi$ is $G$-equivariant and $\\mathbb{R}^2$-translation equivariant when\n",
        "\n",
        "$$\\phi(gx) = \\rho_{\\text{out}}(g) \\phi(x) \\rho_{\\text{in}}(g^{-1})\\ \\forall g \\in G. $$\n",
        "\n",
        "Given $\\rho_\\text{in}$ and $\\rho_\\text{out}$, we can compute a basis for such linear space by solving a linear system. Once we have such a basis, we can express any equivariant filter bank as a linear combination with trainable coefficients: $\\phi = \\sum_{l=1}^L w_l \\phi_l$."
      ],
      "metadata": {
        "id": "J_CY15Um59jV"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Proof:** If $\\phi$ satisfies the steerabile constraint then\n",
        "\n",
        "$$\n",
        "\\begin{align}\n",
        "    \\phi \\star [g \\cdot f_\\text{in}](\\mathbf{y})\n",
        "    &= \\int_{\\mathbb{R}^2} \\phi(\\mathbf{x}-\\mathbf{y}) [g \\cdot f_\\text{in}](\\mathbf{x}) d\\mathbf{x} \\\\\n",
        "    &= \\int_{\\mathbb{R}^2} \\phi(\\mathbf{x}-\\mathbf{y}) \\rho_\\text{in}(g)f_\\text{in}(g^{-1}\\mathbf{x}) d\\mathbf{x} \\\\\n",
        "    &= \\int_{\\mathbb{R}^2} \\rho_{\\text{out}}(g) \\phi(g^{-1}(\\mathbf{x}-\\mathbf{y})) \\rho_{\\text{in}}(g^{-1}) \\rho_\\text{in}(g)f_\\text{in}(g^{-1}\\mathbf{x}) d\\mathbf{x} \\\\\n",
        "    &= \\rho_\\text{out}(g) \\int_{\\mathbb{R}^2} \\phi(g^{-1}\\cdot(\\mathbf{x}-\\mathbf{y}))f_\\text{in}(g^{-1}\\mathbf{x}) d\\mathbf{x} \\\\\n",
        "    &= \\rho_\\text{out}(g) \\int_{\\mathbb{R}^2} \\phi(\\mathbf{z} - g^{-1}\\cdot\\mathbf{y}))f_\\text{in}(\\mathbf{z}) d\\mathbf{z} \\\\\n",
        "    &= \\rho_\\text{out}(g) f_\\text{out}(g^{-1} \\cdot\\mathbf{y}) \\\\\n",
        "    &= [g\\cdot f_\\text{out}](\\mathbf{y})\n",
        "\\end{align}\n",
        "$$"
      ],
      "metadata": {
        "id": "P-99H2tFcKBB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.2 Steerable Convolution Layer"
      ],
      "metadata": {
        "id": "Z5BcLQZKTiL_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Install e2cnn**"
      ],
      "metadata": {
        "id": "FDJnCbeaCm31"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install e2cnn\n",
        "import e2cnn\n",
        "import torch.nn as nn\n",
        "import numpy as np\n",
        "import torch.nn.functional as F\n",
        "from e2cnn.nn.modules.r2_conv.r2convolution import compute_basis_params\n",
        "from e2cnn.nn.modules.r2_conv.basisexpansion_singleblock import block_basisexpansion"
      ],
      "metadata": {
        "id": "mkyWViOOQnfB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.2.1 $\\rho_0 → \\rho_0$"
      ],
      "metadata": {
        "id": "8W13b6gF6HL_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "If both $\\rho_{\\text{in}}$ and $\\rho_{\\text{out}}$ are trivial represenation, then we can write steerable constraint as\n",
        "$$\\phi(gx) = \\phi(x), g \\in SO(2)$$\n",
        "\n",
        "- The special orthogonal group $SO(n)$ is a subgroup of $O(n)$ containing the orthogonal matrices with determinant 1.\n",
        "- The Euclidean group $E(n)$ adds translation and is the semidirect product $E(n) = T(n) \\rtimes O(n)$.\n"
      ],
      "metadata": {
        "id": "7ziOlQT1Ed9s"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**We can visualize the kernel basis with the help of e2cnn**"
      ],
      "metadata": {
        "id": "34MYwyxZQcS8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "kernel_size = 9\n",
        "\n",
        "# Choose the order of rotation group.\n",
        "r2_act = e2cnn.gspaces.Rot2dOnR2(N = 8)\n",
        "\n",
        "# Define the input and output representation\n",
        "feat_type_in = e2cnn.nn.FieldType(r2_act, [r2_act.trivial_repr])\n",
        "feat_type_hid = e2cnn.nn.FieldType(r2_act, [r2_act.trivial_repr])\n",
        "\n",
        "# Obtain sampled basis from E2CNN\n",
        "grid, basis_filter, rings, sigma, maximum_frequency = compute_basis_params(kernel_size = kernel_size)\n",
        "basis = feat_type_in.gspace.build_kernel_basis(feat_type_in._unique_representations.pop(), feat_type_hid._unique_representations.pop(), sigma, rings, maximum_frequency = maximum_frequency)\n",
        "block_expansion = block_basisexpansion(basis, grid, basis_filter, recompute=False).sampled_basis\n",
        "\n",
        "print(block_expansion.shape)\n",
        "# Visualize Steerable Basis\n",
        "fig=plt.figure(figsize=(20, 5))\n",
        "for i in range(block_expansion.shape[0]):\n",
        "    fig.add_subplot(1, block_expansion.shape[0], i+1)\n",
        "    plt.imshow(block_expansion[i,0,0].reshape(9,9))\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "ts0LTOYmx4Nn",
        "outputId": "970e9627-4cb9-461e-f250-e724fcd0f316"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/e2cnn/nn/modules/r2_conv/basisexpansion_singleblock.py:80: UserWarning: indexing with dtype torch.uint8 is now deprecated, please use a dtype torch.bool instead. (Triggered internally at ../aten/src/ATen/native/IndexingUtils.h:27.)\n",
            "  full_mask[mask] = norms.to(torch.uint8)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([5, 1, 1, 81])\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x500 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.2.2 $\\rho_1 → \\rho_1$"
      ],
      "metadata": {
        "id": "WQcZmxTg6RFs"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "When both $\\rho_{\\text{in}}$ and  $\\rho_{\\text{out}}$ are vector representation, then the  kernel constraint for mapping of input feature map $f_\\text{in}: R^2 ->R^2$ to output feature map $f_\\text{out}: R^2 ->R^2$.\n",
        "\n",
        "$$\\phi(gx) = \\rho_1(g)\\phi(x)\\rho_1(g^{-1}), g \\in SO(2)$$\n",
        "\n",
        "The kernel should have the shape of $(2, 2, k, k)$.\n"
      ],
      "metadata": {
        "id": "4CB36siNHJg2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define the input and output representation\n",
        "# irrep(1) corresponds to \\rho_1 representation\n",
        "feat_type_in = e2cnn.nn.FieldType(r2_act, [r2_act.irrep(1)])\n",
        "feat_type_hid = e2cnn.nn.FieldType(r2_act, [r2_act.irrep(1)])\n",
        "\n",
        "# Obtain sampled basis\n",
        "grid, basis_filter, rings, sigma, maximum_frequency = compute_basis_params(kernel_size = kernel_size)\n",
        "basis = feat_type_in.gspace.build_kernel_basis(feat_type_in._unique_representations.pop(), feat_type_hid._unique_representations.pop(), sigma, rings, maximum_frequency = maximum_frequency)\n",
        "block_expansion = block_basisexpansion(basis, grid, basis_filter, recompute=False).sampled_basis\n",
        "\n",
        "visualize_vector_field(block_expansion[7,0].reshape(2,9,9).numpy(), scale = 5)\n",
        "visualize_vector_field(block_expansion[7,1].reshape(2,9,9).numpy(), scale = 5)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 795
        },
        "id": "puq-cNEF3TaA",
        "outputId": "4ecbf4e9-b14b-4d63-fe8c-83ba996c82ca"
      },
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.2.3 Equivariant Activation Functions"
      ],
      "metadata": {
        "id": "OsHgwMtgTj22"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Activation functions should commute with the representation of filters**\n",
        "$$\\sigma(\\rho(h)f(x)) = \\rho(h)\\sigma(f(x))$$"
      ],
      "metadata": {
        "id": "HJBck7ufb7h9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "* **Regular Representation:**  Since this action is\n",
        "just permuting channels of $\\rho_{reg}^G$-fields, it commutes with pointwise nonlinearities like ReLU.\n",
        "\n",
        "* **$\\rho_1$ Representation:**\n",
        "We need to use norm-base activation, such as:\n",
        "$$\n",
        "\\sigma(f(x)) = \\sigma(|f(x)|) * \\frac{f(x)}{|f(x)|},\n",
        "$$\n",
        "where $f(x)$ is feature map of vector type."
      ],
      "metadata": {
        "id": "pADnuR5HIgpp"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "e.g. ReLU does not commute with $\\rho_1$\n",
        "\n",
        "$$\\rho(R_{\\pi})\\text{ReLU}((1,0))=\\rho(R_{\\pi})(1,0) = (-1,0)$$\n",
        "\n",
        "$$\\text{ReLU}(\\rho(R_{\\pi})(1,0))=\\text{ReLU}((-1,0)) = (0,0)$$"
      ],
      "metadata": {
        "id": "SOXh6RX_eMlO"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 3.3.4 Irrep decomposition of the kernel constraint"
      ],
      "metadata": {
        "id": "MpIG83EPO4se"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Representation does not always have nice and clean forms but it can be decomposed into a direct sum of irredubile representations."
      ],
      "metadata": {
        "id": "4pb1mCrFHAba"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Decompose $\\rho_{\\text{out}}$ and $\\rho_{\\text{in}}$ in the kernel constraint $k(gx) = \\rho_{\\text{out}}(g) k(x) \\rho_{\\text{in}}(g^{-1})$ into the direct sum of irreps:\n",
        "\n",
        "$$k(gx) = Q_{out}^{-1}[\\bigoplus_{i\\in I_{out}}\\psi_i(g)]Q_{out}\\;k(x)\\;Q_{in}^{-1}[\\bigoplus_{j\\in I_{in}}\\psi_j^{-1}(g)]Q_{in}$$\n",
        "\n",
        "Define $\\kappa(x) = Q_{out}k(x)Q_{in}^{-1}$:\n",
        "\n",
        "$$\\kappa(gx) = [\\bigoplus_{i\\in I_{out}}\\psi_i(g)]\\;\\kappa(x)\\;[\\bigoplus_{j\\in I_{in}}\\psi_j^{-1}(g)]$$\n",
        "\n",
        "Then we obtain independent constraints on block $\\kappa^{ij}(x)$ in $\\kappa$\n",
        "\n",
        "$$\\kappa^{ij}(gx) = \\psi_i(g)\\; \\kappa^{ij}(x)\\;\\psi_j^{-1}(g)$$"
      ],
      "metadata": {
        "id": "JLIMPDOGP-5r"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![picture](https://drive.google.com/uc?export=view&id=1ceEJ5onRSYtP9WCZFSv_ZXA3yKMnaPh1)"
      ],
      "metadata": {
        "id": "FnLALN0Au42Z"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**The kernel space constraint affects only the angular behavior of the kernels we are free to choose any radial profile. It turns out that each basis element is harmonic and associated to one unique angular frequency. Take a look at Table 2 in [E2CNN paper](https://arxiv.org/pdf/1911.08251.pdf).**"
      ],
      "metadata": {
        "id": "ZGGABPBnHE0l"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3.3 E2CNN and ESCNN"
      ],
      "metadata": {
        "id": "zZWW6QiS6hrK"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**With the help of [E2CNN](https://github.com/QUVA-Lab/e2cnn) & [ESCNN](https://github.com/QUVA-Lab/escnn) & [E3NN](https://e3nn.org/), constructing equivariant models has become much simpler!**\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "lBKXGtyTh_Ci"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "[E2CNN](https://github.com/QUVA-Lab/e2cnn) focuses on exploiting isometries, i.e. distance preserving transformations. Specifically for \\mathbb{R}^2\n",
        " , the isometries are formalized by the Euclidean group  𝐸(2)\n",
        "  which consists of translations, rotations and reflections\n",
        "\n",
        "Since we are using CNNs in this tutorial, all models will be equivariant under translations. We can choose a subgroup $G\\leq O(2)$ to determine the point symmetries (reflections and rotations). In the example below, we consider a the cyclic subgroup $C_4$, which models the $4$ *rotations* which are multiples of $\\frac{\\pi}{2}$, that is, $\\big\\{0, \\frac{\\pi}{2}, \\pi, \\frac{3\\pi}{2}\\big\\}$.\n",
        "\n",
        "We determine both the point group and its action on the space by instantiating a subclass of gspace.GSpace. For the rotational action of  𝐺=𝐶4\n",
        "  on  ℝ2\n",
        "  this is done by:"
      ],
      "metadata": {
        "id": "zMeWh7UusU8c"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "r2_act = e2cnn.gspaces.Rot2dOnR2(N = 4)\n",
        "r2_act"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xaYYgbnisUYq",
        "outputId": "3f27072a-b734-4114-e7fd-0d7b09bac57c"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<e2cnn.gspaces.r2.rot2d_on_r2.Rot2dOnR2 at 0x7c0bc4a6b2e0>"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "When building an equivariant model to a group $G$, we require that the output produced by the model transforms consistently when the input transforms under the action of an element $g \\in G$. We thus have to define the *field types* `feat_type_in` and `feat_type_out` of `nn.FieldType` for every layer of the network.\n"
      ],
      "metadata": {
        "id": "jZbPMLmBuG1i"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Suppose the input is a gray-scale image\n",
        "feat_type_in = e2cnn.nn.FieldType(r2_act, [r2_act.trivial_repr])\n",
        "\n",
        "# we usually use the regular representation for middle layers.\n",
        "feat_type_hid = e2cnn.nn.FieldType(r2_act, [r2_act.regular_repr])\n"
      ],
      "metadata": {
        "id": "iahhB_wEugp8"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Then we can define a convolution layer that maps trivial representation field to regular representation fields.\n",
        "\n",
        "*Each equivariant module has an input and output type. It requires its inputs to transform according to its input type and is guaranteed to return feature fields associated with its output type.*"
      ],
      "metadata": {
        "id": "erCHuy0VuTaQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "e2cnn_layer = e2cnn.nn.R2Conv(feat_type_in, feat_type_hid, kernel_size = 3, padding = 1)"
      ],
      "metadata": {
        "id": "qBdxvcRPvMDq"
      },
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Feature fields, each of its own type $\\rho_l$ can be stacked: should be thought of as the channels in standard CNNs. The sub-vectors or channels in these fields live in their own sub-vector spaces $V_l$ and transform by their own representation $\\rho_l$."
      ],
      "metadata": {
        "id": "v86gAFJaGfbJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# For example, we can stack 10 regular feature fields\n",
        "feat_type_hid = e2cnn.nn.FieldType(r2_act, 10* [r2_act.regular_repr])\n",
        "\n",
        "e2cnn_layer = e2cnn.nn.R2Conv(feat_type_in, feat_type_hid, kernel_size = 3, padding = 1)"
      ],
      "metadata": {
        "id": "hR05f6DnG6A1"
      },
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "In ESCNN or E2CNN, **geometric tensors** are defined as data containers.\n",
        "They wrap a *PyTorch* `torch.Tensor` to augment them with an instance of `FieldType`."
      ],
      "metadata": {
        "id": "L-BnDjK2wpoh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x = torch.randn(4, 1, 32, 32)\n",
        "x = e2cnn.nn.GeometricTensor(x, feat_type_in)\n",
        "# x.type"
      ],
      "metadata": {
        "id": "BEhvMYxpwwda"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Suppose we want to make one-step ahead prediction of ocean current velocity field given historic five frames as input.**"
      ],
      "metadata": {
        "id": "seLD6cBYWThU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class Rot_Equiv_CNN(torch.nn.Module):\n",
        "    def __init__(self,\n",
        "                 input_frames, # number of input frames\n",
        "                 output_frames, # number of output frames\n",
        "                 hidden_dim, # hidden dimensions\n",
        "                 kernel_size,\n",
        "                 N # group order\n",
        "                ):\n",
        "        super(Rot_Equiv_CNN, self).__init__()\n",
        "\n",
        "\n",
        "        # We determine both the point group and its action on the space by instantiating a subclass of gspace.GSpace\n",
        "        # the model is equivariant under rotations by 90 degrees, modelled by C4\n",
        "        r2_act = e2cnn.gspaces.Rot2dOnR2(N = N)\n",
        "\n",
        "        # the input is a stack of vector fields, corresponding to the \\rho_1 representation\n",
        "        feat_type_in = e2cnn.nn.FieldType(r2_act, input_frames*[r2_act.irrep(1)])\n",
        "\n",
        "        # we use the regular representation for middle layers.\n",
        "        feat_type_hid = e2cnn.nn.FieldType(r2_act, hidden_dim*[r2_act.regular_repr])\n",
        "\n",
        "        # the output is a single vector field.\n",
        "        feat_type_out = e2cnn.nn.FieldType(r2_act, output_frames*[r2_act.irrep(1)])\n",
        "        self.feat_type_in = feat_type_in\n",
        "\n",
        "        # Since we use regular representation for the middle layer, we can use element-wise ReLU.\n",
        "        self.model = e2cnn.nn.SequentialModule(\n",
        "            e2cnn.nn.R2Conv(feat_type_in, feat_type_hid, kernel_size = kernel_size, padding = (kernel_size - 1)//2),\n",
        "            e2cnn.nn.ReLU(feat_type_hid),\n",
        "            e2cnn.nn.R2Conv(feat_type_hid, feat_type_hid, kernel_size = kernel_size, padding = (kernel_size - 1)//2),\n",
        "            e2cnn.nn.ReLU(feat_type_hid),\n",
        "            e2cnn.nn.R2Conv(feat_type_hid, feat_type_out, kernel_size = kernel_size, padding = (kernel_size - 1)//2)\n",
        "        )\n",
        "\n",
        "    def forward(self, x):\n",
        "        # wrap the input tensor in a GeometricTensor\n",
        "        # (associate it with the input type)\n",
        "        x = e2cnn.nn.GeometricTensor(x, self.feat_type_in)\n",
        "\n",
        "        out = self.model(x)\n",
        "\n",
        "        # unwrap the output GeometricTensor\n",
        "        # (take the Pytorch tensor and discard the associated representation)\n",
        "        return out.tensor\n",
        "\n"
      ],
      "metadata": {
        "id": "_Mfo1QBVNBDP"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ocean_sample = torch.load(\"ocean_atlantic_sample.pt\")\n",
        "model = Rot_Equiv_CNN(input_frames = 5,\n",
        "                      output_frames = 1,\n",
        "                      hidden_dim = 16,\n",
        "                      kernel_size = 3,\n",
        "                      N = 4)\n",
        "\n",
        "x = ocean_sample[None,:5,:,:64,:64].reshape(1,10,64,64)\n",
        "out = model(x)\n",
        "out.shape"
      ],
      "metadata": {
        "id": "xKUXBYpERWjp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "*when we verify equivariance, we need to be clear about the representation of group action in the input and output spaces.*  "
      ],
      "metadata": {
        "id": "1a1jkdsXHtGS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "## |rf(x) - f(rx)|\n",
        "for i in range(4):\n",
        "    theta = np.pi/2*i\n",
        "    f_x = model(x)\n",
        "    r_f_x = rot_vector(rot_img(f_x.squeeze(1), theta), theta).data.numpy()\n",
        "    r_x = rot_vector(rot_img(x.squeeze(1), theta), theta)\n",
        "    f_r_x = model(r_x).data.numpy()\n",
        "\n",
        "    equiv_error = np.mean(np.abs(r_f_x - f_r_x))\n",
        "    print(\"Equivariance Error with Rotation {:0.0f} Degrees: {:0.5f}\".format(theta/np.pi/2*360, equiv_error))\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2CuyhEKUW5QP",
        "outputId": "d9213486-74c3-4a35-f746-d11e14ae48f4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Equivariance Error with Rotation 0 Degrees: 0.00000\n",
            "Equivariance Error with Rotation 90 Degrees: 0.00000\n",
            "Equivariance Error with Rotation 180 Degrees: 0.00000\n",
            "Equivariance Error with Rotation 270 Degrees: 0.00000\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/torch/nn/functional.py:4236: UserWarning: Default grid_sample and affine_grid behavior has changed to align_corners=False since 1.3.0. Please specify align_corners=True if the old behavior is desired. See the documentation of grid_sample for details.\n",
            "  warnings.warn(\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "titles = [\"$x$\", \"$f(x)$\", \"$gx$\", \"$f(gx))$\", \"$g(f(x))$\"]\n",
        "vis_imgs = [x[0,0], f_x[0,0].data.numpy(), r_x[0,0].data.numpy(), f_r_x[0,0], r_f_x[0,0]]\n",
        "fig=plt.figure(figsize=(10, 5))\n",
        "idx = 0\n",
        "for i in range(6):\n",
        "    if i == 1:\n",
        "        continue\n",
        "    fig.add_subplot(2, 3, i+1)\n",
        "    plt.imshow(vis_imgs[idx])\n",
        "    plt.title(titles[idx], fontsize = 20)\n",
        "    plt.axis('off')\n",
        "    idx += 1\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 453
        },
        "id": "jPJXJV6hXZQa",
        "outputId": "d975a649-23ae-4ca0-b51b-b9a1c8b33a78"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x500 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Exercise: Can a C4-equivariant model able deform a square to a rectangle? Can it transform a rectangle to a square?**"
      ],
      "metadata": {
        "id": "I2Z5gjNnLTzR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "r2_act = e2cnn.gspaces.flipRot2dOnR2(N=8)\n",
        "square_image = torch.zeros((9,9))\n",
        "square_image[2,2] = 1.0\n",
        "square_image[2,6] = 1.0\n",
        "square_image[6,2] = 1.0\n",
        "square_image[6,6] = 1.0\n",
        "\n",
        "rect_image = torch.zeros((9,9))\n",
        "rect_image[1,3] = 1.0\n",
        "rect_image[1,5] = 1.0\n",
        "rect_image[7,3] = 1.0\n",
        "rect_image[7,5] = 1.0\n",
        "\n",
        "# wrap images in geometric tensor\n",
        "feat_type = nn.FieldType(r2_act,[r2_act.trivial_repr])\n",
        "\n",
        "square_image_geom = nn.GeometricTensor(square_image[None,None,:,:],feat_type)\n",
        "rect_image_geom = nn.GeometricTensor(rect_image[None,None,:,:],feat_type)"
      ],
      "metadata": {
        "id": "WhRpft7FLUQv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 4. Physics-Guided Deep Learning for Dynamical Systems"
      ],
      "metadata": {
        "id": "q4uFYLVVDEqM"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "* Rui Wang and Rose Yu; [Physics-Guided Deep Learning for Dynamical Systems: A Survey](https://arxiv.org/abs/2107.01272)\n",
        "\n",
        "* Xuan Zhang et al. [Artificial Intelligence for Science in Quantum, Atomistic, and Continuum Systems](https://arxiv.org/abs/2307.08423)"
      ],
      "metadata": {
        "id": "g-we6BXEqtb8"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4.1 Problem Formulation"
      ],
      "metadata": {
        "id": "GEGdl0FeL9xc"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Dynamical systems are mathematical objects that are used to describe the evolution of phenomena over time and space occurring in nature. Dynamical systems are commonly described with differential equations which are equations related to one or more unknown functions and their derivatives.\n",
        "\n",
        "**Definition**: Fix an integer $k \\geq 1$ and let $U$ denote an open subset of $\\mathbb{R}^n$. Let $u: U \\mapsto \\mathbb{R}^m$ and we write ${u} = (u^1, ..., u^m)$, where $x \\in U$. Then an expression of the form\n",
        "\n",
        "$$F(D^k {u}(x), D^{k-1} {u}(x), ..., D{u}(x), {u}(x), x) = 0$$\n",
        "\n",
        "is called a $k^{\\text{-th}}$-order system of partial differential equation (or ordinary differential equation when $n=1$), where $F: \\mathbb{R}^{mn^k}\\times \\mathbb{R}^{mn^{k-1}}\\times...\\times\\mathbb{R}^{mn}\\times \\mathbb{R}^{m}\\times U \\mapsto \\mathbb{R}^m $.\n"
      ],
      "metadata": {
        "id": "w0eCa6_jMAsd"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 4.1.1 Solving Differential Equations\n",
        "When $F$ is *known* but too complicated to be solvable, researchers tend to directly solve the differential Eq.ations by approximating solution of ${u}(x)$ with a deep neural network, and enforcing the governing equations as a soft constraint on the output of the neural nets during training at the same time. This approach can be formulated as the following optimization problem,\n",
        "\n",
        "$$\\text{min}_{\\theta} \\; {L}({u}) + \\lambda_{F}{L}_{F}({u})$$\n",
        "\n",
        "$L({u})$ denotes the misfit of neural net predictions and the training data points. $\\theta$ denotes the neural net parameters. ${L}_{F}({u})$ is a constraint on the residual of the differential equation system under consideration.\n"
      ],
      "metadata": {
        "id": "zNS2ZyXdL2Hl"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "* Maziar Raissi et al. [Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations](https://www.sciencedirect.com/science/article/pii/S0021999118307125)\n"
      ],
      "metadata": {
        "id": "qL0S76Tgrq0Z"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 4.1.2 Learning Dynamics Residuals\n",
        "When $F$ is *partially known*, we can use neural nets to learn the errors or residuals made by physics-based models. The key is to learn the bias of physics-based models and correct it with the help of deep learning. The final prediction of the state is composed of the simulation from the physics-based models and the residual prediction from neural nets as below,\n",
        "\n",
        "$$\\hat{u} = \\hat{u}_{F} + \\hat{u}_{\\text{NN}}.$$\n",
        "\n",
        "where $\\hat{u}_{F}$ is the prediction obtained by numerically solving ${F}$, $\\hat{u}_{\\text{NN}}$is the prediction from neural networks and $\\hat{u}$ is the final prediction made by hybrid physics-DL models.\n"
      ],
      "metadata": {
        "id": "lTZhQWn5M8qz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "* Yuan Yin et al. [Augmenting Physical Models with Deep Networks for Complex Dynamics Forecasting](https://arxiv.org/abs/2010.04456)\n",
        "\n",
        "* Filipe de Avila Belbute-Peres et al. [Combining Differentiable PDE Solvers and Graph Neural Networks for Fluid Flow Prediction](https://arxiv.org/abs/2007.04439)"
      ],
      "metadata": {
        "id": "4rpl5IOKuKmx"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 4.1.3 Dynamics Forecasting\n",
        "When ${F}$ is *unknown* or numerically solving it requires too much computation, many works studied learning high-dimensional spatiotemporal dynamics by directly predicting the input-output system state mapping and bypassing numerical discretization and integration. If we assume the first dimension $x_1$ of ${u}$ is the time dimension $t$, then the problem of dynamics forecasting can be defined as learning a map $f: \\mathbb{R}^{n \\times k} \\mapsto \\mathbb{R}^{n \\times q} $ that maps a sequence of historic states to future states of the dynamical system,\n",
        "\n",
        "$$f({u}(t-k+1, \\cdot), ..., {u}(t, \\cdot)) = {u}(t+1, \\cdot), ..., {u}(t+q, \\cdot)$$\n",
        "where $k$ is the input length and $q$ is the output length. $f$ is commonly approximated with purely data-driven or physics-guided neural nets and the neural nets are optimized by minimizing the prediction errors of the state ${L}({u})$.\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "k4-DuvYXNfmj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "* Tobias Pfaff et al. [Learning mesh-based simulation with graph networks.](https://arxiv.org/abs/2010.03409)\n",
        "* Emmanuel de Bezenac et al. [Deep learning for physical\n",
        "processes: Incorporating prior scientific knowledge.](https://arxiv.org/abs/1711.07970)"
      ],
      "metadata": {
        "id": "o7f1XgH8vN5W"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 4.1.4 Search for Governing Equations\n",
        "When ${F}$ is *unknown* and it is necessary to determine the precise governing equations to solve practical problems, numerous efforts have been made to discover the exact mathematical formulation of ${F}$. The most common approach is to select from a wide range of possible candidate functions and choose the model that minimizes fitting errors on observation data.\n",
        "\n",
        "More specifically,the goal of discovering governing equations is to find an approximate function $\\hat{F} = \\Phi({u}(x), x) \\boldsymbol{\\theta} \\approx {F}$, where $\\Phi({u}(x), x) = [\\phi_1({u}(x), x), \\phi_2({u}, x), \\ldots, \\phi_p({u}, x)]$ is a library of candidate functions, such as polynomials and trigonometric functions, and $\\boldsymbol{\\theta} \\in \\mathbb{R}^p$ is a sparse vector indicating which candidate functions are active in the dynamics. This problem can be formulated as an optimization problem, where we aim to minimize the following cost function over a set of observed data $\\{{y}_i\\}_{i=1}^n$ of ${u}$:\n",
        "\n",
        "$${L}(\\boldsymbol{\\theta}) = \\sum_{i=1}^n ||\\Phi({y}_i, x) \\boldsymbol{\\theta}||^2 $$"
      ],
      "metadata": {
        "id": "Qm2HCIHLNkAd"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "* Subham Sahoo et al. [Learning equations for extrapolation\n",
        "and control.](https://arxiv.org/abs/1806.07259)\n",
        "\n",
        "* Steven L. Brunton et al. [Discovering governing equations from data: Sparse identification of nonlinear dynamical systems](https://arxiv.org/abs/1509.03580)"
      ],
      "metadata": {
        "id": "iZiznFLS9VyI"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4.2 Long term robust forecasting"
      ],
      "metadata": {
        "id": "qocUhL64LFWB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Suppose we are using a MLP to forecast low-dimensional dynamics."
      ],
      "metadata": {
        "id": "Odmx4f2xIufv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class MLP(nn.Module):\n",
        "  def __init__(self, input_dim, input_length, output_dim, output_length, hidden_dim):\n",
        "    super().__init__()\n",
        "\n",
        "    self.output_length = output_length\n",
        "    self.output_dim = output_dim\n",
        "    self.model = nn.Sequential(\n",
        "        nn.Linear(input_dim*input_length, hidden_dim),\n",
        "        nn.ReLU(),\n",
        "        nn.Linear(hidden_dim, hidden_dim),\n",
        "        nn.ReLU(),\n",
        "        nn.Linear(hidden_dim, output_dim*output_length),\n",
        "        )\n",
        "\n",
        "  def forward(self, inps):\n",
        "    # expected input dims (batch size, sequence length, number of features)\n",
        "    inps = inps.reshape(inps.shape[0], -1)\n",
        "\n",
        "    # reshape it to (batch size, sequence length, number of features)\n",
        "    return self.model(inps).reshape(inps.shape[0], self.output_length, self.output_dim)"
      ],
      "metadata": {
        "id": "mup2b_KpQyOC"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 4.2.1 Autogressive Rollout vs. Direct Mapping vs. Temperal Bundling"
      ],
      "metadata": {
        "id": "v2gONLqtVKBt"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Autoregressive Rollout:** Make one-step prediction at a time and predict the full rollout by applying the trained network autoregressively $𝑇$ times."
      ],
      "metadata": {
        "id": "jIsH4sufRqZz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_length = 8\n",
        "output_length = 10\n",
        "feat_dim = 3\n",
        "\n",
        "model = MLP(input_dim = feat_dim,\n",
        "            input_length = input_length,\n",
        "            output_dim = feat_dim,\n",
        "            output_length = 1, # make one-step ahead prediction\n",
        "            hidden_dim = 16)\n",
        "\n",
        "loss = 0\n",
        "preds = []\n",
        "loss_fun = torch.nn.MSELoss()\n",
        "x = torch.randn(1, input_length, feat_dim)\n",
        "y = torch.randn(1, output_length, feat_dim)\n",
        "for i in range(output_length):\n",
        "    pred = model(x)\n",
        "    x = torch.cat([x[:, 1:], pred], dim = 1)\n",
        "    loss += loss_fun(pred, y[:,i:i+1])\n",
        "    preds.append(pred)\n",
        "total_pred = torch.cat(preds, dim = 1)"
      ],
      "metadata": {
        "id": "Of6rVTVoSgBE"
      },
      "execution_count": 62,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Direct Mapping:** Learn a function that directly map history to all future $T$ steps."
      ],
      "metadata": {
        "id": "fIeUlKmjSjQ-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = MLP(input_dim = feat_dim,\n",
        "            input_length = input_length,\n",
        "            output_dim = feat_dim,\n",
        "            output_length = output_length,\n",
        "            hidden_dim = 16)\n",
        "\n",
        "total_pred = model(x)\n",
        "loss = loss_fun(total_pred, y)"
      ],
      "metadata": {
        "id": "DUCvOLxGXckp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Temporal bundling[(Link to the paper)](https://arxiv.org/pdf/2202.03376.pdf):** the model generates multiple steps in each autoregressive call, to reduce number of solver calls and therefore error propagation speed."
      ],
      "metadata": {
        "id": "2U9uPnPvSueq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_length = 8\n",
        "output_length = 10\n",
        "feat_dim = 3\n",
        "num_steps_per_call = 2\n",
        "\n",
        "model = MLP(input_dim = feat_dim,\n",
        "            input_length = input_length,\n",
        "            output_dim = feat_dim,\n",
        "            output_length = num_steps_per_call,\n",
        "            hidden_dim = 16)\n",
        "\n",
        "loss = 0\n",
        "preds = []\n",
        "loss_fun = torch.nn.MSELoss()\n",
        "for i in range(output_length//num_steps_per_call):\n",
        "    pred = model(x)\n",
        "    x = torch.cat([x[:, num_steps_per_call:], pred], dim = 1)\n",
        "    loss += loss_fun(pred, y[:,i*num_steps_per_call:(i+1)*num_steps_per_call])\n",
        "    preds.append(pred)\n",
        "total_pred = torch.cat(preds, dim = 1)"
      ],
      "metadata": {
        "id": "aJ-cc6NgYNUR"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 4.2.2 Reversible Instance Normalization [(Link to the paper)](https://openreview.net/forum?id=cGDAkQo1C0p)"
      ],
      "metadata": {
        "id": "P_897YiwVU7d"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class RevIN(nn.Module):\n",
        "  \"\"\"Reinversible instance normalization.\n",
        "\n",
        "  Attributes:\n",
        "    num_features: the number of features or channels\n",
        "    eps: a value added for numerical stability\n",
        "    axis: axis to be normalized\n",
        "    affine: if True, RevIN has learnable affine parameters\n",
        "  \"\"\"\n",
        "  def __init__(self, num_features, eps=1e-5, affine=False, axis=1):\n",
        "    super().__init__()\n",
        "    self.num_features = num_features\n",
        "    self.eps = eps\n",
        "    self.axis = axis\n",
        "    self.affine = affine\n",
        "    if self.affine:\n",
        "      self.init_params()\n",
        "\n",
        "  def forward(self, x, mode):\n",
        "    \"\"\"Apply normalization or inverse normalization.\n",
        "\n",
        "    Args:\n",
        "      x: input ts\n",
        "      mode: normalize or denormalize\n",
        "\n",
        "    Returns:\n",
        "      nomarlized or denormalized x\n",
        "\n",
        "    Raises:\n",
        "      NotImplementedError:\n",
        "    \"\"\"\n",
        "\n",
        "    if mode == \"norm\":\n",
        "      self.get_statistics(x)\n",
        "      x = self.normalize(x)\n",
        "    elif mode == \"denorm\":\n",
        "      x = self.denormalize(x)\n",
        "    else:\n",
        "      raise NotImplementedError\n",
        "    return x\n",
        "\n",
        "  def init_params(self):\n",
        "    # initialize RevIN params:\n",
        "    self.affine_weight = nn.Parameter(torch.ones(1, 1, self.num_features))\n",
        "    self.affine_bias = nn.Parameter(torch.zeros(1, 1, self.num_features))\n",
        "\n",
        "  def get_statistics(self, x):\n",
        "    self.mean = torch.mean(x, dim=self.axis, keepdim=True)\n",
        "    self.stdev = torch.sqrt(torch.std(x, dim=self.axis, keepdim=True) + self.eps)\n",
        "\n",
        "  def normalize(self, x):\n",
        "    x = x - self.mean\n",
        "    x = x / self.stdev\n",
        "    if self.affine:\n",
        "      x = x * self.affine_weight\n",
        "      x = x + self.affine_bias\n",
        "    return x\n",
        "\n",
        "  def denormalize(self, x):\n",
        "    if self.affine:\n",
        "      x = x - self.affine_bias\n",
        "      x = x / (self.affine_weight + self.eps * self.eps)\n",
        "    x = x * self.stdev\n",
        "    x = x + self.mean\n",
        "    return x"
      ],
      "metadata": {
        "id": "VuT41jaQUI9o"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = MLP(input_dim = feat_dim,\n",
        "            input_length = input_length,\n",
        "            output_dim = feat_dim,\n",
        "            output_length = 1,\n",
        "            hidden_dim = 16)\n",
        "\n",
        "loss = 0\n",
        "preds = []\n",
        "loss_fun = torch.nn.MSELoss()\n",
        "normalizer = RevIN(num_features=feat_dim, axis=(1, 2))\n",
        "\n",
        "x = torch.randn(1,input_length,feat_dim)\n",
        "y = torch.randn(1,output_length,feat_dim)\n",
        "for i in range(output_length):\n",
        "\n",
        "    # First normalize the input\n",
        "    norm_inp = normalizer.forward(x, mode=\"norm\")\n",
        "    norm_pred = model(norm_inp)\n",
        "    norm_tgt = normalizer.normalize(y[:,i:i+1])\n",
        "    loss += loss_fun(norm_pred, norm_tgt)\n",
        "\n",
        "    # Denormalize the prediction and add it back to the input for autoregressive call.\n",
        "    pred = normalizer.forward(norm_pred, mode=\"denorm\")\n",
        "    x = torch.cat([x[:, 1:], pred], dim = 1)\n",
        "    preds.append(pred)\n",
        "total_pred = torch.cat(preds, dim = 1)"
      ],
      "metadata": {
        "id": "j2hu6eyIa7kM"
      },
      "execution_count": 65,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 4.2.3 Noise Injection"
      ],
      "metadata": {
        "id": "TjIIlxCzW8DH"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "* Tobias Pfaff et al. [Learning mesh-based simulation with graph networks.](https://arxiv.org/abs/2010.03409)"
      ],
      "metadata": {
        "id": "p4lrQLIeEGeR"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Suppose the current position $x_t = 2$ and we need to predict $x_{t+1} = 3$. We add noise to $x_t$ ($x^*_t = 2.1$) and  train the model the predict the velocity instead of positions, then the target velocity $\\dot{x_t} = 1$ will be adjusted to $\\dot{x^*_t} = 0.9$, so that integration, the model prediction $\\hat{x_{t+1}} = 2.1 + 0.9$ matches $x_{t+1}$."
      ],
      "metadata": {
        "id": "ZDtrt3E0F_cl"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_length = 8\n",
        "output_length = 10\n",
        "feat_dim = 1\n",
        "\n",
        "model = MLP(input_dim = feat_dim,\n",
        "            input_length = input_length,\n",
        "            output_dim = feat_dim,\n",
        "            output_length = 1,\n",
        "            hidden_dim = 16)\n",
        "\n",
        "loss = 0\n",
        "preds = []\n",
        "loss_fun = torch.nn.MSELoss()\n",
        "x = torch.randn(1,input_length,feat_dim)\n",
        "y = torch.randn(1,output_length,feat_dim)\n",
        "vel_y = torch.diff(torch.cat([x[:,-1:], y], dim = 1), dim = 1)"
      ],
      "metadata": {
        "id": "8C2qNOSwffx8"
      },
      "execution_count": 66,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for i in range(output_length):\n",
        "    # add noise to last input step\n",
        "    noise = np.random.normal(0, feat_dim)\n",
        "    x[...,-1] += noise\n",
        "\n",
        "    # Predictint velocity instead of acceleration.\n",
        "    pred_vel = model(x)\n",
        "\n",
        "    #velocity_adj = x_{t+1} - (x_{t} + noise)\n",
        "    loss += loss_fun(pred_vel, vel_y[:,i:i+1] - noise)\n",
        "\n",
        "    pred = pred_vel + x[:, -1:]\n",
        "    x = torch.cat([x[:, 1:], pred], dim = 1)\n",
        "    preds.append(pred)\n",
        "total_pred = torch.cat(preds, dim = 1)"
      ],
      "metadata": {
        "id": "lY_LwumdhALN"
      },
      "execution_count": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "ZG1MQPOrRgJQ"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}