{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"$$\n",
"\\newcommand{\\mat}[1]{\\boldsymbol {#1}}\n",
"\\newcommand{\\mattr}[1]{\\boldsymbol {#1}^\\top}\n",
"\\newcommand{\\matinv}[1]{\\boldsymbol {#1}^{-1}}\n",
"\\newcommand{\\vec}[1]{\\boldsymbol {#1}}\n",
"\\newcommand{\\vectr}[1]{\\boldsymbol {#1}^\\top}\n",
"\\newcommand{\\rvar}[1]{\\mathrm {#1}}\n",
"\\newcommand{\\rvec}[1]{\\boldsymbol{\\mathrm{#1}}}\n",
"\\newcommand{\\diag}{\\mathop{\\mathrm {diag}}}\n",
"\\newcommand{\\set}[1]{\\mathbb {#1}}\n",
"\\newcommand{\\cset}[1]{\\mathcal{#1}}\n",
"\\newcommand{\\norm}[1]{\\left\\lVert#1\\right\\rVert}\n",
"\\newcommand{\\pderiv}[2]{\\frac{\\partial #1}{\\partial #2}}\n",
"\\newcommand{\\bb}[1]{\\boldsymbol{#1}}\n",
"\\newcommand{\\E}[2][]{\\mathbb{E}_{#1}\\left[#2\\right]}\n",
"\\newcommand{\\ip}[3]{\\left<#1,#2\\right>_{#3}}\n",
"\\newcommand{\\given}[]{~\\middle\\vert~}\n",
"$$\n",
"\n",
"# CS236605: Deep Learning\n",
"# Tutorial 8: Geometric deep learning"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"## Introduction\n",
"\n",
"In this tutorial, we will cover:\n",
"\n",
"- Filters on graphs\n",
"- Graph Laplacians and their meaning\n",
"- Graph convolution layers\n",
"- Application: semi-supervised node classification"
]
},
{
"cell_type": "code",
"execution_count": 104,
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"running on: cpu\n"
]
}
],
"source": [
"# Setup\n",
"%matplotlib inline\n",
"import os\n",
"import sys\n",
"import math\n",
"import time\n",
"import torch\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"\n",
"plt.rcParams['font.size'] = 20\n",
"data_dir = os.path.expanduser('~/.pytorch-datasets')\n",
"device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
"print(f'running on: {device}')"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"## Theory Reminders: Non-Euclidean domains"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"So far we have heavily utilized convolutional layers in our deep learning models."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Convolutional layers operate on an input tensor $\\vec{x}$ containing $M$ feature maps.\n",
"The $j$-th feature map of the output tensor $\\vec{y}$ can be written as\n",
"\n",
"$$\n",
"\\vec{y}^j = \\sum_{i=1}^{M} \\vec{w}^{ij}\\ast\\vec{x}^i+b^j,\n",
"$$\n",
"\n",
"where $\\ast$ denotes convolution, and $x^i$ is the $i$-th input feature map.\n",
"\n",
"Convolutional layers have some important advantages:"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"- Number of parameters not dependent on the input dimensions.\n",
"- Parameters shared across the spatial dimensions of the input.\n",
"- Output dimension changes based on input dimension."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"### Convolution in the time and frequency domains"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Let's dig deeper into convolutions and what they do. Recall the definition of convolution:\n",
"\n",
"$$\n",
"\\left\\{\\vec{w}\\ast\\vec{x}\\right\\}_j = \\sum_{i} w_{j-i} x_{i}.\n",
"$$\n",
"\n",
"\n",
"\n"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"This is a **linear** operator, which can therefore be represented simply as a matrix multiplication.\n",
"\n",
"Moreover, it's **shift-equivariant** meaning that a shifted input will result in an output shifted by the same amount.\n",
"Due to the shift-equivariance, the matrix representing a convolution is always a **Toeplitz** matrix.\n",
"\n",
""
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"From linear algebra we know that a linear operator can sometimes be **diagonalized** by computing it's **eigendecomposition** (if it exists), and using it's eigenvectors as a basis."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"It turns out that **all Topelitz operators** have the same basis (same eigenvectors).\n",
"\n",
"This is the **Fourier** basis,\n",
"\n",
"$$\n",
"\\vec{\\phi}^{\\vec{\\xi}}_{\\vec{n}} = \\exp(j 2\\pi \\vectr{\\xi}\\vec{n}),\\text{ where } {\\vec{n}\\in\\set{Z}^d}, \\vec{\\xi}\\in[0,1]^d\\triangleq\\blacksquare,\n",
"$$\n",
"\n",
"which is composed of discrete harmonic functions of varying continuous frequency $\\vec{\\xi}$ sampled in space at $\\vec{n}$."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"For any Toeplitz operator $\\cset{W}$, since we can change matrix-product to convolution we have, \n",
"\n",
"$$\n",
"\\begin{align}\n",
"\\left(\\cset{W}\\vec{\\phi}^\\vec{\\xi}\\right)_\\vec{n} =\n",
"\\left(\\vec{w}\\ast\\vec{\\phi}^\\vec{\\xi}\\right)_\\vec{n} &=\n",
"\\sum_{\\vec{m}}\\vec{w}_\\vec{m} \\exp(j 2\\pi \\vectr{\\xi}(\\vec{n-m})) \\\\ &=\n",
"\\exp(j 2\\pi \\vectr{\\xi}\\vec{n})\\sum_{\\vec{m}}\\vec{w}_\\vec{m} \\exp(-j 2\\pi \\vectr{\\xi}\\vec{m}) \\\\ &=\n",
"\\vec{\\phi}^{\\vec{\\xi}}_{\\vec{n}} \\ip{\\vec{w}}{\\vec{\\phi}^{\\vec{\\xi}}}{\\ell^2(\\set{Z}^d)} \\\\ &=\n",
"\\vec{\\phi}^{\\vec{\\xi}}_{\\vec{n}}\\lambda_{\\vec{\\xi}}.\n",
"\\end{align}\n",
"$$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"Thus $\\vec{\\phi}^{\\vec{\\xi}}_{\\vec{n}}$ is an eigenvector of any Toeplitz operator $\\cset{W}$,\n",
"with eigenvalue $\\lambda_{\\vec{\\xi}}=\\hat{w}(\\vec{\\xi})=\\ip{\\vec{w}}{\\vec{\\phi}^{\\vec{\\xi}}}{\\ell^2(\\set{Z}^d)}$."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"We can define the **Fourier transform** $\\cset{F}$ and it's inverse in-terms of inner products with these eigenvectors.\n",
"\n",
"$$\n",
"\\begin{align}\n",
"(\\cset{F}\\vec{x})(\\vec{\\xi}) &= \\ip{\\vec{x}}{\\vec{\\phi}^{\\vec{\\xi}}}{\\ell^2(\\set{Z}^d)}\n",
"= \\sum_{\\vec{n}\\in\\set{Z}^d} \\vec{x}_{\\vec{n}}\\vec{\\phi}^{\\vec{-\\xi}}_{\\vec{n}} = \\hat{x}(\\vec{\\xi}) \\\\\n",
"(\\cset{F^{-1}}\\hat{x})_\\vec{n} &= \\ip{\\hat{x}}{\\vec{\\phi}^{\\ast}_{\\vec{n}}}{L^2(\\blacksquare)}\n",
"= \\int_{\\blacksquare} \\hat{x}(\\vec{\\xi})\\vec{\\phi}^{\\vec{\\xi}}_{\\vec{n}}d\\vec{\\xi}.\n",
"\\end{align}\n",
"$$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"This means we can write any time-domain convolution in terms of frequency-domain product:\n",
"\n",
"$$\n",
"\\begin{align}\n",
"\\vec{y} = \\vec{w}\\ast\\vec{x}\n",
"&= \\cset{W}\\vec{x}\n",
"= \\cset{W} \\int_{\\blacksquare} \\ip{\\vec{x}}{\\vec{\\phi}^{\\vec{\\xi}}}{}\\vec{\\phi}^{\\vec{\\xi}} d\\vec{\\xi}\n",
"= \\int_{\\blacksquare} \\hat{x}(\\vec{\\xi}) \\cset{W} \\vec{\\phi}^{\\vec{\\xi}} d\\vec{\\xi}\n",
"= \\int_{\\blacksquare} \\hat{x}(\\vec{\\xi}) \\hat{w}(\\vec{\\xi}) \\vec{\\phi}^{\\vec{\\xi}} d\\vec{\\xi} \\\\\n",
"&= \\ip{\\hat{x}\\cdot\\hat{w}}{\\vec{\\phi}^{\\vec{-\\xi}}}{L^2(\\blacksquare)} \\\\\n",
"& = \\cset{F^{-1}}\\left\\{\\hat{x}\\cdot\\hat{w}\\right\\}.\n",
"\\end{align}\n",
"$$\n",
"\n",
"This result is known as the **convolution theorem**."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"### Graphs"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"We will focus our discussion here on Graphs, which are an interesting and useful domain for learning. \n",
"They are also general enough to represent many other non-Euclidean domains,\n",
"such as point-clouds and discretized manifolds."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"A graph $\\cset{G} = (\\cset{V},\\cset{E}, \\mat{A}, \\mat{W})$, can be represented with\n",
"\n",
"- A vertex set $\\cset{V} = \\left\\{1,2,\\dots,n\\right\\}$ \n",
"- An edge set $\\cset{E}\\subseteq \\cset{V}\\times {V}$\n",
"- Vertex-weight matrix $\\mat{A}=\\mathrm{diag}\\left\\{a_i\\right\\}_{i=1}^{n}$.\n",
"- Edge-weight matrix $\\mat{W}$, where $W_{ij}=0\\Rightarrow(i,j)\\neq\\cset{E}$.\n",
"\n",
"\n",
"\n"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Signals can be defined with the vertices or edges of a graph as their domain.\n",
"An appropriate inner-product space can be defined for them."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"A **vertex signal** (field) can be defined as a function $f: \\cset{V}\\rightarrow\\set{R}$.\n",
"\n",
"We'll represent these as a vector $\\vec{f}\\in\\set{R}^n$, which can be thought of as a single feature map on the graph.\n",
"\n",
"The inner product for the space of vertex signals is\n",
"$\n",
"\\ip{\\vec{f}}{\\vec{g}}{\\ell^2(\\cset{V})}=\\sum_{i\\in\\cset{V}} a_i f_i g_i^\\ast .\n",
"$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"Similarly, an **edge signal** (flow) can be defined as an anti-symmetric function $F:\\cset{E}\\rightarrow\\set{R}$,\n",
"which can be represented as a matrix $\\mat{F}\\in\\set{R}^{n\\times n}$.\n",
"\n",
"The space of edge signals is equipped with the inner product\n",
"$\n",
"\\ip{\\mat{F}}{\\mat{G}}{\\ell^2(\\cset{E})}=\\sum_{(i,j)\\in\\cset{E}} w_{ij} F_{ij} G^\\ast_{ij} .\n",
"$\n"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"#### Graph Laplacian"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"The graph Laplacian, also known as the Laplace-Beltrami operator is a measure of **smoothness** of a a vertex field, i.e. how quickly it changes between **adjacent** vertices."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"More formally, it can be defined as an operator over vertex fields, returning a new vertex field: $\\Delta:\\ell^2(\\cset{V})\\rightarrow\\ell^2(\\cset{V})$.\n",
"We can write:\n",
"\n",
"$$\n",
"\\begin{align}\n",
"(\\Delta \\vec{f})_i = \\frac{1}{a_i} \\sum_{j\\neq i} w_{ij}\\left(f_i-f_j\\right)\n",
"&= \\frac{1}{a_i}\\left(\\left(\\sum_{j\\neq i} w_{ij}\\right) f_i - \\sum_{j\\neq i}w_{ij}f_j\\right) \\\\\n",
"&= \\frac{1}{a_i}\\left(d_i f_i - \\sum_{j\\neq i}w_{ij}f_j\\right).\n",
"\\end{align}\n",
"$$\n",
"\n",
"Where $d_i$ is the generalized degree of node $i$ (sum of touching-edge weights).\n",
"\n",
"We can thus interpret the Laplacian value at a node as the weighted self-value minus weighted neighbor value, which consolidates the notion of a **smooth** function on a graph."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"We can represent this operator as a Laplacian matrix $\\mat{\\Delta} = \\mat{A}^{-1}\\left(\\mat{D}-\\mat{W}\\right)\\in\\set{R}^{n\\times n}$, where $\\mat{D}$ is a diagonal degree matrix."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"#### Eigendecomposition of Graph Laplacian"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"In the Euclidean domain, the **Laplacian**,\n",
"$\n",
"\\Delta f(\\vec{x}) = -\\mathrm{div}\\nabla f (\\vec{x})= - \\sum_i \\pderiv{{}^2 f}{x_i^2}\n",
"$ is a translation equivariant operator and therefore also has a Fourier eigenbasis!\n",
"\n",
"For example, notice that $\\Delta e^{jkx} = - \\frac{d^2}{dx^2} e^{jkx} = k^2 e^{jkx}.$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"Similarly, for a graph domain, we can compute the eigenvectors of the Laplacian matrix (which always has an eigen decomposition), and use them as a basis for representing any vertex function."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"We can write the eigendecomposition as\n",
"\n",
"$$\n",
"\\mat{\\Delta}\\mat{\\Phi} = \\mat{\\Lambda}\\mat{\\Phi}\n",
"$$\n",
"\n",
"where as usual $\\mat{\\Phi}$ contains the eigenvectors of $\\mat{\\Delta}$ in its columns and $\\mat{\\Lambda}$ is a diagonal matrix of eigenvalues.\n",
"Note that $\\mat{\\Phi}$ is $\\mat{A}$-orthonormal, i.e. $\\mattr{\\Phi}\\mat{A}\\mat{\\Phi}=\\mat{I}$."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"The eigenfunctions of the laplacian form the smoothest-possible basis function over a specific graph (they minimize the Dirichlet energy).\n",
"\n",
""
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Since the Laplacian eigenvectors form a basis for the space $\\ell^2(\\cset{V})$,\n",
"we can represent any vertex field as a linear combination of them:\n",
"\n",
"$$\n",
"\\vec{f}=\\sum_k \\hat{f}_k\\vec{\\phi}_k = \\sum_k \\ip{\\vec{f}}{\\vec{\\phi}_k}{\\ell^2(\\cset{V})} \\vec{\\phi}_k.\n",
"$$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"We can thus define the **graph-specific** Fourier basis as these eigenvectors.\n",
"\n",
"The Fourier transforms for a graph function is then,\n",
"$$\n",
"\\begin{align}\n",
"(\\cset{F}\\vec{f}) &= \\hat{\\vec{f}} = \\mattr{\\Phi}\\vec{f} \\\\\n",
"(\\cset{F^{-1}}\\hat{\\vec{f}}) &= \\vec{f} = \\mat{\\Phi}\\hat{\\vec{f}}.\n",
"\\end{align}\n",
"$$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"### Spectral convolution layer for graphs"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"In the Euclidean case, we have seen that applying a linear filter amounts to time-domain convolution or freq-domain product.\n",
"\n",
"$$\n",
"\\cset{W}\\vec{f} = \\vec{w}\\ast\\vec{f}\n",
"= \\cset{F^{-1}}\\left\\{\\hat{f}\\cdot\\hat{w}\\right\\}\n",
"= \\cset{F^{-1}}\\left\\{(\\cset{F}f)\\cdot\\hat{w}\\right\\}.\n",
"$$\n",
"\n",
"This result was obtained by directly using the definition of the convolution operator and the Fourier basis."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"In a non-euclidean domain, the notion of translation (shift) is undefined, so we can't easily define convolution.\n",
"\n",
"However, we can **define** $\\cset{W}f \\triangleq \\cset{F^{-1}}\\left\\{\\hat{f}\\cdot\\hat{w}\\right\\}$, skipping the time-domain convolution, and plug in the domain-specific definition of the Fourier transform based on the graph Laplacian:\n",
"\n",
"$$\n",
"\\cset{W}f = \\cset{F^{-1}}\\left\\{\\hat{f}\\cdot\\hat{w}\\right\\}\n",
"= \\sum_k \\hat{f}_k \\hat{w}_k \\vec{\\phi}_k\n",
"= \\sum_k \\ip{\\vec{f}}{\\vec{\\phi}_k}{} \\hat{w}_k \\vec{\\phi}_k\n",
"= \\mat{\\Phi}\\mat{\\hat W}\\mattr{\\Phi}\\vec{f},\n",
"$$\n",
"\n",
"where $\\mat{\\hat W}$ is a diagonal matrix representing the filter parameters."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Based on this method for applying a filter on a graph function, we can define our Spectral CNN layer.\n",
"\n",
"- Input: vector-valued vertex field\n",
" $\\mat{X} = (\\vec{x}^1, \\dots, \\vec{x}^m),\\ \\vec{x}^i\\in\\ell^2({\\cset{V}})$.\n",
" Each vertex has $m$ associated features, i.e. our graph has $m$ feature maps.\n",
"- Output: vector-valued vertex field\n",
" $\\mat{Y} = (\\vec{y}^1, \\dots, \\vec{y}^{m'}),\\ \\vec{y}^i\\in\\ell^2({\\cset{V}})$.\n",
" \n",
"Thus,\n",
"\n",
"$$\n",
"\\vec{y}^j = \\varphi\\left( \\sum_{i=1}^{m} \\mat{\\Phi}\\mat{\\hat{W}}^{ij}\\mattr{\\Phi}\\vec{x}^i + b^j \\right).\n",
"$$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Unfortunately, this type of graph convolution layer suffers from severely inhibiting **drawbacks**."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"- Number of learned parameter depends on domain size $n$ and is not fixed.\n",
"- Degree of freedom per eigenvector: potential to create very high frequency (non-smooth) filters. Recall that smoothness in frequency equates to spatial-localization.\n",
"- Need to compute both directions of the Fourier transform, each time with $\\cset{O}(n^2)$ cost.\n",
"- Learned filters are domain-dependent and will not generalize well (i.e. to a different graph)."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"\n",
"\n",
"\n",
"Left: a signal $f$ on a manifold $\\cset{X}$.\n",
"\n",
"Middle: a spectral edge-detection filter, represented as $\\mat{W}$, is applied to $f$.\n",
"\n",
"Right: The same filter applied to the same function but on a different domain $\\cset{Y}$, which is an isometry of $\\cset{X}$. This causes the filter to produce completely different results."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"### Spatial convolution layer for graphs"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"To overcome the drawbacks imposed by the Spectral CNN layer, we will first limit our learned parameters to be smooth interpolations of the Laplacian eigenvalues.\n",
"\n",
"This will simultaneously address the first two drawbacks listed above.\n"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"We'll define our learned filter parameters as $\\hat{\\mat W}^{ij}=\\mathrm{diag}\\{\\mat{B}\\vec{\\alpha}^{ij}\\}$, where\n",
"\n",
"$$\n",
"\\begin{pmatrix}\n",
"\\hat{w}^{ij}_1\\\\\n",
"\\vdots\\\\\n",
"\\hat{w}^{ij}_n\\\\\n",
"\\end{pmatrix}\n",
"=\n",
"\\begin{pmatrix}\n",
"\\beta_1(\\lambda_1) & \\cdots & \\beta_q(\\lambda_1) \\\\\n",
"\\vdots & \\ddots & \\vdots \\\\\n",
"\\beta_1(\\lambda_n) & \\cdots & \\beta_q(\\lambda_n) \\\\\n",
"\\end{pmatrix}\n",
"\\begin{pmatrix}\n",
"\\alpha^{ij}_1\\\\\n",
"\\vdots\\\\\n",
"\\alpha^{ij}_q\\\\\n",
"\\end{pmatrix}.\n",
"$$\n",
"\n",
"The $\\beta_i(\\lambda)$'s are smooth functions such as polynomials."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"The number of parameters no longer depends on domain size. The use of interpolation helps ensure we get smooth spectral coefficients for our filter.\n",
"\n",
"How many parameters do we have in our layer now?"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"$q\\cdot m \\cdot m'$. No dependence on $n$."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"The second improvement comes from discarding the Fourier transform.\n",
"If we select $\\beta_k(\\lambda)=\\lambda^k$, then our spatial filter is\n",
"\n",
"$$\n",
"\\begin{align}\n",
"\\mat{W} = \\mat{\\Phi}\\mat{\\hat W}\\mattr{\\Phi}\n",
"&= \n",
"\\mat{\\Phi}\n",
"\\begin{pmatrix}\n",
"\\sum_{k=1}^{q}\\alpha_k\\lambda_1^k & & \\\\\n",
" & \\ddots & \\\\\n",
" & & \\sum_{k=1}^{q}\\alpha_k\\lambda_n^k \\\\\n",
"\\end{pmatrix}\n",
"\\mattr{\\Phi}\n",
"=\n",
"\\sum_{k=1}^{q}\\alpha_k\n",
"\\mat{\\Phi}\n",
"\\begin{pmatrix}\n",
"\\lambda_1^k & & \\\\\n",
" & \\ddots & \\\\\n",
" & & \\lambda_n^k \\\\\n",
"\\end{pmatrix}\n",
"\\mattr{\\Phi} \\\\\n",
"&= \\sum_{k=1}^{q}\\alpha_k \\mat{\\Phi} \\mat{\\Lambda}^k \\mattr{\\Phi}\n",
"= \\sum_{k=1}^{q}\\alpha_k \\mat{\\Delta}^k.\n",
"\\end{align} \n",
"$$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Finally, this gives us a **spatial** (Fourier-free) graph CNN layer:\n",
"\n",
"$$\n",
"\\vec{y}^j = \\varphi\\left( \\sum_{i=1}^{m} \\sum_{k=1}^{q} \\alpha^{ij}_k \\mat{\\Delta}^k \\vec{x}^i + b^j \\right).\n",
"$$"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"Note:\n",
"\n",
"- The powers of the Laplacian matrix $\\mat{\\Delta}^k$ are local in $k$-rings around each node.\n",
"- We're no longer relying on the Laplacian eigendecomposition, but applying a filter directly in the spatial domain.\n",
"- Time complexity now depends on the graph Laplacian sparsity."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"Instead of using a regular polynomial, we can generalize by applying other functions of the Laplacian. This can be written slightly more generally as\n",
"\n",
"$$\n",
"\\mat{Y}=\\varphi\\left( \\sum_{k=1}^{q} \\beta_k(\\mat{\\Delta}) \\mat{X} \\mat{\\alpha}_k + \\vec{b} \\right).\n",
"$$\n",
"\n",
"Where $\\mat{\\Delta}$ is the $n\\times n$ Laplacian, $\\mat{X}$ is an $n\\times m$ features matrix, and $\\mat{\\alpha}_k$ is an $m\\times m'$ weight matrix."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"## Application: Semi-supervised node classification"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"We'll now face the problem of classifying nodes in a small social-network graph, where we only have true labels for a tiny subset of the nodes.\n",
"\n",
"The approach is based on [Kipf & Welling (2016)](http://arxiv.org/abs/1609.02907)."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
"### Dataset"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"We'll use a toy dataset of a small social network represented as a graph, known as \"Zachary's Karate Club\" (Zachary, W. W., J. Anthr. R., 1977).\n",
"The network has two **communities** which we'll try to detect, by classifying each node as belonging to one of them.\n",
"\n",
"From [Wikipedia](https://en.wikipedia.org/wiki/Zachary%27s_karate_club),\n",
"\n",
"> A social network of a karate club was studied by Wayne W. Zachary for a period of three years from 1970 to 1972.\n",
"> The network captures 34 members of a karate club, documenting links between pairs of members who **interacted outside the club**.\n",
"> During the study a conflict arose between the **administrator** \"John A\" and **instructor** \"Mr. Hi\" (pseudonyms), which led to the split of the club into two.\n",
"> \n",
"> Half of the members formed a new club around Mr. Hi; members from the other part found a new instructor or gave up karate. Based on collected data Zachary correctly assigned all but one member of the club to the groups they actually joined after the split. \n",
"> "
]
},
{
"cell_type": "code",
"execution_count": 106,
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"outputs": [],
"source": [
"# NetworkX is a very useful graph library for python\n",
"import networkx as nx\n",
"\n",
"G = nx.karate_club_graph()\n",
"ID_INSTR = 0\n",
"ID_ADMIN = 33\n",
"ID_MEMBERS = set(G.nodes()) - {ID_ADMIN, ID_INSTR}"
]
},
{
"cell_type": "code",
"execution_count": 107,
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"