We refer to this classifier simply as ChebyNet. Layer. Input Size. Output Size. Convolution ... Table 1: ChebyNet classifier architecture in detail for the.
ChebyNet[Defferrardet al., 2016] introduces a polynomial parametrization to convo-lution kernel, i.e., convolution kernel is taken as a polyno-mial function of the diagonal matrix of eigenvalues. Subse-quently, Kipf and Welling[Kipf and Welling, 2017] proposed graph convolutional network (GCN) via a localized rst-order approximation to ChebyNet.
ChebyNet[Defferrardet al., 2016] introduces a polynomial parametrization to convo-lution kernel, i.e., convolution kernel is taken as a polyno-mial function of the diagonal matrix of eigenvalues. Subse-quently, Kipf and Welling[Kipf and Welling, 2017] proposed graph convolutional network (GCN) via a localized rst-order approximation to ChebyNet.
ChebyNet利用切比雪夫多项式的矩阵形式参数化核卷积 g θ g_{\theta} g θ 和特征值矩阵 Λ \Lambda Λ 的多项式组合,经过一些简单运算,使得卷积定理结果中仅保留了要学习的参数 θ \theta θ 和 L L L 的多项式,大大减少的参数量和计算复杂度,图卷积神经网络变得实用 ...