概
本文在 Attributed (结点带属性) Multiplex (两个结点间可能有多个关系, 边) Heterogeneous (异构) 图上进行讨论. 这里不会讨论方法, 只记录下这些定义.
符号说明
各种定义
Heterogeneous network
假设存在一 node 映射 \(\phi: \mathcal{V} \rightarrow \mathcal{O}\) 将结点映射为某个结点类型, 一 edge 映射 \(\psi: \mathcal{E} \rightarrow \mathcal{R}\) 将边映射为某个边类型.
\(G = (\mathcal{V}, \mathcal{E})\) 为一异构图若 \(|\mathcal{O}| + |\mathcal{R}| > 2\).
Attributed network
令 \(G = (\mathcal{V, E, A})\) 为一 attributed network, 若 \(\mathcal{A} = \{\mathbf{x}_i|v_i \in \mathcal{V}\}\) 赋予每个结点 \(v_i\) 以属性 \(\mathbf{x}_i\).
Attributed multiplex network
称 attributed network \(G = (\mathcal{V, E, A}), \mathcal{E}=\bigcup_{r \in \mathcal{R}} \mathcal{E}_r, |\mathcal{R}| > 1\) 为 attributed multiplex heterogeneous network.
代码
- Representation Heterogeneous Attributed Multiplex Learningrepresentation heterogeneous attributed multiplex recommendation heterogeneous preference learning heterogeneous federated learning yourself representation incremental classifier learning representation sparsification learning robust representation generative synthesis learning image representation continuous learning representation unsupervised degradation learning multiplex attributed