🛣[Deep Learning]Stanford CS224w:Machine Learning with Graphs

---

想说的话🎇

🔝课程网站：http://web.stanford.edu/class/cs224w/

👀一些资源: B站精讲：https://www.bilibili.com/video/BV1pR4y1S7GA/?spm_id_from=333.337.search-card.all.click&vd_source=280e4970f2995a05fdeab972a42bfdd0

https://github.com/TommyZihao/zihao_course/tree/main/CS224W

Slides: http://web.stanford.edu/class/cs224w/slides

Heterogeneous Graphs Transformer(HGT)

Innovation: Decompose heterogeneous graph to Node-type and edge-type dependent attention mechanism

\[ \begin{aligned} AttenHead^i (s,e,t) =& (K^i(s) W^{ATT}_{\phi(e)} Q^i(t)^T) \\ K^i(s) = KLinear^i_{\tau(s)} &(H^{(l-1)[s]}) \\ Q^i(t) = QLinear^i_{\tau(t)} &(H^{(l-1)[t]}) \\ \end{aligned} \]

Each relation \((Type(s), Relation(e), Type(t))\) has a distinct set of projection weights

Understanding Heterogeneous Graph Transformer

阅读地址：Understanding Heterogeneous Graph Transformer

Design space of Herterogeneous GraphNNs

Observation: Each node could receive multiple types of messages from its neighbors, and multiple neighbors may belong to each message type.

\[ h_v^{(l)} = AGG^{(l)}_{all} (AGG^{(l)}_r (\{ \mathbf{m}^{(l)}_u, u \in N_r(v)\}) ) \]

Within each message type, aggregate the messages that belongs to the edge type with \(AGG^{(l)}_r\). Aggregate across the edge types with \(AGG^{(l)}_{all}\)