GCN现有变体不完全汇总（在时空数据挖掘中的应用）

绝地灬酷狼 2023-07-11 05:49 19阅读 0赞

**GCN现有变体汇总（应用篇）**

**Mix Hop（高阶多跳的图特征）融合**

文献：

ICML\_2019

MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixing

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70][]

AAAI\_20: Multi-Range Attentive Bicomponent Graph Convolutional Network for Traffic Forecasting

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 1][]

**2. 两路并行，同时对节点和边的关系建模，形成以边为中心的图网络和以节点为中心的图网络**

NodeNet

EdgeNet

AAAI\_20: Multi-Range Attentive Bicomponent Graph Convolutional Network for Traffic Forecasting （同上）

TKDE\_20: Flow Prediction in Spatio-Temporal Networks Based on Multitask Deep Learning

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 2][]

IJCAI\_19: MR-GNN: Multi-Resolution and Dual Graph Neural Network for Predicting Structured Entity Interactions

两路并行 将GCN的卷积结果和S-LSTM(summary)和I-LSTM（interaction）

1. weighted graph convolution

2. graph-gather layers 经过一层全连接再加起来得到全图的全部信息（和）是表示graph-level的信息

3. 对gt做 graph-state的S-LSTM 也就是对summary graph-gate做 graph-level的LSTM

4. 对gXt和gYt进行连接，再对其做LSTM 就是interaction

5. 最后把得到的结果都concantenate起来 经过全连接 得到1\*k的向量 k表示标注交集的label数。

这个工作得到的都是graph-level的结果，我们也可以拓展到node-level去

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 3][]

**3. Multi-step Prediction: GCN+Seq2Seq**

IJCAI\_19: STG2Seq: Spatial-Temporal Graph to Sequence Model for Multi-step Passenger Demand Forecasting

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 4][]

IJCAI\_19: GSTNet: Global Spatial-Temporal Network for Traffic Flow Prediction

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 5][]

STSGCN [https://github.com/Davidham3/STSGCN][https_github.com_Davidham3_STSGCN] AAAI\_20 【Spatial-Temporal Synchronous Graph Convolutional Networks: A New Framework for Spatial-Temporal Network Data Forecasting】

It is designed for spatio-temporal network data forecasting, which captures complex localized spatial-temporal correlations and heterogeneity with a Spatial-Temporal Synchronous Graph Convolutional Network.

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 6][]

**4. 异质GCN：Hetero-GCN**

KDD\_19: Heterogeneous Graph Neural Network

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 7][]

AAAI\_20: An Attention-based Graph Neural Network for Heterogeneous Structural Learning

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 8][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 9][]

**推荐系统里建模异质网络IntentGC**

**IntentGC: a Scalable Graph Convolution Framework Fusing Heterogeneous Information for Recommendation**

vector-wise/bit-wise

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 10][]

**5. MaskGCN:**

IJCAI\_19: STAR-GCN: Stacked and Reconstructed Graph Convolutional Networks for Recommender Systems

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 11][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 12][]

IJCAI\_19: Masked Graph Convolutional Network

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 13][]

Network embedding就是通过训练特征表示representation来使得图中相邻的节点表征尽可能小，而较远的节点表征尽可能大。或者使得特征表示满足其他的task相关的要求。

Network embedding aims to represent graph nodes in a low dimensional space where the network structure and properties are preserved.

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70]: https://img-blog.csdnimg.cn/20200304110335194.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 1]: https://img-blog.csdnimg.cn/20200304110335676.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 2]: https://img-blog.csdnimg.cn/20200304110337869.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 3]: https://img-blog.csdnimg.cn/20200304110337772.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 4]: https://img-blog.csdnimg.cn/20200304110335698.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 5]: https://img-blog.csdnimg.cn/20200304110336110.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[https_github.com_Davidham3_STSGCN]: https://github.com/Davidham3/STSGCN
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 6]: https://img-blog.csdnimg.cn/20200304110337329.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 7]: https://img-blog.csdnimg.cn/20200304110337692.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 8]: https://img-blog.csdnimg.cn/20200304110337806.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 9]: https://img-blog.csdnimg.cn/20200304110337808.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 10]: https://img-blog.csdnimg.cn/2020030411033766.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 11]: https://img-blog.csdnimg.cn/20200304110337395.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 12]: https://img-blog.csdnimg.cn/20200304110337213.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5_size_16_color_FFFFFF_t_70 13]: https://img-blog.csdnimg.cn/20200304110337211.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p6eXkwOTI5,size_16,color_FFFFFF,t_70