【世界新视野】GNN科研入门计划第二期

2023-06-17 18:46:59 来源:哔哩哔哩

快放假了,来学习吧~

GNN科研入门计划第二期:图神经网络理论、时间序列预测(时空模型)

面向对象:GNN科研入门的大三、大四、研一、博一学生


(资料图片仅供参考)

要求:掌握基本的Python、Pytorch语法 

形式:线上代码讲解+作业布置/讲解+组会+研究方向讨论

图神经网络理论

周期:3个月,5次授课(全是直播形式),2h/次

价格:149元

具体内容:

空间域卷积理论

谱域卷积理论

谱域卷与空间域卷积的关系

消息传递

GCN的底层实现(PyG版)

谱域滤波器设计

项目代码讲解一:GAT、GraphSAGE

项目代码讲解二:BernNet: Learning Arbitrary Graph Spectral Filters via Bernstein Approximation

项目代码讲解三:LON-GNN: Spectral GNNs with Learnable Orthonormal Basis

2. 时间序列预测

周期:3个月,8次授课(2次录播+6次直播),2h/次

价格:199元

注意:项目会有一定难度,都是最新的科研论文项目代码讲解

具体内容:

Graph WaveNet: Graph WaveNet for Deep Spatial-Temporal Graph Modeling

Ada-STNet: Adaptive Spatio-temporal Graph Neural Network for traffic forecasting

AdapGL:  Adaptive Graph Spatial-Temporal Transformer Network for Traffic Flow Forecasting

TSAT: Expressing Multivariate Time Series as Graphs with Time Series Attention

STGODE: Spatial-Temporal Graph ODE Networks for Traffic Flow Forecasting

TLNets: Transformation Learning Networks for long-range time-series prediction

GRAM-ODE: Graph-based Multi-ODE Neural Networks for SpatioTemporal Traffic Forecasting

二选一

GraFITi: Forecasting Irregularly Sampled Time Series using Graphs

STID: Spatial-Temporal Identity: A Simple yet Effective Baseline for Multivariate Time Series Forecasting

标签: