search

d-ailin / GDN

Implementation code for the paper "Graph Neural Network-Based Anomaly Detection in Multivariate Time Series" (AAAI 2021)
MIT License
482 stars 141 forks source link

Question about gnn_layers #42

Open xeroCBW opened 2 years ago

xeroCBW commented 2 years ago

Why not use nodes to create the gnn layer, but instead use edges.

self.gnn_layers = nn.ModuleList([ GNNLayer(input_dim, dim, inter_dim=dim+embed_dim, heads=1) for i in range(edge_set_num) ])

d-ailin commented 2 years ago

The edge_set_num is the number of the edge set, which would be one as we only use one set of edges, which contains fully connected edges. The parameter setting in the gnn layer is based on the input_dim, dim and inter_dim variables.