Adding virtual nodes with a semi-adaptive adjacency matrix improves long-term traffic flow prediction accuracy over a standard STGCN baseline.
Oversquashing in gnns through the lens of information contraction and graph expansion
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Virtual Nodes Improve Long-term Traffic Prediction
Adding virtual nodes with a semi-adaptive adjacency matrix improves long-term traffic flow prediction accuracy over a standard STGCN baseline.