The paper claims node attribute shift matters more than topology shift in graph domain adaptation, but the proof and loss formulation contain critical errors.
Unsupervised domain adaptive graph convolutional networks
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
On the Benefits of Attribute-Driven Graph Domain Adaptation
The paper claims node attribute shift matters more than topology shift in graph domain adaptation, but the proof and loss formulation contain critical errors.