DGCDR applies post-hoc disentanglement to GNN-extracted user embeddings and uses a hierarchical contrastive decoder to supervise the separation, achieving state-of-the-art results in six cross-domain recommendation tasks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Enhancing Transferability and Consistency in Cross-Domain Recommendations via Supervised Disentanglement
DGCDR applies post-hoc disentanglement to GNN-extracted user embeddings and uses a hierarchical contrastive decoder to supervise the separation, achieving state-of-the-art results in six cross-domain recommendation tasks.