Unfolding GCF prediction shows SSM limitations in neighbor pair weighting; NT-SSM adds type-aware dynamics and yields consistent gains over SSM on multiple datasets and models.
eSjv ′ ∥eu∥∥ej∥ # − eSiv′ ∥eu∥∥ei∥ ! . Since∥e u∥−1 is shared across both terms, it cancels, yielding the upweighting condition: eSiv′ ∥ei∥ > α (tv′ ) I Ej∼ˆπu
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Rethinking Contrastive Learning for Graph Collaborative Filtering: Limitations and a Simple Remedy
Unfolding GCF prediction shows SSM limitations in neighbor pair weighting; NT-SSM adds type-aware dynamics and yields consistent gains over SSM on multiple datasets and models.