SUPRA resolves the aggregation dilemma in MAGL by decoupling modality-specific MLP paths from a shared lightweight GNN, achieving SOTA performance with 3.5x lower memory and 4.4x faster training.
Deep semantic graph learning via llm based node enhancement.arXiv preprint arXiv:2502.07982, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Beyond the Aggregation Dilemma: Prior-Retaining Decoupled Learning for Multimodal Graphs
SUPRA resolves the aggregation dilemma in MAGL by decoupling modality-specific MLP paths from a shared lightweight GNN, achieving SOTA performance with 3.5x lower memory and 4.4x faster training.