MAGNET fuses partial multimodal embeddings via dynamic attention and constructs missingness-aware patient graphs for GNN-based cancer classification, outperforming prior fusion methods on three multiomics datasets.
MedFuse: multi-modal fusion with clinical time-series data and chest x-ray images
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
-
Missing-Modality-Aware Graph Neural Network for Cancer Classification
MAGNET fuses partial multimodal embeddings via dynamic attention and constructs missingness-aware patient graphs for GNN-based cancer classification, outperforming prior fusion methods on three multiomics datasets.