In-network linear-trend prediction on SmartNICs reduces communication in multi-server full-graph GNN training by 21 to 45 percent with under 1 percent accuracy loss.
Topology-aware graph pooling networks,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SNI-GNN: SmartNIC-Assisted Full-Graph GNN Training with In-Network Embedding Prediction
In-network linear-trend prediction on SmartNICs reduces communication in multi-server full-graph GNN training by 21 to 45 percent with under 1 percent accuracy loss.