TAAM achieves replay-free, pre-training-free continual graph learning by routing each task to a small node-attentive modulator inserted into a frozen GNN.
• ER-GNN (Zhou and Cao 2021b) is a replay-based method that constructs a memory buffer by storing rep- resentative nodes selected from previous tasks
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
-
Task-Aware Adaptive Modulation: A Replay-Free and Resource-Efficient Approach For Continual Graph Learning
TAAM achieves replay-free, pre-training-free continual graph learning by routing each task to a small node-attentive modulator inserted into a frozen GNN.