
Addressing the variance imbalance between seen and novel classes is crucial for effective open-world semi-supervised learning in graph data.


coremsg

open-world-semi-supervised-learning-for-node-classification-addressing-imbalance-in-graph-data


Open-World Semi-Supervised Learning for Node Classification: Addressing Imbalance in Graph Data



Novel graph embedding algorithms tailored for node classification, focusing on hubs and biased random walks.



Hub-aware Random Walk Graph Embedding Methods for Node Classification
