Graph Transductive Sharpening: Leveraging Unlabeled Predictions in Node Classification
About
In the transductive setting, where the full graph is observed but node labels are only partially available, progress in semi-supervised node classification has largely focused on architectural innovation. In this paper, we revisit an orthogonal axis: the training objective. We start from a simple observation: transductive models produce predictions for every node during training, including nodes without labels. These unlabeled-node predictions may contain useful training signal, but standard supervised objectives discard them because no ground-truth labels are available. Inspired by the decomposition of cross-entropy into a label-dependent alignment term and a label-independent entropy term, we propose prediction confidence as a natural way to extract this signal in the absence of labels. This motivates Transductive Sharpening (TS): a loss-level modification that minimizes prediction entropy on unlabeled nodes while counterbalancing this effect on labeled nodes. We evaluate Transductive Sharpening across a wide range of node-classification benchmarks and observe consistent performance improvements without requiring any changes to the backbone architecture. Code is available at https://github.com/transductive-sharpening/tunedGNN.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Node Classification | Chameleon | Accuracy45.3 | 867 | |
| Node Classification | Squirrel | Accuracy44.57 | 786 | |
| Node Classification | Pubmed | Accuracy80.74 | 627 | |
| Node Classification | Cora | Accuracy85.74 | 583 | |
| Node Classification | ogbn-arxiv (test) | Accuracy73.3 | 497 | |
| Node Classification | Roman-Empire | Accuracy91.66 | 327 | |
| Node Classification | amazon-ratings | Accuracy56.72 | 309 | |
| Node Classification | CS | Accuracy96.29 | 175 | |
| Node Classification | Computer | Accuracy94.12 | 159 | |
| Node Classification | Photo | Accuracy96.69 | 153 |