Graph learning improves node classification by combining feature and structure insights
DualGSL: Prior-Guided Graph Structure Learning with Node-Adaptive Prediction Fusion
Machine Learning
Summary
Predicting information about items connected in a graph can be tricky because different parts of the graph have different clues about the answers. The authors introduce DualGSL, a method that learns how to best combine predictions from two sources—a graph based on features and one based on learned relationships. By adjusting the combination uniquely for each item, the method makes better overall predictions. Tests on malware, text, and image data show this way of combining predictions works better than other approaches.
What this means in practice
- •For cybersecurity analysts: Improve malware attribution by combining feature-based and learned graph relations for more accurate node classification in network data.
- •For natural language processing teams: Enhance text classification systems by integrating prior knowledge and learned graph structures with node-adaptive prediction fusion.
Authors
Yuxia Sun, Yutszyuk Wong, Huihong Chen, Longhao Lv, Chang-Dong Wang, Li Zeng
Abstract
Graph structure learning (GSL) adapts propagation graphs to prediction tasks, but the relative predictive value of feature-derived priors and learned relations can vary across nodes. We propose DualGSL, a prior-guided framework for transductive node classification that learns node-specific weights for combining predictions from separate prior and task-adaptive propagation paths. The prior branch augments a feature graph with virtual class nodes to incorporate training-label information, while the structural branch combines learned affinity graphs with the feature-graph prior. The two branches share encoder parameters while preserving distinct representations and classification logits. A node-adaptive gate fuses the branch logits based on their projected representations, element-wise absolute difference, and prediction confidence. A joint objective integrates fused and branch-specific classification losses with prior-branch class-anchor alignment, coordinating graph construction, representation learning, and prediction fusion. Experiments against seven baselines on the AMG18 malware attribution dataset and the widely used 20News text classification and USPS handwritten-digit recognition benchmarks show that DualGSL achieves the highest cross-dataset average accuracy and Macro-F1 among the evaluated methods. Ablation studies on AMG18 further demonstrate the benefit of combining prior and learned propagation through logit fusion.