# DualGSL: Prior-Guided Graph Structure Learning with Node-Adaptive Prediction Fusion

DevFeed: [DualGSL: Prior-Guided Graph Structure Learning with Node-Adaptive Prediction Fusion](<https://devfeed.tech/articles/dualgsl-prior-guided-graph-structure-learning-with-node-adaptive-prediction-fusion-99614.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.32462>)

Author: Yuxia Sun, Yutszyuk Wong, Huihong Chen, Longhao Lv, Chang-Dong Wang, Li Zeng

Published: 2026-10-26T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [graph theory](<https://devfeed.tech/topics/graph-theory.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Graphs](<https://devfeed.tech/topics/graphs.md>), [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [text-classification](<https://devfeed.tech/topics/text-classification.md>)

Tags: [ablation](<https://devfeed.tech/tags/ablation.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [affinity](<https://devfeed.tech/tags/affinity.md>), [alignment](<https://devfeed.tech/tags/alignment.md>), [anchor](<https://devfeed.tech/tags/anchor.md>), [class](<https://devfeed.tech/tags/class.md>), [classification](<https://devfeed.tech/tags/classification.md>), [combines](<https://devfeed.tech/tags/combines.md>), [confidence](<https://devfeed.tech/tags/confidence.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [feature](<https://devfeed.tech/tags/feature.md>), [framework](<https://devfeed.tech/tags/framework.md>), [fusion](<https://devfeed.tech/tags/fusion.md>), [gate](<https://devfeed.tech/tags/gate.md>), [graph](<https://devfeed.tech/tags/graph.md>), [learning](<https://devfeed.tech/tags/learning.md>), [node](<https://devfeed.tech/tags/node.md>), [structure](<https://devfeed.tech/tags/structure.md>)

## AI overview

DualGSL addresses node classification settings where feature-derived graph priors and learned relations have different value for different nodes. It uses two propagation branches: one augments a feature graph with virtual class nodes to incorporate training labels; the other combines learned affinity graphs with the feature-graph prior. A node-adaptive gate fuses their logits using projected representations, their element-wise absolute difference, and prediction confidence. The branches share encoder parameters but retain distinct representations and logits, and a joint objective combines fused and branch-specific classification losses with class-anchor alignment. Against seven baselines on AMG18, 20News, and USPS, the abstract reports the highest average accuracy and Macro-F1 across the three datasets; AMG18 ablations support the benefit of logit fusion. These are abstract-level claims, with no detailed scores provided.

## Source excerpt

arXiv:2610.32462v1 Announce Type: new 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.