cascaded graph label propagation

This work addresses the high computational cost of graph neural networks (GNNs) in semi-supervised node classification, their reliance on the homophily assumption, and the limited adaptability of existing training-free methods to heterophilous graphs. To overcome these challenges, the authors propose a novel training-free, efficient label propagation framework that unifies handling of both homophilous and heterophilous graph structures. The method introduces the local clustering coefficient into an adaptive propagation kernel for the first time and leverages the geometric median to construct robust class prototypes. Extensive experiments demonstrate that the proposed approach achieves accuracy on par with or even surpassing that of trainable GNNs across multiple benchmark datasets, while significantly improving computational efficiency.

Training graph neural networks on large-scale graphs is computationally expensive, and existing graph condensation methods rely heavily on clean labels, leading to significant performance degradation under label scarcity, noise, or distribution shifts. This work proposes a self-supervised graph condensation framework that generates latent pseudo-labels from node embeddings without requiring ground-truth labels. By jointly optimizing prototypes and node assignments, the method constructs a compact synthetic graph whose structural and feature statistics closely match those of the original graph. Theoretical analysis demonstrates that the approach effectively preserves the original graph structure and ensures embedding alignment. Experiments show that the method matches state-of-the-art supervised approaches on clean data and substantially outperforms all baselines under label noise, exhibiting remarkable robustness in both node classification and link prediction tasks.

Graph neural network (GNN) training on multi-label graph data—such as social and biological networks—is computationally inefficient and resource-intensive; moreover, existing graph compression methods are restricted to single-label settings. Method: This paper pioneers the extension of graph compression to the multi-label regime. We propose a multi-label-aware synthetic graph initialization strategy based on K-Center clustering and an optimization objective using binary cross-entropy loss, thereby relaxing the restrictive single-label assumption. Contribution/Results: We systematically establish the first benchmark for multi-label graph compression. Extensive experiments across eight real-world multi-label datasets demonstrate that our method—GCond augmented with K-Center initialization and binary cross-entropy loss—significantly improves both GNN training efficiency and generalization performance. Our approach establishes a scalable, high-fidelity compression paradigm for large-scale multi-label graph learning.

This paper addresses the fundamental over-smoothing problem in Graph Convolutional Networks (GCNs), arising from the difficulty of jointly modeling graph structure and label information across layers. We theoretically identify its root cause within a unified optimization framework and clarify the essential divergence between conventional graph-based semi-supervised learning (grounded in the cluster assumption) and GCNs in their optimization objectives. Building on this analysis, we propose three novel graph convolution paradigms: (i) supervised OGC, (ii) learning-free structure-preserving GGC, and (iii) multi-scale GGCM—each explicitly unifying label guidance with structural preservation. Experiments demonstrate that our methods significantly mitigate over-smoothing and consistently outperform mainstream models—including GCN and GAT—on benchmark datasets such as Cora and Citeseer. This validates the critical importance of co-modeling structural fidelity and label supervision.

To address the under-smoothing and over-smoothing issues inherent in Graph Convolutional Networks (GCNs) for semi-supervised learning on sparsely labeled graphs, this paper proposes GND-Nets: a single-layer graph neural network architecture. Its core innovation lies in introducing a learnable neural diffusion mechanism—embedding neural modules into linear or nonlinear graph diffusion processes to jointly model local neighborhood structures and global topological information. By integrating differentiable graph propagation with localized and global neighborhood aggregation, GND-Nets achieves a favorable trade-off between expressive power and training stability. Extensive experiments on multiple sparsely labeled graph benchmarks demonstrate that GND-Nets significantly outperforms state-of-the-art methods in node classification accuracy while exhibiting faster convergence.

This work addresses the challenge of effectively modeling complex inter-label dependencies in multi-label node classification on graph-structured data. The authors propose a novel approach that decouples the message-passing mechanism of graph neural networks into distinct propagation and transformation operations, enabling explicit analysis and quantification of positive and negative interactions among labels. Building upon this decomposition, they construct a label influence graph to propagate higher-order label effects and introduce a dynamic adjustment mechanism to optimize the learning process. To the best of our knowledge, this is the first method to systematically model inter-label influence relations in non-Euclidean graph data. Extensive experiments demonstrate that the proposed framework significantly outperforms state-of-the-art baselines across multiple benchmark datasets, achieving substantial improvements in multi-label node classification performance.

This work investigates whether the reported performance gains of existing multi-label node classification methods stem from specialized designs or merely from insufficient optimization of classical baselines. To address this, the authors systematically enhance general-purpose GNN architectures—such as GCN, SSGConv, and GCNII—by integrating standard techniques including normalization, Dropout, and residual connections to construct strong baselines. Extensive experiments demonstrate that these well-tuned classical models outperform current specialized approaches on four out of five benchmark datasets and achieve state-of-the-art results across various experimental settings. These findings underscore the critical importance of employing rigorously optimized baselines in multi-label graph learning research to ensure meaningful methodological comparisons.

This work addresses the high cost of re-annotation in document layout analysis caused by evolving label categories by proposing a plug-and-play pseudo-labeling framework tailored for object detection. It introduces label propagation to document layout analysis for the first time, constructing multimodal object representations through the fusion of visual, textual, and positional embeddings. This enables efficient semi-supervised category propagation using only a small set of annotated samples. Experimental results on the D4LA dataset demonstrate that with merely 10% of the labeled data, the method achieves a mean average precision (mAP) of 54.0%, equivalent to 81.6% of the fully supervised performance, thereby substantially reducing manual annotation effort.

本文探讨了如何利用视觉表示来增强图神经网络的推理和学习能力，提出了视觉与图结合的新方法，并归纳了三个研究方向。

Existing graph construction methods based on feature similarity often introduce semantically irrelevant adjacency relationships, which hinder the performance of semi-supervised image classification. This work proposes a novel approach that leverages large language models (LLMs) to refine the semantic structure of image graphs. Specifically, it first employs a vision-language model together with an LLM to generate textual descriptions for images, then utilizes the LLM to assess pairwise semantic similarity between images. These similarity scores are used to reweight edges in both k-nearest neighbor (kNN) and mutual kNN graphs, followed by text-guided edge pruning to enhance semantic consistency within the graph. The resulting semantically refined graphs significantly improve classification accuracy of graph convolutional networks across multiple backbone architectures, with particularly pronounced gains observed in kNN-based graphs.