The rapid growth of social networks has enabled the widespread deployment of social bots that manipulate public opinion and disseminate misinformation, thereby posing significant cybersecurity risks. Most existing detection methods for social bots primarily focus on individual features and low-order neighbor information, while neglecting the higher-order topological semantics embedded in frequent substructures, or motifs. This oversight limits their ability to effectively identify sophisticated social bots exhibiting complex behaviors. To address this gap, we propose a novel approach called the multi-motif attention fusion network (MMAFN). Our method enhances the ability to capture complex structures by explicitly modeling higher-order topological relationships within social graphs. Specifically, we first extract typical network motifs from the original graph to generate motif networks that preserve higher-order interaction patterns. We then integrate the original graph structure with these motif networks across multiple scales, constructing a composite adjacency relationship that captures multilevel semantics. Following this integration, we design specialized graph convolution operators that perform parallel feature propagation based on the composite adjacency matrix, generating node representations that incorporate multiorder topological information. Finally, we dynamically merge features from different motif views through an attention mechanism, establishing interaction channels between motifs and ultimately predicting the anomaly probability of user accounts. Experimental results on two large-scale datasets demonstrate that our method significantly outperforms state-of-the-art methods, exhibiting superior detection performance and robust generalization capabilities.

Online social media has become a primary channel for the dissemination of fake news, whose rapid diffusion can generate substantial cross-sectoral impacts. However, most structure-based fake news detection methods are essentially static and overlook temporal dynamics, thereby limiting performance and obscuring underlying mechanisms. To bridge this gap, we propose a framework based on temporal propagation motifs (TPMs) that jointly models structural and temporal dependencies in propagation for fake news detection. Specifically, we introduce TPMs, defined as temporal propagation subgraphs with 3–5 nodes at specific time scales, and derive the higher-order temporal motif degree features to quantify local temporal propagation characteristics at a given time scale$\Delta w$. Building on these features, we propose the multiscale higher-order temporal Motif degree (MSTD) method, which aggregates motif degrees across multiple time scales to capture distinct temporal propagation signatures. Experiments on three real-world datasets show that MSTD outperforms state-of-the-art baselines by an average of 3.2% in accuracy and 2.4% in F1 score. To investigate the drivers of these performance gains, we conduct a temporal-structural decoupling analysis that combines null models with a controlled generative model. The results show that temporal factors are the primary drivers of local propagation patterns, while structural factors play a secondary but complementary role, jointly shaping the local temporal evolution of fake news propagation networks.

The identification of influential nodes in multilayer networks is a rapidly growing area in network science. However, insufficient consideration of both inter- and intra-layer weights in existing research has limited the effectiveness of node identification methods. To address this gap, we propose a novel algorithm, coupling weighted intra-layer and inter-layer influence factors (CWIIIF), which accurately identifies nodes that exert significant influence in multilayer networks. The algorithm integrates weighted intra- and inter-layer influence factors, taking into account the unique properties of multilayer network structures. First, we define a set of layer weight influence parameters, including active nodes, active paths, and communication intersections between layers, to determine the weight of each network layer. We then calculate the intra-layer influence of each node using a combination of K-shell and betweenness centrality methods. Finally, we introduce a set of coupled equations that convert the intra-layer influence vectors into scalar values by incorporating the weights of each layer, producing a final influence score for each node. To validate the effectiveness of our algorithm, we conducted four comparative experiments across nine real-world and one synthetic multilayer networks. The results demonstrate that our algorithm significantly outperforms nine classical and state-of-the-art methods for identifying influential nodes.

Despite their excellent performance in graph representation learning, graph convolutional networks have been proved to be vulnerable to adversarial perturbations on the connectivity between nodes in an unnoticed manner. In this work, by looking into the impacts of adversarial attacks on graph data, we empirically find that the dominant edge-addition attacks generally increase the heterophily between connected nodes, which will fool the transductive inference models on node classification task. To defend against such attacks, we develop a Two-Stage Denoising (TSD) method that aims at removing possible malicious edges so as to mitigate the heterophily issue introduced by attacks. In particular, after a rough removal of the links that have quite low feature similarity, our method further spots the potentially heterophilous links by predicting node labels with a multi-view labeling consensus. This design is based on assumption that if the label predictions for the same node from two different views of a graph data are consistent, then we have a high chance to acquire the reliable labeling. The experiments demonstrate that by denoising a graph this way, the robustness of graph convolutional networks on node classification task is remarkably improved, compared to several strong competitive robust graph neural network models.

Time-series prediction is a fundamental problem in various scientific and engineering domains. Recently, attention-based models have shown great promise in long-term time-series forecasting. However, we prove that vanilla attention is equivalent to a one-step random walk on a bipartite graph between the query and the keys, in which the limited number of walks and simplified graph structure could make it less powerful in capturing complex, high-order featural and temporal dependencies. Inspired by how human brains iteratively reactivate memories through reminding, we propose "Alternated Similarity And Proximity Attention," or ASAP-attention. ASAP-attention employs a random walk on two concurrent views (graphs) that, respectively, capture the featural similarity and the temporal proximity between time points. In particular, the random walk alternately visits the two graphs, each time remembering the previous probability configuration to build a coherent chain of distributions to retrieve useful historical data. This dynamic interplay between temporal and featural clues enhances the model's ability to capture implicit and heterogeneous data dependencies without using positional encoding. When incorporating ASAP-attention with encoder-only Transformer architecture, we observed highly promising results against a wide collection of state-of-the-art methods on various benchmark datasets for long time-series forecasts (e.g., weather, electricity, illness, and exchange-rate data). Our source code is available at https://github.com/jychen01/ASAP-attention.

Continuous molecular graph representations are highly useful for effective molecule property predictions. However, learning graph-specific structure information remains challenging. Current graph neural network models typically aggregate all node embeddings of a graph into a single vector, which, while being an efficient compressor, may not capture the high-level function-related substructure information, e.g., motifs. Moreover, the nodes that are loosely relevant to the graph representation may lead to the sub-optimal representation learning. Towards these limitations, in this paper we propose a new approach to enrich molecular graph representation, where the most relevant motif are selected to represent the graph at a high level and the key-node spanned subgraph is constructed to filter out irrelevant nodes. By comparing to recently proposed unimodal graph neural network models and multi-modal methods, we show that our method achieves state-of-the-art performance on molecular property prediction task.

Emotion classification has wide applications in education, robotics, virtual reality, etc.However, identifying subtle differences between fine-grained emotion categories remains challenging.Current methods typically aggregate numerous token embeddings of a sentence into a single vector, which, while being an efficient compressor, may not fully capture their complex semantic and temporal distributions.To solve this problem, we propose SEmantic ANchor Graph Neural Networks (SEAN-GNN) for fine-grained emotion classification.It learns a group of representative, multi-faceted semantic anchors in the token embedding space: using these anchors as global reference, any sentence can be projected onto them to form a "semantic-anchor graph", with node attributes and edge weights quantifying semantic and temporal information, respectively.The graph structure is well aligned across sentences and, importantly, allows for generating comprehensive emotion representations regarding K different anchors.Message passing on the anchor graph can further integrate the semantic and temporal information and refine the learned features.Empirically, SEAN-GNN produces meaningful semantic anchors and discriminative graph patterns, with promising classification results on 6 popular benchmark datasets against state-of-the-arts.

Contrastive learning is a powerful paradigm for representation learning with prominent success in computer vision and NLP, but how to extend its success to high-dimensional tensors remains a challenge. This is because tensor data often exhibit high-order mode-interactions that are hard to profile and with negative samples growing combinatorially faster than second-order contrastive learning; furthermore, many real-world tensors have ordinal entries that necessitate more delicate comparative levels. To solve the challenge, we propose High-Order Contrastive Tensor Completion (HOCTC), an innovative network to extend contrastive learning to sparse ordinal tensor data. HOCTC employs a novel attention-based strategy with query-expansion to capture high-order mode interactions even in case of very limited tokens, which transcends beyond second-order learning scenarios. Besides, it extends two-level comparisons (positive-vs-negative) to fine-grained contrast-levels using ordinal tensor entries as a natural guidance. Efficient sampling scheme is proposed to enforce such delicate comparative structures, generating comprehensive self-supervised signals for high-order representation learning. Extensive experiments show that HOCTC has promising results in sparse tensor completion in traffic/recommender applications.

MOTIVATION: Predicting the properties of molecules is a fundamental problem in drug design and discovery, while how to learn effective feature representations lies at the core of modern deep-learning-based prediction methods. Recent progress shows expressive power of graph neural networks (GNNs) in capturing structural information for molecular graphs. However, we find that most molecular graphs exhibit low clustering along with dominating chains. Such topological characteristics can induce feature squashing during message passing and thus impair the expressivity of conventional GNNs. RESULTS: Aiming at improving node features' expressiveness, we develop a novel chain-aware graph neural network model, wherein the chain structures are captured by learning the representation of the center node along the shortest paths starting from it, and the redundancy between layers are mitigated via initial residual difference connection (IRDC). Then the molecular graph is represented by attentive pooling of all node representations. Compared to standard graph convolution, our chain-aware learning scheme offers a more straightforward feature interaction between distant nodes, thus it is able to capture the information about long-range dependency. We provide extensive empirical analysis on real-world datasets to show the outperformance of the proposed method. AVAILABILITY AND IMPLEMENTATION: The MolPath code is publicly available at https://github.com/Assassinswhh/Molpath.

Graph representation learning aims at integrating node contents with graph structure to learn nodes/graph representations. Nevertheless, it is found that many existing graph learning methods do not work well on data with high heterophily level that accounts for a large proportion of edges between different class labels. Recent efforts to this problem focus on improving the message passing mechanism. However, it remains unclear whether heterophily truly does harm to the performance of graph neural networks (GNNs). The key is to unfold the relationship between a node and its immediate neighbors, e.g., are they heterophilous or homophilious? From this perspective, here we study the role of heterophily in graph representation learning before/after the relationships between connected nodes are disclosed. In particular, we propose an end-to-end framework that both learns the type of edges (i.e., heterophilous/homophilious) and leverage edge type information to improve the expressiveness of graph neural networks. We implement this framework in two different ways. Specifically, to avoid messages passing through heterophilous edges, we can optimize the graph structure to be homophilious by dropping heterophilous edges identified by an edge classifier. Alternatively, it is possible to exploit the information about the presence of heterophilous neighbors for feature learning, so a hybrid message passing approach is devised to aggregate homophilious neighbors and diversify heterophilous neighbors based on edge classification. Extensive experiments demonstrate the remarkable performance improvement of GNNs with the proposed framework on multiple datasets across the full spectrum of homophily level.

Multiple recent studies show a paradox in graph convolutional networks (GCNs)—that is, shallow architectures limit the capability of learning information from high-order neighbors, whereas deep architectures suffer from over-smoothing or over-squashing. To enjoy the simplicity of shallow architectures and overcome their limits of neighborhood extension, in this work we introduce a biaffine technique to improve the expressiveness of GCNs with a shallow architecture. The core design of our method is to learn direct dependency on long-distance neighbors for nodes, with which only 1-hop message passing is capable of capturing rich information for node representation. Besides, we propose a multi-view contrastive learning method to exploit the representations learned from long-distance dependencies. Extensive experiments on nine graph benchmark datasets suggest that the shallow biaffine graph convolutional networks (BAGCN) significantly outperform state-of-the-art GCNs (with deep or shallow architectures) on semi-supervised node classification. We further verify the effectiveness of biaffine design in node representation learning and the performance consistency on different sizes of training data.

Recently, factorization machine and its variants have shown promising results for context-aware recommender systems (CARS), especially when combined with deep neural networks. Among them, convolutional factorization machine (CFM) is a prominent example. The key to the success of CFM is its 3D convolutional architecture for capturing complex interactions on top of embedded features. However, the resultant computational cost can also be demanding. Moreover, the feature embedding scheme of CFM and other factorization models can be potentially vulnerable to noise. To tackle these issues, in this study we propose two models, namely, the fast convolutional factorization machine (FCFM) that slims down the complete pairwise feature interaction for higher computational efficiency, and adversarial fast convolutional factorization machine (AFCFM) that further enhances the robustness of the model by introducing adversarial noise to the feature interaction image generated by the model. Experimental results on four benchmark datasets prove that the proposed FCFM is nearly five times faster than CFM with competitive performance, while AFCFM improves the performance of the state-of-the-art models by about 8\% with higher efficiency than CFM.

In graph learning, it is fundamental to integrate the features from graph structure and node attributes. Towards this end, graph convolution technique has been devised based on the premise that the similarity of node attributes between two nodes is semantically consistent with their topological proximity. However, many real-networks are found to exhibit the semantic inconsistency, i.e., the phenomenon that directly connected nodes are dissimilar in their attributes. This work is concerned with two related issues: how do we quantitatively measure the semantic consistency between node attributes and graph structure? can we leverage this information to facilitate graph representation? To answer those questions, we first introduce a novel metric to evaluate the semantic consistency in a graph, and then we identify a set of key designs to encode the local semantic consistency information into a type of ego's node feature. Then, we fuse this new node feature with the original node attributes by concatenating the two parts using the semantic consistency metric as weight factor. Experiments on real-world datasets show that linear classifier (e.g. multilayer perceptrons) based on our unsupervised feature learning scheme achieves strong performance across the datasets, especially on the datasets with low semantic consistency, compared to the popular supervised GCNs and other competitive unsupervised graph representation learning models.

Recent spatial-temporal graph-based deep learning methods for Traffic Flow Prediction (TFP) problems have shown superior performance in modeling higher-level spatial interactions and temporal correlations. However, most of these methods suffer from post-fusion efficiency difficulty caused by separate explorations of the spatial communications and the temporal dependencies, which could result in delayed and biased predictions. To address that, we propose a Traffic Gated Graph Neural Networks (Traffic-GGNN) for real-time-fused spatial-temporal representation modeling. Firstly, we adopt bidirectional message passing to capture the location-wise spatial interactions. Secondly, we apply a GRU-based module to explore and aggregate the spatial interactions with the temporal correlations in a real-time fusion way. Lastly, we introduce a self-attention mechanism to reweight the location-based importance and produce the final prediction. Moreover, our proposed model allows end-to-end training thus it is easy to scale to diverse types of traffic datasets and yield better efficiency and effectiveness on three real-world datasets (SZ-taxi, Los-loop, and PEMS-BAY).

Many network analysis tasks like classification over nodes require careful efforts in engineering features used by learning algorithms. Most of recent studies have been made and succeeded in the field of static network representation learning. However, real-world networks are often dynamic and little work has been done on how to describe dynamic networks. In this work, we pose the problem of condensing dynamic networks and introduce SuRep, an encoding-decoding framework which utilizes matrix factorization technique to derive a succinct representation of a dynamic network in any stationary phase. We show that the succinct representation method can uncover the invariant structural properties in the network evolution and derive dense feature representations of the nodes as the byproduct. This method can be easily extended to dynamic attribute networks. For experiments on detecting change points in dynamic networks and network classification with real-world datasets we demonstrate SuRep's potential for capturing latent patterns among nodes.