Unfair recommendations stem from user-sensitive attributes and information transmission biases. Graph-structured data can provide more balanced information for fair recommendations by capturing multidimensional user–item interactions. However, graph-based fair recommendation still faces some challenges: Traditional graphs rely on static edge-connected topology, struggling to dynamically update many-to-many relationships, which impairs the long-term fairness modeling; Most existing graph mining algorithms overlook individual differences arising from filtered sensitive information, thereby exacerbating the fairness-accuracy tradeoff; Hypergraph neural networks’ propagation relies on structural density, while sparse connections reduce it, leading to inaccurate representations in sparse regions and uneven diffusion. To address these issues, we propose a structure-aware fair recommendation approach based on counterfactual dynamic hypergraphs (FairCH). First, we propose a multidimensional user fairness model that captures many-to-many higher-order user–item relationships and their preference-fairness co-evolution via dynamic hypergraphs. Second, sensitive information is filtered through adversarial learning, and counterfactual hyperedges is reconstructed by counterfactual reasoning, compensating for information loss. Finally, a cross-hierarchy structure-aware model is proposed, which extracts counterfactual fairness layers, global preference layers, and shared evolution layers from hypergraphs and integrates them via an inter-layer interactive attention mechanism to enhance information propagation and mitigate structural biases. Experimental results demonstrate that FairCH exhibits superior recommendation performance to the baselines.

Sequential recommendation systems have become essential for personalized services in e-commerce and content platforms. While recent research has extended these systems with multi-modal features, existing approaches face three major challenges. First, they inadequately model fine-grained temporal interval distributions, failing to discriminate between high-frequency short intervals and low-frequency long intervals. Second, uniform fusion in the time domain leads to semantic misalignment across modalities because it ignores their inherent differences in the frequency domain. Third, rigid fusion strategies without self-supervised constraints lead to limited representation quality and semantic drift from pretrained embeddings. To address these issues, we propose Adaptive Temporal Expert Routing with Hierarchical Wavelet Enhancement (ATHWE) framework. ATHWE employs exponential saturation time mapping to generate temporally adaptive embeddings. These embeddings guide a sparse mixture of experts to model multi-scale user behavior dynamics. A hierarchical wavelet decomposition with band-specific gating selectively fuses complementary frequency components across modalities. Furthermore, contrastive learning and cluster-preserving objectives preserve semantic information during multi-modal fusion. Extensive experiments on multiple datasets validate the effectiveness of our framework. Our code is available at https://github.com/lulusiyuyu/ATHWE .

Feature selection, an essential technique in data mining, is often confined to batch learning or online idealization of data scenarios despite its significance. Existing online feature selection methods have specific assumptions regarding the data stream, such as requiring a fixed feature space with an explicit pattern and complete labeling of samples. Unfortunately, data streams generated in many real scenarios commonly exhibit arbitrarily incomplete feature spaces and scarcity labels, making existing approaches unsuitable for real applications. To fill these gaps, this study proposes a new problem called Online Feature Selection with Varying Features Spaces (OFSVF). OFSVF has a three-fold main idea: 1) it leverages Gaussian Copula to model the incomplete feature correlation in a complete latent space, encoded by continuous variables, 2) it employs a novel tree-ensemble-based approach to select the most informative features on-the-fly, and 3) it develops the underlying geometric structure of instances to establish the relationship between unlabeled and labels. Experimental results are documented to demonstrate the feasibility and effectiveness of our proposed method.

Recent advances in pre-trained language models (PLMs), such as GPT, BERT, and their larger variants, have significantly enhanced the ability to generate synthetic datasets for zero-shot learning, where models can generate data without the need for human annotations. However, existing methods often focus on generating domain-specific datasets, which restricts their generalization capabilities when applied to new, unseen domains. Furthermore, many of these methods rely on a single, static dataset that is either too broad or too narrowly tailored to a specific domain, making them less adaptable to diverse downstream domains. In this paper, we propose a dual prompt data generation framework, called Adaptive Domain-Specific Classification via Zero-shot Dataset Generation (AdaGen), which decouples universal representation learning from domain-specific adaptation. The framework consists of two components: domainagnostic training set generation via a universal prompt, followed by domain-specific validation set generation for weight learning. By combining these two components, our method achieves an optimal trade-off between cross-domain robustness and domainspecific accuracy, allowing a small task-aware model to effectively adapt across diverse domains while preserving generalization. This novel method overcomes the limitations of traditional methods by introducing dynamic feedback, enabling efficient zeroshot learning with improved adaptability. Extensive experiments in seven real-world datasets demonstrate the effectiveness of AdaGen, which outperforms the compared baseline models.

In recent years, Large Language Models (LLMs) have gained increasing attention in personalized recommendation systems due to their strong capabilities to understand contextual and semantic information. However, most existing LLM-based recommendation methods primarily rely on textual features, often neglecting the unique structural information embedded in user and item IDs, which are critical for capturing personalized preferences. To bridge this gap, we propose a novel method called CL 3M - Rec, which enhances the semantic representation of user and item IDs through Contrastive Learning (CL) and optimizes LLM recommendation performance via prompt construction. By effectively integrating structured ID information with textual semantics, our approach enables the LLMs to better model individual user interests and improve personalized prediction. Experiments on four public datasets demonstrate that CL3M-Rec consistently outperforms existing LLM-based and CL-based recommendation approaches, achieving superior predictive accuracy and robust generalization across diverse recommendation scenarios. The code of our method is available at https://github.com/Lewis44zhou/CL-3M-Rec.

Graph Neural Networks (GNNs) have demonstrated remarkable proficiency in handling a range of graph analytical tasks across various domains, such as e-commerce and social networks. Despite their versatility, GNNs face significant challenges in transferability, limiting their utility in real-world applications. Existing research in GNN transfer learning overlooks discrepancies in distribution among various graph datasets, facing challenges when transferring across different distributions. How to effectively adopt a well-trained GNN to new graphs with varying feature and structural distributions remains an under-explored problem. Taking inspiration from the success of Low-Rank Adaptation (LoRA) in adapting large language models to various domains, we propose GraphLoRA, an effective and parameter-efficient method for transferring well-trained GNNs to diverse graph domains. Specifically, we first propose a Structure-aware Maximum Mean Discrepancy (SMMD) to align divergent node feature distributions across source and target graphs. Moreover, we introduce low-rank adaptation by injecting a small trainable GNN alongside the pre-trained one, effectively bridging structural distribution gaps while mitigating the catastrophic forgetting. Additionally, a structure-aware regularization objective is proposed to enhance the adaptability of the pre-trained GNN to target graph with scarce supervision labels. Extensive experiments on eight real-world datasets demonstrate the effectiveness of GraphLoRA against fourteen baselines by tuning only 20% of parameters, even across disparate graph domains. The code is available at https://github.com/AllminerLab/GraphLoRA.

Abstract Recommendation system technologies predominantly focus on user-item interaction data, which are mapped into shared vector spaces for digital representation. These representations are then analyzed to uncover the relationships between users and items. As recommendation technologies have seen widespread adoption, a novel challenge has emerged in supply-demand matching contexts: the dual-triangular recommendation problem, involving four key entities, i.e., users with their demands, and suppliers with their offered items, forming a heterogeneous information network. In this work, we introduce the concept of dual-triangular recommendation and formally define this scientific problem. We propose a dual-triangular recommendation algorithm, enhanced by large language models, which utilizes knowledge graph encoder and LLM-augmented encoder to generate embedding representations for the four entities. A multi-task framework is employed to enable the sharing of underlying parameters across multiple recommendation tasks within the dual-triangular context. Through extensive experiments conducted on a real-world technology commercialization platform dataset, patent transfer dataset, and talent recruitment dataset, we demonstrate the effectiveness of our approach, offering a feasible and scalable solution to the dual-triangular recommendation problem.

The anchor-based multi-view subspace clustering (AMSC) has turned into a favorable tool for large-scale multi-view clustering. However, there still exist some limitations to the current AMSC approaches. First, they typically recover anchor graph structure in the original linear space, restricting their feasibility for nonlinear scenarios. Second, they usually overlook the potential benefits of jointly capturing the inter-view and intra-view information for enhancing the anchor representation learning. Third, these approaches mostly perform anchor-based subspace learning by a specific matrix norm, neglecting the latent high-order correlation across different views. To overcome these limitations, this article presents an efficient and effective approach termed Large-Scale Tensorized Multi-View Kernel Subspace Clustering (LTKMSC). Different from the existing AMSC approaches, our LTKMSC approach exploits both inter-view and intra-view awareness for anchor-based representation building. Concretely, the low-rank tensor learning is leveraged to capture the high-order correlation (i.e., the inter-view complementary information) among distinct views, upon which the \(l_{1,2}\) norm is imposed to explore the intra-view anchor graph structure in each view. Moreover, the kernel learning technique is leveraged to explore the nonlinear anchor–sample relationships embedded in multiple views. With the unified objective function formulated, an efficient optimization algorithm that enjoys low computational complexity is further designed. Extensive experiments on a variety of multi-view datasets have confirmed the efficiency and effectiveness of our approach when compared with the other competitive approaches.

Online learning, characterized by its feature space’s adaptability over time, has emerged as a flexible learning paradigm that has attracted widespread attention. However, existing online learning methods often overlook the distributional differences between instances and the presence of label noise in streaming data, thus significantly hindering the effectiveness and robustness of these algorithms. To overcome these challenges, we propose an online confidence learning algorithm for noisy labeled features, which aims to achieve robustness against arbitrary data streams and noisy labels. It employs two new strategies: online confidence inference, which applies the principle of empirical risk minimization to identify inconsistencies in spatial distributions, and geometric structure learning, which utilizes dynamic instance confidence to compute disparities between instances and their labels. Empirical findings demonstrate that our label correction mechanism enhances classification accuracy more effectively across various types of noisy labels (i.e., symmetric, asymmetric, and flipped). Additionally, a case study on image datasets was conducted to illustrate in detail the effectiveness of our OLNLS algorithm. Code is released in https://github.com/Zhuosd/OLNLS .

Effectively and efficiently mining valuable clustering patterns is a challenging problem when handling large-scale data from diverse sources. Existing approaches adopt anchor graph learning or binary representation embedding to reduce computational complexity. Normally, anchor graph learning can not directly obtain the clustering assignment except adopt the post-processing stage, such as graph cut or k-means clustering. The binary representation embedding neglects the structure information in Hamming space. In order to overcome these limitations, this paper proposes a novel, effective, and efficient angular reconstructive discrete embedding method with fusion similarity for a multi-view clustering (AFMC) that can jointly learn the global and local structure preserving binary representation and clustering assignment. Specifically, we propose to use angular reconstructive error minimization to maintain the global similarity correlation of binary representations of heterogeneous features in a common Hamming space. Moreover, we design a multi-view discrete ridge regression with fusion similarity term to handle the out-of-sample problem and preserve the local manifold structure. In addition, we propose an efficient optimization algorithm with linear computational complexity to solve the non-convex and non-smooth objective function. The experimental results demonstrate that AFMC outperforms several state-of-the-art large-scale multi-view clustering methods.

Patent data generally includes information from different perspectives or different types, and its heterogeneous attributes can be greatly beneficial to data clustering analysis. However, the existing patent analysis method always focus on the patent text cues, and such a strategy merely depends on the feature information to capture the data characteristics, failing to multi-type informative patent representation. Therefore, in this paper, to model the underlying structure/relationships of patent data, we employ the knowledge graph to depict the heterogeneous attributes of patent, and propose a novel Knowledge Graph-based Patent Clustering (KGPC) method, where the relationship reconstruction in knowledge graph as well as clustering-oriented representation refinement for patent clustering are jointly considered. With this model, there are three components, i.e., entity representation refinement, relationship reconstruction and self-supervised entity clustering. Given a patent knowledge graph as input, the entity representation refinement can be mutually boosted by the relationship reconstruction and self-supervised clustering objective, thereby leading to a balanced clustering-oriented output. Extensive experiments on several real-world patent knowledge graph datasets validate the effectiveness of KGPC while compared with the state-of-the-art.

Due to the emergency of multi-view graph data, considerable attention is focused on the multi-view graph clustering. Although great efforts have been made in developing the multi-view graph clustering methods, most of them implicitly follow the homophily assumption, where the connected nodes with edges tend to be in the same category. As a matter of fact, such an ideal assumption is hard to be satisfied in the real-world graph data, and there are some heterogeneous edges connecting dissimilar nodes in graph. How to well consider the homophily and refine the noisy/heterogeneous edges in multi-view graph clustering still remains an under-explored challenge. Therefore, in this paper, we propose a Homophily-enhanced Adaptive Graph Refinement for Multi-view Graph Clustering (HomoMGC) method, where an adaptive graph refinement strategy is seamlessly designed. Specifically, a feature-oriented graph is constructed based on the shared feature, and an integrated graph is computed by averagely fusing all the input adjacent graphs. Then, the feature-oriented graph and integrated graph are stacked into a graph tensor with a low-rank tensor constraint, where a refined affinity probability matrix can be adaptively recovered from the integrated graph by considering multiple graph information as well as the semantics features. Extensive experiments on several benchmark datasets demonstrate the superiority of HomoMGC compared with the state-of-the-art graph clustering methods. For the code reproducibility, the source code of HomoMGC is public available at https://github.com/ManshengChen/Code-for-HomoMGc-master.

In the evolving landscape of sequential recommendation systems, the application of Large Language Models (LLMs) is increasingly prominent. However, current attempts typically utilize general-purpose LLMs, which present a mismatch in capability and a large semantic gap relative to the specialized needs of recommendation tasks. To tackle these issues, we introduce RecCoder, an innovative model that reformulates sequential recommendation as a code completion task. This approach leverages the superior reasoning capability of code LLMs as a backbone, aligning well with the requirements of recommendation systems. To bridge the semantic gap, RecCoder creates extra tokens for each item and employs item content to initialize token embeddings. Furthermore, we have developed a suite of Semantic Adaptation Fine-tuning tasks, tailored to enhance the model's acquisition of both content and collaborative semantic information, thus aligning the model's intrinsic capabilities with the unique demands of recommendation tasks. Through extensive testing on three public datasets, RecCoder has shown remarkable improvements over existing models in terms of recommendation accuracy and efficiency. This success highlights the substantial yet previously underexplored potential of code LLMs in improving recommendation accuracy and efficiency, suggesting a promising new direction for future research in this area. The implementation code is accessible at https://github.com/AllminerLab/Code-for-RecCoder-master.

In next basket recommendation, baskets are usually formed through a large number of user interactions with items in the early stage. In general, the existing methods for next basket recommendation primarily focus on historical purchase behavior of users, assuming that user purchase interests are static, and overlook the dynamic and diverse changes in user purchase interests. In order to fully capture dynamic user interests and provide users with more diverse recommendations, we propose our method, Dynamic Heterogeneous Graph Prompt (DHGP), for next basket recommendation. By constructing a dynamic heterogeneous graph, we can adequately consider the influence of various interactive behaviors on the user's baskets at different times. Furthermore, we introduce a periodic dynamic heterogeneous prompt strategy to capture the interest directions between baskets from different users and provide users with more diverse interest directions. Extensive experimental validation on six real world datasets demonstrates that our method shows strong applicability across datasets under various conditions and outperforms several state-of-the-art recommendation methods. To the best of our knowledge, DHGP is the first next basket recommendation method that effectively combines dynamic and heterogeneous information. The implementation code is accessible at https://github.com/AllminerLab.

With the explosive growth of information, recommendation systems have emerged to alleviate the problem of information overload. In order to improve the performance of recommendation systems, many existing methods introduce Large Language Models to extract textual information from description text. However, Large Language Models are trained on large-scale generic textual data and may face a semantic gap for downstream recommendation tasks. To address the above issues, we propose Contrastive Learning for Adapting Language Model to Sequential Recommendation (CLA-Rec). In CLA-Rec, we first extract text embeddings from description text using Large Language Models and align the text embeddings learned by Large Language Models with the collaborative information through contrastive learning to obtain high-quality item representations. Through semantic alignment, we bridge the semantic gap between Large Language Models and the recommendation task. To map textual information and collaborative information into user representations, we utilize a Transformer model to learn user representations and capture user preferences by combining the semantically aligned item representations. Extensive experiments on three public datasets demonstrate that our method outperforms state-of-the-art approaches on multiple evaluation metrics, illustrating the effectiveness of the CLA-Rec model in adapting Large Language Models to recommendation tasks.

Next-basket recommendation (NBR) infers a set of items that a user will interact with in the next basket. Existing methods often struggle with the data sparsity problem, particularly when the number of baskets is significantly large due to diverse user behaviors. Cross-domain recommendation (CDR) can effectively alleviate this problem in NBR by transferring knowledge across different domains. Nevertheless, these methods often rely on the similarities of overlapping users, which leads to the negative transfer problem and ignores the overlapping items that are general in real-world scenarios like chain stores. In this paper, we provide a clear symbolic definition of cross-store recommendation (CSR) and distinguish it from CDR. We also propose a novel CSNBR model for cross-store next-basket recommendation task. To fully model the transferable collaborative information between two stores, we learn the embeddings of users, baskets, and items by two intra-store bipartite graphs, and use an inter-store unified bipartite graph to transfer the previously learned knowledge. Furthermore, to alleviate the negative transfer problem, we propose to reconstruct the inter-store unified bipartite graph by utilizing user embeddings obtained from the transfer layer and the disentanglement layer. We also employ two sequence encoders to model the historical sequential information at basket-level and item-level. Extensive experiments conducted on real-world datasets demonstrate the effectiveness of the CSNBR model.

Abstract Recently, with graph neural networks (GNNs) becoming a powerful technique for graph representation, many excellent GNN-based models have been proposed for processing heterogeneous graphs, which are termed Heterogeneous graph neural networks (HGNNs). However, existing HGNNs tend to aggregate information from either direct neighbors or those connected by short metapaths, thereby neglecting the higher-order information and global feature similarity information in heterogeneous graphs. In this paper, we propose a Multi-View Heterogeneous graph neural network (MV-HGNN) to aggregate these information. Firstly, two auxiliary views, specifically a global feature similarity view and a graph diffusion view, are generated from the original heterogeneous graph. Secondly, MV-HGNN performs two message-passing strategies to get the representation of different views. Subsequently, a transformer-based aggregator is used to get the semantic information. Subsequently, the representations of the three views are fused into a final composite representation. We evaluate our method on the node classification task over three commonly used heterogeneous graph datasets, and the results demonstrate that our proposed MV-HGNN significantly outperforms state-of-the-art baselines.

Most existing large-scale multiview clustering algorithms attempt to capture data distribution in multiple views by selecting view-wise anchor representations beforehand with$k$-means, or by direct matrix factorization on the original observations. Despite impressive performance, few of them have paid attention to the semantic correlations between anchor bases and cluster centroids, or even the underlying relations between clusters and data samples. In view of this, we propose aConceptFactorization basedMultiviewClustering for Large-scale Data (CFMC) method with nearly linear complexity. The anchor bases learning, coefficient expression with clear semantic cues and partitioning are integrated together in this unified model. Meanwhile, explicit connections among multiview data, anchor bases and clusters are modeled via coefficient representations with semantic meanings. A four-step alternate minimizing algorithm is designed to handle the optimization problem, which is proved to have linear time complexityw.r.t.the sample size. Extensive experiments conducted on several challenging large-scale datasets confirm the superiority of the method compared with the state-of-the-art methods.

Graph contrastive learning (GCL) has become the de-facto approach to conducting self-supervised learning on graphs for its superior performance. However, non-semantic graph augmentation methods prevent it from achieving better performance, and it suffers from vulnerability to graph attacks. To deal with these problems, we propose AEGCL to leverage graph AutoEncoder in Graph Contrastive Learning which directly targets graph property reconstruction to boost GCL effectiveness and robustness. Specifically, AEGCL has two distinctive characteristics, (1) a novel adaptive augmentation strategy based onmotifcentrality is proposed, which leverages semantic significant higher-order graph property; (2) the original attributed graph is decoupled into feature graph and topology graph to extract their dedicated information, and a simpleAttnFuseis proposed to combine the two augmented graphs and the two decoupled graphs. Graph autoencoder can thus be applied to the topology domain and raw attribute domain. Empirically, extensive experiments on benchmark graph datasets show that AEGCL outperforms existing baseline methods in terms of classification accuracy and robustness.

Next-basket recommendation methods focus on the inference of the next basket by considering the corresponding basket sequence. Although many methods have been developed for the task, they usually suffer from data sparsity. The number of interactions between entities is relatively small compared to their huge bases, so it is crucial to mine as much hidden information as possible from the limited historical interactions for prediction. However, the existing methods mainly just treat the next-basket recommendation task as a single-view sequential prediction problem, which leads to the inadequate mining of the information hidden in multiple views, and the mining of other patterns in the historical interactions is neglected, thus making it difficult to learn high-quality representations and limiting the recommendation effect. To alleviate the above issues, we propose a novel method named HapCL for next-basket recommendation, which mines information from multiple views and patterns with the help of polar contrastive learning. A hierarchical module is designed to mine multiple patterns of historical interactions from different views at two levels. In order to mine self-supervised signals, we design a polar contrastive learning module with a novel graph-based augmentation approach. Experiments on three real-world datasets validate the effectiveness of HapCL.

With the rapid development of wireless communication and localization technologies, the easier collection of trajectory data can bring potential data-driven value. Recently, there has been an increasing interest in how to publish trajectory dataset without revealing personal information. However, since the large-scale and real-world sequential trajectory dataset presents a heterogeneous regional distribution, the existing study ignores the relationship between privacy budget allocation and spatial characteristics, resulting in unreasonable continuity and mapping distortion, and thus lowering the utility of the synthetic dataset. To address this problem, we propose a probability distribution model named Adaptive grid-based Weighted Differential Privacy (AWDP). First, trajectories are adaptively discretized into the multi-resolution grid structures to make trajectories more uniformly distributed and less disturbed by the noise. Second, we allocate different weighted budgets for different grids according to density-based regional characteristics. Third, a spatio-temporal continuity maintenance method is designed to solve unrealistic direction- and density-based continuity deviations of synthetic trajectories. An application system is developed for demonstration purposes which is available online athttp://qgailab.com/awdp/. The extensive experiments on three datasets demonstrate that AWDP performs significantly better than the state-of-the-art model in preserving the density distribution of the original trajectories with differential privacy guarantee and high utility.

Feature selection, an essential technique in data mining, is often confined to batch learning or online idealization of data scenarios despite its significance. Existing online feature selection methods have specific assumptions regarding the data stream, such as requiring a fixed feature space with an explicit pattern and complete labeling of samples. Unfortunately, data streams generated in many real scenarios commonly exhibit arbitrarily incomplete feature spaces and scarcity labels, making existing approaches unsuitable for real applications. To fill these gaps, this study proposes a new problem calledOnline Feature Selection with Varying Features Spaces(OFSVF). OFSVF has a three-fold main idea: 1) it leverages Gaussian Copula to model the incomplete feature correlation in a complete latent space, encoded by continuous variables, 2) it employs a novel tree-ensemble-based approach to select the most informative features on-the-fly, and 3) it develops the underlying geometric structure of instances to establish the relationship between unlabeled and labels. Experimental results are documented to demonstrate the feasibility and effectiveness of our proposed method.

In recent years, the gaming industry has flourished. Therefore, game manufacturers need to strive to improve the gaming experience of users in the game. Social recommendation tasks in game scenes have become increasingly important. In this work, we focus on community recommendation scenario. A distinctive feature of game community recommendation is that each member can only belong to one gang for a certain duration, which we refer to as uniqueness of communities. The problem caused by uniqueness is that for users to be recommended, there are no positive samples available for training. The challenge caused by uniqueness is that there are no positive samples available for training when users are recommended. Therefore, the collaborative filtering information between the user and the community is very sparse. Meanwhile, existing methods fail to fully model communities and users based on their features and profiles. To address these problems, we propose Hypergraph Attribute Attention Network (HATT) framework. In order to fully exploit user profiles and similarity between users, we discretize user features into entity nodes and model the heterogeneous relationships between users and communities by hyperedge. We propose a hypergraph attention-based message passing mechanism to capture the high-order relationships and obtain embedding with more semantics. At last, we design contrastive learning paradigms to enhance the model’s representation ability and apply a multi task training strategy to train the model. Extensive experiments on two real world game datasets are conducted and the results demonstrate the superiority of our method in community recommendation.

Bundle recommendation focuses on recommending users with associated item sets at once. Recently, some works utilize Graph Neural Network (GNN), which has a solid power for mining information behind the topological structure, to enhance bundle recommendation performance. The previous GNN-based methods focus on designing the bundle-item association mechanism and fusing the extra information from the item view into the final prediction. However, the crucial component in GNN, namely the neighborhood aggregation mechanism, is yet to be explored under the bundle recommendation scenario. In this work, we propose a bundle-specific neighborhood aggregation mechanism named Auto Graph Filtering (AGF). The AGF refines the neighborhood aggregation mechanism from two aspects. (1) AGF utilizes the 2-hop meta paths in the bundle recommendation scenario instead of user interactions directly, which alleviates the extreme sparsity in the user-bundle graph. (2) AGF automatically reweights all the meta paths during the training. With training procedure completed, AGF optimizes the user-bundle graph to meet the bundle recommendation requirement. The experimental results show that our simplest AGF version, AGFN, consistently outperforms all the baselines. Moreover, the user-bundle graph learned by AGFN could also boost the existing GNN-based methods to achieve a better performance.

With the rapid development of modern communication technology, telecom fraud has been increasing year by year. If fraudsters can be accurately identified before they carry out their scams, it can not only protect people from potential losses but also increase trust in telecom operators. Therefore, in recent years, telecom fraud detection has garnered widespread attention in both academia and industry. Although existing methods for telecom fraud detection have achieved good performance, there are still many unresolved issues for real-world telecom operators. First, existing methods only focus on a single telecom scenario, while real-world telecom scenarios are diverse. Utilizing the characteristics of these different telecom scenarios can improve the effectiveness of telecom fraud detection. Second, existing methods usually use Graph Neural Networks (GNNs) to aggregate neighbor information. However, real-world telecom operators can’t obtain information of users from other operators, resulting in the lacking destination node attributes, which degenerates the performance of GNNs. To address the above issues, in this paper, we propose a new model for Telecom Fraud Detection Based on Feature binning and Autoencoder (TFD-FA). In TFD-FA, a feature binning framework is designed to partition users into different telecom scenarios in order to reflect their unique characteristics. An autoencoder component is also designed to aggregate neighbor information. Furthermore, an imbalance classifier component is constructed to solve the problem of the significantly lower number of fraudsters compared to normal users. Extensive experiments in a real-world dataset demonstrate the effectiveness of TFD-FA, which outperforms the compared baseline models.

With the rapid growth of online platforms and the abundance of available information, personalized recommender systems have become essential for assisting users in discovering relevant and interesting content. Among the various methods, knowledge-aware recommendation model has achieved notable success by leveraging the rich semantic information encoded in knowledge graphs. However, it overlooks the fact that users’ historical click sequences can better reflect their preferences within a period of time, thus imposing certain limitations on the recommendation performance. On the other hand, the application of pre-trained language models in recommender systems has demonstrated increasingly significant potential, as they can capture sequential patterns and dependencies within users’ historical click sequences and effectively capture contextual information in user-item interactions. To this end, we propose a hybrid recommendation model that leverages Pre-training in the collaborative Knowledge graph Attention neTwork (PKAT), to extract both the high-order connectivity information in collaborative knowledge graphs and the contextual information in users’ historical click sequences captured by Bidirectional Encoder Representations from Transformers (BERT). The collaborative knowledge graph attention network enables the model to effectively capture the intricate relationships between users, items, and knowledge entities, thus enhancing the representation learning process. Furthermore, what sets PKAT apart from other state-of-the-art knowledge-aware recommendation methods is the incorporation of the BERT language model. This integration allows PKAT to capture the contextual sequence information of user behavior, enabling it to generate more accurate and personalized recommendations. Extensive experiments are conducted on multiple benchmark datasets. And the results demonstrate that our PKAT model outperforms several state-of-the-art baselines.

In recent years, various recommendation methods are proposed to capture user preferences more accurately, with the assumption that different types of records reflects the positive intention of users to buy items in different degree. However, the records of different channels may denote positive or negative impacts on users’ willingness to buy items in multi-channel scenario, which is a salient features of games. Making recommendation only with the records of buy channel makes it difficult to capture cross-channel impact of items. To solve the issue, this paper proposes a multi-channel gated gifts recommendation method, named MCRec, which is able to mine the impact of acquisition in different channels on buy channel from multi-channel records and generate personalized gifts for users. The MCRec method extracts channel-aware correlation of items from channel-aware item-item graphs. The contribution of different items in various sessions is distinguished with item gate, and the impact of the context of different channels on buy channel is measured with channel gate. Finally, personalized gifts will be generated hierarchically with different purposes. Extensive experiments are conducted on two datasets that are constructed with the data collected from two massively multiplayer online (MMO) games. The results demonstrate the superiority of our MCRec over state-of-the-art recommendation methods in gifts recommendation. Further ablation studies validate the effectiveness of the design of MCRec in modeling cross-channel impact of items.

Class imbalance is the phenomenon that some classes have much fewer instances than others, which is ubiquitous in real-world graph-structured scenarios. Recent studies find that off-the-shelf Graph Neural Networks (GNNs) would under-represent minor class samples. We investigate this phenomenon and discover that the subspaces of minor classes being squeezed by those of the major ones in the latent space is the main cause of this failure. We are naturally inspired to enlarge the decision boundaries of minor classes and propose a general framework GraphSHA by Synthesizing HArder minor samples. Furthermore, to avoid the enlarged minor boundary violating the subspaces of neighbor classes, we also propose a module called SemiMixup to transmit enlarged boundary information to the interior of the minor classes while blocking information propagation from minor classes to neighbor classes. Empirically, GraphSHA shows its effectiveness in enlarging the decision boundaries of minor classes, as it outperforms various baseline methods in class-imbalanced node classification with different GNN backbone encoders over seven public benchmark datasets. Code is avilable at https://github.com/wenzhilics/GraphSHA.

Contrastive learning (CL) has become the de-facto learning paradigm in self-supervised learning on graphs, which generally follows the "augmenting-contrasting'' learning scheme. However, we observe that unlike CL in computer vision domain, CL in graph domain performs decently even without augmentation. We conduct a systematic analysis of this phenomenon and argue that homophily, i.e., the principle that "like attracts like'', plays a key role in the success of graph CL. Inspired to leverage this property explicitly, we propose HomoGCL, a model-agnostic framework to expand the positive set using neighbor nodes with neighbor-specific significances. Theoretically, HomoGCL introduces a stricter lower bound of the mutual information between raw node features and node embeddings in augmented views. Furthermore, HomoGCL can be combined with existing graph CL models in a plug-and-play way with light extra computational overhead. Extensive experiments demonstrate that HomoGCL yields multiple state-of-the-art results across six public datasets and consistently brings notable performance improvements when applied to various graph CL methods. Code is avilable at https://github.com/wenzhilics/HomoGCL.

Nowadays, it is more and more convenient for people to participate in group activities. Therefore, providing some recommendations to groups of individuals is indispensable. Group recommendation is the task of suggesting items or events for a group of users in social networks or online communities. In this work, we study group recommendation in a particular scenario, namely occasional group recommendation, which has few or no historical directly interacted items. Existing group recommendation methods mostly adopt attention-based preference aggregation strategies to capture group preferences. However, these models either ignore the complex high-order interactions between groups, users and items or greatly reduce the efficiency by introducing complex data structures. Moreover, occasional group recommendation suffers from the problem of data sparsity due to the lack of historical group-item interactions. In this work, we focus on addressing the aforementioned challenges and propose a novel group recommendation model called Self-Supervised Group Graph Collaborative Filtering (SGGCF). The goal of the model is capturing the high-order interactions between users, items and groups and alleviating the data sparsity issue in an efficient way. First, we explicitly model the complex relationships as a unified user-centered heterogeneous graph and devise a base group recommendation model. Second, we explore self-supervised learning on the graph with two kinds of contrastive learning module to capture the implicit relations between groups and items. At last, we treat the proposed contrastive learning loss as supplementary and apply a multi-task strategy to jointly train the BPR loss and the proposed contrastive learning loss. We conduct extensive experiments on three real-world datasets, and the experimental results demonstrate the superiority of our proposed model in comparison to the state-of-the-art baselines.

Abstract Graph collaborative filtering methods have shown great performance improvements compared with deep neural network-based models. However, these methods suffer from data sparsity and data noise problems. To address these issues, we propose a new contrastive learning-based graph collaborative filtering method to learn more robust representations. The proposed method is called signal contrastive enhanced graph collaborative filtering (SC-GCF), which conducts contrastive learning on graph signals. It has been proved that graph neural networks correspond to low-pass filters on the graph signals from the graph convolution perspective. Different from the previous contrastive learning-based methods, we first pay attention to the diversity of graph signals to directly optimize the informativeness of the graph signals. We introduce a hypergraph module to strengthen the representation learning ability of graph neural networks. The hypergraph learning module utilizes a learnable hypergraph structure to model the latent global dependency relations that graph neural networks cannot depict. Experiments are conducted on four public datasets, and the results show significant improvements compared with the state-of-the-art methods, which confirms the importance of considering signal-level contrastive learning and hypergraph learning.

Graph-oriented multi-view clustering methods have achieved impressive performances by employing relationships and complex structures hidden in multi-view data. However, most of them still suffer from the following two common problems. (1) They target at studying a common representation or pairwise correlations between views, neglecting the comprehensiveness and deeper higher-order correlations among multiple views. (2) The prior knowledge of view-specific representation can not be taken into account to obtain the consensus indicator graph in a unified graph construction and clustering framework. To deal with these problems, we propose a novel Low-rank Tensor Based Proximity Learning (LTBPL) approach for multi-view clustering, where multiple low-rank probability affinity matrices and consensus indicator graph reflecting the final performances are jointly studied in a unified framework. Specifically, multiple affinity representations are stacked in a low-rank constrained tensor to recover their comprehensiveness and higher-order correlations. Meanwhile, view-specific representation carrying different adaptive confidences is jointly linked with the consensus indicator graph. Extensive experiments on nine real-world datasets indicate the superiority of LTBPL compared with the state-of-the-art methods.

Despite significant progress, there remain three limitations to the previous multi-view clustering algorithms. First, they often suffer from high computational complexity, restricting their feasibility for large-scale datasets. Second, they typically fuse multi-view information via one-stage fusion, neglecting the possibilities in multi-stage fusions. Third, dataset-specific hyperparameter-tuning is frequently required, further undermining their practicability. In light of this, we propose afastmulti-viewclustering viaensembles (FastMICE) approach. Particularly, the concept of random view groups is presented to capture the versatile view-wise relationships, through which the hybrid early-late fusion strategy is designed to enable efficient multi-stage fusions. Withmultipleviews extended tomanyview groups, three levels of diversity (w.r.t. features, anchors, and neighbors, respectively) are jointly leveraged for constructing the view-sharing bipartite graphs in the early-stage fusion. Then, a set of diversified base clusterings for different view groups are obtained via fast graph partitioning, which are further formulated into a unified bipartite graph for final clustering in the late-stage fusion. Notably, FastMICE has almost linear time and space complexity, and is free of dataset-specific tuning. Experiments on 22 multi-view datasets demonstrate its advantages in scalability (for extremely large datasets), superiority (in clustering performance), and simplicity (to be applied) over the state-of-the-art. Code available:https://github.com/huangdonghere/FastMICE.

Anomaly detection on attributed networks has received an increasing amount of attention in recent years. Despite the success, most of the existing methods only focus on detecting the abnormal nodes while fail to detect the abnormal subgraphs. In this paper, we define a new problem of hybrid-order anomaly detection on attributed networks, which aims to detect both of the abnormal nodes and subgraphs. To this end, a new deep learning model called Hybrid-Order Graph Attention Network (HO-GAT) is developed, which is able to simultaneously detect the abnormal nodes and motif instances in an attributed network. In order to model the mutual influence between nodes and motif instances, the learning procedures of the node representation and the motif instance representation are integrated into a unified graph attention network with a novel hybrid-order self-attention mechanism. After learning the node representation and the motif instance representation, two decoders are respectively designed to reconstruct the attribute information of the nodes and motif instances, and the hybrid-order topological structure among nodes and motif instances. And finally, the reconstruction errors are utilized as the abnormal score of nodes and motif instances respectively. Extensive experiments conducted on real-world datasets have confirmed the effectiveness of the HO-GAT method.

Graph neural network (GNN) has shown its prominent performance in representation learning of graphs but it has not been fully considered for heterogeneous graphs which contain more complex structures and rich semantics. The rich semantic information of heterogeneous graph can be usually revealed by meta-paths. Therefore, most of the existing GNN models designed for heterogeneous graphs utilize the meta-path based neighborhood sampler to divide a heterogeneous graph into multiple homogeneous subgraphs according to various meta-paths so that the homogeneous GNN can be applied to investigate heterogeneous graphs. Nevertheless, the way of embedding semantic information of meta-paths into multiple homogeneous graphs isimplicitand ineffective, which cannot accurately capture the semantics of heterogeneous graphs. In this paper, we propose a novel semi-supervised GNN model namedExplicitMessage-Passing Heterogeneous Graph Neural Network (EMP), which executes the process ofexplicitmessage-passing along the meta-paths. Besides, we also propose a split method for meta-paths and consider mutual effect between various meta-paths in advance in the proposed model, so that the semantic information of the whole set of meta-paths can be captured accurately. Extensive experiments conducted on three real-world datasets demonstrate the superiority of the proposed model.

Most existing recommendation research has been concentrated on unidirectional recommendation, i.e. only recommending items to users. However, in many real-world scenarios, the platform needs to achieve bi-directional recommendation. For example, in an online recruitment scenario, the recommender system not only needs to recommend positions to candidates, but also recommend candidates to enterprises. In this paper, we first formalize a new recommendation problem called bi-directional recommendation and contribute a new bidirectional recommendation model named BiROR (Bi-directional Recommendation for Online Recruitment). In BiROR, an encoder component is utilized to learn the text embeddings, and a graph learning component is designed to learn the graph embeddings. In addition, a multi-task learning framework is designed to achieve bi-directional recommendation. In the multi-task learning framework, we share the text embeddings and graph embeddings to alleviate the problems of data sparsity and data asymmetry in online recruitment. Extensive experiments in a real-world task show that BiROR outperforms the state-of-the-art methods, verifying the effectiveness of the designs of our model.

Multi-view subspace clustering targets at clustering data lying in a union of low-dimensional subspaces. Generally, an n X n affinity graph is constructed, on which spectral clustering is then performed to achieve the final clustering. Both graph construction and graph partitioning of spectral clustering suffer from quadratic or even cubic time and space complexity, leading to difficulty in clustering large-scale datasets. Some efforts have recently been made to capture data distribution in multiple views by selecting key anchor bases beforehand with k-means or uniform sampling strategy. Nevertheless, few of them pay attention to the algebraic property of the anchors. How to learn a set of high-quality orthogonal bases in a unified framework, while maintaining its scalability for very large datasets, remains a big challenge. In view of this, we propose an Efficient Orthogonal Multi-view Subspace Clustering (OMSC) model with almost linear complexity. Specifically, the anchor learning, graph construction and partition are jointly modeled in a unified framework. With the mutual enhancement of each other, a more discriminative and flexible anchor representation and cluster indicator can be jointly obtained. An alternate minimizing strategy is developed to deal with the optimization problem, which is proved to have linear time complexity w.r.t. the sample number. Extensive experiments have been conducted to confirm the superiority of the proposed OMSC method. The source codes and data are available at https://github.com/ManshengChen/Code-for-OMSC-master.

Abstract Multi-view clustering (MVC) has attracted more and more attention in the recent few years by making full use of complementary and consensus information between multiple views to cluster objects into different partitions. Although there have been two existing works for MVC survey, neither of them jointly takes the recent popular deep learning-based methods into consideration. Therefore, in this paper, we conduct a comprehensive survey of MVC from the perspective of representation learning. It covers a quantity of multi-view clustering methods including the deep learning-based models, providing a novel taxonomy of the MVC algorithms. Furthermore, the representation learning-based MVC methods can be mainly divided into two categories, i.e., shallow representation learning-based MVC and deep representation learning-based MVC, where the deep learning-based models are capable of handling more complex data structure as well as showing better expression. In the shallow category, according to the means of representation learning, we further split it into two groups, i.e., multi-view graph clustering and multi-view subspace clustering. To be more comprehensive, basic research materials of MVC are provided for readers, containing introductions of the commonly used multi-view datasets with the download link and the open source code library. In the end, some open problems are pointed out for further investigation and development.

Recently, hybrid-order community detection has been proposed for addressing the hypergraph fragmentation issue suffered by the motif-based higher-order community detection. However, the existing attempts of hybrid-order community detection inadvertently damage the lower-order connectivity pattern and the higher-order connectivity pattern when constructing the fusion model. Additionally, like the higher-order community detection approaches, they also adopt a two-phase strategy that separately applies the existing graph node clustering methods to the proximity matrix derived from the lower-order connectivity pattern and the higher-order connectivity pattern. Therefore, the higher-order connectivity pattern is only utilized for constructing proximity matrix and hence has no direct effect on the final community results. In this paper, to address the above issues, we propose a Generative model for Hybrid-Order Community detection (GHOC). The main idea lies in defining a likelihood function of a generative model that finds the optimal community membership strength vectors of nodes, based on which the original lower-order connectivity pattern and the higher-order connectivity pattern can be directly reconstructed simultaneously. From the community membership strength vectors, the final community structure can be derived. Extensive experiments have been conducted on several data sets, and the results have confirmed the superiority of the proposed GHOC method.

Most proximity-based multi-view clustering methods are sensitive to the initial proximity matrix, where the clustering performance is quite unstable when using different initial proximity matrixes. This problem is defined as the initial value sensitivity problem. Since clustering is an unsupervised learning task, it is unrealistic to tune the initial proximity matrix. Thus, how to overcome the initial value sensitivity problem is a significant but unsolved issue in the proximity-based multi-view clustering. To this end, this paper proposes a novel multi-view proximity learning method, named multi-view consensus proximity learning (MCPL). On the one hand, by integrating the information of all views in a self-weighted manner and giving a rank constraint on the Laplacian matrix, the MCPL method learns the consensus proximity matrix to directly reflect the clustering result. On the other hand, different from most multi-view proximity learning methods, in the proposed MCPL method, the data representatives rather than the original data objects are adopted to learn the consensus proximity matrix. The data representatives will be updated in the process of the proximity learning so as to weaken the impact of the initial value on the clustering performance. Extensive experiments are conducted to demonstrate the effectiveness of the proposed method.

Recently, some attempts have been made in introducing deep neural networks (DNNs) to recommender systems for generating more accurate prediction due to the nonlinear representation learning capability of DNNs. However, they inevitably result in high computational and storage costs. Worse still, due to the relatively small number of ratings that can be fed into DNNs, they may easily suffer from the overfitting issue. To tackle these issues, we propose a novel recommendation framework based on Back Propagation (BP) neural network with attention mechanism, namely BPAM++. In particular, the BP neural network is utilized to learn the complex relationship between the target user and his/her neighbors and the complex relationship between the target item and its neighbors. Compared with DNNs, the shallow neural network, i.e., BP neural network, can not only reduce the computational and storage costs, but also alleviate the overfitting issues in DNNs caused by a relatively small number of ratings. In addition, an attention mechanism is designed to capture the global impact of the nearest users of the target user on their nearest target user sets. Extensive experiments conducted on eight benchmark datasets confirm the effectiveness of the proposed model.

Many real-world graphs can be viewed as Heterogeneous Networks or Heterogeneous Information Networks (HINs) for that they comprise a diversity of node types and relation types. Due to the efficient representation ability of Graph Neural Network and the idea of random walk, many recent studies apply graph representation learning to HINs and achieve satisfactory results. However, these works either treat different node types in a metapath equally, which is inconsistent with the original graph semantic information for that different node types should have different statuses, or only consider the first-order (node-level) and second-order (metapath-level, a.k.a. link-level) information aggregation while ignoring the higher-order relations. To tackle these two problems, we propose a novel Star-Shaped Hierarchical Graph Attentional Network (StarGAT) model to boost representation learning in HINs. Specifically, we assume nodes in HINs can be categorized into a star-shaped structure including one center node type and a bunch of auxiliary node types in a specific task; and we encode node-level, link-level and motif-level attentions in a hierarchical manner to capture richer semantic information. Extensive experiments on three datasets illustrate the model effectiveness.

Most of the existing methods define recommendation as regression or classification for user-item interactions and apply discriminative models. However, recommender systems suffer from interaction data sparsity and data noise problems in reality. Recent Generative Adversarial Network-based recommender systems have the potential to solve the aforementioned problems. The negative sampling methods use the generator to collect effective signals from a large amount of unlabeled data to alleviate the data sparsity problem, while they suffer from sparse rewards in the policy gradient training process. The vector reconstruction methods generate user-related vectors for data augmentation to enhance robustness, but they lead to redundant calculation and only take the user as a condition and ignore information conveyed by items. To alleviate the limitations of these methods, we propose a novel framework termed Personalized Recommendation with Conditional Generative Adversarial Networks (PRGAN) to consider both the user and the item subset as conditions and formulate conditional rating vector generation as a user-item matching problem. The sparsity of conditional rating vectors can be controlled in our method, which simplifies the discriminator’s learning task. Experiments are conducted on four datasets to evaluate the effectiveness of the proposed framework.

As a critical task in intelligent traffic systems, traffic prediction has received a large amount of attention in the past few decades. The early efforts mainly model traffic prediction as the time-series mining problem, in which the spatial dependence has been largely ignored. As the rapid development of deep learning, some attempts have been made in modeling traffic prediction as the spatio-temporal data mining problem in a road network, in which deep learning techniques can be adopted for modeling the spatial and temporal dependencies simultaneously. Despite the success, the spatial and temporal dependencies are only modeled in a regionless network without considering the underlying hierarchical regional structure of the spatial nodes, which is an important structure naturally existing in the real-world road network. Apart from the challenge of modeling the spatial and temporal dependencies like the existing studies, the extra challenge caused by considering the hierarchical regional structure of the road network lies in simultaneously modeling the spatial and temporal dependencies between nodes and regions and the spatial and temporal dependencies between regions. To this end, this article proposes a new Temporal Hierarchical Graph Attention Network (TH-GAT). The main idea lies in augmenting the original road network into a region-augmented network, in which the hierarchical regional structure can be modeled. Based on the region-augmented network, the region-aware spatial dependence model and the region-aware temporal dependence model can be constructed, which are two main components of the proposed TH-GAT model. In addition, in the region-aware spatial dependence model, the graph attention network is adopted, in which the importance of a node to another node, of a node to a region, of a region to a node, and of a region to another region, can be captured automatically by means of the attention coefficients. Extensive experiments are conducted on two real-world traffic datasets, and the results have confirmed the superiority of the proposed TH-GAT model.

Recently, network embedding has received a large amount of attention in network analysis. Although some network embedding methods have been developed from different perspectives, on one hand, most of the existing methods only focus on leveraging the plain network structure, ignoring the abundant attribute information of nodes. On the other hand, for some methods integrating the attribute information, only the lower-order proximities (e.g., microscopic proximity structure) are taken into account, which may suffer if there exists the sparsity issue and the attribute information is noisy. To overcome this problem, the attribute information and mesoscopic community structure are utilized. In this article, we propose a novel network embedding method termed Attributed Network Embedding with Micro-Meso structure, which is capable of preserving both the attribute information and the structural information including the microscopic proximity structure and mesoscopic community structure. In particular, both the microscopic proximity structure and node attributes are factorized by Nonnegative Matrix Factorization (NMF), from which the low-dimensional node representations can be obtained. For the mesoscopic community structure, a community membership strength matrix is inferred by a generative model (i.e., BigCLAM) or modularity from the linkage structure, which is then factorized by NMF to obtain the low-dimensional node representations. The three components are jointly correlated by the low-dimensional node representations, from which two objective functions (i.e., ANEM_B and ANEM_M) can be defined. Two efficient alternating optimization schemes are proposed to solve the optimization problems. Extensive experiments have been conducted to confirm the superior performance of the proposed models over the state-of-the-art network embedding methods.

Multi-layer network community detection has drawn an increasing amount of attention recently. Despite success, the existing methods mainly focus on the lower-order connectivity structure at the level of individual nodes and edges. And the higher-order connectivity structure has been largely ignored, which contains better signature of community compared with edges. The main challenges in utilizing higher-order structure for multi-layer network community detection are that the most representative higher-order structure may vary from one layer to another and the connectivity structure formed by the same node subset may exhibit different higher-order connectivity patterns in different layers. To this end, this paper proposes a novel higher-order structure, termed harmonic motif, which is a dense subgraph having on average the largest statistical significance in each layer. Based on the harmonic motif, a primary layer is constructed by integrating higher-order structural information from all layers. Additionally, the higher-order structural information of each individual layer is taken as the auxiliary information. A coupling is established between the primary layer and each auxiliary layer. Accordingly, a harmonic motif modularity is designed to generate the community structure. Extensive experiments on eleven real-world multi-layer network datasets have been conducted to confirm the effectiveness of the proposed method.

Graph Neural Networks (GNNs), which extend deep neural networks to graph-structured data, have attracted increasing attention. They have been proven to be powerful for numerous graph related tasks such as graph classification, link prediction, and node classification. To adapt GNNs to graph classification, recent works aim to learn graph-level representation through a hierarchical pooling procedure. One major direction is to select important nodes to hierarchically coarsen the input graph and gradually reduce the information into the graph representation. However, most of the existing methods only select important nodes, which can be redundant and cannot represent the original graph well. Meanwhile, the information of non-selected nodes is often overlooked when generating a new coarser graph, which may lead to the tremendous loss of important structural and node feature information. In this paper, we propose a novel pooling operator RepPool to learn hierarchical graph representations. Specifically, we introduce the concept of representativeness that is combined with the importance for node selection and we provide a learnable way to integrate non-selected nodes. By combining the RepPool operator with conventional GCN convolutional layers, a hierarchical graph classification architecture is developed. Extensive experiments on various public benchmarks have demonstrated the effectiveness of the proposed method. The implementation of the proposed framework is available.

Community detection (or graph clustering) is crucial for unraveling the structural properties of complex networks. As an important technique in community detection, label propagation has shown the advantage of finding a good community structure with nearly linear time complexity. However, despite the progress that has been made, there are still several important issues that have not been properly addressed. First, the label propagation typically proceeds over the lower order structure of the network and only the direct one-hop connections between nodes are taken into consideration. Unfortunately, the higher order structure that may encode design principle of the network and be crucial for community detection is neglected under this regime. Second, the stability of the identified community structure may also be seriously affected by the inherent randomness in the label propagation process. To tackle the above issues, this article proposes a Motif-Aware Weighted Label Propagation method for community detection. We focus on triangles within the network, but our technique extends to other kinds of motifs as well. Specifically, the motif-based higher order structure mining is conducted to capture structural characteristics of the network. First, the motif of interest (locally meaningful pattern) is identified, and then, the motif-based hypergraph can be constructed to encode the higher order connections. To further utilize the structural information of the network, a re-weighted network is designed, which unifies both the higher order structure and the original lower order structure. Accordingly, a novel voting strategy termed NaS (considering both Number and Strength of connections) is proposed to update node labels during the label propagation process. In this way, the random label selection can be effectively eliminated, yielding more stable community structures. Experimental results on multiple real-world datasets have shown the superiority of the proposed method.

This paper focuses on scalability and robustness of spectral clustering for extremely large-scale datasets with limited resources. Two novel algorithms are proposed, namely, ultra-scalable spectral clustering (U-SPEC) and ultra-scalable ensemble clustering (U-SENC). In U-SPEC, a hybrid representative selection strategy and a fast approximation method for K-nearest representatives are proposed for the construction of a sparse affinity sub-matrix. By interpreting the sparse sub-matrix as a bipartite graph, the transfer cut is then utilized to efficiently partition the graph and obtain the clustering result. In U-SENC, multiple U-SPEC clusterers are further integrated into an ensemble clustering framework to enhance the robustness of U-SPEC while maintaining high efficiency. Based on the ensemble generation via multiple U-SEPC's, a new bipartite graph is constructed between objects and base clusters and then efficiently partitioned to achieve the consensus clustering result. It is noteworthy that both U-SPEC and U-SENC have nearly linear time and space complexity, and are capable of robustly and efficiently partitioning 10-million-level nonlinearly-separable datasets on a PC with 64 GB memory. Experiments on various large-scale datasets have demonstrated the scalability and robustness of our algorithms. The MATLAB code and experimental data are available at https://www.researchgate.net/publication/330760669.

Recently, multi-view outlier detection has emerged as a challenging research topic in outlier detection because of complex distributions of data across different views. There are mainly three types of outliers, i.e., attribute outliers, class outliers and class-attribute outliers. Most existing multi-view outlier detection approaches only detect part of the three types of outliers in a pairwise manner across different views, which is not able to accomplish the task of multi-view outlier detection comprehensively and uniformly. Outlier detection in a pairwise manner across different views also leads to time-consuming computation. We propose a new algorithm termed Multi-view Outlier Detection in Deep Intact Space (MODDIS) to find all the three types of outliers simultaneously and avoid comparing different views in a pairwise manner. Rather than leveraging subspace clustering, the performance of which is seriously affected by the dependence of subspaces on most real datasets, neural networks are employed in MODDIS in that neural networks have a stronger representation learning ability. Meanwhile, based on the view insufficiency assumption, a multi-view intact outlierness space assumption is proposed. Based on this assumption, a multi-view latent intact space is constructed to encode outlierness information of all views, where outlierness in any view is a snapshot from some perspective. Finally, an outlier detection measurement is defined in the latent intact space. Experiments are conducted on several UCI datasets and the empirical results demonstrate the effectiveness of our proposed method.

Graph Learning has emerged as a promising technique for multi-view clustering, and has recently attracted lots of attention due to its capability of adaptively learning a unified and probably better graph from multiple views. However, the existing multi-view graph learning methods mostly focus on the multi-view consistency, but neglect the potential multi-view inconsistency (which may be incurred by noise, corruptions, or view-specific characteristics). To address this, this paper presents a new graph learning-based multi-view clustering approach, which for the first time, to our knowledge, simultaneously and explicitly formulates the multi-view consistency and the multi-view inconsistency in a unified optimization model. To solve this model, a new alternating optimization scheme is designed, where the consistent and inconsistent parts of each single-view graph as well as the unified graph that fuses the consistent parts of all views can be iteratively learned. It is noteworthy that our multi-view graph learning model is applicable to both similarity graphs and dissimilarity graphs, leading to two graph fusion-based variants, namely, distance (dissimilarity) graph fusion and similarity graph fusion. Experiments on various multi-view datasets demonstrate the superiority of our approach. The MATLAB source code is available at https://github.com/youweiliang/ConsistentGraphLearning.

Recommender system is widely used in e-commercial platforms to recommend users suitable items according to users's preferences. In recent years, an increasing amount of attention has been paid to the application of recommender system in education. There are many online learning systems that can recommend students suitable courses according to students' learning performances. However, there are few universities using recommender system to recommend students suitable elective courses. It is generally known that students in higher grade take the courses earlier than those in lower grade. Therefore, the elective course scores of sophomores can be predicted by using the course score information from students of higher grades. However, the unbalanced distribution of course-enrollment data makes it hard to predict the scores of the courses that are in a low selection rate. Therefore, we propose a Constrained Matrix Factorization (ConMF) algorithm to predict sophomores' elective course scores, which integrates the course average score into the objective function so as to make up the prediction deviation caused by the unbalanced course selection rate and make more accurate prediction than the traditional Matrix Factorization (MF) approach. The experimental results show that our proposed model outperforms the state-of-the-art methods in the task of university students' course score prediction.

Network community detection is a hot research topic in network analysis. Although many methods have been proposed for community detection, most of them only take into consideration the lower-order structure of the network at the level of individual nodes and edges. Thus, they fail to capture the higher-order characteristics at the level of small dense subgraph patterns, e.g., motifs. Recently, some higher-order methods have been developed but they typically focus on the motif-based hypergraph which is assumed to be a connected graph. However, such assumption cannot be ensured in some real-world networks. In particular, the hypergraph may become fragmented. That is, it may consist of a large number of connected components and isolated nodes, despite the fact that the original network is a connected graph. Therefore, the existing higher-order methods would suffer seriously from the above fragmentation issue, since in these approaches, nodes without connection in hypergraph can't be grouped together even if they belong to the same community. To address the above fragmentation issue, we propose an Edge enhancement approach for Motif-aware community detection (EdMot ). The main idea is as follows. Firstly, a motif-based hypergraph is constructed and the top K largest connected components in the hypergraph are partitioned into modules. Afterwards, the connectivity structure within each module is strengthened by constructing an edge set to derive a clique from each module. Based on the new edge set, the original connectivity structure of the input network is enhanced to generate a rewired network, whereby the motif-based higher-order structure is leveraged and the hypergraph fragmentation issue is well addressed. Finally, the rewired network is partitioned to obtain the higher-order community structure. Extensive experiments have been conducted on eight real-world datasets and the results show the effectiveness of the proposed method in improving the community detection performance of state-of-the-art methods.

Mood disorders are common and associated with significant morbidity and mortality. Early diagnosis has the potential to greatly alleviate the burden of mental illness and the ever increasing costs to families and society. Mobile devices provide us a promising opportunity to detect the users' mood in an unobtrusive manner. In this study, we use a custom keyboard which collects keystrokes' meta-data and accelerometer values. Based on the collected time series data in multiple modalities, we propose a deep personalized mood prediction approach, called dpMood, by integrating convolutional and recurrent deep architectures as well as exploring each individual's circadian rhythm. Experimental results not only demonstrate the feasibility and effectiveness of using smart-phone meta-data to predict the presence and severity of mood disturbances in bipolar subjects, but also show the potential of personalized medical treatment for mood disorders.

During the past several years, multi-layer network community detection has drawn an increasing amount of attention and many approaches have been developed from different perspectives. Despite the success, they mainly rely on the lower-order connectivity structure at the level of individual nodes and edges. However, the higher-order connectivity structure plays the essential role as the building block for multiplex networks, which may contain better signature of community than edge. The main challenge in utilizing higher-order structure for multi-layer network community detection is that the most representative higher-order structure may vary from one layer to another. In this paper, we propose a higher-order structural approach for multi-layer network community detection, termed harmonic motif modularity (HM-Modularity). The key idea is to design a novel higher-order structure, termed harmonic motif, which is able to integrate higher-order structural information from multiple layers to construct a primary layer. The higher-order structural information of each individual layer is also extracted, which is taken as the auxiliary information for discovering the multi-layer community structure. A coupling is established between the primary layer and each auxiliary layer. Finally, a harmonic motif modularity is designed to generate the community structure. By solving the optimization problem of the harmonic motif modularity, the community labels of the primary layer can be obtained to reveal the community structure of the original multi-layer network. Experiments have been conducted to show the effectiveness of the proposed method.

Missing value is common in many machine learning problems and much effort has been made to handle missing data to improve the performance of the learned model. Sometimes, our task is not to train a model using those unlabeled/labeled data with missing value but process examples according to the values of some specified features. So, there is an urgent need of developing a method to predict those missing values. In this paper, we focus on learning from the known values to learn missing value as close as possible to the true one. It's difficult for us to predict missing value because we do not know the structure of the data matrix and some missing values may relate to some other missing values. We solve the problem by recovering the complete data matrix under the three reasonable constraints: feature relationship, upper recovery error bound and class relationship. The proposed algorithm can deal with both unlabeled and labeled data and generative adversarial idea will be used in labeled data to transfer knowledge. Extensive experiments have been conducted to show the effectiveness of the proposed algorithms.

Expert finding is an important technique to obtain the user authority ranking in community question answering (CQA) websites. ZhihuRank is a topic-sensitive expert finding algorithm, which is based on both LDA and PageRank. Currently, with the amount of participants and documents increasing rapidly in CQA websites, how to parallel expert finding algorithms for big data analysis has received significant attention. In this paper, we find that the Spark framework is more suitable for paralleling expert finding algorithms than the MapReduce framework, which is a memory-based parallel computing model to support complicated iterative algorithms. As an example, we parallel ZhihuRank using MLlib's LDA and GraphX's PageRank in Spark. Experiments have been conducted on large-scale real data from Zhihu1(the most popular CQA website in China). And the experimental results confirmed the effectiveness and scalability of our proposed approach.

Point of interest (POI) recommendation, a service which can help people discover useful and interesting locations has emerged rapidly with the development of location-based social networks (LBSNs), like Foursquare, Gowalla and Wechat. The large number of check-in histories make it possible to mine the preference of each user and then to provide accurate personalized POI recommendation. In real-world applications, apart from check-in data, there are some other useful information available for making better POI recommendation, such as social relationship among users and geographical influence. In this paper, a new POI recommendation method called Social and Geographical Fusing Model (SGFM) is designed. The basic idea is summarized as follows. Firstly, the users' check-in records and social influence are integrated in a combinative model. Then the global user impact factors generated by the PageRank algorithm are used to improve the combinative model. Secondly, a geographical influence measurement is used to capture the users' physical check-in characters. Finally, the enhanced combinative model and geographical influence are combined together to form a new framework. Extensive experiments have been conducted on a famous dataset, namely Gowalla. The comparison results confirm that the proposed framework outperforms state-of-the-art POI recommendation methods significantly.

Although many successful ensemble clustering approaches have been developed in recent years, there are still two limitations to most of the existing approaches. First, they mostly overlook the issue of uncertain links, which may mislead the overall consensus process. Second, they generally lack the ability to incorporate global information to refine the local links. To address these two limitations, in this paper, we propose a novel ensemble clustering approach based on sparse graph representation and probability trajectory analysis. In particular, we present the elite neighbor selection strategy to identify the uncertain links by locally adaptive thresholds and build a sparse graph with a small number of probably reliable links. We argue that a small number of probably reliable links can lead to significantly better consensus results than using all graph links regardless of their reliability. The random walk process driven by a new transition probability matrix is utilized to explore the global information in the graph. We derive a novel and dense similarity measure from the sparse graph by analyzing the probability trajectories of the random walkers, based on which two consensus functions are further proposed. Experimental results on multiple real-world datasets demonstrate the effectiveness and efficiency of our approach.

The availability of many heterogeneous but related views of data has arisen in numerous clustering problems. Different views encode distinct representations of the same data, which often admit the same underlying cluster structure. The goal of multi-view clustering is to properly combine information from multiple views so as to generate high quality clustering results that are consistent across different views. Based on max-product belief propagation, we propose a novel multi-view clustering algorithm termed multi-view affinity propagation (MVAP). The basic idea is to establish a multi-view clustering model consisting of two components, which measure the within-view clustering quality and the explicit clustering consistency across different views, respectively. Solving this model is NP-hard, and a multi-view affinity propagation is proposed, which works by passing messages both within individual views and across different views. However, the exemplar consistency constraint makes the optimization almost impossible. To this end, by using some previously designed mathematical techniques, the messages as well as the cluster assignment vector computations are simplified to get simple yet functionally equivalent computations. Experimental results on several real-world multi-view datasets show that MVAP outperforms existing multi-view clustering algorithms. It is especially suitable for clustering more than two views.

Recently, discovering dynamic communities has become an increasingly important task. Many algorithms have been proposed, most of which only use linkage structure. However, rich information is encoded in the content of social networks such as node content and edge content, which is essential to discover topically meaningful communities. Therefore, to detect both structurally and topically meaningful communities, linkage structure, node content and edge content should be integrated. The main challenge lies in how to integrate them dynamically in a seamless way. This paper proposes a novel transformation of content-based network into a Node-Edge Interaction (NEI) network where linkage structure, node content and edge content are embedded seamlessly. A differential activity based approach is proposed to incrementally maintain the NEI network as the content-based network evolves. To capture the semantic effect of different edge types, a transition probability matrix is devised for the NEI network. Based on this, heterogeneous random walk is applied to discover dynamic communities, leading to a new dynamic community detection method termed NEIWalk (NEI network based random Walk). Theoretical analysis shows that the proposed NEIWalk method gets a bounded accuracy loss due to the random walk sampling. Experimental results demonstrate the effectiveness and efficiency of NEIWalk.

In this paper, we aim to tackle the problem of discovering dynamic communities in weighted graph streams, especially when the underlying social behavior of individuals varies considerably over different graph regions. To tackle this problem, a novel structure termed Local Weighted-Edge-based Pattern (LWEP) Summary is proposed to describe a local homogeneous region. To efficiently compute LWEPs, some statistics need to be maintained according to the principle of preserving maximum weighted neighbor information with limited memory storage. To this end, the proposed approach is divided into online and offline components. During the online phase, we introduce some statistics, termed top-k neighbor lists and top-k candidate lists, to track. The key is to maintain only the top-k neighbors with the largest link weights for each node. To allow for less active neighbors to transition into top-k neighbors, an auxiliary data structure termed top-k candidate list is used to identify emerging active neighbors. The statistics can be efficiently maintained in the online component. In the offline component, these statistics are used at each snapshot to efficiently compute LWEPs. Clustering is then performed to consolidate LWEPs into high level clusters. Finally, mapping is made between clusters of consecutive snapshots to generate temporally smooth communities. Experimental results are presented to illustrate the effectiveness and efficiency of the proposed approach.

In this paper, we propose a novel data stream clustering algorithm, termed SVStream, which is based on support vector domain description and support vector clustering. In the proposed algorithm, the data elements of a stream are mapped into a kernel space, and the support vectors are used as the summary information of the historical elements to construct cluster boundaries of arbitrary shape. To adapt to both dramatic and gradual changes, multiple spheres are dynamically maintained, each describing the corresponding data domain presented in the data stream. By allowing for bounded support vectors (BSVs), the proposed SVStream algorithm is capable of identifying overlapping clusters. A BSV decaying mechanism is designed to automatically detect and remove outliers (noise). We perform experiments over synthetic and real data streams, with the overlapping, evolving, and noise situations taken into consideration. Comparison results with state-of-the-art data stream clustering methods demonstrate the effectiveness and efficiency of the proposed method.

Kernel-based clustering is one of the most popular methods for partitioning nonlinearly separable dataset. However, exhaustive search for the global optimum is NP-hard. Iterative procedure such as k-means can be used to seek one of the local minima. Unfortunately, it is easily trapped into degenerate local minima when the prototypes of clusters are ill-initialized. In this paper, we restate the optimization problem of kernel-based clustering in an on-line learning framework, whereby a conscience mechanism is easily integrated to tackle the ill-initialization problem and faster convergence rate is achieved. Thus, we propose a novel approach termed conscience on-line learning (COLL). For each randomly taken data point, our method selects the winning prototype based on the conscience mechanism to bias the ill-initialized prototype to avoid degenerate local minima, and efficiently updates the winner by the on-line learning rule. Therefore, it can more efficiently obtain smaller distortion error than k-means with the same initialization. Experimental results on synthetic and large-scale real-world datasets, as well as that in the application of video clustering, have demonstrated the significant improvement over existing kernel clustering methods.