Diffusion models (DMs) have recently gained significant interest for their exceptional potential in recommendation tasks. This stems primarily from their prominent capability in distilling, modeling, and generating comprehensive user preferences. However, previous work fails to examine DMs in recommendation tasks through a rigorous lens. In this paper, we first experimentally investigate the completeness of recommender models from a thermodynamic view. We reveal that existing DM-based recommender models operate by maximizing the energy, while classic recommender models operate by reducing the entropy. Based on this finding, we propose a minimalistic diffusion framework that incorporates both factors via the maximization of Helmholtz free energy. Meanwhile, to foster the optimization, our reverse process is armed with a well-designed denoiser to maintain the inherent anisotropy, which measures the user-item cross-correlation in the context of bipartite graphs. Finally, we adopt an Acceptance-Rejection Gumbel Sampling Process (AR-GSP) to prioritize the far-outnumbered unobserved interactions for model robustness. AR-GSP integrates an acceptance-rejection sampling to ensure high-quality hard negative samples for general recommendation tasks, and a timestep-dependent Gumbel Softmax to handle an adaptive sampling strategy for diffusion models. Theoretical analyses and extensive experiments demonstrate that our proposed framework has distinct superiority over baselines in terms of accuracy and efficiency.

Despite their success in Collaborative Filtering (CF), Graph Convolutional Networks (GCNs) are often viewed as Low-pass Graph Filters (LGFs) with task-specific supervision, while the influence of the underlying graph's spectral properties on LGF performance remains underexplored. Our analysis reveals that the performance of LGFs is strongly affected by algebraic connectivity. When connectivity is strong, LGFs tend to perform well; when it is weak, their effectiveness diminishes noticeably. This spectral sensitivity highlights an important limitation of existing models. To address this limitation, we propose Graph Booster, a learnable module that adaptively improves graph connectivity by reweighting edges. Unlike heuristic preprocessing, Graph Booster identifies bottleneck edges via spectral embeddings and adjusts their weights with a monotonic network guided by a lightweight graph connectivity regularizer. Integrated into LightGCN framework, our model BoostGCN achieves improvements over state-of-the-art methods, underscoring the significance of algebraic connectivity for graph-based CF.

The rapid development of natural language processing (NLP) has enabled automatic emotion recognition in conversations (ERC). However, most existing models train on isolated one-hot emotion labels that overlook emotional connections between utterances and cannot effectively convey the extent and complexity of nuanced emotional perception similar to human understanding, resulting in reduced performance. To address this, we propose a self-supervisedSituation-awareEmotionLabel (SEL) generation method, which can automatically and effectively compute weights of all emotions for every utterance in a self-supervised way, utilizing no extra emotional annotated data. To generate SELs, we propose a self-supervised situation-aware method to first fusesituationinformation into utterances to generate emotion-perceptive representations.Situationis defined as the combination of the topic and surrounding people, with which emotions are widely thought to be more predictable. Then, we combine similarities of the emotion-perceptive representations with prior human knowledge to compute SELs. SEL can be deployed in most existing ERC models that train on one-hot emotion labels to enhance their performance. Experiments on six ERC models and four benchmark datasets indicate that SEL can effectively boost the performance of base models and attain state-of-the-art outcomes. Ablation study, case study, and quantitative analysis further show how SEL works.

Temporal graph representation learning seeks to capture the intrinsic evolution of nodes in temporal graphs for various applications. While existing models primarily learn node representations by aggregating temporal information from historical interactions of nodes, they often overlook the critical structural impacts arising from these interactions. To address this issue, we propose a Structure-aware model for Temporal Graph representation learning (STG), a framework that explicitly incorporates the impacts of evolving structural roles to enhance the learned node representations. Specifically, STG encodes distinct structural roles of nodes by extracting both single-unit and multi-unit interaction patterns. These roles are then transformed into the Fourier domain for a deeper analysis of the complex structural dynamics. To capture the structural impacts on future node interactions, we design a dynamic filter to process these roles. The filter is equipped with a personalized weight coefficient generator to perform the interaction-specific analysis. Finally, we employ a mixer to collaboratively aggregate the temporal and structural information to obtain structure-aware temporal node representations. Extensive experiments conducted on several real-world temporal graph datasets demonstrate the superior performance of our model in dynamic link prediction tasks under both transductive and inductive settings.

The structural heterogeneity of user-item interaction graphs poses a fundamental challenge for graph-based recommender systems. While Graph Convolutional Networks (GCNs) have achieved remarkable success in collaborative filtering, their uniform low-pass filtering nature often fails to accommodate the varying spectral needs of nodes with different local structures, resulting in suboptimal performance. To address this issue, we propose a Structurally Sensitive Adaptive Graph Filter, dubbed as SSAGF, a novel framework that enables structure-aware filtering on user-item graphs. SSAGF first clusters nodes based on local structural properties, then learns customized filters per group by reinterpreting the convolutional depth in GCNs. This adaptive mechanism ensures that nodes with distinct structural roles are treated appropriately, enhancing both accuracy and fairness without significantly increasing model complexity. To further improve scalability, SSAGF avoids costly eigenvalue decompositions by approximating spectral filters through Maclaurin series expansion, transforming the convolution into a pooling-like operation over standard GCN outputs. Extensive experiments on four benchmark datasets demonstrate that SSAGF consistently outperforms competitive baselines, especially in scenarios with high structural heterogeneity, offering a principled and efficient solution for structure-aware recommendation.

In graph-based multimedia recommendation, accurately modeling item-item semantic similarity is crucial for constructing high-quality semantic structures. However, multimodal content often exhibits semantic inconsistencies across modalities, resulting in noisy or misleading similarity signals. We refer to this as modality mismatching, where unaligned representations, such as an image conveying semantics unrelated to its accompanying text, undermine the reliability of feature-based similarity estimation. Importantly, modality consistency is context sensitive, varying with the underlying semantic environment in which modalities are interpreted. This highlights the necessity of jointly modeling modality reliability and contextual semantics. To address this challenge, we propose RGSLMRec, a robust graph structure learning framework that models and exploits the semantic reliability of multimodal features under diverse contexts. At its core, RGSLMRec builds on the invariant learning paradigm and introduces two key innovations: (i) it simulates multiple perturbed semantic environments and employs environment-specific monotonic networks to estimate reliability; and (ii) it adopts a risk-invariant objective based on Variance Risk Extrapolation to enforce the learning of invariant reliability across environments. On top of this, (iii) RGSLMRec constructs reliability-guided item-item graphs and captures collaborative and semantic signals via a hybrid early-late fusion strategy. Extensive experiments on several real-world datasets and additional synthetically perturbed datasets demonstrate that RGSLMRec not only outperforms strong baselines but also exhibits superior robustness to modality mismatching.

The knowledge tracing (KT) model based on deep learning has been proven to be superior to the traditional knowledge tracing model, eliminating the need for artificial engineering features. However, there are still problems, such as insufficient interpretability of the learning and answering processes. To address these issues, we propose a new approach in knowledge tracing with attention-based embedding and forgetting curve integration, namely KVFKT. Firstly, the embedding representation module is responsible for embedding the questions and computing the attention vector of knowledge concepts (KCs) when students answer questions and when answer time stamps are collected. Secondly, the forgetting quantification module performs the pre-prediction update of the student’s knowledge state matrix. This quantification involves calculating the interval time and associated forgetting rate of relevant KCs, following the forgetting curve. Thirdly, the answer prediction module generates responses based on students’ knowledge status, guess coefficient, and question difficulty. Finally, the knowledge status update module further refines the students’ knowledge status according to their answers to the questions and the characteristics of those questions. In the experiment, four real-world datasets are used to test the model. Experimental results show that KVFKT better traces students’ knowledge state and outperforms state-of-the-art models.

In the current digital media environment, micro-video platforms such as TikTok and Kuaishou have become major channels for content consumption and social interaction. Micro-video classification is crucial for the performance of personalized recommendation systems. However, traditional micro-video classification methods rely on a large amount of labeled data, making it challenging to handle the dynamic emergence of new categories and topics. Therefore, zero-shot micro-video classification has become an important research topic. This paper proposes a Dual Alignment Topic Model (DATM) model that achieves knowledge transfer from known categories to unknown categories through dual topic mining and alignment constraints. Specifically, we use a pre-trained language model to generate word embeddings for class semantic descriptions (CSD) and perform topic modeling on both videos and CSDs. By employing distribution alignment and reconstruction alignment constraints, we effectively mitigate the transfer problem between videos and CSDs. Extensive experiments on real-world industrial datasets validate the effectiveness of our model, significantly improving the performance of zero-shot micro-video classification.

Graph Neural Networks (GNNs) have shown promising performance in various tasks on graph-structured data. However, recent studies demonstrate that GNNs are vulnerable to adversarial attacks. In this paper, we study a realistic graph adversarial attack scenario: node injection attack (NIA). In the NIA setting, malicious nodes will be injected into the original graph instead of modifying the existing edge or node features. We present an analysis of the node fragility of GNNs under the NIA setting, based on which we propose GraphDC, which has two stages Detecting and Confusing. Specifically, GraphDC first detects the node fragility in the graph and proposes the neighbor-based and logits-based fragility measurement to select the original node connected to the injected node. Then, to confuse the decision boundary, GraphDC designs a contrastive adversarial loss to optimize the injected node features, it aims to mislead the model into classifying original nodes into closely related but incorrect classes, thereby enhancing the effectiveness of adversarial attacks. Extensive experiments conducted on four datasets demonstrate the effectiveness and better performance compared with state-of-the-art baselines.

With the widespread adoption of online education platforms, an increasing number of students are gaining new knowledge through Massive Open Online Courses (MOOCs). Exercise recommendation have made strides toward improving student learning outcomes. However, existing methods not only struggle with high dropout rates but also fail to match the diverse learning pace of students. They frequently face difficulties in adjusting to inactive students' learning patterns and in accommodating individualized learning paces, resulting in limited accuracy and diversity in recommendations. To tackle these challenges, we propose Neural Re-ranking for Diversified Exercise Recommendation (in short, NR4DER). NR4DER first leverages the mLSTM model to improve the effectiveness of the exercise filter module. It then employs a sequence enhancement method to enhance the representation of inactive students, accurately matches students with exercises of appropriate difficulty. Finally, it utilizes neural re-ranking to generate diverse recommendation lists based on individual students' learning histories. Extensive experimental results indicate that NR4DER significantly outperforms existing methods across multiple real-world datasets and effectively caters to the diverse learning pace of students.

Contemporary datasets sourced from the web often adopt a multiview format, collecting data from diverse sources, domains, or modules.Existing methodologies employed to analyze such datasets frequently overlook or inaccurately allocate the view-weights, pivotal metrics reflecting each view's significance.This work introduces EVA-MVC, a simple yet effective algorithm designed for Equitable View-weight Allocation (EVA) seamlessly integrated with arbitrary Multi-view Clustering (MVC) methods.Within the EVA module, we establish theoretical connections between view supplementarity and Multi-view Subspace Learning (MSL), leading to the partition of views into View Communities (VCs) based on these foundational principles.These VCs exhibit internal supplementarity similarities, facilitating Equitable View-weights Allocation through VCspecific MSL.The proposed EVA process precedes and operates independently of traditional or SOTA MVC approaches, requiring no additional processing or specialized design, making it an ideal preprocessing step for MVC applications.Through comprehensive evaluations across diverse multi-view datasets, our findings reveal that our EVA significantly enhances the effectiveness of mainstream MVC frameworks, resulting in a notable performance improvement.

Intrusion detection is crucial for safeguarding intelligent vehicle systems, aiming to identify abnormal network traffic and operational anomalies. Traditional methods primarily focus on spatial features of attacks, often neglecting temporal dynamics essential for detecting complex, evolving threats. Additionally, the effectiveness of existing techniques is limited by the scope and quality of available datasets, reducing their ability to detect novel, unseen attacks. To address these challenges, this article introduces a Transformer-based transfer learning intrusion detection system (TIDS), designed to capture and analyze spatiotemporal sequence features from vehicle data. TIDS generates high-dimensional feature representations of intricate intrusion patterns, improving the detection of known attack types through instance-based transfer learning, enhancing domain adaptability. Moreover, we proposed a novel self-supervised box classification method that enhances the system’s capability to detect previously unknown attacks, thereby increasing the overall robustness of the intrusion detection process. Comparative experiments demonstrate that TIDS outperforms traditional methods in detection speed and accuracy across various intrusion scenarios, effectively responding to emerging threats in intelligent vehicle networks.

Recommending suitable exercises and providing the reasons for these recommendations is a highly valuable task, as it can significantly improve students' learning efficiency. Nevertheless, the extensive range of exercise resources and the diverse learning capacities of students present a notable difficulty in recommending exercises. Collaborative filtering approaches frequently have difficulties in recommending suitable exercises, whereas deep learning methods lack explanation, which restricts their practical use. To address these issue, this paper proposes KG4EER, an explainable exercise recommendation with a knowledge graph. KG4EER facilitates the matching of various students with suitable exercises and offers explanations for its recommendations. More precisely, a feature extraction module is introduced to represent students' learning features, and a knowledge graph is constructed to recommend exercises. This knowledge graph, which includes three primary entities - knowledge concepts, students, and exercises - and their interrelationships, serves to recommend suitable exercises. Extensive experiments conducted on three real-world datasets, coupled with expert interviews, establish the superiority of KG4EER over existing baseline methods and underscore its robust explainability.

Nowadays, as human activities are shifting to social media, fake news detection has been a crucial problem. Existing methods ignore the classification difference in online news and cannot take full advantage of multi-classification knowledges. For example, when coping with a post “A mouse is frightened by a cat,” a model that learns “computer” knowledge tends to misunderstand “mouse” and give a fake label, but a model that learns “animal” knowledge tends to give a true label. Therefore, this research proposes a multi-classification division-aggregation framework to detect fake news, namedCKA, which innovatively learns classification knowledges during training stages and aggregates them during prediction stages. It consists of three main components: a news characterizer, an ensemble coordinator, and a truth predictor. The news characterizer is responsible for extracting news features and obtaining news classifications. Cooperating with the news characterizer, the ensemble coordinator generates classification-specifical models for the maximum reservation of classification knowledges during the training stage, where each classification-specifical model maximizes the detection performance of fake news on corresponding news classifications. Further, to aggregate the classification knowledges during the prediction stage, the truth predictor uses the truth discovery technology to aggregate the predictions from different classification-specifical models based on reliability evaluation of classification-specifical models. Extensive experiments prove that our proposedCKAoutperforms state-of-the-art baselines in fake news detection.

Unlabeled text representation learning (UTRL), encompassing static word embeddings such as Word2Vec and contextualized word embeddings such as bidirectional encoder representations from transformer (BERT), aims to capture semantic word relationships in a low-dimensional space without the need for manual labeling. These word embeddings are invaluable for downstream tasks such as document classification and clustering. However, the surge of short texts generated daily on social media platforms results in sparse word cooccurrences, compromising UTRL outcomes. Contextualized models such as recurrent neural network (RNN) and BERT, while impressive, often struggle with predicting the next word due to sparse word sequences in short texts. To address this, we introduce CGraphNet, a contrastive graph context prediction model designed for UTRL. This approach converts short texts into graphs, establishing links between sequentially occurring words. Information from the next word and its neighbors informs the target prediction, a process referred to as graph context prediction, mitigating sparse word cooccurrence issues in brief sentences. To minimize noise, an attention mechanism assigns importance to neighbors, while a contrastive objective encourages more distinctive representations by comparing the target word with its neighbors. Our experiments demonstrate CGraphNet's superior performance over other baselines, particularly in classification and clustering tasks on real-world datasets.

Generalized few-shot node classification (GFS-NC) is a very important challenge for graph-based algorithms, as it requires to identify novel classes and base classes simultaneously. Although there are several methods that try to combine metalearning or metric learning with graph neural networks to solve few-shot problem, most of them assume that test samples only come from the novel classes, which is impractical in reality. Besides, they overlook the relationship among classes, which can provide additional information for the novel classes classification. In this article, we propose a graph-based knowledge distillation network (GraphKD) to extract the class relationship and learn better nodes representations for nodes from novel classes in GFS-NC task. GraphKD consists of two modules: balanced pretraining module and class-relation transferring module. Balanced pretraining can optimize network parameters to a suitable manifold for subsequent initialization. The class-relation transferring module leverages a knowledge distillation model, where a teacher model generates soft labels containing interclass relationships and then transfer them to the student model. The student model is optimized to fit both the soft labels and hard labels concurrently. This relationship information can help the student model better understand the similarities and differences between classes, thereby improving its classification performance. In addition, we employee information entropy to distinguish the samples locate at the boundary of a base class and novel class and then assign them larger weights in the student model to enhance its expressive capacity for novel nodes. Our experiments show that the proposed method outperforms state-of-the-art baselines on various few-shot node classification datasets.

Online social networks often enforce a restricted web interface to allow third part users to access their data objects in one-by-one manner along links. The heavy cost of collecting data from them makes it very difficult to train a graph neural networks (GNNs) algorithm by using valuable data resources from those social network services. In this article, we endeavor to harness the potential of collected data in order to cultivate robust node representations, all while taking into account the prevalent class-imbalanced scenario. To this end, we propose a framework based on the reinforcement learning technique that can implement training while sampling. We employee deep Q-learning network (DQN) as the basic reinforcement learning framework which maintains an experience replay buffer to store experiences in order to improve efficiency and generalization ability of the algorithm. The proposed framework can guide the training process according to the label distribution of the currently collected training data. To maximize the utility of the collected data, we introduce two ways to guide the agent to learn more robust node representations under the natural class-imbalanced scenario. Specifically, effective-number GNNs (ENG) try to employ a reward function with the technique of effective number and resampling GNNs (RSG) attempts to modify the state transition function motivated by resampling technique. Extensive experiments are conducted on several real-world imbalanced datasets and the proposed methods significantly outperform the state-of-the-art methods in node classification task, accomplishing higher performance while utilizing a reduced volume of data.

In the domain of Multi-view Subspace Clustering (MSC) in Latent Embedding Space (LES), existing methods aim to capture and leverage critical multi-view information by mapping it into a low-dimensional LES. However, several aspects can be further improved: (i) Fusion Strategy: Existing methods adopt either early fusion or late fusion to integrate multi-view information, limiting the effectiveness of the fusion. (ii) Diversity: Current methods often overlook the inherent diversity in the multi-view data by focusing on a single LES. (iii) Efficiency: LES-based methods exhibit high computational complexity, with cubic time and quadratic space requirements based on the number of samples. To address these issues, we propose a novel framework called MSC-DOLES (Multi-view Subspace Clustering in Diverse Orthogonal Latent Embedding Spaces), a novel framework designed to tackle these challenges. MSC-DOLES incorporates a two-stage fusion approach that generates and learns from multiple LES to maximize cross-view diversity. Orthogonality constraints on individual LES ensure view-internal diversity, resulting in a set of Diverse Orthogonal Latent Embedding Spaces (DOLES). The DOLES are then fused into a consensus anchor graph using learnable anchors. The final clustering is induced by partitioning the obtained graph without pre-processing. We develop an eight-step optimization algorithm for MSC-DOLES, which exhibits nearly linear time and space complexities relative to the number of samples. Extensive experiments demonstrate the superiority of MSC-DOLES over state-of-the-art methods.

Graph-based semi-supervised learning (GSSL) has long been a research focus. Traditional methods are generally shallow learners, based on the cluster assumption. Recently, graph convolutional networks (GCNs) have become the predominant techniques for their promising performance. However, a critical question remains largely unanswered: why do deep GCNs encounter the oversmoothing problem, while traditional shallow GSSL methods do not, despite both progressing through the graph in a similar iterative manner? In this article, we theoretically discuss the relationship between these two types of methods in a unified optimization framework. One of the most intriguing findings is that, unlike traditional ones, typical GCNs may not effectively incorporate both graph structure and label information at each layer. Motivated by this, we propose three simple but powerful graph convolution methods. The first, optimized simple graph convolution (OGC), is a supervised method, which guides the graph convolution process with labels. The others are two "no-learning" unsupervised methods: graph structure preserving graph convolution (GGC) and its multiscale version GGCM, both aiming to preserve the graph structure information during the convolution process. Finally, we conduct extensive experiments to show the effectiveness of our methods.

Graphs in the real world keep evolving with the integration of new nodes, and it is often infeasible to manually label all the new nodes promptly. In this case, graph learning algorithms can come in handy and perform classification on these newly emerging nodes. Typically, if unseen classes exist (i.e., no training samples from these classes), one can perform zero-shot learning (ZSL) or generalized ZSL (GZSL). During testing, ZSL aims to classify samples within unseen classes, whereas GZSL aims to classify samples within both seen and unseen classes, which is even more challenging. In our previous work, we proposed a decomposed graph prototype network (DGPN) to decompose the graph convolution operation for handling the zero-shot node classification (ZNC) problem. However, DGPN is not well-suited for the generalized ZNC (GZNC) problem. To this end, in this article, we propose a novel graph generative model, multihop reconstruction graph autoencoder (MHR-GAE). Unlike DGPN, MHR-GAE utilizes a multihop encoder with class semantic descriptions (CSDs) (as condition signals) to reconstruct the information and generate nodes of unseen classes. Thus, it can handle both the ZNC and GZNC problems and obtain competitive performance. We evaluate our model on real-world datasets, and the experimental results demonstrate that MHR-GAE outperforms other baseline methods.

Multiview clustering (MVC) has been widely studied in machine learning and data mining for its capability of improving clustering performance by fusing the information from multiview data. In the past decade, a large number of MVC methods have made impressive progress, but most of them suffer from computational burdens, especially in large-scale tasks. Binary MVC (BMVC) is proposed to address this issue by representing the large-scale high-dimensional dataset as a group of consensus and low-dimensional binary codes. However, current BMVC-based approaches generate the clustering by executing binary k-means on the obtained binary codes, which fail to capture the embedded geometric information, leading to poor clustering performance. In addition, parameter selection is another "mission impossible" in unsupervised learning tasks including MVC. To tackle these challenges, a framework of multiview clustering via partitioning the signed prototype graph (SPGMVC) is proposed in this work. The SPGMVC framework offers several contributions. First, SPGMVC is designed as a unified framework for MVC. It combines effective technologies, such as consensus binary coding, code compression (CC), signed prototype graph (SPG) partitioning, and prototype-based cluster assignment. Second, SPGMVC partitions the signed graph (SG) based on the relationships between positive and negative edges. By capturing the underlying structure of the data, this partitioning strategy improves clustering accuracy (ACC). CC techniques are applied to reduce the graph's scale, enabling further partitioning and enhancing computational efficiency. Third, SPGMVC employs an alternate minimizing strategy to efficiently handle the optimization problem. This strategy has nearly linear time and space complexity with respect to the data volume, making it suitable for large-scale tasks. Fourth, SPGMVC proposes an automatic parameter selection strategy, eliminating the need for extensive parameter exploration. Comprehensive experiments illustrate the superiority of our model. The implementation of SPGMVC is available at: https://github.com/gepingyang/PSGMVC.

Traffic prediction is essential for the efficient resource allocation and management of smart cities. However, many cities face challenges in accessing sufficient traffic data due to high collection costs and privacy concerns, which impacts the accuracy of predictions. To address the issue of data sparsity, transfer learning methods have been explored with promising outcomes. Nonetheless, key challenges remain in determining which knowledge to transfer and how to transfer it effectively. In this paper, we propose a Knowledge Transfer Service for multi-granular traffic prediction in smart cities, leveraging Multi-Granular Spatiotemporal Pattern Transfer Learning (MGSTPAT). MGSTPAT utilizes meta-learning to acquire rich, multi-granularity spatiotemporal knowledge from data-abundant source cities and initializes robust models. Additionally, our Embedded Adaptive Clustering module generates task-specific multi-granularity patterns, while the Granularity-Aware Pattern Attention mechanism enables target cities with limited data to select the most relevant patterns, enhancing prediction accuracy. Extensive experiments on real-world datasets demonstrate the effectiveness of our approach in addressing traffic prediction challenges across cities.

Sequential Recommendation plays a significant role in daily recommendation systems, such as e-commerce platforms like Amazon and Taobao. However, even with the advent of large models, these platforms often face sparse issues in the historical browsing records of individual users due to new users joining or the introduction of new products. As a result, existing sequence recommendation algorithms may not perform well. To address this, sequence-based data augmentation methods have garnered attention. Existing sequence enhancement methods typically rely on augmenting existing data, employing techniques like cropping, masking prediction, random reordering, and random replacement of the original sequence. While these methods have shown improvements, they often overlook the exploration of the deep embedding space of the sequence. To tackle these challenges, we propose a Sparse Enhanced Network (SparseEnNet), which is a robust adversarial generation method. SparseEnNet aims to fully explore the hidden space in sequence recommendation, generating more robust enhanced items. Additionally, we adopt an adversarial generation method, allowing the model to differentiate between data augmentation categories and achieve better prediction performance for the next item in the sequence. Experiments have demonstrated that our method achieves a remarkable 4-14% improvement over existing methods when evaluated on the real-world datasets. (https://github.com/junyachen/SparseEnNet)

Social recommendation systems leverage the social relations among users to deal with the inherent cold-start problem in user-item interactions. However, previous models only treat the social graph as the static auxiliary to the user-item interaction graph, rather than dig out the hidden essentials and optimize them for better recommendations. Thus, the potential of social influence is still under-explored. In this paper, we will fill this gap by proposing a novel model for social influence learning to derive the essential influence patterns within the user relationships. Our model views the social influence from the perspectives of (1) the diversity of neighborhood's influence on the users, (2) the disentanglement of neighborhood's influence on the users, and (3) the exploration of underlying implicit social influence. To this end, we first employ a novel layerwise graph-enhanced variational autoencoder for the reconstruction of neighborhoods' representations, which aims to learn the pattern of social influence as well as simulate the social profile of each user for overcoming the sparsity issue in social relation data. Meanwhile, we introduce a layerwise graph attentive network for capturing the most influential scope of neighborhood. Finally, we adopt a dual sampling process to generate new social relations for enhancing the social recommendation. Extensive experiments have been conducted on three widely-used benchmark datasets, verifying the superiority of our proposed model compared with the representative approaches.

Multi-Modal Large Language Models (MM-LLMs) have demonstrated powerful reasoning abilities in various visual question-answering tasks. However, they face the challenge of lacking rigorous reasoning and precise arithmetic, when solving geometry questions. To address this challenge, we propose a novel prompting method, namely Reason-and-Execute (R&E), to enhance the accuracy of solving geometry questions by MM-LLMs. Specifically, the R&E method includes two templates: reasoning template and execution template. We first adopt a reverse-thinking approach to construct a rigorous reasoning template so that it guides MM-LLMs to start reasoning from the most relevant domain knowledge of the question and ultimately identify the arithmetic requirements. We then make use of program-assisted thought to construct execution template in order to guide MM-LLMs to understand the arithmetic requirements from reasoning template and generate executable code block. The answer is finally obtained by executing the code block. We evaluate our prompting method on 9 models in answering questions on 6 datasets (including four geometry datasets and two science datasets) compared to Chain-of-Thought (CoT) and Program-Aided Language (PAL) prompting methods. R&E method shows up to 12.8% improvement compared to CoT and PAL, proving strong reasoning and arithmetic abilities for solving geometry questions of our method. Moreover, we further analyze the answering accuracy from the different perspectives on solving geometric questions, including domain knowledge, geometry shapes, question length, and language. Through multiple analysis, our method is able to enhance the ability of MM-LLMs to solve geometry questions.

In this paper, we propose a novel self-supervised Gaussian Restricted Boltzmann Machine with contrastive learning (CL-GRBM), which fuses contrastive representation learning and contrastive divergence to optimize and enhance the representation of GRBM. Built upon the concept of contrastive representation learning, CL-GRBM aims to enhance the representation capacity of the GRBM. Firstly, a pair of positive samples are constructed by one times Gibbs sampling with the original data. Then, the contrastive loss is used to pull the positive samples closer and push other samples further apart, making similar representations closer and different representations farther apart. In summary, during the training process of CL-GRBM using the CD algorithm, the objective is to align the sampling probability distribution of the visible layer in CL-GRBM as closely as possible with the empirical distribution of the original data. In the feature space of the sampling probabilities in the hidden layer, the distance between positive samples is minimized to capture the intrinsic structure of the data. According to the experimental verification, the proposed CL-GRBM shows better performance than contrastive models. As a shallow self-supervised model, it has even better performance than some excellent deep self-supervised models.

In unsupervised domain adaptation (UDA), a classifier is applied to the target domain without or with limited labels, when the target domain has no or few labels. Recently, inspired by their capabilities of long-distance feature dependencies, vision transformer (ViT)-based methods have been used in UDA, however, they ignore the fact that ViT lacks strength in extracting local feature details. To handle the above problems, the purpose of this article is to demonstrate how to take advantage of both convolutional operations and transformer mechanisms for adversarial UDA by using a hybrid network structure called transformer fused convolution (TFC). TFC integrates local features with global features to boost the representation capacity for UDA which can enhance the discrimination between foreground and background. Moreover, to improve the robustness of the TFC, we leverage an uncertainty penalty loss to make incorrect classes have consistently lower scores. Extensive experiments validate the significant performance gains compared to conditional adversarial domain adaptation (CDAN) on all five datasets including DomainNet ($\uparrow ~8.5$%), VisDA-2017 ($\uparrow ~14.9$%), Office-Home ($\uparrow ~18.9$%), Office-31 ($\uparrow ~11.5$%), and ImageCLEF-DA ($\uparrow ~5.5$%).

The rapid development of heterogeneous networks has proposed new challenges to the long-standing link prediction problem. Existing models trained on the verified edge samples from different types usually learn type-specific knowledge, and their type-specific predictions may be contradictory for unverified edge samples with uncertain types. This challenge is termed edge-type disturbance in link prediction in heterogeneous networks. To address this challenge, we develop a disturbance-resilient prediction method ( DRPM ) comprising a structural characterizer, a type differentiator, and a resilient predictor. The structural characterizer is responsible for learning edge representations for link prediction. Concurrently, the type differentiator distinguishes type-specific edge representations to generate diverse type experts while maximizing their link prediction performances on specific types. Furthermore, the resilient predictor evaluates the reliability weights of different type experts to develop a resilient prediction mechanism to aggregate discriminable predictions. Extensive experiments conducted on various real-world datasets demonstrate the importance of the explainable introduction of the edge-type disturbance and the superiority of DRPM over state-of-the-art methods.

In recent years, graph embedding techniques have exhibited great potential for various downstream tasks, which can leverage both topological structures and the temporal dependencies of nodes in their representations, leading to remarkable achievements. However, the multi-role nature of nodes during their temporally interacting is neglected. To tackle this problem, we propose a novel model, Temporal graph Multi-Aspect Embedding (TMAE), to capture the latent multi-aspect characteristics of nodes in temporal graphs, thereby enhancing the quality of graph embeddings. Specifically, we propose to learn the aspect embeddings of nodes and their weights at different timestamps separately for a better adaptation. In contrast to the conventional fixed aspect number assumption, a Hierarchical Dirichlet Process-based approach is employed to dynamically determine the weight of aspects for nodes at different times. Through this framework, we effectively learn the multi-aspect information through Time-reversed Temporal Walks (TTWs). Extensive experiments performed across eight publicly accessible datasets have demonstrated the significant improvements of the proposed TMAE model over state-of-the-art algorithms by taking advantage of the multi-aspect nature.

Graph Neural Networks (GNNs) have achieved remarkable success in graph-based tasks. However, learning unbiased node representations under class-imbalanced training data remains challenging. Existing solutions may face overfitting due to extensive reuse of those limited labeled data in minority classes. Furthermore, many works address the class-imbalanced issue based on the embeddings generated from the biased GNNs, which make models intrinsically biased towards majority classes. In this paper, we propose a novel data augmentation strategy GraphGLS for semi-supervised class-imbalanced node classification, which aims to select informative unlabeled nodes to augment minority classes with consideration of both global and local information. Specifically, we first design a Global Selection module to learn global information (pseudo-labels) for unlabeled nodes and then select potential ones from them for minority classes. The Local Selection module further conducts filtering over those potential nodes by comparing their neighbor distributions with minority classes. To achieve this, we further design a neighbor distribution auto-encoder to learn a robust node-level neighbor distribution for each node. Then, we define class-level neighbor distribution to capture the overall neighbor characteristics of nodes within the same class. We conduct extensive experiments on multiple datasets, and the results demonstrate the superiority of GraphGLS over state-of-the-art baselines.

Graph neural networks (GNNs) have achieved great success in node classification tasks. However, existing GNNs naturally bias towards the majority classes with more labelled data and ignore those minority classes with relatively few labelled ones. The traditional techniques often resort over-sampling methods, but they may cause overfitting problem. More recently, some works propose to synthesize additional nodes for minority classes from the labelled nodes, however, there is no any guarantee if those generated nodes really stand for the the corresponding minority classes. In fact, improperly synthesized nodes may result in insufficient generalization of the algorithm. To resolve the problem, in this paper we seek to automatically augment the minority classes from the massive unlabelled nodes of the graph. Specifically, we propose \textit{GraphSR}, a novel self-training strategy to augment the minority classes with significant diversity of unlabelled nodes, which is based on a Similarity-based selection module and a Reinforcement Learning(RL) selection module. The first module finds a subset of unlabelled nodes which are most similar to those labelled minority nodes, and the second one further determines the representative and reliable nodes from the subset via RL technique. Furthermore, the RL-based module can adaptively determine the sampling scale according to current training data. This strategy is general and can be easily combined with different GNNs models. Our experiments demonstrate the proposed approach outperforms the state-of-the-art baselines on various class-imbalanced datasets.

Feature learning is a crucial phase machine learning [1] – [3] . How to obtain appropriate features distribution without any background is still a hard problem in machine learning. In this paper, we present a novel unsupervised feature learning architecture (see Fig. 1 ), which consists of a multi-clustering integration module and a variant of RBM termed multi-clustering integration RBM (MIRBM). In the multi-clustering integration module, we choose three clusterers to obtain three different global clustering partitions (CPs). Then, an unanimous voting strategy is used to generate the local clustering partition (LCP) of visible layer data. Hence, the LCP only has partial visible layer data. The novel MIRBM model is a core feature encoding part of the proposed unsupervised feature learning architecture. The novelty of it is that the LCP as an unsupervised guidance is integrated into the CD 1 learning to guide the distribution of the hidden layer features. For the instance in the same LCP cluster, the hidden and reconstructed hidden layer features of the MIRBM model in the proposed architecture tend to constrict together in the training process. Meanwhile, each LCP center tends to disperse from each other as much as possible in the hidden and reconstructed hidden layer during training. This work has three main contributions: 1) A novel unsupervised feature learning architecture is proposed, which consists of a multi-clustering integration module and an MIRBM model. 2) In the multi-clustering integration module of the proposed architecture, three unsupervised algorithms are employed to obtain three different global CPs without any background knowledge or label. 3) The MIRBM model in the proposed architecture uses the LCP as an unsupervised guidance to guide the distribution of the hidden layer features by integrating the LCP into the CD 1 learning.

THE popularity of location-aware devices such as smart phones makes users freely share their activities through various location-based social networks (LBSNs), such as Foursquare and Yelp. A large amount of user-contributed data enable to develop effective point-of-interest (POI) recommender systems. It not only guides users to explore more interesting attractions, but also helps the location service providers deliver targeted advertising. Now most of existing studies focus on recommending POIs in the same city or region, named as traditional POI recommender systems. However, they fail to deal with the increasingly popular case: users travel to new cities to explore more attractions. This raises the problem that how we shall recommend POIs in a target city to a new visitor based on her/his check-in records in source cities. We refer to this problem as crossing-city POI recommendations. Compared with traditional POI recommender systems, crossing-city POI recommender systems are more challenging due to the following aspects:

Open set domain adaptation aims to transfer knowledge in the presence of unknown samples in the target domain. Previous approaches use additional classifiers or threshold-based methods to identify unknown samples and try to investigate the information of class diversity within the unknown samples. Despite achieving excellent adaptation results, these methods ignore those samples that lie on the cluster boundaries, especially the clustering-based methods. In this paper, we propose a novel Neighbor Prototype Contrastive Clustering (NPC2) method, which uses the Local Semantic Structure (LSS) to help these low-confidence samples located on the boundary of clusters to return to their own clusters. Further, we propose Local Semantic Consistency (LSC) to evaluate the clustering result and apply it to the domain adaptation process as a metric to assess the reliability of the samples. Results on four benchmarks show that our NPC2significantly outperforms most state-of-the-art methods with higher LSC.

Abstract User-generated content is daily produced in social media, as such user interest summarization is critical to distill salient information from massive information for recommendation tasks. While the interested messages (e.g., tags or posts) from a single user are usually sparse becoming a bottleneck for existing methods, we propose a neural inference method (NIGraphNet) by mining user social interest for item recommendation. It can unearth user latent topics combined with user relation learning. Specifically, we exploit a neural variational inference approach to learn the distributions between user interests and hidden topics. (We denote it as interest-topic distributions in the following.) Then, we adopt a unified graph-based training loss that jointly learns the hidden topics and user relations for item recommendation. Experiments on two datasets collected from well-known social media platforms demonstrate the superior performance of our model in the tasks of user interest summarization and item recommendation. Further discussions also show that exploiting the latent topic representations and user relations is conducive to the user’s automatic language understanding.

The instability is shown in the existing methods of representation learning based on Euclidean distance under a broad set of conditions. Furthermore, the scarcity and high cost of labels prompt us to explore more expressive representation learning methods which depends on as few labels as possible. To address above issues, the small-perturbation ideology is firstly introduced on the representation learning model based on the representation probability distribution. The positive small-perturbation information (SPI) which only depend on two labels of each cluster is used to stimulate the representation probability distribution and then two variant models are proposed to fine-tune the expected representation distribution of Restricted Boltzmann Machine (RBM), namely, Micro-supervised Disturbance Gaussian-binary RBM (Micro-DGRBM) and Micro-supervised Disturbance RBM (Micro-DRBM) models. The Kullback-Leibler (KL) divergence of SPI is minimized in the same cluster to promote the representation probability distributions to become more similar in Contrastive Divergence (CD) learning. In contrast, the KL divergence of SPI is maximized in the different clusters to enforce the representation probability distributions to become more dissimilar in CD learning. To explore the representation learning capability under the continuous stimulation of the SPI, we present a deep Micro-supervised Disturbance Learning (Micro-DL) framework based on the Micro-DGRBM and Micro-DRBM models and compare it with a similar deep structure which has no external stimulation. Experimental results demonstrate that the proposed deep Micro-DL architecture shows better performance in comparison to the baseline method, the most related shallow models and deep frameworks for clustering.

With the rapid development of the Internet of Things technology, the concept of smart cities that aims to help residents improve their quality of life has raised much attention in several application areas. In the context of smart cities, the provision of point of interest (POI) recommendations become an important requirement because a wide range of POIs are available for urban dwellers. Location-based social networks (LBSNs) such as Foursquare and Gowalla provide a massive volume of user check-in records that can assist users in choosing new POIs. However, user trajectories are mostly sparse in the real world. For example, users only check in a few POIs, and this makes it difficult to provide recommendations based on limited history trajectories. Though some attempts have adopted auxiliary geographical information to enhance POI recommendation, they still encounter the following problems: 1) the geographical trajectories of users are usually sparse in real-world datasets; 2) users may be more interested in the remote POIs; and 3) the previous models inherently perform transductive learning that cannot handle well the recommendation of unseen users and POIs. To address these problems, we propose an inductive representation learning model (IRLM) for location recommendation. IRLM contains two parts, namely geographic feature extraction and inductive representation learning. IRLM first captures global geographical influences among POIs through a standard Gaussian mixture model (GMM). Then IRLM adopts an attention neural network for the recommendation. Experimental results indicate that our proposed model can achieve superior performance over state-of-the-art models.

Recently, nonshared-and-imbalanced unsupervised domain adaption has been proposed to fix domain shift from Big Data source domain with long-tail distribution to specific small target domain with imbalanced distribution, including two challenges: 1) nonshared classes sharing in big data with long-tail distribution; and 2) imbalanced domain adaptation. Prior approaches explore knowledge sharing between classes to improve performance of unsupervised domain adaption methods. However these methods have inductive bias for prior tree or graph. And previous contrastive domain adaptation methods take center-based prototypes as positive samples which only coarsely characterize the domain structure, and fail to depict the local data structure. To fix these problems, we propose a novel framework called contrastive memory feature transfer (CMFT). To solve nonshared data sharing without inductive bias, we build a centroid memory baseddirected memory transfermechanism to enhance imbalanced class features with similar nonshared class centroid. To address the imbalanced domain adaptation, we design a fault-tolerant and fine-grainedneighborhood prototypefor the contrastive learning which can narrow the domain shift. The proposed CMFT outperforms previous methods on most benchmarks.

Accurate traffic prediction is one of the most important techniques in building a smart city. Many works, especially deep learning models, have made great progress in traffic prediction based on rich historical data. However, many cities still suffer from the problem of data scarcity in many aspects. Some works use transfer learning to solve this kind of problem, but what and how to transfer is still an important problem. In this article, we propose a novel Cross-city Multi-Granular Adaptive Transfer Learning method named MGAT for traffic prediction with only a few data in the target city. We first use the meta-learning algorithm to train the model on multiple source cities to get a good initialization. And at the same time, the multi-granular regional characteristics of each source city will be obtained based on our model structure. Then we design an Adaptive Transfer module mainly composed of Spatial-Attention and Multi-head Attention mechanism to automatically select the most appropriate features from the multi-granular features trained from multiple source cities, to achieve the best transfer effect. We conduct extensive experiments on two kinds of real-world traffic datasets cross several cities. Experimental results with other state-of-the-art models demonstrate the effectiveness of the proposed model.

Despite the massive surge of evolving social network analysis in popularity, existing research usually represent the observed social interactions among individuals as completely credible edges. However, due to information inaccuracy, individual non-response and dropout, and sampling biases in observations, the evolving noisy social network that coexists true edges and spurious edges is pervasive in actual applications, where the ignoration of credibility otherness of observed edges could lead to the wrong estimates of social properties and misleading conclusions. To discover credible edge information to shape correct social interactions among individuals, we propose a universal and explainable multiple-neighbor evolutional filtering method (MEFM) to evaluate how credible of observed edges to ‘truly’ exist in the evolving noisy social network.MEFMconsists of an evolutional extractor and a filtering evaluator. To resist the noisy disturbance, the evolutional extractor exploits the evolutional states of edges from the perspective of evolution mechanisms within multiple-neighbor ranges, which applies different link prediction algorithms to fit the evolution mechanism in the formation of each edge. Further, the filtering evaluator reconstructs Kalman filter to predict and refine the evolutional states of edges based on their evolving local structures. As a result,MEFMcombines the evolutional extractor and the filtering evaluator to analyze the evolutional fluctuations of the observed edges to evaluate their credibility. Extensive experiments on real-world datasets demonstrate that our proposedMEFMcan effectively and reasonably evaluate edge credibility in evolving noisy social networks.

As travel efficiency matters to the work productivity of cities, shortening passengers' travel time for metros is therefore a pressing need. To this end, we study a strategy by dynamically scheduling dwell time for trains. Developing such a strategy is challenging because of three aspects: 1) Optimizing the average travel time of passengers needs to properly balance passengers' waiting time at platforms and journey time on trains, as well as considering long-term impacts; 2) Capturing dynamic spatio-temporal (ST) correlations of incoming passengers for metro stations is difficult; and 3) For each train, the dwell time scheduling is affected by other trains, which is hard to measure. To tackle these challenges, we propose a novel deep neural network, entitled AutoDwell. Specifically, AutoDwell optimizes the long-term rewards of dwell time settings in terms of passengers' waiting and journey time by a reinforcement learning framework. Next, AutoDwell employs gated recurrent units and graph attention networks to extract the ST correlations of the passenger flows among metro stations. Moreover, attention mechanisms are leveraged in AutoDwell for capturing the interactions between the trains. Extensive experiments on two real-world datasets demonstrate the superior performance of AutoDwell over several baselines, capable of saving passengers' travel time significantly.

Spectral Clustering (SC) is an effective clustering method for its excellent performance in partitioning non-linearly distributed data. On the other hand, Ensemble Clustering (EC), a different clustering technology, can promote cluster quality by ensembling the results of base clusterings. In this work, we concentrate on an EC framework that utilizes SC as the base method. Nevertheless, SC suffers from scalability due to its high computational complexity in constructing the Laplacian graph and computing the corresponding eigendecomposition. In the past decades, many efforts have been made to it. However, SC suffers from the scalability issue in processing extensive data, especially in web-scale scenarios. Additionally, EC requires multiple clustering results as the ensemble bases, which further aggravates resource consumption. To address this issue, LiteWSEC, a simple yet efficient Lightweight Framework for Web-scale Spectral Ensemble Clustering, is proposed to cluster web-scale data with limited resource requirements. It adopts the Web-scale Spectral Clustering (WSC) as the base method, which has minimal space overhead without computing overall embedding explicitly. LiteWSEC is highly flexible in the memory requirement, which is adaptive to the available resource. It can partition web-scale data (e.g.,$n $= 8,000 k) in an resource-limited host (e.g., memory is restricted to 1 GB). Experiments on real-world, large-scale, and web-scale datasets demonstrate both the efficiency and effectiveness of LiteWSEC over state-of-the-art SC and EC methods.

Graph embedding is regarded as one of the most advanced techniques for graph data analyses due to its significant performance. However, the majority of existing works only focus on static graphs while ignoring the ubiquitous dynamic graphs. In fact, the temporal evolution of edges in a dynamic graph sets a harsh challenge for the traditional embedding algorithms. To solve the problem, in this paper we propose a Dynamic Graph Multi-Aspect Embedding (DGMAE) to automatically learn the proper number of aspects and their distributions in each temporal duration based on a distance dependent Chinese Restaurant Process. The proposed method can encode the inherent property of varying interactions among nodes along the time and present different aspect-influences to nodes embedding. Our extensive experiments on several public datasets show the performance improvement over state-of-the-art works.

User representation learning plays an essential role in Internet applications, such as recommender systems. Though developing a universal embedding for users is demanding, only few previous works are conducted in an unsupervised learning manner. The unsupervised method is however important as most of the user data is collected without specific labels. In this paper, we harness the unsupervised advantages of Variational Autoencoders (VAEs), to learn user representation from large-scale, high-dimensional, and multi-field data. We extend the traditional VAE by developing Field-aware VAE (FVAE) to model each feature field with an independent multinomial distribution. To reduce the complexity in training, we employ dynamic hash tables, a batched softmax function, and a feature sampling strategy to improve the efficiency of our method. We conduct experiments on multiple datasets, showing that the proposed FVAE significantly outperforms baselines on several tasks of data reconstruction and tag prediction. Moreover, we deploy the proposed method in real-world applications and conduct online A/B tests in a look-alike system. Results demonstrate that our method can effectively improve the quality of recommendation. To the best of our knowledge, it is the first time that the VAE-based user representation learning model is applied to real-world recommender systems.

Citation recommendation plays an important role in the context of scholarly big data, where finding relevant papers has become more difficult because of information overload. Applying traditional collaborative filtering (CF) to citation recommendation is challenging due to the cold start problem and the lack of paper ratings. To address these challenges, in this article, we propose a collaborative filtering with network representation learning framework for citation recommendation, namely CNCRec, which is a hybrid user-based CF considering both paper content and network topology. It aims at recommending citations in heterogeneous academic information networks. CNCRec creates the paper rating matrix based on attributed citation network representation learning, where the attributes are topics extracted from the paper text information. Meanwhile, the learned representations of attributed collaboration network is utilized to improve the selection of nearest neighbors. By harnessing the power of network representation learning, CNCRec is able to make full use of the whole citation network topology compared with previous context-aware network-based models. Extensive experiments on both DBLP and APS datasets show that the proposed method outperforms state-of-the-art methods in terms of precision, recall, and MRR (Mean Reciprocal Rank). Moreover, CNCRec can better solve the data sparsity problem compared with other CF-based baselines.

The abundant sequential documents such as online archival, social media, and news feeds are streamingly updated, where each chunk of documents is incorporated with smoothly evolving yet dependent topics. Such digital texts have attracted extensive research on dynamic topic modeling to infer hidden evolving topics and their temporal dependencies. However, most of the existing approaches focus on single-topic-thread evolution and ignore the fact that a current topic may be coupled with multiple relevant prior topics. In addition, these approaches also incur the intractable inference problem when inferring latent parameters, resulting in a high computational cost and performance degradation. In this work, we assume that a current topic evolves from all prior topics with corresponding coupling weights, forming the multi-topic-thread evolution . Our method models the dependencies between evolving topics and thoroughly encodes their complex multi-couplings across time steps. To conquer the intractable inference challenge, a new solution with a set of novel data augmentation techniques is proposed, which successfully discomposes the multi-couplings between evolving topics. A fully conjugate model is thus obtained to guarantee the effectiveness and efficiency of the inference technique. A novel Gibbs sampler with a backward–forward filter algorithm efficiently learns latent time-evolving parameters in a closed-form. In addition, the latent Indian Buffet Process compound distribution is exploited to automatically infer the overall topic number and customize the sparse topic proportions for each sequential document without bias. The proposed method is evaluated on both synthetic and real-world datasets against the competitive baselines, demonstrating its superiority over the baselines in terms of the low per-word perplexity, high coherent topics, and better document time prediction.

In this paper, we present a novel unsupervised feature learning architecture, which consists of a multi-clustering integration module and a variant of RBM termed multi-clustering integration RBM (MIRBM). In the multi-clustering integration module, we apply three clusterers (K-means, affinity propagation and spectral clustering algorithms) to obtain three different clustering partitions (CPs) without any background knowledge or label. Then, an unanimous voting strategy is used to generate a local clustering partition (LCP). The novel MIRBM model is a core feature encoding part of the proposed unsupervised feature learning architecture. The novelty of it is that the LCP as an unsupervised guidance is integrated into one step contrastive divergence (${\mathtt{{CD}}}_{1}$) learning to guide the distribution of the hidden layer features. For the instance in the same LCP cluster, the hidden and reconstructed hidden layer features of the MIRBM model in the proposed architecture tend to constrict together in the training process. Meanwhile, each LCP center tends to disperse from each other as much as possible in the hidden and reconstructed hidden layer during training. The experiments demonstrate that the proposed unsupervised feature learning architecture has more powerful feature representation and generalization capability than the state-of-the-art models for clustering tasks in the Microsoft Research Asia Multimedia (MSRA-MM)2.0 dataset.

With the popularity of location-aware devices (e.g., smart phones), large amounts of location-based social media data such as check-ins are generated. This stimulates plenty of studies for POI recommendations by applying machine learning techniques. However, most of the existing studies focus on POI recommendations in the same city or region, and fail to recommend POIs for users when they travel to a new city. In this paper, we propose a novel deep neural network, named as ST-TransRec, for crossing-city POI recommendations. It integrates the deep neural network, transfer learning technique, and density-based resampling method into a unified framework. In this model, the deep neural network is used to capture users’ preferences for POIs and learn the embeddings of POIs. Besides, the transfer learning technique is employed to bridge the gap between cities that results from the city-dependent features. As the distributions over POIs are imbalanced, we design a density-based spatial resampling model which enables POIs to be well matched across cities. We conduct extensive experiments on two real-world datasets. The experimental results show the advantages of ST-TransRec over the state-of-the-art methods for crossing-city POI recommendations.

Network embedding (NE) aims to encode the relations of vertices into a low-dimensional space. After NE, we can obtain the learned vectors of vertices that preserve the proximity of network structures for subsequent applications, e.g., vertex classification and link prediction. In existing NE models, they usually exploit the skip-gram with a negative sampling method to optimize their objective functions. Generally, this method learns the vertex representation only from the local connectivity of vertices (i.e., neighbors). However, there is a larger scope of vertex connectivity in real-world scenarios: a vertex may have multifaceted aspects and should belong to overlapping communities. Taking a social network as the overlapping example, a user may subscribe to the channels of politics, economy, and sports simultaneously, but the politics share more common attributes with the economy and less with the sports. In this article, we propose an adversarial learning approach (ACNE) for modeling overlapping communities of vertices. Specifically, we map the association between communities and vertices into an embedding space. Moreover, we take further research on enhancing our ACNE with the following two operations. First, in the initialization stage, we adopt a walking strategy with perception to obtain paths containing more possible boundary vertices to improve overlapping community detection. Then, after representation learning with ACNE, we use soft community assignments from a simple classifier as supervision to update the weights of ACNE. This self-training mechanism referred to as ACNE-ST can help ACNE to achieve better performance. Experimental results demonstrate that the proposed methods, including ACNE and ACNE-ST, can outperform the state-of-the-art models on the subsequent tasks of vertex classification and overlapping community detection.

Network representation learning (NRL) has shown its effectiveness in many tasks, such as vertex classification, link prediction, and community detection. In many applications, vertices of social networks contain textual information, e.g., citation networks, which form a text corpus and can be applied to the typical representation learning methods. The global context in the text corpus can be utilized by topic models to discover the topic structures of vertices. Nevertheless, most existing NRL approaches focus on learning representations from the local neighbors of vertices and ignore the global structure of the associated textual information in networks. In this article, we propose a unified model based on matrix factorization (MF), named collaborative representation learning (CRL), which: 1) considers complementary global and local information simultaneously and 2) models topics and learns network embeddings collaboratively. Moreover, we incorporate the Fletcher-Reeves (FR) MF, a conjugate gradient method, to optimize the embedding matrices in an alternative mode. We call this parameter learning method as AFR in our work that can achieve convergence after a few numbers of iterations. Also, by evaluating CRL on topic coherence and vertex classification using several real-world data sets, our experimental study shows that this collaborative model not only can improve the performance of topic discovery over the baseline topic models but also can learn better network representations than the state-of-the-art context-aware NRL models.

Network representation learning (NRL) has far-reaching effects on data mining research, showing its importance in many real-world applications. NRL, also known as network embedding, aims at preserving graph structures in a low-dimensional space. These learned representations can be used for subsequent machine learning tasks, such as vertex classification, link prediction, and data visualization. Recently, graph convolutional network (GCN)-based models, e.g., GraphSAGE, have drawn a lot of attention for their success in inductive NRL. When conducting unsupervised learning on large-scale graphs, some of these models employ negative sampling (NS) for optimization, which encourages a target vertex to be close to its neighbors while being far from its negative samples. However, NS draws negative vertices through a random pattern or based on the degrees of vertices. Thus, the generated samples could be either highly relevant or completely unrelated to the target vertex. Moreover, as the training goes, the gradient of NS objective calculated with the inner product of the unrelated negative samples and the target vertex may become zero, which will lead to learning inferior representations. To address these problems, we propose an adversarial training method tailored for unsupervised inductive NRL on large networks. For efficiently keeping track of high-quality negative samples, we design a caching scheme with sampling and updating strategies that has a wide exploration of vertex proximity while considering training costs. Besides, the proposed method is adaptive to various existing GCN-based models without significantly complicating their optimization process. Extensive experiments show that our proposed method can achieve better performance compared with the state-of-the-art models.

Distantly supervised relation extraction is widely used in the construction of knowledge bases due to its high efficiency.However, the automatically obtained instances are of low quality with numerous irrelevant words.In addition, the strong assumption of distant supervision leads to the existence of noisy sentences in the sentence bags.In this paper, we propose a novel Multi-Layer Revision Network (MLRN) which alleviates the effects of wordlevel noise by emphasizing inner-sentence correlations before extracting relevant information within sentences.Then, we devise a balanced and noise-resistant Confidence-based Multi-Instance Learning (CMIL) method to filter out noisy sentences as well as assign proper weights to relevant ones.Extensive experiments on two New York Times (NYT) datasets demonstrate that our approach achieves significant improvements over the baselines.

Node classification is a central task in graph data analysis. Scarce or even no labeled data of emerging classes is a big challenge for existing methods. A natural question arises: can we classify the nodes from those classes that have never been seen?

Natural language processing technologies, such as topic models, have been proven to be effective for scholarly recommendation tasks with the ability to deal with content information. Recently, venue recommendation is becoming an increasingly important research task due to the unprecedented number of publication venues. However, traditional methods focus on either the author’s local network or author-venue similarity, where the multiple relationships between scholars and venues are overlooked, especially the venue–venue interaction. To solve this problem, we propose an author topic model–enhanced joint graph modeling approach that consists of venue topic modeling, venue-specific topic influence modeling, and scholar preference modeling. We first model the venue topic with Latent Dirichlet Allocation. Then, we model the venue-specific topic influence in an asymmetric and low-dimensional way by considering the topic similarity between venues, the top-influence of venues, and the top-susceptibility of venues. The top-influence characterizes venues’ capacity of exerting topic influence on other venues. The top-susceptibility captures venues’ propensity of being topically influenced by other venues. Extensive experiments on two real-world datasets show that our proposed joint graph modeling approach outperforms the state-of-the-art methods.

The online event discovery in social media based documents is useful, such as for disaster recognition and intervention. However, the diverse events incrementally identified from social media streams remain accumulated, ad hoc, and unstructured. They cannot assist users in digesting the tremendous amount of information and finding their interested events. Further, most of the existing work is challenged by jointly identifying incremental events and dynamically organizing them in an adaptive hierarchy. To address these problems, this article proposes d ynamic and h ierarchical C ategorization M odeling (dhCM) for social media stream. Instead of manually dividing the timeframe, a multimodal event miner exploits a density estimation technique to continuously capture the temporal influence between documents and incrementally identify online events in textual, temporal, and spatial spaces. At the same time, an adaptive categorization hierarchy is formed to automatically organize the documents into proper categories at multiple levels of granularities. In a nonparametric manner, dhCM accommodates the increasing complexity of data streams with automatically growing the categorization hierarchy over adaptive growth. A sequential Monte Carlo algorithm is used for the online inference of the dhCM parameters. Extensive experiments show that dhCM outperforms the state-of-the-art models in terms of term coherence, category abstraction and specialization, hierarchical affinity, and event categorization and discovery accuracy.

With the widely used Internet of Things, 5G, and smart city technologies, we are able to acquire a variety of vehicle trajectory data. These trajectory data are of great significance which can be used to extract relevant information in order to, for instance, calculate the optimal path from one position to another, detect abnormal behavior, monitor the traffic flow in a city, and predict the next position of an object. One of the key technology is to cluster vehicle trajectory. However, existing methods mainly rely on manually designed metrics which may lead to biased results. Meanwhile, the large scale of vehicle trajectory data has become a challenge because calculating these manually designed metrics will cost more time and space. To address these challenges, we propose to employ network representation learning to achieve accurate vehicle trajectory clustering. Specifically, we first construct the k-nearest neighbor-based internet of vehicles in a dynamic manner. Then we learn the low-dimensional representations of vehicles by performing dynamic network representation learning on the constructed network. Finally, using the learned vehicle vectors, vehicle trajectories are clustered with machine learning methods. Experimental results on the real-word dataset show that our method achieves the best performance compared against baseline methods.

With the advancement of communication technology and location acquisition technology in the context of modern smart cities, the sharing bike systems offer users the great autonomy and convenience for the last/first-kilometer trip. Meanwhile, we can now able to collect, store, and analyze a large amount of sharing bike data. How to effectively use these massive data to provide better services is an emerging task. However, due to the skewed and imbalanced bike usages for stations located at different places, it is of great significance yet very challenging to predict the potential destinations of each individual trip beforehand so that the service providers can better schedule manual bike re-dispatch in advance. To address this issue, this paper proposes an attention-based deep learning framework for trip destination prediction (AFTER). AFTER first learns the low-dimension representations of users and sharing bike stations via negative sampling strategies. Then, a convolution neural network with an attention mechanism is utilized to predict the future trip destination. Experimental results on a real-world dataset indicate that the proposed framework outperforms several state-of-the-art approaches in terms of precision, recall, and F1.

While scientific collaboration is critical for a scholar, some collaborators can be more significant than others, e.g., lifetime collaborators. It has been shown that lifetime collaborators are more influential on a scholar’s academic performance. However, little research has been done on investigating predicting such special relationships in academic networks. To this end, we propose Scholar2vec, a novel neural network embedding for representing scholar profiles. First, our approach creates scholars’ research interest vector from textual information, such as demographics, research, and influence. After bridging research interests with a collaboration network, vector representations of scholars can be gained with graph learning. Meanwhile, since scholars are occupied with various attributes, we propose to incorporate four types of scholar attributes for learning scholar vectors. Finally, the early-stage similarity sequence based on Scholar2vec is used to predict lifetime collaborators with machine learning methods. Extensive experiments on two real-world datasets show that Scholar2vec outperforms state-of-the-art methods in lifetime collaborator prediction. Our work presents a new way to measure the similarity between two scholars by vector representation, which tackles the knowledge between network embedding and academic relationship mining.

Mining topics on social media (e.g., Twitter and Facebook) is an important task for various applications, such as hot topic discovery, advertising, and promotion activities. Topic modeling techniques are helpful to find out topics that people are talking about. However, current full-analysis models cannot perform well on a focused analysis task-find out all topics related to one particular area in short documents. One reason is that the targeted topic is usually sparse in the corpus of short texts. Another one is, during clustering, even minor errors may compound and render the model useless. This article studies these problems and proposes a targeted analysis model (TAM) with reinforcement learning (RL) to extract any specific topic in a given corpus and perform fine-grained topic generation. In this work, we design a reward function of RL to prevent the false propagation problem induced by Gibbs sampling during the clustering. We amend the targeted topic modeling techniques to the case of RL and use policy search combined with the Gibbs EM algorithm for parameter estimation. Metrics of F1 score and the proposed normalized mutual information-F1 are exploited for the evaluation of clustering and topic generation, respectively. Our experiments have demonstrated that TAM can outperform state-of-the-art models-specifically achieving 25.7% improvement on the F1 score for binary clustering on average.

With the advancement of Cyber Physic Systems and Social Internet of Things, the tourism industry is facing challenges and opportunities. We can now able to collect, store, and analyze large amounts of travel data. With the help of data science and artificial intelligence, smart tourism enables tourists with great autonomy and convenience for an intelligent trip. It is of great significance to make full use of these massive data to provide better services for smart tourism. However, due to the skewed and imbalanced visiting for point of interest located at different places, it is of great significance to predict the tourist flow of each place, which can help the service providers for designing a better schedule visiting strategy in advance. Against this background, this article proposes a multi-graph convolutional network framework, named AMOUNT, for tourist flow prediction. To capture the diverse relationships among POIs, AMOUNT first constructs three subgraphs, including the geographical graph, interaction graph, and the co-relation graph. Then, a multi-graph convolution network is utilized to predict the future tourist flow. Experimental results on two real-world datasets indicate that the proposed AMOUNT model outperforms all other baseline tourist flow prediction approaches.

Finding both efficient and effective quantitative representations for scholars in scientific digital libraries has been a focal point of research. The unprecedented amounts of scholarly datasets, combined with contemporary machine learning and big data techniques, have enabled intelligent and automatic profiling of scholars from this vast and ever-increasing pool of scholarly data. Meanwhile, recent advance in network embedding techniques enables us to mitigate the challenges of large scale and sparsity of academic collaboration networks. In real-world academic social networks, scholars are accompanied with various attributes or features, such as co-authorship and publication records, which result in attributed collaboration networks. It has been observed that both network topology and scholar attributes are important in academic relationship mining. However, previous studies mainly focus on network topology, whereas scholar attributes are overlooked. Moreover, the influence of different scholar attributes are unclear. To bridge this gap, in this work, we present a novel framework of Attributed Collaboration Network Embedding (ACNE) for academic relationship mining. ACNE extracts four types of scholar attributes based on the proposed scholar profiling model, including demographics, research, influence, and sociability. ACNE can learn a low-dimensional representation of scholars considering both scholar attributes and network topology simultaneously. We demonstrate the effectiveness and potentials of ACNE in academic relationship mining by performing collaborator recommendation on two real-world datasets and the contribution and importance of each scholar attribute on scientific collaborator recommendation is investigated. Our work may shed light on academic relationship mining by taking advantage of attributed collaboration network embedding.

Recently, dynamic social network research has attracted a great amount of attention, especially in the area of anomaly analysis that analyzes the anomalous change in the evolution of dynamic social networks. However, most of the current research focused on anomaly analysis of the macro representation of dynamic social networks and failed to analyze the nodes that have anomalous structural changes at a micro level. To identify and evaluate anomalous structural change-based nodes in generalized dynamic social networks that only have limited structural information, this research considers undirected and unweighted graphs and develops a multiple-neighbor superposition similarity method ( ), which mainly consists of a multiple-neighbor range algorithm ( ) and a superposition similarity fluctuation algorithm ( ). introduces observation nodes, characterizes the structural similarities of nodes within multiple-neighbor ranges, and proposes a new multiple-neighbor similarity index on the basis of extensional similarity indices. Subsequently, maximally reflects the structural change of each node, using a new superposition similarity fluctuation index from the perspective of diverse multiple-neighbor similarities. As a result, based on and , not only identifies anomalous structural change-based nodes by detecting the anomalous structural changes of nodes but also evaluates their anomalous degrees by quantifying these changes. Results obtained by comparing with state-of-the-art methods via extensive experiments show that can accurately identify anomalous structural change-based nodes and evaluate their anomalous degrees well.

Many achievements have been made by studying how to visualize large-scale and high-dimensional data in typically 2D or 3D space. Normally, such a process is performed through a non-parametric (unsupervised) approach which is limited in handling the unseen data. In this work, we study the parametric (supervised) model which is capable to learn a mapping between high-dimensional data space Rd and low-dimensional latent space Rs with similarity structure in Rd preserved where s l d. The GNNVis is proposed, a framework that applies the idea of Graph Neural Networks (GNNs) to the parametric learning process and the learned mapping serves as a Visualizer (Vis) to compute the low-dimensional embeddings of unseen data online. In our framework, the features of data nodes, as well as the (hidden) information of their neighbors are fused to conduct Dimension Reduction. To the best of our knowledge, none of the existing visualization works have studied how to combine such information into the learning representation. Moreover, the learning process of GNNVis is designed as an end-to-end manner and can easily be extended to arbitrary Dimension Reduction methods if the corresponding objective function is given. Based on GNNVis, several typical dimension reduction methods t-SNE, LargeVis, and UMAP are investigated. As a parametric framework, GNNVis is an inherently efficient Visualizer capable of computing the embeddings of large-scale unseen data. To guarantee its scalability in the Training Stage, a novel training strategy with Subgraph Negative Sampling (SNS) is conducted to reduce the corresponding cost. Experimental results in real datasets demonstrate the advantages of GNNVis. The visualization quality of GNNVis outperforms the state-of-the-art parametric models, and is comparable to that of the non-parametric models.

Network Embedding (NE) aims at modeling network graph by encoding vertices and edges into a low-dimensional space. These learned vectors which preserve proximities can be used for subsequent applications, such as vertex classification and link prediction. Skip-gram with negative sampling is the most widely used method for existing NE models to approximate their objective functions. However, this method only focuses on learning representation from the local connectivity of vertices (i.e., neighbors). In real-world scenarios, a vertex may have multifaceted aspects and should belong to overlapping communities. For example, in a social network, a user may subscribe to political, economic and sports channels simultaneously, but the politics share more common attributes with the economy and less with the sports. In this paper, we propose an adversarial learning approach for modeling overlapping communities of vertices. Each community and vertex are mapped into an embedding space, while we also learn the association between each pair of community and vertex. The experimental results show that our proposed model not only can outperform the state-of-the-art (including GANs-based) models on vertex classification tasks but also can achieve superior performances on overlapping community detection.

In this paper, we study an information exchange process in which a network of individuals exchanges a binary opinion.In the process, the individuals change their opinions only if a majority of their neighbours have the opposite opinion and they do it synchronously.Motivated by applications in multiagent systems, distributed computing, and social science, our goal is to derive graphtheoretic features of the network that guarantee whenever a majority of individuals initially have the same opinion, they will eventually spread the opinion to all individuals.We tackle the problem by first introducing a graph-theoretic notion called controlling set which is capable of characterising the information exchange process and, by exploiting the notion, we obtain a series of lower and upper bounds on the in-degree of vertices as well as lower bound on the size of certain neighbourhoods for guaranteeing the majority to unanimity behaviour.

With the development of Internet of Things technology, the will to make cities smarter is growing. In the context of smart cities, where people are surrounded by a tremendous number of points of interests (POIs), POI recommendation is of great significance. The massive amount of user check-in data collected by location-based social networks (LBSNs) can help users explore new places via POI recommendations. With rich auxiliary data becomeing available in LBSNs, purely exploiting users’ check-in information for POI recommendation is not sufficient. Although several attempts have been done to employ geographical influence to enhance POI recommendation, they simply use a certain distribution function to measure the geographical influence between POIs, which may lead to biased results. To this end, this article proposes a geography-aware inductive matrix completion (GAIMC) approach for personalized POI recommendation. Specifically, the GAIMC model consists of two parts, including geographic feature extraction via a Gaussian mixture model (GMM) and inductive matrix completion for recommendation. The GAIMC model first captures the geographical influence among users and POIs by the GMM, which can mine clustering information with hierarchical structures based on the users’ check-in data and POIs’ location information. Then, a matrix-completion method called inductive matrix completion, which can incorporate geographical features with the user-POI association metric, is utilized to recommend POIs. The experimental results on two real-world LBSN data sets demonstrate that our proposed model can achieve the best recommendation performance in comparison with the state-of-the-art counterparts.

Predicting the user's trajectory behavior sequence based on point of interests (POIs) recommendation is of great significance in the realization of the smart city with the emerging of social Internet of Things technology. One of the widely adopted frameworks is the user-based collaborative filtering, where the explicit POI rating is calculated based on similar users’ preference. However, the trust between users is seldom considered. We believe that if two users show similar preferences or personality traits, the trust level between them should be high. To this end, we propose to calculate the trust-enhanced user similarity in user-based collaborative filtering based on network representation learning. Meanwhile, due to the significance of geographic influence and temporal influence, we integrate these two factors into POI recommendation by a fusion model. Therefore, our proposed POI recommendation system is unified collaborative recommendation framework, which fuses trust-enhanced users’ preferences to potential POIs with geographic influences and temporal influence for POI recommendation. Finally, we conduct extensive experiments on two real-world datasets by comparing with several state-of-the-art methods in terms of precision@k and recall@k. Experimental results indicate that our proposed trust-enhanced collaborative filtering method outperforms other recommendation approaches.

In this article, we present a novel solution to automating the procurement of food materials by using electronic scales, which can automatically identify the food materials along weighing them. Although the CNN model is regarded as one of the most effective solutions to image recognition, the traditional techniques cannot handle the mismatch problem between the lab training data and the real world data. To solve the problem, we propose to embed a partial-and-imbalanced domain adaptation technique (tree adaptation network) in the deep learning model, which can borrow knowledge from sibling classes, to overcome the imbalance problem, and transfer knowledge from the source domain to the target domain, to fight the mismatch problem between the lab training data and the real world data. Experiments show that the proposed approach outperforms state-of-the-art algorithms. Furthermore, the proposed techniques have already been used in practice.

With the popularity of location-aware devices (e.g., smart phones), large amounts of location-based social media data (i.e., user check-in data) are generated, which stimulate plenty of works on personalized point of interest (POI) recommendations using machine learning techniques. However, most of the existing works could not recommend POIs in a new city to a user where the user and his/her friends have never visited before. In this paper, we propose a common topic transfer learning graphical model-the common-topic transfer learning model (CTLM)-for crossing-city POI recommendations. The proposed model separates the city-specific topics (or features) of each city from the common topics (or features) shared by all cities, to enable the users' real interests in the source city to be transferred to the target city. By doing so, the ill-matching problem between users and POIs from different cities can be well addressed by preventing the real interests of users from being influenced by the city-specific features. Furthermore, we incorporate the spatial influence into our proposed model by introducing the regions' accessibility. As a result, the co-occurrence patterns of users and POIs are modeled as the aggregated result from these factors. To evaluate the performance of the CTLM, we conduct extensive experiments on Foursquare and Twitter datasets, and the experimental results show the advantages of CTLM over the state-of-the-art methods for the crossing-city POI recommendations.

Ride-sharing (RS) has great values in saving energy and alleviating traffic pressure. In this paper, we propose a new ride-sharing model, where each driver requires that the shared route percentage (SRP, the ratio of the shared route's distance to the driver's total traveled distance) exceeds her expected rate (e.g., 0.8) when sharing with a rider. We consider two variants of this problem. The first considers multiple drivers and multiple riders, and aims to compute a set of driver-rider pairs to maximize the overall SRP. We model this problem as the maximum weighted bigraph matching problem. We propose an effective exact algorithm, and an efficient approximate solution with error-bound guarantee. The second considers multiple drivers and a single rider and aims to find the top-k drivers for the rider with the largest SRP. We devise pruning techniques and propose a best-first algorithm to progressively selects drivers with high probability to be in the top-k results.

Mining the most influential location set finds$k$locations, traversed by the maximum number of unique trajectories, in a given spatial region. These influential locations are valuable for resource allocation applications, such as selecting charging stations for electric automobiles and suggesting locations for placing billboards. This problem is NP-hard and usually calls for an interactive mining processes involving a user's input, e.g., changing the spatial region and$k$, or removing some locations (from the results in the previous round) that are not eligible for an application according to the domain knowledge. Efficiency is the major concern in conducting this human-in-the-loop mining. To this end, we propose a complete mining framework, which includes an optimal method for the light setting (i.e., small region and$k$) and an approximate method for the heavy setting (i.e., large region and$k$). The optimal method leveragesvertex groupingandbest-first pruningtechniques to expedite the mining process. The approximate method can provide the performance guarantee by utilizing the greedy heuristic, and it is comprised ofefficient updating strategy,index partitionandworkload-based optimizationtechniques. We evaluate the efficiency and effectiveness of our methods based on two taxi datasets from China, and one check-in dataset from New York.

Ride-sharing (RS) has great values in saving energy and alleviating traffic pressure. Existing studies can be improved for better efficiency. Therefore, we propose a new ride-sharing model, where each driver has a requirement that if the driver shares a ride with a rider, the shared route percentage (i.e., the ratio of the shared route's distance to the driver's total travel distance) exceeds an expectation rate of the driver, e.g., 0.8. We consider two variants of this problem. The first considers multiple drivers and multiple riders and aims to compute driver-rider pairs to maximize the overall shared route percentage (SRP). We model this problem as the maximum weighted bigraph matching problem, where the vertices are drivers and riders, edges are driver-rider pairs, and edge weights are driver-rider's SRP. However, it is rather expensive to compute the SRP values for large numbers of driver-rider pairs on road networks. To address this problem, we propose an efficient method to prune many unnecessary driver-rider pairs and avoid computing the SRP values for every pair. To improve the efficiency, we propose an approximate method with error bound guarantee. The basic idea is that we compute an upper bound and a lower bound for each driver-rider pair in constant time. Then, we estimate an upper bound and a lower bound of the graph matching. Next, we select some driver-rider pairs, compute their real shortest-route distance, and update the lower and upper bounds of the maximum graph matching. We repeat above steps until the ratio of the upper bound to the lower bound is not larger than a given approximate rate. The second considers multiple drivers and a single rider and aims to find the top-$k$drivers for the rider with the largest SRP. We first prune a large number of drivers that cannot meet the SRP requirements. Then, we propose a best-first algorithm that progressively selects the drivers with high probability to be in the top-$k$results and prunes the drivers that cannot be in the top-$k$results. Extensive experiments on real-world datasets demonstrate the superiority of our method.

Malicious phone call is a plague, in which unscrupulous salesmen or criminals make to acquire money illegally from the victims. As a result, there has been broad interest in deveploing systems to make the end-users vigilant when receiving such phone calls. Typically, these systems justify the phone numbers either by the crowd-generated blacklist or exploiting the features via machine learning techniques. However, the former is frail due to the rare and lazy crowd, while the later suffers from the scarcity of effective features. In this work, we propose a solution named DeMalC to address those problems by applying the machine learning algorithmm on a novel set of discriminative features. These features consist of properties and behaviors that are powerful enough to characterize phone numbers from different perspectives. We extensively evaluated our solution, i.e., DeMalC, using massive call detail records. The experimental result shows the effectiveness of our extracted features. Capable of achieving 91.86% overall accuracy and 79.34% F1-score on the detection of malicious phone numbers, the DeMalC has been deployed online and demonstrated to be a competitive solution for detecting malicious calls.

20th International Conference on Extending Database Technology, EDBT 2017, Venice, Italy, 21-24 March 2017

In this paper, we propose a novel online event discovery model DP-density to capture various events from the social media data. The proposed model can flexibly accommodate the incremental arriving of the social documents in an online manner by leveraging Dirichlet Process, and a density based technique is exploited to deduce the temporal dynamics of events. The spatial patterns of events are also incorporated in the model by a mixture of Gaussians. To remove the bias caused by the streaming process of the documents, Sequential Monte Carlo is used for the parameter inference. Our extensive experiments over two different real datasets show that the proposed model is capable to extract interpretable events effectively in terms of perplexity and coherence.

Point of Interests (POI) identification using social media data (e.g. Flickr, Microblog) is one of the most popular research topics in recent years. However, there exist large amounts of noises (POI irrelevant data) in such crowd-contributed collections. Traditional solutions to this problem is to set a global density threshold and remove the data point as noise if its density is lower than the threshold. However, the density values vary significantly among POIs. As the result, some POIs with relatively lower density could not be identified. To solve the problem, we propose a technique based on the local drastic changes of the data density. First we define the local maxima of the density function as the Urban POIs, and the gradient ascent algorithm is exploited to assign data points into different clusters. To remove noises, we incorporate the Laplacian Zero-Crossing points along the gradient ascent process as the boundaries of the POI. Points located outside the POI region are regarded as noises. Then the technique is extended into the geographical and textual joint space so that it can make use of the heterogeneous features of social media. The experimental results show the significance of the proposed approach in removing noises.

Crowdsourced query processing is an emerging processing technique that tackles computationally challenging problems by human intelligence. The basic idea is to decompose a computationally challenging problem into a set of human friendly microtasks (e.g., pairwise comparisons) that are distributed to and answered by the crowd. The solution of the problem is then computed (e.g., by aggregation) based on the crowdsourced answers to the microtasks. In this work, we attempt to revisit the crowdsourced processing of the top-k queries, aiming at (1) securing the quality of crowdsourced comparisons by a certain confidence level and (2) minimizing the total monetary cost. To secure the quality of each paired comparison, we employ two statistical tools, Student's t-distribution estimation and Stein's estimation, to estimate the confidence interval of the underlying mean value, which is then used to draw a conclusion to the comparison. Based on the pairwise comparison process, we attempt to minimize the monetary cost of the top-k processing within a Select-Partition-Rank framework. Our experiments, conducted on four real datasets, demonstrate that our stochastic method outperforms other existing top-k processing techniques by a visible difference.

Ranking techniques have been widely used in ubiquitous applications like recommendation, information retrieval, etc. For ranking computation hostile but human friendly items, crowdsourcing is considered as an emerging technique to process the ranking by human power. However, there is a lack of an easy-to-use toolkit for answering crowdsourced top- k query with minimal effort. In this work, we demonstrate an interactive programming toolkit that is a unified solution for answering the crowd-sourced top- k queries. The toolkit employs a new confidence-aware crowdsourced top- k algorithm, SPR. The whole progress of the algorithm is monitored and visualized to end users in a timely manner. Besides the visualized result and the statistics, the system also reports the estimation of the monetary cost and the breakdown of each phase. Based on the estimation, end users can strike a balance between the budget and the quality through the interface of this toolkit.

SimRank is a widely adopted similarity measure for objects modeled as nodes in a graph, based on the intuition that two objects are similar if they are referenced by similar objects. The recursive nature of SimRank definition makes it expensive to compute the similarity score even for a single pair of nodes. This defect limits the applications of SimRank. To speed up the computation, some existing works replace the original model with an approximate model to seek only rough solution of SimRank scores. In this work, we propose a novel solution for computing all-pair SimRank scores. In particular, we propose to convert SimRank to the problem of solving a linear system in matrix form, and further prove that the system is non-singular, diagonally dominate, and symmetric definite positive (for undirected graphs). Those features immediately lead to the adoption of Conjugate Gradient (CG) and Bi-Conjugate Gradient (BiCG) techniques for efficiently computing SimRank scores. As a result, a significant improvement on the convergence rate can be achieved; meanwhile, the sparsity of the adjacency matrix is not damaged all the time. Inspired by the existing common neighbor sharing strategy, we further reduce the computational complexity of the matrix multiplication and resolve the scalable issues. The experimental results show our proposed algorithms significantly outperform the state-of-the-art algorithms.

In this work, a three-stage social event detection (SED) framework is proposed to discover events from Flickr-like data. First, multiple bipartite graphs are constructed for the heterogeneous feature modalities to achieve fused features. Furthermore, considering the geometrical structures of dictionary and data, a dual structure constrained multimodal feature coding model is designed to learn discriminative feature codes by incorporating corresponding regularization terms into the objective. Finally, clustering models utilizing density or label knowledge and data recovery residual models are devised to discover real-world events. The proposed SED approach achieves the highest performance on the MediaEval 2014 SED dataset.

The availability of geographical and temporal tagged documents enables many location and time based mining tasks. Event discovery is one of such tasks, which is to identify interesting happenings in the geographical and temporal space. In recent years, several techniques have been proposed. However, no existing work has provided a nonparametric algorithm for detecting events in the joint space crossing geographical and temporal dimensions. Furthermore, though some prior works proposed to capture the periodicities of topics in their solutions, some restrictions on the temporal patterns are often placed and they usually ignore the spatial patterns of the topics. To break through such limitations, in this paper we propose a novel nonparametric model to identify events in the geographical and temporal space, where any recurrent patterns of events can be automatically captured. In our approach, parameters are automatically determined by exploiting a Dirichlet Process. To reduce the influence from noisy terms in the detection, we distinguish its event role from its background role using a Bernoulli model in the solution. Experimental results on three real world datasets show the proposed algorithm outperforms previous state-of-the-art approaches.

Mining the most influential k-location set finds k locations, traversed by the maximum number of unique trajectories, in a given spatial region. These influential locations are valuable for resource allocation applications, such as selecting charging stations for electric automobiles and suggesting locations for placing billboards. This problem is NP-hard and usually calls for an interactive mining processes, e.g., changing the spatial region and k, or removing some locations (from the results in the previous round) that are not eligible for an application according to the domain knowledge. Thus, efficiency is the major concern in addressing this problem. In this paper, we propose a system by using greedy heuristics to expedite the mining process. The greedy heuristic is efficient with performance guarantee. We evaluate the performance of our proposed system based on a taxi dataset of Tianjin, and provide a case study on selecting the locations for charging stations in Beijing.

With the rapid development of mobile Internet technology, Internet users are shifting from desktop to mobile devices. Modern mobile devices (e.g., smartphones and tablets) are equipped with GPS, which can help users to easily obtain their locations, and location-based services (LBS) have been widely deployed. LBS users are generating more and more spatio-textual data which contains both textual descriptions and geographical locations. In user-generated data, a spatiotextual entity may have different representations, possibly due to GPS deviations or typographical errors [6], [2], and it calls for effective methods to integrate the spatio-textual data from different data sources. A spatio-textual similarity join is an important operation in spatio-textual data integration, which, given two sets of spatio-textual objects, finds all similar pairs from the two sets, where the similarity can be quantified by combining spatial proximity and textual relevancy. There are many applications in spatio-textual similarity joins, e.g., user recommendation in location-based social networks, image duplication detection using spatio-textual tags, spatio-textual advertising, and location-based market analysis [6], [2]. For example, a house rental agency (e.g., rent.com) wants to perform a similarity join on the spatio-textual data of house requirements from renters and the data of house properties from owners. For another example, a startup company, e.g., Factual (factual.com), crawls spatio-textual records to generate points of interest (POIs). As the records are from multiple sources and may contain many duplicates, It needs to run similarity joins to remove the duplicates.

To the best of our knowledge, this is the first work that offers very low latency (0.1ms) per kNN query on 10-million-node networks on a commodity machine. This translates to a query throughput of 10,000 queries per second per commodity machine. Our experimental studies on large scale road networks show that our solutions are 1–3 orders of magnitudes faster than existing methods while our indexes are compact and can fit into main memory. In the future, we plan to further enhance the performance of kNN search with keywords.

With rapidly growing popularity, Location Based Services (LBS), e.g., Google Maps, Yahoo Local, WeChat, FourSquare, etc., started offering web-based search features that resemble a kNN query interface. Specifically, for a user-specified query location q, these websites extract from the objects in their backend database the top-k nearest neighbors to q and return these k objects to the user through the web interface. Here k is often a small value like 50 or 100. For example, McDonald [1] returns the top 25 nearest restaurants for a user-specified location through its locations search webpage.

With the development of location-based services (LBS), LBS users are generating more and more spatio-textual data, e.g., checkins and attraction reviews. Since a spatio-textual entity may have different representations, possibly due to GPS deviations or typographical errors, it calls for effective methods to integrate the spatio-textual data from different data sources. In this paper, we study the problem of top-$k$spatio-textual similarity join (Topk-STJoin), which identifies the$k$most similar pairs from two spatio-textual data sets. One big challenge inTopk-STJoinis to efficiently identify the top-$k$similar pairs by considering both textual relevancy and spatial proximity. Traditional join algorithms that consider only one dimension (textual or spatial) are inefficient because they cannot utilize the pruning ability on the other dimension. To address this challenge, we propose a signature-based top-$k$join framework. We first generate a spatio-textual signature set for each object such that if two objects are in the top-$k$similar pairs, their signature sets must overlap. With this property, we can prune large numbers of dissimilar pairs without common signatures. We find that the order of accessing the signatures has a significant effect on the performance. So, we compute an upper bound for each signature and propose a best-first accessing method that preferentially accesses signatures with large upper bounds while those pairs with small upper bounds can be pruned. We prove the optimality of our best-first accessing method. Next, we optimize the spatio-textual signatures and propose progressive signatures to further improve the pruning power. Experimental results on real-world datasets show that our algorithm achieves high performance and good scalability, and significantly outperforms baseline approaches.

Many websites offering Location Based Services (LBS) provide a$k$NN search interface that returns the top-$k$nearest-neighbor objects (e.g., nearest restaurants) for a given query location. This paper addresses the problem of crawling all objects efficiently from an LBS website, through the public$k$NN web search interface it provides. Specifically, we develop crawling algorithm for 2D and higher-dimensional spaces, respectively, and demonstrate through theoretical analysis that the overhead of our algorithms can be bounded by a function of the number of dimensions and the number of crawled objects, regardless of the underlying distributions of the objects. We also extend the algorithms to leverage scenarios where certain auxiliary information about the underlying data distribution, e.g., the population density of an area which is often positively correlated with the density of LBS objects, is available. Extensive experiments on real-world datasets demonstrate the superiority of our algorithms over the state-of-the-art competitors in the literature.

Many online social networks feature restrictive web interfaces that only allow the query of a user’s local neighborhood. To enable analytics over such an online social network through its web interface, many recent efforts use Markov Chain Monte Carlo (MCMC) methods such as random walks to sample users in the social network and thereby support analytics based on the samples. The problem with such an approach, however, is the large amount of queries often required for a random walk to converge to a desired (stationary) sampling distribution. In this article, we consider a novel problem of enabling a faster random walk over online social networks by “rewiring” the social network on-the-fly. Specifically, we develop a Modified TOpology Sampling (MTO-Sampling) scheme that, by using only information exposed by the restrictive web interface, constructs a “virtual” random-walk-friendly overlay topology of the social network while performing a random walk and ensures that the random walk follows the modified overlay topology rather than the original one. We describe in this article instantiations of MTO-Sampling for various types of random walks, such as Simple Random Walk (MTO-SRW), Metropolis-Hastings Random Walk (MTO-MHRW), and General Random Walk (MTO-GRW). We not only rigidly prove that MTO-Sampling improves the efficiency of sampling, but we also demonstrate the significance of such improvement through experiments on real-world online social networks such as Google Plus, Epinion, Facebook, etc.

Emergency medical service provides a variety of services for those in need of emergency care. One of the major challenges encountered by emergency service providers is selecting the appropriate locations for ambulance stations. Prior works measure spatial proximity under Euclidean space or static road network. In this paper, we focus on locating the ambulance stations by using the real traffic information so as to minimize the average travel-time to reach the emergency requests. To this end, we estimate the travel-time of road segments using real GPS trajectories and propose an efficient PAM-based refinement for the location problem. We conduct extensive experimental evaluations using real emergency requests collected from Tianjin, and the result shows that the proposed solution can reduce the travel-time to reach the emergency requests by 29.9% when compared to the original locations of ambulance stations.

Discovering motifs in sequence databases has been receiving abundant attentions from both database and data mining communities, where the motif is the most correlated pair of subsequences in a sequence object. Motif discovery is expensive for emerging applications which may have very long sequences (e.g., million observations per sequence) or the queries arrive rapidly (e.g., per 10 seconds). Prior works cannot offer fast correlation computations and prune subsequence pairs at the same time, as these two techniques require different orderings on examining subsequence pairs. In this work, we propose a novel framework named Quick-Motif which adopts a two-level approach to enable batch pruning at the outer level and enable fast correlation calculation at the inner level. We further propose two optimization techniques for the outer and the inner level. In our experimental study, our method is up to 3 orders of magnitude faster than the state-of-the-art methods.

Peer reviewing is a standard process for assessing the quality of submissions at academic conferences and journals. A very important task in this process is the assignment of reviewers to papers. However, achieving an appropriate assignment is not easy, because all reviewers should have similar load and the subjects of the assigned papers should be consistent with the reviewers' expertise. In this paper, we propose a generalized framework for fair reviewer assignment. We first extract the domain knowledge from the reviewers' published papers and model this knowledge as a set of topics. Then, we perform a group assignment of reviewers to papers, which is a generalization of the classic Reviewer Assignment Problem (RAP), considering the relevance of the papers to topics as weights. We study a special case of the problem, where reviewers are to be found for just one paper (Journal Assignment Problem) and propose an exact algorithm which is fast in practice, as opposed to brute-force solutions. For the general case of having to assign multiple papers, which is too hard to be solved exactly, we propose a greedy algorithm that achieves a 1/2-approximation ratio compared to the exact solution. This is a great improvement compared to the 1/3-approximation solution proposed in previous work for the simpler coverage-based reviewer assignment problem, where there are no weights on topics. We theoretically prove the approximation bound of our solution and experimentally show that it is superior to the current state-of-the-art.

Peer reviewing is a widely accepted mechanism for assessing the quality of submitted articles to scientific conferences or journals. Conference management systems (CMS) are used by conference organizers to invite appropriate reviewers and assign them to submitted papers. Typical CMS rely on paper bids entered by the reviewers and apply simple matching algorithms to compute the paper assignment. In this paper, we demonstrate our Reviewer Assignment System (RAS), which has advanced features compared to broadly used CMSs. First, RAS automatically extracts the profiles of reviewers and submissions in the form of topic vectors. These profiles can be used to automatically assign reviewers to papers without relying on a bidding process, which can be tedious and error-prone. Second, besides supporting classic assignment models (e.g., stable marriage and optimal assignment), RAS includes a recently published assignment model by our research group, which maximizes, for each paper, the coverage of its topics by the profiles of its reviewers. The features of the demonstration include (1) automatic extraction of paper and reviewer profiles, (2) assignment computation by different models, and (3) visualization of the results by different models, in order to assess their effectiveness.

The k nearest neighbor (kNN) search on road networks is an important function in web mapping services. These services are now dealing with rapidly arriving queries, that are issued by a massive amount of users. While overlay graph-based indices can answer shortest path queries efficiently, there have been no studies on utilizing such indices to answer kNN queries efficiently. In this paper, we fill this research gap and present two efficient kNN search solutions on overlay graph-based indices. Experimental results show that our solutions offer very low query latency (0.1 ms) and require only small index sizes, even for 10-million-node networks.

A large number of web data repositories are hidden behind restrictive web interfaces, making it an important challenge to enable data analytics over these hidden web databases. Most existing techniques assume a form-like web interface which consists solely of categorical attributes (or numeric ones that can be discretized). Nonetheless, many real-world web interfaces (of hidden databases) also feature checkbox interfaces-e.g., the specification of a set of desired features, such as A/C, navigation, etc., for a car-search website like Yahoo! Autos. We find that, for the purpose of data analytics, such checkbox-represented attributes differ fundamentally from the categorical/numerical ones that were traditionally studied. In this paper, we address the problem of data analytics over hidden databases with checkbox interfaces. Extensive experiments on both synthetic and real datasets demonstrate the accuracy and efficiency of our proposed algorithms.

Location-based services have been widely adopted in many systems. Existing works employ a pull model or user-initiated model, where a user issues a query to a server which replies with location-aware answers. To provide users with instant replies, a push model or server-initiated model is becoming an inevitable computing model in the next-generation location-based services. In the push model, subscribers register spatio-textual subscriptions to capture their interests, and publishers post spatio-textual messages. This calls for a high-performance location-aware publish/subscribe system to deliver publishers' messages to relevant subscribers. In this paper, we address the research challenges that arise in designing a location-aware publish/subscribe system. We propose an R-tree based index by integrating textual descriptions into R-tree nodes. We devise efficient filtering algorithms and effective pruning techniques to achieve high performance. Our method can support both conjunctive queries and ranking queries. We discuss how to support dynamic updates efficiently. Experimental results show our method achieves high performance which can filter 500 messages in a second for 10 million subscriptions on a commodity computer.

In the recent decades, we have witnessed the rapidly growing popularity of location-based systems. Three types of location-based queries on road networks, single-pair shortest path query, k nearest neighbor (kNN) query, and keyword-based kNN query, are widely used in location-based systems. Inspired by R-tree, we propose a height-balanced and scalable index, namely G-tree, to efficiently support these queries. The space complexity of G-tree is O(|V|log|V|) where |V| is the number of vertices in the road network. Unlike previous works that support these queries separately, G-tree supports all these queries within one framework. The basis for this framework is an assembly-based method to calculate the shortest-path distances between two vertices. Based on the assembly-based method, efficient search algorithms to answer kNN queries and keyword-based kNN queries are developed. Experiment results show G-tree's theoretical and practical superiority over existing methods.

Heterogeneous device services generated by various devices in different contexts prevent users from efficiently and correctly consuming device services. This seriously hinders the development of Internet of Things. This paper addresses the problems appearing in device discovery and device interaction. It devises a user interoperability framework (UIF) to enable device users to interoperate with heterogeneous devices of different contexts with consistent syntax and semantics. In this framework, a new separation strategy is provided; a device representation method for real, common, and virtual devices is devised; and a device transformability model is proposed to guarantee the proper transformation of device syntax and semantics. To demonstrate the correctness of UIF, a UIF prototype is implemented and several experiment methods are compared to determine which one should be adopted as semantic relatedness computing tools in device discovery for device users and in common device publishing for device providers.

Deducing trip-related information from web-scale datasets has received large amounts of attention recently. Identifying points of interest (POIs) in geo-tagged photos is one of these problems. The problem can be viewed as a standard clustering problem of partitioning two-dimensional objects. In this work, we study spectral clustering, which is the first attempt for the identification of POIs. However, there is no unified approach to assigning the subjective clustering parameters, and these parameters vary immensely in different metropolitans and locations. To address this issue, we study a self-tuning technique that can properly determine the parameters for the clustering needed. Besides geographical information, web photos inherently store other rich information. Such heterogenous information can be used to enhance the identification accuracy. Thereby, we study a novel refinement framework that is based on the tightness and cohesion degree of the additional information. We thoroughly demonstrate our findings by web-scale datasets collected from Flickr.

The online shortest path problem aims at computing the shortest path based on live traffic circumstances. This is very important in modern car navigation systems as it helps drivers to make sensible decisions. To our best knowledge, there is no efficient system/solution that can offer affordable costs at both client and server sides for online shortest path computation. Unfortunately, the conventional client-server architecture scales poorly with the number of clients. A promising approach is to let the server collect live traffic information and then broadcast them over radio or wireless network. This approach has excellent scalability with the number of clients. Thus, we develop a new framework called live traffic index (LTI)which enables drivers to quickly and effectively collect the live traffic information on the broadcasting channel. An impressive result is that the driver can compute/update their shortest path result by receiving only a small fraction of the index. Our experimental study shows that LTI is robust to various parameters and it offers relatively short tune-in cost (at client side), fast query response time (at client side), small broadcast size (at server side), and light maintenance time (at server side)for online shortest path problem.

Many online social networks feature restrictive web interfaces which only allow the query of a user's local neighborhood through the interface. To enable analytics over such an online social network through its restrictive web interface, many recent efforts reuse the existing Markov Chain Monte Carlo methods such as random walks to sample the social network and support analytics based on the samples. The problem with such an approach, however, is the large amount of queries often required (i.e., a long “mixing time”) for a random walk to reach a desired (stationary) sampling distribution. In this paper, we consider a novel problem of enabling a faster random walk over online social networks by “rewiring” the social network on-the-fly. Specifically, we develop Modified TOpology (MTO)-Sampler which, by using only information exposed by the restrictive web interface, constructs a “virtual” overlay topology of the social network while performing a random walk, and ensures that the random walk follows the modified overlay topology rather than the original one. We show that MTO-Sampler not only provably enhances the efficiency of sampling, but also achieves significant savings on query cost over real-world online social networks such as Google Plus, Epinion etc.

In this paper, we propose a novel automatic image annotation model by mining the web. In our approach, the terms or words appearing in the associated text are extracted and filtered as labels or annotations for the corresponding web images. Sure, much noise exists in those selected labels. In order to reduce the influence caused by the noisy labels, for each label or potential word, we improve web image-word relationships using Mixture Gaussian Distribution Model. By doing so, the relationships between words and images are re-weighted both in terms of sematic relevance and in terms of visual feature similarity. In fact, all the words associated to an image are not semantically independent. We use co-occurrences between two words to describe their semantic relevance. Thus, we further use a method, called Word Promotion, to co-enhance the weights of all the words associated to a given image based on their co-occurrences. Our experiments are conducted in several ways and the results show that our annotation method can achieve a satisfactory performance in respects of system scalability and sematic evolution.

Most existing work on sequence databases use correlation (e.g., Euclidean distance and Pearson correlation) as a core function for various analytical tasks. Typically, it requires users to set a length for the similarity queries. However, there is no steady way to define the proper length on different application needs. In this work we focus on discovering longest-lasting highly correlated subsequences in sequence databases, which is particularly useful in helping those analyses without prior knowledge about the query length. Surprisingly, there has been limited work on this problem. A baseline solution is to calculate the correlations for every possible subsequence combination. Obviously, the brute force solution is not scalable for large datasets. In this work we study a space-constrained index that gives a tight correlation bound for subsequences of similar length and offset by intra-object grouping and inter-object grouping techniques. To the best of our knowledge, this is the first index to support normalized distance metric of arbitrary length subsequences. Extensive experimental evaluation on both real and synthetic sequence datasets verifies the efficiency and effectiveness of our proposed methods.

In general, users may influence each other by their activities in social networks. It is interesting to explore hidden relationships of users based on their social activities. In this paper, we define and study a novel community detection problem that is to discover the hidden community structure in large social networks based on their common interests. We observe that the users typically pay more attention to those users who share similar interests, which enable a way to partition the users into different communities according to their common interests. We propose two algorithms to detect influential communities using common interests in large social networks efficiently and effectively. We conduct our experimental evaluation using a dataset from Epinions, which demonstrates that our method achieves 4%-11.8% accuracy improvement over the state of the art method.

For the multilingual semantic interoperations in cross-organizational enterprise systems and e-commerce systems, semantic consistency is a research issue that has not been well resolved. This paper contributes to improving multilingual semantic interoperation by proposing a concept-connected near synonym (NSG) framework for concept disambiguation. NSG framework provides a vocabulary preprocessing process of collaborative vocabulary editing, which further ensures semantically consistent vocabulary for building semantically consistent business processes and documents between context-different information systems. The vocabulary preprocessing offered by NSG automates the process of finding potential near synonym sets and identifying collaboratively editable near synonym sets. The realization of NSG framework includes a probability model that computes concept values between concepts based on a newly introduced semantic relatedness method-SRCT. In this paper, SRCT-based methods are implemented and compared with some existing semantic relatedness methods. Experiments have shown that SRCT-based methods outperform the existing methods. This paper has made an improvement on the existing methods of semantic relatedness and reduces the collaboration cost of collaborative vocabulary editing.

An electronic marketplace (e-marketplace) is a common business information space populated with many entities of different system types. Each of them has its own context of how to process activities. This leads to heterogeneous e-marketplace activities, which are difficult to make interoperable and inferred from one entity to another. This study solves this problem by proposing a concept of separation strategy and implementing it through providing a semantic inference engine with a novel inference algorithm. The solution, called the RuleXPM approach, enables one to semantically infer a next e-marketplace activity across multiple contexts/domains. Experiments show that the cross-context/cross-domain semantic inference is achievable. This paper is an understanding of many aspects related to heterogeneous activity inference.

Deducing trip related information from web-scale datasets has received very large amounts of attention recently. Identifying points of interest (POIs) in geo-tagged photos is one of these problems. The problem can be viewed as a standard clustering problem of partitioning two dimensional objects. In this work, we study spectral clustering which is the first attempt for the POIs identification. However, there is no unified approach to assign the clustering parameters; especially the features of POIs are immensely varying in different metropolitans and locations. To address this, we are intent to study a self-tuning technique which can properly assign the parameters for the clustering needed. Besides geographical information, web photos inherently store rich information. These information are mutually influenced each others and should be taken into trip related mining tasks. To address this, we study reinforcement which constructs the relationship over multiple sources by iterative learning. At last, we thoroughly demonstrate our findings by web scale datasets collected from Flickr.

In this paper, we provide a novel image annotation model by mining the Web. In our approach, the concepts or words appearing in the associated text are extracted and filtered as the semantic annotations for the corresponding Web images. In order to alleviate the influence caused by the noise images, for each semantic concept, we improve Web image-word relationships using Mixture Gaussian Distribution Model. By doing so, the concepts or words relevant to any image are re-weighed by both considering their relevance to the image in term of text and in term of visual feature. In fact, all the words associated to an image are not semantically independent. We use co-occurrences between two words to describe their semantic relevance. Thus, we further use a method, called Word Promotion, to co-enhance the weights of all the words associated to a given image based on their co-occurrences. Our experiments are conducted in several ways and the results show that our annotation method can achieve a satisfactory performance.

An analysis on the existing e-marketplaces shows there are seven types of technical construction methods for e-marketplaces. They are e-catalogue, e-shop, e-portal, e-hub, e-switch, e-integrator and e-merger. The quality of these e-marketplaces can be measured based on a quality matrix of accuracy, reach and richness.

E-marketplace activities initiated by users in general require representing the user requirements and preferences matched with a set of offerings. One issue of these activities is the heterogeneity among them, which asks for semantic consistency maintenance. This paper solves this problem by applying collaborative concept exchange technology and developing a novel RuleXPM approach. This approach transforms XPM reified documents into rule-based RuleXPM documents that are suitable for making cross-domain inference on heterogeneous e-marketplace activities, using defeasible reasoning on a newly designed RuleXPM inference engine made from a RuleXPM inference algorithm.

As World Wide Web is becoming the most important information repository, increasing amount of information is available. Currently, web search engines can only provide document oriented searches. In order to fully make use of information from the web, some effective and efficient extraction algorithms are definitely desirable. In this paper, some existing achievements are investigated firstly. Then our current technique on web information extraction is discussed in detail. In our approach, rules and patterns are extracted from sample pages through training process, with human involvements. We use both keywords and regular expressions to represent rules and patterns in our system. The keywords work as anchors to locate the positions of the potential information and regular expressions work as validations of the values. In our system, all the extracted information is represented in XML format.