Large language models (LLMs) have transformed recommender systems through strong semantic understanding and generalization. However, the design of item identifiers remains a critical bottleneck that directly affects recommendation quality. Traditional metadata-based identifiers introduce length variability and semantic ambiguity, whereas existing collaborative indexing (CID) approaches often neglect item attributes, show limited cross-dataset generalizability, and incur high computational cost at scale. To address these limitations, we propose a Graph Neural Network (GNN)–based item indexing framework with three coordinated innovations. First, we construct attribute-enriched co-occurrence graphs and use a GNN encoder to fuse item features with collaborative signals, yielding semantically informed representations that work well for attribute-rich catalogs. Second, we replace recursive spectral clustering with hierarchical agglomerative clustering on GNN embeddings, enabling direct control of index length via tree depth and reducing hyperparameter tuning across datasets. Third, we exploit localized message passing rather than global eigendecomposition, which provides considerably better runtime efficiency and is amenable to mini-batch training, supporting online index updates as interactions evolve. Across five benchmarks, GID achieves strong average ranking performance, showing larger improvements on sparse and attribute-rich datasets while remaining competitive in dense settings. The framework is robust under both seen and unseen prompt templates, which supports practical LLM-based recommendation. On sequential recommendation, GID improves HR@10 by 7.9% on average over the strongest baseline in each dataset.

Graph Anomaly Detection (GAD) identifies deviant patterns including anomalous nodes, edges, and subgraphs in graph data, with significant applications in social networks, cybersecurity, and financial risk control. While spectral methods have proven effective for homogeneous graph anomaly detection, their application to heterogeneous graphs remains challenging due to structural complexity and semantic richness. Existing heterogeneous graph anomaly detection methods either rely on manually designed meta-paths or decompose the graph into homogeneous subgraphs, leading to limited flexibility or loss of structural integrity. To address these limitations, we propose the Graph Wavelet Transformer (GWT), a novel spectral-based approach that integrates global graph properties and spectral analysis without requiring meta-path information. GWT employs a three-stage process: heterogeneous-to-homogeneous graph conversion, global dependency modeling via graph transformers, and spectral-aware feature enhancement focusing on frequency band components. Extensive experiments on multiple benchmarks demonstrate that GWT significantly outperforms ten baseline methods, providing a new paradigm for heterogeneous graph anomaly detection that preserves structural completeness while achieving computational efficiency.

Graph-level out-of-distribution (OOD) detection, which attempts to identify OOD graphs originated from an unknown distribution, is a vital building block for safety-critical applications in Web and society. Current approaches concentrate on how to learn better graph representations, but fail to provide any statistically guarantee on detection results, therefore impeding their deployments in the scenario where detection errors would result in serious consequences. To overcome this critical issue, we propose the Conformal Graph-level Out-of-distribution Detection (CGOD), extending the theory of conformal prediction to graph-level OOD detection with a rigorous control over the false positive rate. In CGOD, we develop a new aggregated non-conformity score function based on the proposed adaptive data augmentation. Through the guidance from two designed metrics, i.e., score consistency and representation diversity, our augmentation strategy can generate multiple non-conformity scores, and aggregating these generated non-conformity scores together is robust to the misleading information. Meanwhile, our score function can perceive the subsequent process of conformal inference, enabling the aggregated non-conformity score to be adaptive to different input graphs and deriving a more accurate score estimation. We conduct experiments on multiple real-world datasets with different empirical settings. Extensive results and model analyses demonstrate the superior performance of our approach over several competitive baselines.

Message passing scheme is a general idea for Graph Neural Networks (GNNs) to learn node representations. During message passing, given a target node, we transform and aggregate the feature vectors of its neighbors and generate a representation vector for the target node. However, real-world graph data is usually constructed from complicated scenarios based on manually pre-defined rules; it is often the case that noisy information gets involved in message passing, thereby resulting in sub-optimal performance for GNNs and also impacting their trustworthiness and reliability. In this study, we present an effective learnable threshold technique that explicitly optimizes heterogeneous graph structure with the goal to maximize performance improvement of GNNs for downstream tasks. We give an explanation about the design of the learnable threshold and show the ability that our model can be applied to large-scale graphs. Experiments on seven datasets show that our model has a powerful ability to deal with homogeneous graphs with low homophily ratio and dense graphs. With the verification of robustness analysis, our model can resist the noisy information, which proves the robustness of our model.

Methods based on graph neural networks for solving combinatorial optimization (CO) problems have exhibited promising results in tackling a range of NP-hard problems, eliminating the necessity for reliance on manually created domain knowledge. Existing models including reinforcement learning (RL) framework assume that combinatorial instances in the training set contribute equally during training. Nevertheless, there is considerable variation in the quality of training instances, and the performance of models may suffer from the inclusion of low-quality training instances. This paper expands the current scope of neural solvers for CO problems through the incorporation of curriculum learning (CL). To alleviate the adverse impact of low-quality training instances, we propose CL4CO which utilizes CL strategy, a selective training method, to train models based on the rank of instances' quality in neural Combinatorial Optimization framework. Also, we introduce several candidate topology-aware metrics based on heterophily ratio and evaluation of clustering for the training scheduler. Furthermore, it is noteworthy to emphasize that it has potential to enhance the generalization capacity of RL-based baselines and we give a experimental validation. This enhancement plugin from the fact that CL empowers the acquired RL-based solver to effectively leverage commonly shared features within the same class of CO. Empirically, we conduct a case study on MaxCut, a classical discrete Oil-vector CO, to verify our findings and our results demonstrate that CL4CO is efficient and superiority with good generalization ability.

Decision-focused learning (DFL) combines both machine learning and combinatorial optimization so as to enhance the quality of decision-making. In general, DFL adds an optimization layer after the neural network and solves a focused combinatorial optimization problem. The optimization layer is usually based on KKT conditions or surrogate functions. However, the optimizer associated with the DFL is inflexible and requires a large amount of expert knowledge. Furthermore, it has been shown that the optimizer exhibits a lack of robustness and differentiability when confronted with complex tasks, especially in the graph domain. To solve this problem, we study a more generic situation by taking the optimization layer as a black-box operator on graphs. Then, we design a mixed zeroth-order optimization to differentiate the layer. Specifically, we combine DFL with graph neural networks and present a decision-focused graph neural network named ZO4Graph based on a mixed zeroth-order optimizer. Empirically, we conduct extensive experiments compared with two-stage models on community detection. Numerical results demonstrate that our proposed framework outperforms its peers.

Anomaly detection in multi-type event sequences is a crucial and challenging problem with important applications in various domains, including cybersecurity, finance and healthcare. Temporal point process has emerged as a powerful technique for modeling event sequences and has gained considerable attention in the field of anomaly detection. However, existing temporal point process approaches are either inapplicable to multi-type event sequence data or incur the loss of valuable information in subsequences associated with specific event types. To this end, we propose a novel Multiple Hypothesis Testing based Anomaly Detection method (MultiAD) to detect anomalous multi-type event sequences. The basic idea of MultiAD is to capture the underlying distribution of normal sequences using a neural multivariate point process, based on which the original hypothesis testing problem can be converted into a multiple hypothesis testing using the multivariate time rescaling theorem. By conducting multiple hypothesis tests on the time-rescaled subsequences, MultiAD makes full use of the valuable information contained within individual subsequences. Moreover, we claim that the existing test statistic ignores the sequential information of inter-event time intervals and propose new statistics to address this shortcoming. Finally, we employ the kernel method to obtain a smooth estimator of the distribution of the proposed statistics under the null hypothesis. This ensures a more accurate and reliable computation of the p-value, providing robust statistical inference. Extensive experimental results demonstrate that MultiAD significantly outperforms the state-of-the-art methods on both synthetic and real-world data.

Graph neural networks (GNNs) are powerful models to handle graph data and can achieve state-of-the-art in many critical tasks including node classification and link prediction. However, existing graph neural networks still face both challenges of over-smoothing and over-squashing based on previous literature. To this end, we propose a new Curvature-based topology-aware Dropout sampling technique named CurvDrop, in which we integrate the Discrete Ricci Curvature into graph neural networks to enable more expressive graph models. Also, this work can improve graph neural networks by quantifying connections in graphs and using structural information such as community structures in graphs. As a result, our method can tackle the both challenges of over-smoothing and over-squashing with theoretical justification. Also, numerous experiments on public datasets show the effectiveness and robustness of our proposed method. The code and data are released in https://github.com/liu-yang-maker/Curvature-based-Dropout.

Data augmentation has been widely studied as it can be used to improve the generalizability of graph representation learning models. However, existing works focus only on the data augmentation on homogeneous graphs. Data augmentation for heterogeneous graphs remains under-explored. Considering that heterogeneous graphs contain different types of nodes and links, ignoring the type information and directly applying the data augmentation methods of homogeneous graphs to heterogeneous graphs will lead to suboptimal results. In this paper, we propose a novel Multi-Aspect Heterogeneous Graph Augmentation framework named MAHGA. Specifically, MAHGA consists of two core augmentation strategies: structure-level augmentation and metapath-level augmentation. Structure-level augmentation pays attention to network schema aspect and designs a relation-aware conditional variational auto-encoder that can generate synthetic features of neighbors to augment the nodes and the node types with scarce links. Metapath-level augmentation concentrates on metapath aspect, which constructs metapath reachable graphs for different metapaths and estimates the graphons of them. By sampling and mixing up based on the graphons, MAHGA yields intra-metapath and inter-metapath augmentation. Finally, we conduct extensive experiments on multiple benchmarks to validate the effectiveness of MAHGA. Experimental results demonstrate that our method improves the performances across a set of heterogeneous graph learning models and datasets.

Graph neural networks (GNNs) are popularly used to analyze non-Euclidean graph data. Despite their successes, the design of graph neural networks requires heavy manual work and rich domain knowledge. Recently, neural architecture search algorithms are widely used to automatically design neural architectures for CNNs and RNNs. Inspired by the success of neural architecture search algorithms, we present a graph neural architecture search algorithm GraphNAS that enables automatic design of the best graph neural architecture based on reinforcement learning. Specifically, GraphNAS uses a recurrent network as the controller to generate variable-length strings that describe the architectures of graph neural networks, and trains the recurrent network with policy gradient to maximize the expected accuracy of the generated architectures on a validation data set. Moreover, based on GraphNAS, we design a new GraphNAS++ model using distributed neural architecture search. Compared with GraphNAS that generates and evaluates only one candidate architecture at each iteration, GraphNAS++ generates a mini-batch of candidate architectures and evaluates them in a distributed computing environment until convergence. Experiments on real-world datasets demonstrate that GraphNAS can design a novel network architecture that rivals the best human-invented architecture. Moreover, GraphNAS++ can speed up the design process at least five times by using the distributed training framework with GPUs.

Heterogeneous graphs are commonly used to describe networked data with multiple types of nodes and edges. Heterogeneous Graph Neural Networks (HGNNs) are powerful tools for analyzing heterogeneous graphs. However, designing neural architectures of HGNNs requires extensive domain knowledge and time-consuming manual work. Recently, neural architecture search algorithms have become popular in automatically designing neural architectures for homogeneous graph neural networks. In this paper, we present a Heterogeneous Graph Neural Architecture Search algorithm (HGNAS for short) which allows the automatic design of heterogeneous graph neural architectures. Specifically, HGNAS first designs a new search space based on existing popular HGNNs. Then, HGNAS uses a policy network as the controller to sample and find the best neural architecture from the designed search space by maximizing the expected accuracy of the selected architectures on a given validation dataset. Moreover, we design a new method HGNAS++ to improve the efficiency of HGNAS by training the RNN controller within a generative adversarial learning framework. The basic idea of HGNAS++ is to embed a pairwise ranker into the reinforcement learning based architecture search algorithm. The pairwise ranker can be taken as a discriminator which selects more accurate architectures between pairs of candidate architectures. Then, the RNN controller can be updated more efficiently by only using a relatively small number of candidate architectures selected by the pairwise ranker. Experiments on real-world heterogeneous graph datasets show that HGNAS is capable of designing novel HGNNs that beat the best human-invented HGNNs. On the benchmark datasets, HGNAS++ improves HGNAS in terms of evaluation cost, with a reduction of 50% of the evaluated candidate architectures and a decrease of 24% in search time on average. As a byproduct, HGNAS++ can find sparse yet powerful neural architectures for HGNNs.

Structured sequences are a popular data representation, used to model complex data such as traffic networks. A key machine learning task for structured sequences is node classification, that is predicting the class labels of unlabeled nodes. Though many node classification models were proposed, they assume a closed world setting, that all class labels appear in the training data. But in the real-world, the presence of never-before-seen class labels in testing data can considerably degrade a classifier’s accuracy. A promising solution to this issue is to build classifiers for an open-world setting, where samples with unknown class labels are continuously observed such that training and testing data may have different class label spaces. Several approaches have been proposed for open-world learning problems in computer vision and natural language processing, but they cannot be applied directly to structured sequences due to the complexity of their non-Euclidean properties and their dynamic nature. This paper addresses this important research gap by proposing a novel Open-world Structured Sequence node Classification (OSSC) model, to learn from structured sequences in an open-world setting. OSSC captures the structural and temporal information via a GCN-based dynamic variational framework. A latent distribution sequence is learned for each node using both stochastic states and deterministic states, to capture the evolution of node attributes and topology, followed by a sampling process to generate node representations. An open-world classification loss is further adopted to ensure that node representations are sensitive to unknown classes. And a combination of Openmax and Softmax is utilized to recognize nodes from unknown classes and to classify others to one of the known classes. Experiments on real-world datasets show that the proposed OSSC method is capable of learning accurate open-world node classifiers from structured sequence data.

Guaranteed Display (GD) is widely used in e-commerce marketing for advertisers to acquire an agreed-upon number of impressions with target audiences. With the main objective to maximize the contract delivery rate under contract constraints, user interest (such as click-through rate and conversion rate) is also essential to improve the long-time return on investment for advertisers and the e-commerce platform. In this paper, we design an adaptive unified allocation framework (AUAF) by not only considering supply of audience impressions in request-level but also avoiding over-allocation of audience impressions. Specifically, our allocation model simultaneously optimizes the contract delivery and the match between advertisements and user interests with explicit constraint to prevent unnecessary allocation. Facing the challenge of serving billion-scale requests per day, a parameter-server based parallel optimization algorithm is also developed, enabling the proposed allocation model to be efficiently optimized and incrementally updated in minutes. Thus, the offline optimization results and the online decisions can be synchronized for real-time serving. In other words, our approach can achieve adaptive pacing that is consistent with the optimal allocation solution. Our extensive experimental results demonstrate that the proposed AUAF framework can improve both contract delivery rate and average click-through rate (CTR), which we use to measure the user interest in this paper. The improvements on CTR are statistically significant in comparison with existing methods. Moreover, since March 2020, AUAF has been deployed in the guaranteed display advertising system of Alibaba, bringing more than 10% increase on CTR without loss of contract delivery rate, which has resulted in significant value creation for the business.

In this demo, we present an online multi-modal community query engine (MQE) on Alibaba's billion-scale heterogeneous network. MQE has two distinct features in comparison with existing community query engines. Firstly, MQE supports multimodal community search on heterogeneous graphs with keyword and image queries. Secondly, to facilitate community understanding in real business scenarios, MQE generates natural language descriptions for the retrieved community in combination with other useful demographic information. The distinct features of MQE benefit many downstream applications in Alibaba's e-commerce platform like recommendation. Our experiments confirm the effectiveness and efficiency of MQE on graphs with billions of edges.

On-demand food delivery (OFD) platforms have greatly impacted the food service industry, where OFD recommendation systems play a central role in enhancing user experience and raising revenues. OFD recommendation, compared with existing online e-commerce recommendation systems, needs to put more emphasis on fulfillment time related variables, because the order fulfillment cycle time (OFCT) which refers to the time elapsed between a user placing a food order and receiving the food significantly influences a user's choice from the recommended items. In this paper, we investigate the OFCT related information and propose a Fulfillment-Time-Aware Personalized Ranking (FTAPR) method for recommendation. FTAPR mainly consists of three components. First, Transformers are used to estimate OFCT based on a large amount of user order sequences. Then, the predicted OFCT and other OFCT related features are fused and encoded by a deep & cross network to learn fulfillment time related feature representation. At the last step, the time bias representation from the deep & cross network is integrated into the ranking system to deliver final search results. Extensive offline and online experiments on real-world datasets collected from one of China's largest OFD platforms Ele.me show the superiority of our model, e.g., an online A/B testing shows that FTAPR brings 1.3% and 2.5% gains in CTR and CVR compared with baselines.

Malicious Web crawlers threaten information system due to heavily taking up bandwidth resources and stealing private user data. Ele.me, a prevalent on-demand food delivery platform in China, suffers from the negative impact of crawlers. The crawler detection systems face two major challenges: spatial patterns of the crawler behaviors and limited labeled data for training. In this paper, we present efficient solutions to tackle these challenges. Specifically, we propose a new Attributed Action Net (AANet for short) model to detect Location-Based Services~(LBS) crawlers and a three-stage learning framework to train the model. AANet consists of three different embedding modules, including the action token sequence, temporal-spatial attributes of users, and the context information of the raw data. We have deployed the model at Ele.me, and both offline experiments and online A/B tests show that the proposed method is superior to the state-of-the-art models for sequence data classification on the food delivery platform.

Heterogeneous Graph Neural networks (HGNNs) have been popularly used in processing complicated networks such as academic networks, social networks, and knowledge graphs. Despite their success, the design of the neural architectures of HGNNs still requires rich domain knowledge and heavy manual work. In this paper, we propose a Heterogeneous Graph Neural Architecture Search algorithm (HGNAS for short) which enables automatic design of the best neural architectures with minimal human effort. Specifically, HGNAS first defines a general HGNN framework on top of existing popular HGNNs. A search space of HGNAS is designed based on the general framework that includes multiple groups of message encoding and aggregation functions. Then, HGNAS uses a policy network as the controller to sample and find the best neural architecture from the designed search space by maximizing the expected accuracy of the selected architectures on a validation dataset. Moreover, we introduce effective methods to improve HGNAS from three aspects, i.e., the optimization of hyper-parameters, the improvement of search space, and the selection of message receptive fields. Experiments on public datasets show that HGNAS is capable of designing novel HGNNs that rival the best human-invented HGNNs. More interestingly, HGNAS finds some sparse yet powerful neural architectures for HGNNs on the benchmark datasets.

Multivariate Hawkes processes (MHPs) are classic methods to learn temporal patterns in event sequences of different entities. Traditional MHPs with explicit parametric intensity functions are friendly to model interpretability. However, recent Deep MHPs which employ various variants of recurrent neural networks are hardly to understand, albeit more expressive towards event sequences. The lack of model interpretability of Deep MHPs leads to a limited comprehension of complicated dynamics between events. To this end, we present a new Disentangled Deep Multivariate Hawkes Process $(\mathrm{D}^{2}$ MHP) to enhance model expressiveness and meanwhile maintain model interpretability. $\mathrm{D}^{2}$ MHP achieves state disentanglement by disentangling the latent representation of an event sequence into static and dynamic latent variables, and matches these latent variables to interpretable factors in the intensity function. Moreover, considering that an entity typically has multiple identities, $\mathrm{D}^{2}$ MHP further splits these latent variables into factorized representations, each of which is associated with a corresponding identity. Experiments on real-world datasets show that $\mathrm{D}^{2}$ MHP yields significant and consistent improvements over state-of-the-art baselines. We also demonstrate model interpretability via the detailed analysis.

Live-streaming platforms have recently gained significant popularity by attracting an increasing number of young users and have become a very promising form of online shopping. Similar to the traditional online shopping platforms such as Taobao, live-streaming platforms also suffer from online malicious fraudulent behaviors where many transactions are not genuine. The existing anti-fraud models proposed to recognize fraudulent transactions on traditional online shopping platforms are inapplicable on live-streaming platforms. This is mainly because live-streaming platforms are characterized by a unique type of heterogeneous live-streaming networks where multiple heterogeneous types of nodes such as users, live-streamers, and products are connected with multiple different types of edges associated with edge features. In this paper, we propose a new approach based on a heterogeneous graph neural network for LIve-streaming Fraud dEtection (called LIFE). LIFE designs an innovative heterogeneous graph learning model that fully utilizes various heterogeneous information of shopping transactions, users, streamers, and items from a given live-streaming platform. Moreover, a label propagation algorithm is employed within our LIFE framework to handle the limited number of labeled fraudulent transactions for model training. Extensive experimental results on a large-scale Taobao live-streaming platform demonstrate that the proposed method is superior to the baseline models in terms of fraud detection effectiveness on live-streaming platforms. Furthermore, we conduct a case study to show that the proposed method is able to effectively detect fraud communities for live-streaming e-commerce platforms.

The large-scale recommender system mainly consists of two stages: matching and ranking. The matching stage (also known as the retrieval step) identifies a small fraction of relevant items from billion-scale item corpus in low latency and computational cost. Item-to-item collaborative filtering (item-based CF) and embedding-based retrieval (EBR) have been long used in the industrial matching stage owing to its efficiency. However, item-based CF is hard to meet personalization, while EBR has difficulty in satisfying diversity. In this paper, we propose a novel matching architecture, Path-based Deep Network (named PDN), through incorporating both personalization and diversity to enhance matching performance. Specifically, PDN is comprised of two modules: Trigger Net and Similarity Net. PDN utilizes Trigger Net to capture the user's interest in each of his/her interacted item. Similarity Net is devised to evaluate the similarity between each interacted item and the target item based on these items' profile and CF information. The final relevance between the user and the target item is calculated by explicitly considering user's diverse interests, \ie aggregating the relevance weights of the related two-hop paths (one hop of a path corresponds to user-item interaction and the other to item-item relevance). Furthermore, we describe the architecture design of the proposed PDN in a leading real-world E-Commerce service (Mobile Taobao App). Based on offline evaluations and online A/B test, we show that PDN outperforms the existing solutions for the same task. The online results also demonstrate that PDN can retrieve more personalized and more diverse items to significantly improve user engagement. Currently, PDN system has been successfully deployed at Mobile Taobao App and handling major online traffic.

In the emerging business of food delivery, rider traffic accidents raise financial cost and social traffic burden. Although there has been much effort on traffic accident forecasting using temporal-spatial prediction models, none of the existing work studies the problem of detecting the takeaway rider accidents based on food delivery trajectory data. In this article, we aim to detect whether a takeaway rider meets an accident on a certain time period based on trajectories of food delivery and riders’ contextual information. The food delivery data has a heterogeneous information structure and carries contextual information such as weather and delivery history, and trajectory data are collected as a spatial-temporal sequence. In this article, we propose a TakeAway Rider Accident detection fusion network TARA-Net to jointly model these heterogeneous and spatial-temporal sequence data. We utilize the residual network to extract basic contextual information features and take advantage of a transformer encoder to capture trajectory features. These embedding features are concatenated into a pyramidal feed-forward neural network. We jointly train the above three components to combine the benefits of spatial-temporal trajectory data and sparse basic contextual data for early detecting traffic accidents. Furthermore, although traffic accidents rarely happen in food delivery, we propose a sampling mechanism to alleviate the imbalance of samples when training the model. We evaluate the model on a transportation mode classification dataset Geolife and a real-world Ele.me dataset with over 3 million riders. The experimental results show that the proposed model is superior to the state-of-the-art.

Understanding the target audience for popular television series is valuable for online video platform to manage advertising sales, purchase video copyrights, and compete with other video service platforms. Existing studies in this domain generally focus on using data mining and machine learning techniques to recommend television series to individual users or predict the popularity of television series. Knowing only the popularity of television series may, however, limit our ability to answer more in-depth questions and develop more intelligent applications. In this paper, we develop a data-driven framework to model and predict audience competition patterns for popular online television series. Specifically, we first construct a sequence of dynamic competition networks of television series by mining the detailed viewership records. Then, we design the Dynamic Deep Network Factorization (DDNF), a hybrid modeling framework for predicting the future competition networks. Our framework adopts the deep neural network (DNN) and the knowledge-base (KB) embedding to incorporate static features, and integrates the Long Short-Term Memory (LSTM) network to learn dynamic features of the television series. Finally, extensive experiments on real-world data sets validate the effectiveness of our approach compared with state-of-the-art baselines in predicting the audience competition for existing and new television series.

Attributed network embedding enables joint representation learning of node links and attributes. Existing attributed network embedding models are designed in continuous Euclidean spaces which often introduce data redundancy and impose challenges to storage and computation costs. To this end, we present a Binarized Attributed Network Embedding model (BANE for short) to learn binary node representation. Specifically, we define a new Weisfeiler-Lehman proximity matrix to capture data dependence between node links and attributes by aggregating the information of node attributes and links from neighboring nodes to a given target node in a layer-wise manner. Based on the Weisfeiler-Lehman proximity matrix, we formulate a new Weisfiler-Lehman matrix factorization learning function under the binary node representation constraint. The learning problem is a mixed integer optimization and an efficient cyclic coordinate descent (CCD) algorithm is used as the solution. Node classification and link prediction experiments on real-world datasets show that the proposed BANE model outperforms the state-of-the-art network embedding methods.

Graph classification is an important tool for analyzing data with structure dependency, where subgraphs are often used as features for learning. In reality, the dimension of the subgraphs crucially depends on the threshold setting of the frequency support parameter, and the number may become extremely large. As a result, subgraphs may be incrementally discovered to form a feature stream and require the underlying graph classifier to effectively discover representative subgraph features from the subgraph feature stream. In this paper, we propose a primal-dual incremental subgraph feature selection algorithm (ISF) based on a max-margin graph classifier. The ISF algorithm constructs a sequence of solutions that are both primal and dual feasible. Each primal-dual pair shrinks the dual gap and renders a better solution for the optimal subgraph feature set. To avoid bias of ISF algorithm on short-pattern subgraph features, we present a new incremental subgraph join feature selection algorithm (ISJF) by forcing graph classifiers to join short-pattern subgraphs and generate long-pattern subgraph features. We evaluate the performance of the proposed models on both synthetic networks and real-world social network data sets. Experimental results demonstrate the effectiveness of the proposed methods.

Social recommendation explores social information to improve the quality of a recommender system. It can be further divided into explicit and implicit social network recommendation. The former assumes the existence of explicit social connections between users in addition to the rating data. The latter one assumes the availability of only the ratings but not the social connections between users since the explicit social information data may not necessarily be available and usually are binary decision values (e.g., whether two people are friends), while the strength of their relationships is missing. Most of the works in this field use only rating data to infer the latent social networks. They ignore the dynamic nature of users that the preferences of users drift over time distinctly. To this end, we propose a new Implicit Dynamic Social Recommendation (IDSR) model, which infers latent social network from cascade data. It can sufficiently mine the information contained in time by mining the cascade data and identify the dynamic changes in the users in time by using the latest updated social network to make recommendations. Experiments and comparisons on three real-world datasets show that the proposed model outperforms the state-of-the-art solutions in both explicit and implicit scenarios.

In this paper, we study a new problem of continuous learning from doubly-streaming data where both data volume and feature space increase over time. We refer to the doubly-streaming data as trapezoidal data streams and the corresponding learning problem as online learning from trapezoidal data streams. The problem is challenging because both data volume and data dimension increase over time, and existing online learning[1],[2], online feature selection[3], and streaming feature selection algorithms[4],[5]are inapplicable. We propose a new Online Learning with Streaming Features algorithm (OL$_{SF}$for short) and its two variants, which combine online learning[1],[2]and streaming feature selection[4],[5]to enable learning from trapezoidal data streams with infinite training instances and features. When a new training instance carrying new features arrives, a classifier updates the existing features by following the passive-aggressive update rule[2]and updates the new features by following the structural risk minimization principle. Feature sparsity is then introduced by using the projected truncation technique. We derive performance bounds of the OL$_{SF}$algorithm and its variants. We also conduct experiments on real-world data sets to show the performance of the proposed algorithms.

Graph classification is an important tool for analysing structured and semi-structured data, where subgraphs are commonly used as the feature representation. However, the number and size of subgraph features crucially depend on the threshold parameters of frequent subgraph mining algorithms. Any improper setting of the parameters will generate many trivial short-pattern subgraph fragments which dominate the feature space, distort graph classifiers and bury interesting long-pattern subgraphs. In this paper, we propose a new Subgraph Join Feature Selection (SJFS) algorithm. The SJFS algorithm, by forcing graph classifiers to join short-pattern subgraph fragments, can defrag trivial subgraph features and deliver long-pattern interesting subgraphs. Experimental results on both synthetic and real-world social network graph data demonstrate the performance of the proposed method.

The problem of modeling topics on user behavior data in social networks has been widely studied in social marketing and social emotion analysis, where latent topic models are commonly used as the solutions. The user behavior data are highly related in time and space, which demands new latent topic models that consider both temporal and spatial distances. However, existing topic models either fail to model these two factors simultaneously, or cannot handle the high order dependence among user behaviors. In this paper we present a new nonparametric Bayesian model Time and Space Dependent Chinese Restaurant Processes (TSD-CRP for short). TSD-CRP can auto-select the number of topics and model high-order temporal and spatial dependence behind user behavior data. Empirical results on real-world data sets demonstrate the effectiveness of the proposed method.

We study a new problem of learning from doubly-streaming data where both data volume and feature space increase over time. We refer to the problem as mining trapezoidal data streams. The problem is challenging because both data volume and feature space are increasing, to which existing online learning, online feature selection and streaming feature selection algorithms are inapplicable. We propose a new Sparse Trapezoidal Streaming Data mining algorithm (STSD) and its two variants which combine online learning and online feature selection to enable learning trapezoidal data streams with infinite training instances and features. Specifically, when new training instances carrying new features arrive, the classifier updates the existing features by following the passive-aggressive update rule used in online learning and updates the new features with the structural risk minimization principle. Feature sparsity is also introduced using the projected truncation techniques. Extensive experiments on the demonstrated UCI data sets show the performance of the proposed algorithms.

Ensemble learning is a common tool for data stream classification, mainly because of its inherent advantages of handling large volumes of stream data and concept drifting. Previous studies, to date, have been primarily focused on building accurate ensemble models from stream data. However, a linear scan of a large number of base classifiers in the ensemble during prediction incurs significant costs in response time, preventing ensemble learning from being practical for many real-world time-critical data stream applications, such as Web traffic stream monitoring, spam detection, and intrusion detection. In these applications, data streams usually arrive at a speed of GB/second, and it is necessary to classify each stream record in a timely manner. To address this problem, we propose a novel Ensemble-tree (E-tree for short) indexing structure to organize all base classifiers in an ensemble for fast prediction. On one hand, E-trees treat ensembles as spatial databases and employ an R-tree like height-balanced structure to reduce the expected prediction time from linear to sub-linear complexity. On the other hand, E-trees can be automatically updated by continuously integrating new classifiers and discarding outdated ones, well adapting to new trends and patterns underneath data streams. Theoretical analysis and empirical studies on both synthetic and real-world data streams demonstrate the performance of our approach.

Influence maximization, defined as finding a small subset of nodes that maximizes spread of influence in social networks, is NP-hard under both Independent Cascade (IC) and Linear Threshold (LT) models, where many greedy-based algorithms have been proposed with the best approximation guarantee. However, existing greedy-based algorithms are inefficient on large networks, as it demands heavy Monte-Carlo simulations of the spread functions for each node at the initial step [7]. In this paper, we establish new upper bounds to significantly reduce the number of Monte-Carlo simulations in greedy-based algorithms, especially at the initial step. We theoretically prove that the bound is tight and convergent when the summation of weights towards (or from) each node is less than 1. Based on the bound, we propose a new Upper Bound based Lazy Forward algorithm (UBLF in short) for discovering the top-k influential nodes in social networks. We test and compare UBLF with prior greedy algorithms, especially CELF [30]. Experimental results show that UBLF reduces more than 95 percent Monte-Carlo simulations of CELF and achieves about 2-10 times speedup when the seed set is small.

Max-margin matrix factorization (M3F) has been popularly applied to collaborative filtering for personalized recommendations. The nonparametric M3F model represents the latest progress of the M3F methods, which can auto-select the number of factors by using nonparametric techniques. However, existing non-parametric M3F methods assume a collection of user rating data can be fully obtained before training, and they are inapplicable for on-the-fly recommender systems where user rating data arrive continuously. In this paper, we present a new efficient online nonparametric 3F model for flexible recommendation. Specifically, we design an online nonparametric M3F model (OnM3F) based on the online Passive-Aggressive learning and solve the corresponding optimization problem by using the online stochastic gradient descent. Empirical studies on two large real-world data sets verify the effectiveness of the proposed method.

Detecting outliers from big data plays an important role in network security. Previous outlier detection algorithms are generally incapable of handling big data. In this paper we present an parallel outlier detection method for big data, based on a new parallel auto-encoder method. Specifically, we build a replicator model of the input data to obtain the representation of sample data. Then, the replicator model is used to measure the replicability of test data, where records having higher reconstruction errors are classified as outliers. Experimental results show the performance of the proposed parallel algorithm.

In this paper, we study a new problem on social network influence maximization. The problem is defined as, given a target user $w$, finding the top-k most influential nodes for the user. Different from existing influence maximization works which aim to find a small subset of nodes to maximize the spread of influence over the entire network (i.e., global optima), our problem aims to find a small subset of nodes which can maximize the influence spread to a given target user (i.e., local optima). The solution is critical for personalized services on social networks, where fully understanding of each specific user is essential. Although some global influence maximization models can be narrowed down as the solution, these methods often bias to the target node itself. To this end, in this paper we present a local influence maximization solution. We first provide a random function, with low variance guarantee, to randomly simulate the objective function of local influence maximization. Then, we present efficient algorithms with approximation guarantee. For online social network applications, we also present a scalable approximate algorithm by exploring the local cascade structure of the target user. We test the proposed algorithms on several real-world social networks. Experimental results validate the performance of the proposed algorithms.

Influence maximization, defined as finding a small subset of nodes that maximizes spread of influence in social networks, is NP-hard under both Linear Threshold (LT) and Independent Cascade (IC) models, where a line of greedy/heuristic algorithms have been proposed. The simple greedy algorithm [14] achieves an approximation ratio of 1-1/e. The advanced CELF algorithm [16], by exploiting the sub modular property of the spread function, runs 700 times faster than the simple greedy algorithm on average. However, CELF is still inefficient [4], as the first iteration calls for N times of spread estimations (N is the number of nodes in networks), which is computationally expensive especially for large networks. To this end, in this paper we derive an upper bound function for the spread function. The bound can be used to reduce the number of Monte-Carlo simulation calls in greedy algorithms, especially in the first iteration of initialization. Based on the upper bound, we propose an efficient Upper Bound based Lazy Forward algorithm (UBLF in short), by incorporating the bound into the CELF algorithm. We test and compare our algorithm with prior algorithms on real-world data sets. Experimental results demonstrate that UBLF, compared with CELF, reduces more than 95% Monte-Carlo simulations and achieves at least 2-5 times speed-raising when the seed set is small.

Behavior targeting (BT) is a promising tool for online advertising. The state-of-the-art BT methods, which are mainly based on regression models, have two limitations. First, learning regression models for behavior targeting is difficult since user clicks are typically several orders of magnitude fewer than views. Second, the user interests are not fixed, but often transient and influenced by media and pop culture. In this paper, we propose to formulate behavior targeting as a classification problem. Specifically, we propose to use an SVM ensemble for behavior prediction. The challenge of using ensemble SVM for BT stems from the computational complexity (it takes 53 minutes in our experiments to predict behavior for 32 million users, which is inadequate for online application). To this end, we propose a fast ensemble SVM prediction framework, which builds an indexing structure for SVM ensemble to achieve sub-linear prediction time complexity. Experimental results on real-world large scale behavior targeting data demonstrate that the proposed method is efficient and outperforms existing linear regression based BT models.

Data stream classification has drawn increasing attention from the data mining community in recent years, where a large number of stream classification models were proposed. However, most existing models were merely focused on mining from single-label data streams. Mining from multi-label data streams has not been fully addressed yet. On the other hand, although some recent work touched the multi-label stream mining problem, they never consider the expensive labeling cost issue, preventing them from real-world applications. To this end, we study, in this paper, a challenging problem that mining from multi-label data streams with limited labeling resource. Specifically, we propose an ensemble-based active learning framework to handle the large volume of stream data, expensive labeling cost and concept drifting problems on multi-label data streams. Experiments on both synthetic and real world data sets demonstrate the performance of the proposed method.

Mining frequent patterns from data streams has drawn increasing attention in recent years. However, previous mining algorithms were all focused on a single data stream. In many emerging applications, it is of critical importance to combine multiple data streams for analysis. For example, in real-time news topic analysis, it is necessary to combine multiple news report streams from dierent media sources to discover collaborative frequent patterns which are reported frequently in all media, and comparative frequent patterns which are reported more frequently in a media than others. To address this problem, we propose a novel frequent pattern mining algorithm Hybrid-Streaming, H-Stream for short. H-Stream builds a new Hybrid-Frequent tree to maintain historical frequent and potential frequent itemsets from all data streams, and incrementally updates these itemsets for efficient collaborative and comparative pattern mining. Theoretical and empirical studies demonstrate the utility of the proposed method.

Cloud computing represents one of the most important research directions for modern computing systems. Existing research efforts on Cloud computing were all focused on designing advanced storage and query techniques for static data. None of them consider the problem that data in a Cloud may appear as continuous and rapid data streams. To address this problem, in this paper we propose a new LCN-Index framework to handle continuous data stream queries in the Cloud. LCN-Index uses the Map-Reduce computing paradigm to process all the queries. In the Mapping stage, it divides all the queries into a batch of predicate sets which are then deployed onto mapping nodes using interval predicate index. In the reducing stage, it merges results from the mapping nodes using multi attribute hash index. In so doing, a data stream can be efficiently evaluated by traversing through the LCN-Index framework. Experiments demonstrate the utility of the proposed method.

Lazy learning, such as k-nearest neighbor learning, has been widely applied to many applications. Known for well capturing data locality, lazy learning can be advantageous for highly dynamic and complex learning environments such as data streams. Yet its high memory consumption and low prediction efficiency have made it less favorable for stream oriented applications. Specifically, traditional lazy learning stores all the training data and the inductive process is deferred until a query appears, whereas in stream applications, data records flow continuously in large volumes and the prediction of class labels needs to be made in a timely manner. In this paper, we provide a systematic solution that overcomes the memory and efficiency limitations and enables fast lazy learning for concept drifting data streams. In particular, we propose a novel Lazy-tree (Ltree for short) indexing structure that dynamically maintains compact high-level summaries of historical stream records. L-trees are M-Tree [5] like, height-balanced, and can help achieve great memory consumption reduction and sub-linear time complexity for prediction. Moreover, L-trees continuously absorb new stream records and discard outdated ones, so they can naturally adapt to the dynamically changing concepts in data streams for accurate prediction. Extensive experiments on real-world and synthetic data streams demonstrate the performance of our approach.

Ensemble learning has become a common tool for data stream classification, being able to handle large volumes of stream data and concept drifting. Previous studies focus on building accurate prediction models from stream data. However, a linear scan of a large number of base classifiers in the ensemble during prediction incurs significant costs in response time, preventing ensemble learning from being practical for many real world time-critical data stream applications, such as Web traffic stream monitoring, spam detection, and intrusion detection. In these applications, data streams usually arrive at a speed of GB/second, and it is necessary to classify each stream record in a timely manner. To address this problem, we propose a novel Ensemble-tree (E-tree for short) indexing structure to organize all base classifiers in an ensemble for fast prediction. On one hand, E-trees treat ensembles as spatial databases and employ an R-tree like height-balanced structure to reduce the expected prediction time from linear to sub-linear complexity. On the other hand, E-trees can automatically update themselves by continuously integrating new classifiers and discarding outdated ones, well adapting to new trends and patterns underneath data streams. Experiments on both synthetic and real-world data streams demonstrate the performance of our approach.

Missing data commonly occurs in many applications. While many data imputation methods exist to handle the missing data problem for large scale databases, when applied to concept drifting data streams, these methods face some common difficulties. First, due to large and continuous data volumes, we are unable to maintain all stream records to form a candidate pool and estimate missing values, as most existing methods commonly do. Second, even if we could maintain all complete stream records using a summary structure, the concept drifting problem would make some information obsolete, and thus deteriorate the imputation accuracy. Third, in data streams, it is necessary to develop a fast yet accurate algorithm to find the most similar data for imputation. Fourth, due to the dynamic and sophisticated data collection environments, the missing rate of most stream data may be much higher than that in generic static databases, so the imputation method should be able to accommodate high missing rate in the data. To tackle these challenges, we propose, in this paper, a Streaming k-Nearest-Neighbors Imputation Framework (SKIF) for concept drifting data streams. To handle concept drifting and large volume problems in data streams, SKIF first summarizes historical complete records in some micro-resources (which are high-level statistical data structures), and maintains these micro-resources in a candidate pool as benchmark data. After that, SKIF employs a novel hybrid-kNN imputation procedure, which uses a hybrid similarity search mechanism, to find the most similar micro-resources from the large scale candidate pool efficiently. Experimental results demonstrate the effectiveness of the proposed SKIF framework for data stream imputation tasks.

Ensemble learning is a commonly used tool for building prediction models from data streams, due to its intrinsic merits of handling large volumes stream data. Despite of its extraordinary successes in stream data mining, existing ensemble models, in stream data environments, mainly fall into the ensemble classifiers category, without realizing that building classifiers requires labor intensive labeling process, and it is often the case that we may have a small number of labeled samples to train a few classifiers, but a large number of unlabeled samples are available to build clusters from data streams. Accordingly, in this paper, we propose a new ensemble model which combines both classifiers and clusters together for mining data streams. We argue that the main challenges of this new ensemble model include (1) clusters formulated from data streams only carry cluster IDs, with no genuine class label information, and (2) concept drifting underlying data streams makes it even harder to combine clusters and classifiers into one ensemble framework. To handle challenge (1), we present a label propagation method to infer each cluster's class label by making full use of both class label information from classifiers, and internal structure information from clusters. To handle challenge (2), we present a new weighting schema to weight all base models according to their consistencies with the up-to-date base model. As a result, all classifiers and clusters can be combined together, through a weighted average mechanism, for prediction. Experiments on real-world data streams demonstrate that our method outperforms simple classifier ensemble and cluster ensemble for stream data mining.

In this paper, we propose a framework to build prediction models from data streams which contain both labeled and unlabeled examples. We argue that due to the increasing data collection ability but limited resources for labeling, stream data collected at hand may only have a small number of labeled examples, whereas a large portion of data remain unlabeled but can be beneficial for learning. Unleashing the full potential of the unlabeled instances for stream data mining is, however, a significant challenge, consider that even fully labeled data streams may suffer from the concept drifting, and inappropriate uses of the unlabeled samples may only make the problem even worse. To build prediction models, we first categorize the stream data into four different categories, each of which corresponds to the situation where concept drifting may or may not exist in the labeled and unlabeled data. After that, we propose a relational k-means based transfer semi-supervised SVM learning framework (RK-TS3VM), which intends to leverage labeled and unlabeled samples to build prediction models. Experimental results and comparisons on both synthetic and real-world data streams demonstrate that the proposed framework is able to help build prediction models more accurate than other simple approaches can offer.

A new kernel-based learning algorithm called kernel affine subspace nearest point (KASNP) approach is proposed in this paper. Inspired by the geometrical explanation of support vector machines (SVMs) and its nearest point problem in convex hulls, we extend the convex hull of each class to its corresponding affine subspace in high dimensional space induced by kernel. In two class affine subspaces, KASNP finds the nearest points and then constructs a separating hyperplane, which bisects the line segment joining them. The nearest point problem of KASNP is only an unconstrained optimal problem whose solution can be directly computed. Compared with SVM, KASNP avoids solving convex quadratic programming. Experiments on two-spiral dataset, two UCI credit datasets, and face recognition datasets show that our proposed KASNP is effective for data classification.

In this paper, we identify a new research problem on cleansing noisy data streams which contain incorrectly labeled training examples. The objective is to accurately identify and remove mislabeled data, such that the prediction models built from the cleansed streams can be more accurate than the ones trained from the raw noisy streams. For this purpose, we first use bias-variance decomposition to derive a maximum variance margin (MVM) principle for stream data cleansing. Following this principle, we further propose a local and global filtering (LgF) framework to combine the strength of local noise filtering (within one single data chunk) and global noise filtering (across a number of adjacent data chunks) to identify erroneous data. Experimental results on six data streams (including two real-world data streams) demonstrate that LgF significantly outperforms simple methods in identifying noisy examples.

Mining concept drifting data streams is a defining challenge for data mining research. Recent years have seen a large body of work on detecting changes and building prediction models from stream data, with a vague understanding on the types of the concept drifting and the impact of different types of concept drifting on the mining algorithms. In this paper, we first categorize concept drifting into two scenarios: Loose Concept Drifting (LCD) and Rigorous Concept Drifting (RCD), and then propose solutions to handle each of them separately. For LCD data streams, because concepts in adjacent data chunks are sufficiently close to each other, we apply kernel mean matching (KMM) method to minimize the discrepancy of the data chunks in the kernel space. Such a minimization process will produce weighted instances to build classifier ensemble and handle concept drifting data streams. For RCD data streams, because genuine concepts in adjacent data chunks may randomly and rapidly change, we propose a new Optimal Weights Adjustment (OWA) method to determine the optimum weight values for classifiers trained from the most recent (up-to-date) data chunk, such that those classifiers can form an accurate classifier ensemble to predict instances in the yet-to-come data chunk. Experiments on synthetic and real-world datasets will show that weighted instance approach is preferable when the concept drifting is mainly caused by the changing of the class prior probability; whereas the weighted classifier approach is preferable when the concept drifting is mainly triggered by the changing of the conditional probability.

In this paper, we address a new research problem on active learning from data streams where data volumes grow continuously and labeling all data is considered expensive and impractical. The objective is to label a small portion of stream data from which a model is derived to predict newly arrived instances as accurate as possible. In order to tackle the challenges raised by data streams' dynamic nature, we propose a classifier ensembling based active learning framework which selectively labels instances from data streams to build an accurate classifier. A minimal variance principle is introduced to guide instance labeling from data streams. In addition, a weight updating rule is derived to ensure that our instance labeling process can adaptively adjust to dynamic drifting concepts in the data. Experimental results on synthetic and real-world data demonstrate the performances of the proposed efforts in comparison with other simple approaches.