Node classification is a key task in temporal graph learning (TGL). Real-life temporal graphs often introduce new node classes over time, but existing TGL methods assume a fixed set of classes. This assumption brings limitations, as updating models with full data is costly, while focusing only on new classes results in forgetting old ones. Graph continual learning (GCL) methods mitigate forgetting using old-class subsets but fail to account for their evolution. We define this novel problem as temporal graph continual learning (TGCL), which focuses on efficiently maintaining up-to-date knowledge of old classes. To tackle TGCL, we propose a selective learning framework that substitutes the old-class data with its subsets, Learning Towards the Future (LTF). We derive an upper bound on the error caused by such replacement and transform it into objectives for selecting and learning subsets that minimize classification error while preserving the distribution of the full old-class data. Experiments on three real-world datasets show that LTF effectively addresses the TGCL challenge.

Graph neural networks (GNNs) have been gaining a reputation for effective modeling of graph data. Yet, it is challenging to train GNNs efficiently. Many frameworks have been proposed but most of them suffer from high batch preparation cost and data transfer cost for mini-batch training. In addition, existing works have limitations on the device utilization pattern, which results in fewer opportunities for pipeline parallelism. In this paper, we present DAHA, a GNN training framework with data and hardware aware execution planning to accelerate end-to-end GNN training. We first propose a data and hardware aware cost model that is lightweight and gives accurate estimates on per-operation time cost for arbitrary input and hardware settings. Based on the cost model, we further explore the optimal execution plan for the data and hardware with three optimization strategies with pipeline parallelism: (1) group-based in-turn pipelining of batch preparation neural training to explore more optimization opportunities and prevent batch preparation bottlenecks; (2) data and hardware aware rewriting for intra-batch execution planning to improve computation efficiency and create more opportunities for pipeline parallelism; and (3) inter-batch scheduling to further boost the training efficiency. Extensive experiments demonstrate that DAHA can consistently and significantly accelerate end-to-end GNN training and generalize to different message-passing GNN models.

Many real-world binary relations can be modelled as bipartite graphs, which can be inherently temporal and each edge is associated with a timestamp. The ( α, β )-core, a popular structure that requires minimum degrees over two layers of vertices, is useful for understanding the organisation of bipartite networks. However, the temporal property has rarely been considered in cohesive subgraph mining in bipartite graphs. This gap prevents the finding of time-sensitive ( α, β )-cores in real-world applications. In this paper, we aim at finding ( α, β )-cores within any time window over a temporal bipartite graph. To address this problem, we propose a novel DAG (Directed Acyclic Graph)-like hierarchy with qualified time windows to describe the temporal containment property of the ( α, β )-core. Furthermore, we construct the superior-optimized index which significantly optimizes space complexity and guarantees efficient query performance. We also propose a maintenance approach that can efficiently update the index by removing stale information and incorporating newly inserted temporal edges. Extensive experiments are conducted on eight real-world graphs and the results show the effectiveness and efficiency of our indexes.

Subgraph matching is a fundamental problem in graph theory and has wide applications in areas like sociology, chemistry, and social networks. Due to its NP-hardness, the basic approach is a brute-force search over the whole search space. Some pruning strategies have been proposed to reduce the search space. However, they are either space-inefficient or based on assumptions that the graph has specific properties. In this paper, we propose SUFF, a general and powerful structure filtering framework, which can accelerate most of the existing approaches with slight modifications. Specifically, it builds a set of filters using matching results of past queries, and uses them to prune the search space for future queries. By fully utilizing the relationship between matches of two queries, it ensures that such pruning is sound. Furthermore, several optimizations are proposed to reduce the computation and space cost for building, storing, and using filters. Extensive experiments are conducted on multiple real-world data sets and representative existing approaches. The results show that SUFF can achieve up to 15X speedup with small overheads.

Differential privacy (DP) has become the de facto standard of privacy protection. For graphs, there are two widely used definitions of differential privacy, namely, edge differential privacy (edge-DP) and node differential privacy (node-DP), and node-DP is preferred when the minimal unit of interest is a node. To preserve node-DP, one can develop different methods to answer each specific graph query, or develop a graph publishing method to answer all graph queries. However, no existing works worked on such graph publishing methods. In this work, we propose two methods for publishing graphs under node-DP. One is the node-level perturbation algorithm which modifies the input graph by randomly inserting and removing nodes. The other one is the edge-level perturbation algorithm which randomly removing edges and inserting nodes. Both methods can achieve a flexible privacy guarantee by adjusting the running parameters. We conduct extensive experiments on both real-world and synthetic graphs to show the effectiveness and efficiency of proposed algorithms.

Graph Neural Networks (GNNs) have attracted much research interest due to their successful applications on graph-structured data. Despite the effectiveness, due to the data dependency, GNNs are confronted with the neighborhood explosion and over-smoothing problems. Many sampling-based and level-down methods have been proposed to solve the neighborhood explosion problem to boost efficiency. However, they suffer from either poor accuracy or considerable runtime overheads. Moreover, the over-smoothing problem prevents GNNs from exploring more distant neighborhoods effectively. In this paper, we present a Community-and-Contraction-based Graph Neural Network (CC-GNN), which leverages community and contraction to boost the time and space efficiency of GNNs. Specifically, CC-GNN first performs contraction and retrieves the communities as super nodes, and connects them using a tailored similarity function to obtain an informative community contracted graph (CC-Graph). CC-GNN then learns the representations of the super nodes in the CC-Graph, which are used to reconstruct the representations of the original nodes. Finally, CC-GNN explores more distant neighborhoods without additional convolution layers, implicitly alleviating the over-smoothing problem. Since CC-GNN conducts the costly training on a much smaller contracted graph, the efficiency is boosted significantly. Most importantly, we have proved that the information loss of node representations caused by the CC-Graph is bounded. Extensive experimental studies verify the efficiency boost and the effectiveness.

Mining cohesive subgraphs on bipartite graphs is an important task. The k -bitruss is one of many popular cohesive subgraph models, which is the maximal subgraph where each edge is contained in at least k butterflies. The bitruss decomposition problem is to find all k -bitrusses for k ≥ 0. Dealing with large graphs is often beyond the capability of a single machine due to its limited memory and computational power, leading to a need for efficiently processing large graphs in a distributed environment. However, all current solutions are for a single machine and a centralized environment, where processors can access the graph or auxiliary indexes randomly and globally. It is difficult to directly deploy such algorithms on a shared-nothing model. In this paper, we propose distributed algorithms for bitruss decomposition. We first propose SC-HBD as the baseline, which uses H -function to define bitruss numbers and computes them iteratively to a fix point in parallel. We then introduce a subgraph-centric peeling method SC-PBD, which peels edges in batches over different butterfly complete subgraphs. We then introduce local indexes on each fragment, study the butterfly-aware edge partition problem including its hardness, and propose an effective partitioner. Finally we present the bitruss butterfly-complete subgraph concept, and divide and conquer DC-BD method with optimization strategies. Extensive experiments show the proposed methods solve graphs with 30 trillion butterflies in 2.5 hours, while existing parallel methods under shared-memory model fail to scale to such large graphs.

Heterogeneous information networks (HINs) are usually used to model information systems with multi-type objects and relations. In contrast, graphs that have a single type of nodes and edges, are often called homogeneous graphs. Measuring similarities among objects is an important task in data mining applications, such as web search, link prediction, and clustering. Currently, several similarity measures are defined for HINs. Most of these measures are based on meta-paths, which show sequences of node classes and edge types along the paths between two nodes. However, meta-paths, which are often designed by domain experts, are hard to enumerate and choose w.r.t. the quality of similarity scores. This makes using existing similarity measures in real applications difficult. To address this problem, we extend SimRank, a well-known similarity measure on homogeneous graphs, to HINs, by introducing the concept of the decay graph. The newly proposed similarity measure is called HowSim, which has the property of being meta-path free, and capturing the structural and semantic similarity simultaneously. The generality and effectiveness of HowSim, and the efficiency of our proposed algorithms for computing HowSim scores, are demonstrated by extensive experiments.

Measuring similarities among different nodes is important in graph analysis tasks, such as link prediction, and recommendation. Among different similarity measures, SimRank is one of the most popular and promising ones, and has received a lot of research attention. While most current studies focus on single-pair, single-source/top-k, and all-pairs SimRank computation, few of them have studied finding similar pairs given a set of node pairs, which has attractive applications in personalized search and recommendation tasks. In this paper, we present Carmo, an efficient algorithm for retrieving the top-k similarities from an arbitrary set of pairs. In addition, we introduce two types of indexes to boost the efficiency of Carmo: one is hub-based, the other is tree-based. We show the effectiveness and efficiency of our proposed methods by extensive experiments.

Learning network embeddings has attracted growing attention in recent years. However, most of the existing methods focus on homogeneous networks, which cannot capture the important type information in heterogeneous networks. To address this problem, in this paper, we propose TransN, a novel multi-view network embedding framework for heterogeneous networks. Compared with the existing methods, TransN is an unsupervised framework which does not require node labels or user-specified meta-paths as inputs. In addition, TransN is capable of handling more general types of heterogeneous networks than the previous works. Specifically, in our framework TransN, we propose a novel algorithm to capture the proximity information inside each single view. Moreover, to transfer the learned information across views, we propose an algorithm to translate the node embeddings between different views based on the dual-learning mechanism, which can both capture the complex relations between node embeddings in different views, and preserve the proximity information inside each view during the translation. We conduct extensive experiments on real-world heterogeneous networks, whose results demonstrate that the node embeddings generated by TransN outperform those of competitors in various network mining tasks.

Heterogeneous information networks (HINs) are usually used to model information systems with multi-type objects and relations. Measuring the similarity among objects is an important task in data mining applications. Currently, several similarity measures are defined for HIN. Most of these measures are based on meta-paths, which show sequences of node classes and edge types along the paths between two nodes. However, meta-paths, which are often designed by domain experts, are hard to enumerate and choose w.r.t. the quality of the similarity scores. This makes the existing similarity measures difficult to use in real applications. To address this problem, we extend SimRank, a well-known similarity measure for homogeneous graphs, to HINs, by introducing the concept of decay graph. The newly proposed relevance measure is called HowSim, which has the property of being meta-path free, and capturing the structural and semantic similarity simultaneously. The generality and effectiveness of HowSim, are demonstrated by extensive experiments.

Recent years witnessed the emergence of various applications on knowledge graphs, which are often represented as RDF graphs. However, due to the lack of data schema and the complexity of SPARQL language, there is usually a gap between the user's real desire and the actual meaning of a SPARQL query, especially when the query itself is complicated. In this paper, we try to narrow this gap by modifying a given query with a set of modifiers, so that its result approaches a user-provided example set. Specifically, we model this problem as two individual sub-problems, query-restricting, and query-relaxing, both of which are shown to be NP-hard. We further prove that unless P=NP, query-restricting has no polynomial-time approximation scheme (PTAS), and query-relaxing has no polynomial-time constant-factor approximation algorithm. Despite their hardness, we propose a (1-1/ε)-approximation method for query-restricting and 2 heuristics for query-relaxing. Extensive experiments have been conducted on real-world knowledge graphs to evaluate the effectiveness and efficiency of our proposed solutions.

Community search in heterogeneous information networks (HINs) has attracted much attention in graph analysis. Given a vertex, the goal is to find a densely-connected sub-graph that contains the vertex. In practice, the user may need to restrict the number of connections between vertices, but none of the existing methods can handle such queries. In this paper, we propose the relational constraint that allows the user to specify fine-grained connection requirements between vertices. Base on this, we define the relational community as well as the problems of detecting and searching relational communities, respectively. For the detection problem, we propose an efficient solution that has near-linear time complexity. For the searching problem, although it is shown to be NP-hard and even hard-to-approximate, we devise two efficient approximate solutions. We further design the round index to accelerate the searching algorithm and show that it can handle dynamic graphs by its nature. Extensive experiments on both synthetic and real-world graphs are conducted to evaluate both the effectiveness and efficiency of our proposed methods.

For decades, the crowdsourcing has gained much attention from both academia and industry, which outsources a number of tasks to human workers. Typically, existing crowdsourcing platforms include CrowdFlower, Amazon Mechanical Turk (AMT), and so on, in which workers can autonomously select tasks to do. However, due to the unreliability of workers or the difficulties of tasks, workers may sometimes finish doing tasks either with incorrect/incomplete answers or with significant time delays. Existing studies considered improving the task accuracy through voting or learning methods, they usually did not fully take into account reducing the latency of the task completion. This is especially critical, when a task requester posts a group of tasks (e.g., sentiment analysis), and one can only obtain answers of all tasks after the last task is accomplished. As a consequence, the time delay of even one task in this group could delay the next step of the task requester's work from minutes to days, which is quite undesirable for the task requester. Inspired by the importance of the task accuracy and latency, in this paper, we will propose a novel crowdsourcing framework, namely Fast and Reliable crOwdsourcinG framework (FROG), which intelligently assigns tasks to workers, such that the latencies of tasks are reduced and the expected accuracies of tasks are met. Specifically, our FROG framework consists of two important components, task scheduler and notification modules. For the task scheduler module, we formalize a FROG task scheduling (FROG-TS) problem, in which the server actively assigns workers to tasks to achieve high task reliability and low task latency. We prove that the FROG-TS problem is NP-hard. Thus, we design two heuristic approaches, request-based and batch-based scheduling. For the notification module, we define an efficient worker notifying (EWN) problem, which only sends task invitations to those workers with high probabilities of accepting the tasks. To tackle the EWN problem, we propose a smooth kernel density estimation approach to estimate the probability that a worker accepts the task invitation. Through extensive experiments, we demonstrate the effectiveness and efficiency of our proposed FROG platform on both real and synthetic data sets.

Modern networks are of huge sizes as well as high dynamics, which challenges the efficiency of community detection algorithms. In this paper, we study the problem of overlapping community detection on distributed and dynamic graphs. Given a distributed, undirected and unweighted graph, the goal is to detect overlapping communities incrementally as the graph is dynamically changing. We propose an efficient algorithm, called randomized Speaker-Listener Label Propagation Algorithm (rSLPA), based on the Speaker-Listener Label Propagation Algorithm (SLPA) by relaxing the probability distribution of label propagation. Besides detecting high-quality communities, rSLPA can incrementally update the detected communities after a batch of edge insertion and deletion operations. To the best of our knowledge, rSLPA is the first algorithm that can incrementally capture the same communities as those obtained by applying the detection algorithm from the scratch on the updated graph. Extensive experiments are conducted on both synthetic and real-world datasets, and the results show that our algorithm can achieve high accuracy and efficiency at the same time.

Graph similarity search is a common and fundamental operation in graph databases. One of the most popular graph similarity measures is the Graph Edit Distance (GED) mainly because of its broad applicability and high interpretability. Despite its prevalence, exact GED computation is proved to be NP-hard, which could result in unsatisfactory computational efficiency on large graphs. However, exactly accurate search results are usually unnecessary for real-world applications especially when the responsiveness is far more important than the accuracy. Thus, in this paper, we propose a novel probabilistic approach to efficiently estimate GED, which is further leveraged for the graph similarity search. Specifically, we first take branches as elementary structures in graphs, and introduce a novel graph similarity measure by comparing branches between graphs, i.e., Graph Branch Distance (GBD), which can be efficiently calculated in polynomial time. Then, we formulate the relationship between GED and GBD by considering branch variations as the result ascribed to graph edit operations, and model this process by probabilistic approaches. By applying our model, the GED between any two graphs can be efficiently estimated by their GBD, and these estimations are finally utilized in the graph similarity search. Extensive experiments show that our approach has better accuracy, efficiency and scalability than other comparable methods in the graph similarity search over real and synthetic data sets.

Recently, with the rapid development of mobile devices and the crowdsourcing platforms, the spatial crowdsourcing has attracted much attention from the database community. Specifically, spatial crowdsourcing refers to sending a location-based request to workers according to their positions, and workers need to physically move to specified locations to conduct tasks. Many works have studied task assignment problems in spatial crowdsourcing, however, their problem settings are different from each other. Thus, it is hard to compare the performances of existing algorithms on task assignment in spatial crowdsourcing. In this paper, we present a comprehensive experimental comparison of most existing algorithms on task assignment in spatial crowdsourcing. Specifically, we first give general definitions about spatial workers and spatial tasks based on definitions in the existing works such that the existing algorithms can be applied on the same synthetic and real data sets. Then, we provide a uniform implementation for all the tested algorithms of task assignment problems in spatial crowdsourcing (open sourced). Finally, based on the results on both synthetic and real data sets, we discuss the strengths and weaknesses of tested algorithms, which can guide future research on the same area and practical implementations of spatial crowdsourcing systems.

Word embeddings have attracted much attention recently.Different from alphabetic writing systems, Chinese characters are often composed of subcharacter components which are also semantically informative.In this work, we propose an approach to jointly embed Chinese words as well as their characters and fine-grained subcharacter components.We use three likelihoods to evaluate whether the context words, characters, and components can predict the current target word, and collected 13,253 subcharacter components to demonstrate the existing approaches of decomposing Chinese characters are not enough.Evaluation on both word similarity and word analogy tasks demonstrates the superior performance of our model.

Community search problem, which is to find good communities given a set of query nodes in a graph, has attracted increasing research interest recently. Though various measurement models have been proposed to define and solve community search problem. Few of them could define a community concisely and have good quality of query results. They either involve additional constraints for modeling communities, such as size and diameter, or suffer from the free rider effect, i.e., include irrelevant subgraphs. In this paper, we propose a new k-plex based community model for community search. We show that our model not only is simple and clear, but also meets with basic requirements of defining a community search problem. We formulate the maximum k-plex community query (MCKPQ) problem, that is, given a set of query nodes Q , searching for optimal k-plex containing Q . We prove that MCKPQ is NP-hard, and it is hard to approximate in any constant factor. We first give exact solutions. Then, we propose an efficient branch-and-bound (B&B) method and design an effective upper bound function and a pruning strategy. Furthermore, we optimize the basic B&B by fast candidate generation. We also give a fast heuristic solution, which produces high-quality results in practice. The effectiveness of our model of community and the efficiency of our methods are verified by elaborate experiments.

In the initial publication, first name and family name of the second author Xun Jian were switched around. The original article has been corrected.