Attributed heterogeneous information networks (AHINs) encode rich semantics through diverse node and edge types. Recent learning-based community search methods on AHINs have shown promising performance but face two major limitations: i) difficulty scaling to large graphs due to memory-intensive neighbor-based propagation (e.g., GNNs and node-level attention), and ii) reliance on explicit community-level labels, which are often unavailable or costly to obtain. To address these issues, we propose a scalable Semi-supervised Community Search framework on AHINs (SCSAH), enabling scalability and efficiency, while eliminating the need for community-level labels by leveraging readily available node classification labels. Specifically, we devise MvSF2Token to extract Multi-view Semantic Features (MvSFs) as compact subgraph-level tokens before training, significantly reducing model propagation complexity. We then design a View-Aware Semantic Graph Transformer (VASGhormer) to effectively encode MvSFs by capturing cross-view dependencies and fusing semantic features. The combination of MvSF2Token and VASGhormer ensures scalability, efficiency, and robust performance. Furthermore, we design a View-Aware Contrastive Learner to train VASGhormer without requiring community-level supervision. Extensive experiments on five real-world datasets show that SCSAH outperforms state-of-the-art methods, achieving 18.06% higher performance and 10.43 times faster training.

Community search is a fundamental graph-based retrieval problem that aims to identify a query-dependent subgraph whose nodes exhibit strong internal connectivity. While recent learning-based methods improve retrieval effectiveness via graph representation learning, they follow a ''one-use-one-train'' paradigm that requires retraining or fine-tuning for each target graph, leading to high data dependency, high training costs, and limited generalization. To handle this, we propose OFA-CS, a ''one-for-all'' community search framework trained once on source datasets and directly deployed to arbitrary unseen graphs without retraining or fine-tuning, while preserving strong performance. Specifically, we introduce a Spectral-Aware Feature Alignment module to unify feature dimensionality and align cross-domain semantics in a community-aware manner. We further develop a Graph Diffusion Tokenized Transformer that constructs hybrid token sequences from local and global structural contexts for Transformer encoding, and applies diffusion-based refinement to mitigate distribution shifts on unseen graphs. With the unified representations, communities are efficiently retrieved via a modularity-driven search procedure. Extensive experiments on diverse real-world graphs demonstrate that OFA-CS achieves strong cross-domain generalization and competitive retrieval effectiveness against state-of-the-art methods, without requiring target-domain supervision.

Next point-of-interest (POI) recommendation plays a crucial role in enhancing user travel experiences and driving platform revenues by suggesting potentially appealing locations to users. Recent advancements have focused on capturing the general preferences and dynamic interests of users by modeling long- and short-term trajectories. However, existing long-term models struggle to accurately capture periodic user behaviors beyond simple distinctions such as weekdays/weekends or seasons. Meanwhile, short-term models often follow the assumption that users prefer to visit nearby locations, thereby overlooking the personalized spatial preferences. Furthermore, the interaction between the long- and short-term preferences remains underexplored. To address these gaps, we propose a novel model: personalized short-term and periodic long-term preferences modeling with contrastive learning for next POI recommendation. This model leverages the inherent similarities between a user’s periodic long-term and distance-based short-term preferences while distinguishing the travel preferences of different users, ultimately improving the accuracy of next POI predictions. Specifically, we introduce a spatial span graph (S${}^{2}$graph) to model the personalized distance span preferences. Additionally, we employ Mamba-based and discrete wavelet transform-based methods to capture long-term periodic patterns. Extensive experiments conducted on three real-world datasets demonstrate the superiority of our proposed model.

A corpus of recent work has revealed that authenticated query services have been under the spotlight due to the untrustworthiness of outsourced service provider. To enrich scalable functionality, there is an increasing demand for dynamically authenticated query. However, when implementing query and update simultaneously, traditional approaches heavily suffer from the inconsistency between verification digest and requested index and therefore are infeasible in reality. Moreover, the efficiency of storage, query, verification, and update is still a huge hinder when processing large scale data. To address these challenging issues, in this paper, we propose a novel idea of authenticated learned index that is carefully designed and actively optimized for authenticated query processing. Specifically, we first propose a version control update mechanism for consistency guarantee by maintaining historical index versions. Following this, we propose two basic authenticated learned indexes, i.e., query-friendly PVL-tree and update-friendly PVLB-tree, to support efficient scalable authenticated range query. Furthermore, to improve the efficiency, we introduce a hybrid index framework HPVL-tree based on two basic indexes. Extensive theoretical and experimental analysis demonstrate that our proposed HPVL-tree outperforms the state-of-the-art approaches by up to$2.28\times, 3.96\times$, and$2.51\times$in search time, update time, and verification time, respectively. Moreover, the storage overhead and communication overhead occupy only 38 % and 2.25 % of existing approach, respectively.

Entity alignment (EA) aims to identify equivalent entities across different knowledge graphs (KGs). While existing approaches leverage KG neighborhood structures for alignment, they often fail to effectively distinguish relevant from irrelevant neighbors due to insufficient handling of neighbor heterogeneity. Additionally, entity enhancement strategies remain underutilized. To address these issues, we propose a novel Relationship Structure Line Graph for Semi-supervised Entity Alignment Based on Edge Weight Adjustment, named RSGEA. It first enhances entity representations by deeply analyzing relational connectivity structures in KGs, capturing key relational information from second-order and triangular-ring structures. It then employs an attention mechanism to dynamically adjust edge weights, mitigating the impact of noisy edges during information propagation. Finally, we employ the Sinkhorn algorithm to refine the similarity matrix, improving alignment accuracy. Furthermore, we introduce an unsupervised version to accommodate diverse scenarios. Extensive experiments on five cross-lingual datasets validate the effectiveness and robustness of the RSGEA, demonstrating significant performance improvements.

SimRank is a significant metric to measure the similarity of nodes in graph data analysis. The problem of SimRank computation has been studied extensively, however there is no existing work that can provide one unified algorithm to support the SimRank computation both on static and temporal graphs. In this work, we first propose CrashSim, an index-free algorithm for single-source SimRank computation in static graphs. CrashSim can provide provable approximation guarantees for the computational results in an efficient way. In addition, as the reallife graphs are often represented as temporal graphs, CrashSim enables efficient computation of SimRank in temporal graphs. We formally define two typical SimRank queries in temporal graphs, and then solve them by developing an efficient algorithm based on CrashSim, called CrashSim-T. From the extensive experimental evaluation using five real-life and synthetic datasets, it can be seen that the CrashSim algorithm and CrashSim-T algorithm substantially improve the efficiency of the state-of-the-art SimRank algorithms by about 30%, while achieving the precision of the result set with about 97%.