Continual Knowledge Graph Embedding (CKGE) aims to continually learn embeddings for new knowledge, i.e., entities and relations, while retaining previously acquired knowledge. Most existing CKGE methods mitigate catastrophic forgetting via regularization or replaying old knowledge. They conflate new and old knowledge of an entity within the same embedding space to seek a balance between them. However, entities inherently exhibit multi-faceted semantics that evolve dynamically as their relational contexts change over time. A shared embedding fails to capture and distinguish these temporal semantic variations, degrading lifelong link prediction accuracy across snapshots. To address this, we propose a Multi-Faceted CKGE framework (MF-CKGE) for semantic-aware link prediction. During offline learning, MF-CKGE separates temporal old and new knowledge into distinct embedding spaces to prevent knowledge entanglement and employs semantic decoupling to reduce semantic redundancy, thereby improving space efficiency. During online inference, MF-CKGE adaptively identifies semantically query-relevant entity embeddings by quantifying their semantic importance, reducing interference from query-irrelevant noise. Experiments on eight datasets show that MF-CKGE achieves an average (maximum) improvement of 1.7% (2.7%) and 1.4% (3.8%) in MRR and Hits@10, respectively, over the best baseline. Our source code and datasets are available at: https://anonymous.4open.science/r/MF-CKGE-04E5.

Community search on attributed graphs (CSAG) is a fundamental topic in graph data mining. Given an attributed graph G and a query node q , CSAG seeks a structural- and attribute-cohesive subgraph from G that contains q . Exact methods based on graph traversal are time-consuming, especially for large graphs. Approximate methods improve efficiency by pruning the search space with heuristics but still take hundreds of milliseconds to tens of seconds to respond, hindering their use in time-sensitive applications. Moreover, pruning strategies are typically tailored to specific algorithms and their cohesiveness metrics, making them difficult to generalize. To address this, we study a general approach to accelerate various CSAG methods. We first present a proximity graph-based, cohesiveness-aware hierarchical index that accommodates different cohesiveness metrics. Then, we present two optimizations to enhance the index's navigability and reliability. Finally, we design a compressed storage structure for space-efficient indexing. Experiments on real-world datasets show that integrating our index with existing mainstream CSAG methods results in an average 30.7× speedup while maintaining a comparable or even better attribute cohesiveness.

Attributed graphs are extensively utilized in marketing, friend recommendations, disease prediction, etc. In attributed graphs, nodes are associated with attributes to enrich the graph representation, while edges indicate relationships between nodes. However, ensuring data privacy when publishing attributed graphs is a significant challenge due to the sensitive nature of both attributes and relationships. Existing methods fail to preserve graph structures effectively and neglect correlations among node attributes, leading to diminished utility for published synthetic graphs. To address these issues, we propose PrivAGS, a framework for publishing attributed graphs with Rényi Differential Privacy (RDP) guarantees. PrivAGS reconstructs graph structures and attributes based on community structures to capture tightly connected features. We propose a bounded Gaussian threshold mechanism to preserve attribute correlations and utilize probabilistic graph models with optimized inference structures to infer distributions and release node attributes. Additionally, PrivAGS introduces a new structural model, MCEG, to capture clustering structures and enable efficient graph reconstruction. Extensive experiments on five real-world datasets show that PrivAGS generates privacy-preserving, high-utility synthetic data.

Given a graph G and a query node q, community search (CS) aims to find a structurally cohesive subgraph from G that contains q. CS is widely used in many real-world applications, such as online recommendation and expert finding. Recently, the rise of learning-based CS methods has garnered extensive research interests, showcasing the promising potential of neural solutions. However, there remains room for optimization: (1) They initialize node features via classical methods, e.g., one-hot, random, and position encoding, which may fall short in capturing valuable community cohesiveness-related features. (2) The reliance on GCN or GCN-like models poses challenges in scaling to large graphs. (3) Existing methods do not adapt well to dynamic graphs, often requiring retraining from scratch. To handle this, we present CSFormer, a scalable CS based on Graph Transformer. First, we present a novel l-hop neighborhood community vector based on n-order h-index to represent each node's community features, generating a sequence of feature vectors by varying the neighborhood scope l. Then, we build a Transformer backbone to learn a good graph embedding that carries rich community features, based on which we perform a prediction-filtering-based online CS to efficiently return a community of q. We extend CSFormer to dynamic graphs and various community models. Extensive experiments on seven real-world graphs show our solution's superiority on effectiveness, e.g., we attain an average improvement of 20.6% in F1-score compared to the latest competitors.

High-dimensional vector similarity search (HVSS) is gaining prominence as a powerful tool for various data science and AI applications. As vector data scales up, in-memory indexes pose a significant challenge due to the substantial increase in main memory requirements. A potential solution involves leveraging disk-based implementation, which stores and searches vector data on high-performance devices like NVMe SSDs. However, implementing HVSS for data segments proves to be intricate in vector databases where a single machine comprises multiple segments for system scalability. In this context, each segment operates with limited memory and disk space, necessitating a delicate balance between accuracy, efficiency, and space cost. Existing disk-based methods fall short as they do not holistically address all these requirements simultaneously. In this paper, we present Starling, an I/O-efficient disk-resident graph index framework that optimizes data layout and search strategy within the segment. It has two primary components: (1) a data layout incorporating an in-memory navigation graph and a reordered disk-based graph with enhanced locality, reducing the search path length and minimizing disk bandwidth wastage; and (2) a block search strategy designed to minimize costly disk I/O operations during vector query execution. Through extensive experiments, we validate the effectiveness, efficiency, and scalability of Starling. On a data segment with 2GB memory and 10GB disk capacity, Starling can accommodate up to 33 million vectors in 128 dimensions, offering HVSS with over 0.9 average precision and top-10 recall rate, and latency under 1 millisecond. The results showcase Starling's superior performance, exhibiting 43.9x higher throughput with 98% lower query latency compared to state-of-the-art methods while maintaining the same level of accuracy.

Expert finding is crucial for a wealth of applications in both academia and industry. Given a user query and trove of academic papers, expert finding aims at retrieving the most relevant experts for the query, from the academic papers. Existing studies focus on embedding-based solutions that consider academic papers’ textual semantic similarities to a query via document representation and extract the top- n experts from the most similar papers. Beyond implicit textual semantics, however, papers’ explicit relationships (e.g., co-authorship) in a heterogeneous graph (e.g., DBLP) are critical for expert finding, because they help improve the representation quality. Despite their importance, the explicit relationships of papers generally have been ignored in the literature. In this article, we study expert finding on heterogeneous graphs by considering both the explicit relationships and implicit textual semantics of papers in one model. Specifically, we define the cohesive ( k , 𝒫)-core community of papers w.r.t. a meta-path 𝒫 (i.e., relationship) and propose a ( k , 𝒫)-core based document embedding model to enhance the representation quality. Based on this, we design a proximity graph-based index (PG-Index) of papers and present a threshold algorithm (TA)-based method to efficiently extract top- n experts from papers returned by PG-Index. We further optimize our approach in two ways: (1) we boost effectiveness by considering the ( k , 𝒫)-core community of experts and the diversity of experts’ research interests, to achieve high-quality expert representation from paper representation; and (2) we streamline expert finding, going from “extract top- n experts from top- m ( m> n ) semantically similar papers” to “directly return top- n experts”. The process of returning a large number of top- m papers as intermediate data is avoided, thereby improving the efficiency. Extensive experiments using real-world datasets demonstrate our approach’s superiority.

Finding relevant experts in specified areas is often crucial for a wide range of applications in both academia and industry. Given a user input query and a large amount of academic knowledge (e.g., academic papers), expert finding aims to find and rank the experts who are most relevant to the given query, from the academic knowledge. Existing studies mainly focus on the embedding-based solutions that (1) consider academic papers' textual semantic similarities to a given query through document representation models and (2) extract the$\mathbf{top-}n$experts with the greatest similarities. Beyond the implicit textual semantics of papers, however, the papers' explicit relationships (e.g., co-authorship, citation, and same-topic relationship) in a heterogeneous academic graph (e.g., DBLP) are critical for document representation, insofar as they help improve the expert finding quality. Despite their importance, the explicit relationships of papers generally have been ignored in the literature. In this paper, we study the academic expert finding on heterogeneous graphs by considering the explicit relationships besides the implicit textual semantics of papers in one representation model. Specifically, we first define the$(k,\mathcal{P})$-core to denote a cohesive community of papers that are closely connected via a meta-path$\mathcal{P}(\mathcal{P}$indicates the different relationships of papers). We then propose an offline$(k,\mathcal{P})$-core based document embedding model to capture papers' various explicit relationships for representation. Moreover, by using papers' embeddings, we present an online threshold algorithm (TA)-based method to efficiently return top-n experts via a carefully designed proximity graph-based index (PG-Index). We extend our approach to support multiple relationships simultaneously for representation. Extensive experiments over real-world datasets demonstrate the effectiveness and efficiency of our approach.

Approximate nearest neighbor search (ANNS) constitutes an important operation in a multitude of applications, including recommendation systems, information retrieval, and pattern recognition. In the past decade, graph-based ANNS algorithms have been the leading paradigm in this domain, with dozens of graph-based ANNS algorithms proposed. Such algorithms aim to provide effective, efficient solutions for retrieving the nearest neighbors for a given query. Nevertheless, these efforts focus on developing and optimizing algorithms with different approaches, so there is a real need for a comprehensive survey about the approaches' relative performance, strengths, and pitfalls. Thus here we provide a thorough comparative analysis and experimental evaluation of 13 representative graph-based ANNS algorithms via a new taxonomy and fine-grained pipeline. We compared each algorithm in a uniform test environment on eight real-world datasets and 12 synthetic datasets with varying sizes and characteristics. Our study yields novel discoveries, offerings several useful principles to improve algorithms, thus designing an optimized method that outperforms the state-of-the-art algorithms. This effort also helped us pinpoint algorithms' working portions, along with rule-of-thumb recommendations about promising research directions and suitable algorithms for practitioners in different fields.