Dynamic recommendation systems aim to provide personalized suggestions by modeling temporal user-item interactions across time-series behavioral data. Recent studies have leveraged pre-trained dynamic graph neural networks (GNNs) to learn user-item representations over temporal snapshot graphs. However, fine-tuning GNNs on these graphs often results in generalization issues due to temporal discrepancies between pre-training and fine-tuning stages, limiting the model’s ability to capture evolving user preferences. To address this, we propose TarDGR, a task-aware retrieval-augmented framework designed to enhance generalization capability by incorporating task-aware model and retrieval-augmentation. Specifically, TarDGR introduces a Task-Aware Evaluation Mechanism to identify semantically relevant historical subgraphs, enabling the construction of task-specific datasets without manual labeling. It also presents a Graph Transformer-based Task-Aware Model that integrates semantic and structural encodings to assess subgraph relevance. During inference, TarDGR retrieves and fuses task-aware subgraphs with the query subgraph, enriching its representation and mitigating temporal generalization issues. Experiments on multiple large-scale dynamic graph datasets demonstrate that TarDGR consistently outperforms state-of-the-art methods, with extensive empirical evidence underscoring its superior accuracy and generalization capabilities.

Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language processing tasks, yet their application to graph structure analysis, particularly in community search, remains underexplored. Community search, a fundamental task in graph analysis, aims to identify groups of nodes with dense interconnections, which is crucial for understanding the macroscopic structure of graphs. In this paper, we propose GraphCS, a comprehensive benchmark designed to evaluate the performance of LLMs in community search tasks. Our experiments reveal that while LLMs exhibit preliminary potential, they frequently fail to return meaningful results and suffer from output bias. To address these limitations, we introduce CS-Agent, a dual-agent collaborative framework to enhance LLM-based community search. CS-Agent leverages the complementary strengths of two LLMs acting as Solver and Validator. Through iterative feedback and refinement, CS-Agent dynamically refines initial results without fine-tuning or additional training. After the multi-round dialogue, Decider module selects the optimal community. Extensive experiments demonstrate that CS-Agent significantly improves the quality and stability of identified communities compared to baseline methods. To our knowledge, this is the first work to apply LLMs to community search, bridging the gap between LLMs and graph analysis while providing a robust and adaptive solution for real-world applications.

Approximate Vector Retrieval (AVR), which aims to efficiently retrieve the most similar items from a large dataset, is a fundamental task in a variety of applications such as information retrieval, recommender systems, and large language models. Advances in representation learning and multimodal neural models have enabled diverse data types (e.g., text, images, audio) to be embedded into a shared vector space, facilitating similarity-based retrieval in AVR. While single-modal AVR assumes query and database embeddings follow the same distribution (In-Distribution, ID), cross-modal AVR introduces a distribution shift, where query vectors (e.g., text) are Out-of-Distribution (OOD) relative to the database (e.g., images). This mismatch complicates retrieval and degrades accuracy, making it a key challenge in AVR. Existing methods typically focus on either ID or OOD queries but struggle to handle both within a unified framework.

Public transportation plays a vital role in mitigating traffic congestion and reducing carbon emissions. The Top-k Nearest Neighbor ( k NN) search in public transportation networks is a fundamental problem in location-based services, which aims to find k nearest objects from a given query point. The traditional method, Dijkstra's algorithm has been employed to tackle the k NN problem, however, it is notably inefficient in processing queries. While other works precompute an index to speed up query processing. However, they are still slow in processing queries. Furthermore, they cannot scale to large graphs due to their reliance on resource-intensive path indexes. To address these limitations, we introduce a novel index-based approach that utilizes a simple yet effective index structure to handle k NN queries with a near-optimal time complexity. The index does not rely on a path index, making it efficient to construct and scalable to large graphs. Extensive experiments are conducted on real-world datasets to demonstrate the efficiency and scalability of our approach. The results show that our approach outperforms existing solutions by up to four orders of magnitude in query processing and two orders of magnitude in index construction.

With the ever-increasing prevalence of graph data in a wide spectrum of applications, it becomes essential to analyze structural trends in dynamic graphs on a continual basis. The shortest cycle is a fundamental pattern in graph analytics. In this paper, we investigate the problem of shortest cycle counting for a given vertex in dynamic graphs in light of its applicability to problems such as fraud detection. To address such queries efficiently, we propose a 2-hop labeling based algorithm called Counting Shortest Cycle (CSC for short). Additionally, techniques for dynamically updating the CSC index are explored. Comprehensive experiments are conducted to demonstrate the efficiency and effectiveness of our method. In particular, CSC enables query evaluation in a few hundreds of microseconds for graphs with millions of edges, and improves query efficiency by two orders of magnitude when compared to the baseline solutions. Also, the update algorithm could efficiently cope with edge insertions (deletions).