Advancements in model algorithms, the growth of foundational models, and access to high-quality datasets have propelled the evolution of Artificial Intelligence Generated Content (AIGC). Despite its notable successes, AIGC still faces hurdles such as updating knowledge, handling long-tail data, mitigating data leakage, and managing high training and inference costs. Retrieval-augmented generation (RAG) has recently emerged as a paradigm to address such challenges. In particular, RAG introduces the information retrieval process, which enhances the generation process by retrieving relevant objects from available data stores, leading to higher accuracy and better robustness. In this paper, we comprehensively review existing efforts that integrate RAG techniques into AIGC scenarios. We first classify RAG foundations according to how the retriever augments the generator, distilling the fundamental abstractions of the augmentation methodologies for various retrievers and generators. This unified perspective encompasses all RAG scenarios, illuminating advancements and pivotal technologies that help with potential future progress. We also summarize additional enhancement methods for RAG, facilitating effective engineering and implementation of RAG systems. Then from another view, we survey practical applications of RAG across different modalities and tasks, offering valuable references for researchers and practitioners. Furthermore, we introduce the benchmarks for RAG, discuss the limitations of current RAG systems, and suggest potential directions for future research.

Knowledge graph-based recommendation systems have strong capabilities in deep association mining and structured reasoning, which effectively alleviate data sparsity and cold-start problems in recommendation. However, traditional knowledge graph-based recommendation considers each graph relation in isolation, failing to capture the potential semantic correlations between different relation types, which leads to incomplete semantic representations. In addition, existing methods generally ignore negative feedback signals in users' historical interactions, resulting in biased preference modeling. To overcome the above challenges, we introduce Sign-aware Recommendation based on Virtual Semantic Knowledge Graph(SRVSKG), which enhances recommendation performance by combining virtual semantic collaborative representations with sign-aware learning techniques. Firstly, we propose a virtual semantic subgraph collaborative representation module to learn user and item embeddings. In this module, we build virtual semantic subgraphs through relation clustering based on latent semantic similarity measurement. Then a hierarchical feature extraction mechanism based on graph attention network is applied to virtual semantic subgraphs, which captures semantic associations across relations and enriches item embedding. At the same time, we emphasize user preference for attributes to enrich user embedding. Secondly, we design a sign-aware learning module, which constructs a user-item signed graph, applies Laplacian matrix factorization to simultaneously model the topological features of positive and negative feedback, and introduces the transformer architecture to dynamically fuse signed information, effectively utilizing negative feedback information to eliminate bias in user preference modeling. Finally, we establish a multi-feature fusion mechanism that deeply combines features from the virtual semantic subgraph collaborative representation module and the sign-aware learning module to enable recommendation. Experiment results on three public datasets demonstrate that SRVSKG outperforms state-of-the-art recommendation baselines.

Opinion maximization has attracted much attention in viral marketing. It selects an initial seed user set to disseminate user opinions on the target product and finally produces more positive opinions in social networks. In earlier studies, a critical but not studied problem is the fairness of information dissemination in groups with sensitive characteristic (such as age or race). People prefer to promote products for target users (majority groups) rather than in sensitive characteristic groups (minority groups). That leads to the differences in information dissemination between minority groups. In addition, social networks have the in-depth structural information. Therefore, in this paper, we design opinion maximization based on fairness framework (OMBF) using graph attention networks (GAT) to exploit more network information, and only consider the fairness in minority groups. OMBF composes of three parts: (1) the determination of candidate nodes according to node representations, (2) the dynamic changes in opinions, (3) the selection of final seed nodes. Firstly, we utilize GAT to obtain node representations and to determinate candidate nodes and design a node opinion formation model to model the dynamic changes in opinions. Then, we use the fair constraint value to ensure the fairness in the information dissemination process of minority groups. Based on above, final seed nodes are selected. We conduct experiments on synthetic and real-world datasets to show the effectiveness of our approach. The results indicate that the total opinions of active nodes in all nodes and fair values in minority groups are better than the chosen state-of-the-art benchmarks.

The high sampling frequency of highway structural health monitoring systems brings a heavy burden on data storage. However, existing dynamic response identification approaches can guarantee either reduced data volume after identification or high accuracy of dynamic response identification. Motivated by this, we propose a real-time dynamic response identification method to filter meaningless data. Our method not only selects effective features from highway structural health monitoring data, but also designs a training data generation strategy for machine learning models within the dynamic response identification framework. Experimental results on real highway structural health monitoring data demonstrate that our proposed approach spends 0.4 ms to process the monitoring data generated in 1 s and saves around 91.63% storage space. Also, the recall value of our method achieves 0.91 on average.

Recommender systems are widely applied in numerous online platforms such as shopping and social media platforms. They typically utilize large embedding tables that map users and items to dense vectors of uniform sizes. As the number of users and items continues to grow, this design leads to significant memory consumption and computational inefficiencies. This challenge is particularly pronounced in scenarios such as federated learning, where model parameters are updated locally on edge devices with limited computational resources before being transmitted to a central server for aggregation. Numerous approaches have been proposed to address this issue, among which embedding pruning methods have emerged as a compelling solution. Compared to parameter-sharing and variable-size embedding techniques, embedding pruning methods offer lower training costs and leverage sparse embeddings for improved efficiency. Notably, embedding pruning methods based on the Dynamic Sparse Training (DST) paradigm maintain consistent sparsity throughout training and provide a controllable memory budget, establishing them as state-of-the-art lightweight embedding solutions for resource-constrained environments. However, embedding pruning methods are not without limitations. First, despite the use of sparse embeddings during forward passes, dense gradients are still computed in backward passes, introducing inefficiencies. Second, DST's weight exploration mechanism tends to prioritize users or items from the most recent batch, reactivating pruned parameters that do not necessarily enhance overall performance. In this work, we introduce SparseRec, a lightweight embedding method designed to overcome these obstacles. SparseRec accumulates gradients to better identify inactive parameters that, when reactivated, contribute more meaningfully to model performance. Additionally, SparseRec avoids dense gradient computation during backpropagation by selectively sampling key vectors. Gradients are calculated only for parameters in this subset, ensuring sparsity throughout both forward and backward passes. Experiments on three benchmark datasets show that SparseRec achieves up to 11.79% performance gains across three base recommenders and multiple density configurations, highlighting its effectiveness in optimizing memory-constrained recommendation systems.

Keyword search, which identifies transactions associated with specified keywords across historical blocks, is a critical query type in blockchain analytics. However, existing approaches, such as on-chain indexing and off-chain synchronization, may lead to significant space overhead or challenges in maintaining data freshness. To address these challenges, we propose BlockSketch, a novel probabilistic data structure (PDS) that adopts a differentiated encoding strategy, aimed at resolving the trade-off between query performance and storage overhead in blockchain indexing. BlockSketch features a hierarchical filtering architecture that combines Bloom filters and Sketches within a binary tree framework, enabling dynamic structural maintenance. Keywords are categorized as "hot" or "cold" based on their on-chain frequency and encoded into the most suitable component to achieve resource-efficient storage and accurate querying. In addition, Block-Sketch integrates two distinct query rules, namely "level-down" and "jump," to balance query accuracy and efficiency when processing keywords with varying frequencies. Furthermore, we enhance the query efficiency of BlockSketch by merging inefficient lower-level nodes into more compact ones and pruning redundant node checks during query execution. Extensive experiments on a real-world dataset demonstrate that BlockSketch delivers up to 73% faster query processing, achieves 44.56% of the average false positive rate of baselines at low multiplicity and as low as 1.52% at high multiplicity, and saves 79% in storage compared to state-of-the-art methods.

Improving the performance of data systems for join operations has long been a critical challenge. Recently, substantial attention has been focused on optimizing multi-way join performance, particularly in reducing the overhead caused by generating intermediate tuples that do not contribute to the final result. In this paper, we propose a novel algorithm called SieveJoin, which extends the established Bloomjoin approach to support multi-way joins. SieveJoin sets a new benchmark for the efficiency of join query execution. A key innovation of SieveJoin is its ability to propagate Bloom filters along the join path, allowing the system to terminate early and avoid producing superfluous intermediate results. The primary design objective of SieveJoin is to efficiently estimate join results using Bloom filters, while maintaining minimal memory overhead. We analyze the bottlenecks associated with deferred multi-way joins and detail how Bloom filters are utilized to suppress the creation of redundant intermediate tuples. To assess the effectiveness of SieveJoin, we conduct a comprehensive experimental evaluation using the TPC-H benchmark, citation datasets, and a synthetic dataset. Our results compare SieveJoin with a state-of-the-art column-store database and a worst-case optimal join algorithm, highlighting its advantages in both response time and memory usage.

The rapid proliferation of images on online platforms has made emotion analysis a task of paramount significance. However, these images are often privacy-sensitive, making Federated Learning (FL) a compelling paradigm over traditional centralized methods. A critical yet largely unaddressed challenge in applying FL to this domain is the severe concept drift stemming from the subjective and culturally diverse nature of emotional expression, which causes conventional FL algorithms to fail. In this paper, we propose CAFL (Conditional Attention Federated Learning) to fill this gap. CAFL empowers clients to learn collaboratively yet personally. It intelligently routes information through an adaptive gate that separates features into a personalized stream and a global stream. These streams are then processed by dedicated local and global prediction heads. Crucially, collaboration is guided by a conditional attention mechanism, where the server computes a personalized reference model for each client based on an attention-weighted aggregation of peer models, promoting knowledge sharing among kindred clients. Extensive experiments on various lightweight foundation models show that CAFL consistently outperforms existing FL methods, demonstrating its robustness and superior performance as a solution for distributed, privacy-sensitive image emotion analysis.

Dynamic time warping (DTW), a typical elastic similarity measure that compares one-to-many points, has been proven effective for various time-series data mining tasks. However, it requires a quadratic time complexity O(n²) proportional to the length of time-series data, which undermines its applications involving long time series. In this paper, a representation-based similarity measure called Dynamic Sub-Sequence Warping (DSSW) is proposed. Instead of working on the raw data directly, we perform data representation to extract the distributional features of time series. Then, the similarity between two time series is measured by aligning the corresponding sub-sequences composed of the extracted features. We evaluate the proposed method through a supervised learning task on extensive real-world datasets. The results show that DSSW outperforms the prevalent DTW-based methods in terms of precision, and achieves one order of magnitude faster execution time on average compared with DTW.

Real-world datasets frequently exhibit evolving data distributions, reflecting temporal variations and underlying shifts. Overlooking this phenomenon, known as concept drift, can substantially degrade the predictive performance of the model. Furthermore, the presence of hyperparameters in online models exacerbates this issue, as these parameters are typically fixed and lack the flexibility to dynamically adjust to evolving data. This paper introduces "OLR-WAA: An Adaptive and Drift-Resilient Online Regression with Dynamic Weighted Average", a hyperparameter-free model designed to tackle the challenges of non-stationary data streams and enable effective, continuous adaptation. The objective is to strike a balance between model stability and adaptability. OLR-WAA incrementally updates its base model by integrating incoming data streams, utilizing an exponentially weighted moving average. It further introduces a unique optimization mechanism that dynamically detects concept drift, quantifies its magnitude, and adjusts the model based on real-time data characteristics. Rigorous evaluations show that it matches batch regression performance in static settings and consistently outperforms or rivals state-of-the-art online models, confirming its effectiveness. Concept drift datasets reveal a performance gap that OLR-WAA effectively bridges, setting it apart from other online models. In addition, the model effectively handles confidence-based scenarios through a conservative update strategy that prioritizes stable, high-confidence data points. Notably, OLR-WAA converges rapidly, consistently yielding higher R² values compared to other online models.