Retrieval-Augmented Generation (RAG) has significantly enhanced Large Language Models' ability to access external knowledge, yet current graph-based RAG approaches face two critical limitations in managing hierarchical information: they impose rigid layer-specific compression quotas that damage local graph structures, and they prioritize topological structure while neglecting semantic content. We introduce T-Retriever, a novel framework that reformulates attributed graph retrieval as tree-based retrieval using a semantic and structure-guided encoding tree. Our approach features two key innovations: (1) Adaptive Compression Encoding, which replaces artificial compression quotas with a global optimization strategy that preserves the graph's natural hierarchical organization, and (2) Semantic-Structural Entropy (S²-Entropy), which jointly optimizes for both structural cohesion and semantic consistency when creating hierarchical partitions. Experiments across diverse graph reasoning benchmarks demonstrate that T-Retriever significantly outperforms state-of-the-art RAG methods, providing more coherent and contextually relevant responses to complex queries.

Vision-language model (VLM) based web agents demonstrate impressive autonomous GUI interaction but remain vulnerable to deceptive interface elements.Existing approaches either detect deception without task integration or document attacks without proposing defenses.We formalize deception-aware web agent defense and propose DUDE (Deceptive UI Detector & Evaluator), a two-stage framework combining hybrid-reward learning with asymmetric penalties and experience summarization to distill failure patterns into transferable guidance.We introduce RUC (Real UI Clickboxes), a benchmark of 1,407 scenarios spanning four domains and deception categories.Experiments show DUDE reduces deception susceptibility by 53.8% while maintaining task performance, establishing an effective foundation for robust web agent deployment. 1

With the widespread adoption of cloud-based services and Service-Oriented Computing, efficient service recommendation has become pivotal for optimizing service discovery and composition in large-scale ecosystems. While recent diffusion-based recommendation methods have achieved impressive results in service-oriented scenarios, existing approaches predominantly treat user-service interactions as isolated events, overlooking the potential of higher-order collaborative signals between users and services. Such signals, which encapsulate richer and more nuanced relationships, can be naturally captured using graph-based data structures. To address this limitation, we extend diffusion-based service recommendation methods to the graph domain by directly modeling user-service bipartite graphs with diffusion models. This enables better modeling of the higher-order connectivity inherent in complex interaction dynamics. However, this extension introduces two primary challenges: (1) Noise Heterogeneity, where interactions are influenced by various forms of continuous and discrete noise, and (2) Relation Explosion, referring to the high computational costs of processing large-scale graphs. To tackle these challenges, we propose a Graph-based Diffusion Model for Service Recommendation (GDMSR). To address noise heterogeneity, we introduce a multi-level noise corruption mechanism that integrates both continuous and discrete noise, effectively simulating real-world interaction complexities. To mitigate relation explosion, we design a user-active guided diffusion process that selectively focuses on the high-value edges and active users, reducing inference costs while preserving critical service-level dependencies. Extensive experiments on six real-world service datasets demonstrate that GDMSR consistently outperforms state-of-the-art methods, highlighting its effectiveness in capturing higher-order collaborative signals and improving service recommendation performance.

Graph anomaly detection faces significant challenges due to the scarcity of reliable anomaly-labeled datasets, driving the development of unsupervised methods. Graph autoencoders (GAEs) have emerged as a dominant approach by reconstructing graph structures and node features while deriving anomaly scores from reconstruction errors. However, relying solely on reconstruction error for anomaly detection has limitations, as it increases the sensitivity to noise and overfitting. To address these issues, we propose Graph Evidential Learning (GEL), a probabilistic framework that redefines the reconstruction process through evidential learning. By modeling node features and graph topology using evidential distributions, GEL quantifies two types of uncertainty: graph uncertainty and reconstruction uncertainty, incorporating them into the anomaly scoring mechanism. Extensive experiments demonstrate that GEL achieves state-of-the-art performance while maintaining high robustness against noise and structural perturbations.

Large Language Models (LLMs) face significant limitations when applied to large-scale graphs, struggling with context constraints and inflexible reasoning. We introduce GraphChain, a novel framework enabling LLMs to analyze large graphs by orchestrating dynamic sequences of specialized tools, mimicking human exploratory processes. GraphChain incorporates two core technical contributions: (1) Progressive Graph Distillation, a reinforcement learning approach that learns to generate tool sequences balancing task relevance and intermediate state compression, thereby overcoming LLM context limitations. (2) Structure-aware Test-Time Adaptation (STTA), a mechanism using a lightweight, self-supervised adapter conditioned on graph spectral properties to efficiently adapt a frozen LLM policy to diverse graph structures via soft prompts without retraining. Experiments show GraphChain significantly outperforms prior methods, enabling scalable and adaptive LLM-driven graph analysis.

Graph anomaly detection (GAD) has become a critical research area, with successful applications in financial fraud and telecommunications. Traditional Graph Neural Networks (GNNs) face significant challenges: at the topology level, they suffer from over-smoothing that averages out anomalous signals; at the feature level, discriminative models struggle when fraudulent nodes obfuscate their features to evade detection. In this paper, we propose a Conditional Graph Anomaly Diffusion Model (CGADM) that addresses these issues through the iterative refinement and denoising reconstruction properties of diffusion models. Our approach incorporates a prior-guided diffusion process that injects a pre-trained conditional anomaly estimator into both forward and reverse diffusion chains, enabling more accurate anomaly detection. For computational efficiency on large-scale graphs, we introduce a prior confidence-aware mechanism that adaptively determines the number of reverse denoising steps based on prior confidence. Experimental results on benchmark datasets demonstrate that CGADM achieves state-of-the-art performance while maintaining significant computational advantages for large-scale graph applications.

Pedestrian trajectory prediction is a well-established task with significant recent advancements. However, existing datasets are unable to fulfill the demand for studying minute-level long-term trajectory prediction, mainly due to the lack of high-resolution trajectory observation in the wide field of view (FoV). To bridge this gap, we introduce a novel dataset named GigaTraj, featuring videos capturing a wide FoV with ~4 ×104m2and high-resolution imagery at the gigapixel level. Furthermore, GigaTraj in-cludes comprehensive annotations such as bounding boxes, identity associations, world coordinates, group/interaction relationships, and scene semantics. Leveraging these multimodal annotations, we evaluate and validate the state-of-the-art approaches for minute-level long-term trajectory prediction in large-scale scenes. Extensive experiments and analyses have revealed that long-term prediction for pedestrian trajectories presents numerous challenges, indicating a vital new direction for trajectory research. The dataset is available at WWW.gigavision ai.

Driven by Service-Oriented Computing, time-aware service recommendation aims to support personalized mashup development, adapting to the rapid shifts of users’ dynamic preferences. Recently, users’ social connections have shown significant benefits to time-aware service recommendation, and graph neural networks have demonstrated great success in learning the pattern of information flow among users. However, the current paradigm always presumes a given social network, which is not necessarily consistent with the similarities of service preferences among users and is expensive to collect for most service platforms. We propose a novel idea to learn the graph structure among historical mashups and make time-aware service recommendation for dynamic mashup creation collectively in a coupled framework. This idea raises two challenges, i.e., scalability and accuracy. To solve both challenges simultaneously, we introduce the Dynamic Relation Graph Learning (DRGL) framework for time-aware service recommendation. For scalability, our framework has a coarse-to-fine recalling strategy to learn the graph structure among the mashups, which enables the exploration of potential links among all historical mashups while maintaining a tractable amount of computation. For accuracy, we leverage recent advances in self-attention mechanisms to the mashup modeling and propose a transformer-based mashup encoder, which considers long-range dependencies in dense mashups for more accurate mashup representations. Extensive experiments show that the DRGL model consistently outperforms the state-of-the-art methods in terms of prediction accuracy for mashup creation.

As the adoption of Service-Oriented Computing continues to grow, the number of web services has increased significantly, which makes service recommendation become an essential tool to assist users in selecting suitable services. However, a single service cannot satisfy the complex requirements of users, which has led to the emergence of a new technique known as Mashup, which combines services as reusable components to create value-added service compositions. Along with mashup, mashup recommendation has also become an indispensable and important component of service platforms. On service platforms, there are many heterogeneous entities and complex relationships between them. We divide these interaction into three different views: Mashup-Invocation view, Service-Consumption view, and Mashup-Composition view. As user preferences and characteristics of services and mashups are distributed across different views, their cooperation is crucial for accurate mashup recommendation. Therefore, we propose Cross-view Graph Alignment (CGA), a framework that captures the collaborative associations dispersed across different views and enhances the representation learning of users and mashups. This the first study to jointly tackle structure- and representation-level collaboration on the service platforms for better mashup recommendation. Experiments on two real-world service datasets show that CGA outperforms state-of-the-art methods and can better improve the mashup recommendation.

Graph augmentation plays a crucial role in achieving good generalization for contrastive graph self-supervised learning. However, mainstream Graph Contrastive Learning (GCL) often favors random graph augmentations, by relying on random node dropout or edge perturbation on graphs. Random augmentations may inevitably lead to semantic information corruption during the training, and force the network to mistakenly focus on semantically irrelevant environmental background structures. To address these limitations and to improve generalization, we propose a novel self-supervised learning framework for GCL, which can adaptively screen the semantic-related substructure in graphs by capitalizing on the proposed gradient-based Graph Contrastive Saliency (GCS). The goal is to identify the most semantically discriminative structures of a graph via contrastive learning, such that we can generate semantically meaningful augmentations by leveraging on saliency. Empirical evidence on 16 benchmark datasets demonstrates the exclusive merits of the GCS-based framework. We also provide rigorous theoretical justification for GCS’s robustness properties. Code is available at https://github.com/GCS2023/GCS .

Highly skewed long-tail item distribution commonly hurts model performance on tail items in recommendation systems, especially for graph-based recommendation models. We propose a novel idea to learn relations among items as an auxiliary graph to enhance the graph-based representation learning and make recommendations collectively in a coupled framework. This raises two challenges, 1) the long-tail downstream information may also bias the auxiliary graph learning, and 2) the learned auxiliary graph may cause negative transfer to the original user-item bipartite graph. We innovatively propose a novel Meta Graph Learning framework for long-tail recommendation (MGL) for solving both challenges. The meta-learning strategy is introduced to the learning of an edge generator, which is first tuned to reconstruct a debiased item co-occurrence matrix, and then virtually evaluated on generating item relations for recommendation. Moreover, we propose a popularity-aware contrastive learning strategy to prevent negative transfer by aligning the confident head item representations with those of the learned auxiliary graph. Experiments on public datasets demonstrate that our proposed model significantly outperforms strong baselines for tail items without compromising the overall performance.

Next location recommendation services play a pivotal role in Location-Based Social Networks (LBSNs) due to their ability to provide personalized recommendations of attractive destinations, resulting in substantial benefits for both users and service providers. Recent research indicates that these services are influenced by both sequential and geographical factors. However, we argue that most of these services fail to fully exploit the latent multi-group knowledge of location semantics and user preferences, resulting in suboptimal performance. Therefore, we propose STMGCL, a novel spatial-temporal multi-group contrastive learning-based method to discover intrinsic multi-group information for improving next location recommendation services. Specifically, STMGCL designs Spatial Group Contrastive Learning (SGCL) to extract multiple group knowledge regarding location semantics. Additionally, it develops Temporal Group Contrastive Learning (TGCL) to explore multiple user preference group information through a self-attention based encoder. Finally, we leverage a multi-task learning strategy and a generalized Expectation Maximization (EM) algorithm to ensure that STMGCL is optimized end-to-end with guaranteed convergence. Extensive experiments conducted on four real-world datasets demonstrate the superior performance of STMGCL over baselines.

Driven by Service-Oriented Computing techniques, time-aware service recommendation aims to support personalized mashup development, adapting to the rapid shifts of users' dynamic preferences. Recent studies have revealed that users' social connections may help better model their dynamic preferences. However, two phenomena exist to influence users' dynamic preferences of service selection. First, users and their friends may only share preferences in certain services, which means not every service in the friends' consumed mashups has the same impact on a target user's dynamic preference. Second, for a target user, friends in his social network with similar interests and behaviors may contribute more influence intensities. To cover the above phenomena synergistically, this paper proposes a Social-powered Graph Hierarchical Attention Network (SGHAN), as a deep learning model capable of learning similar behaviors from proper friends during mashup development. SGHAN is powered by the reciprocity between its two core components: a service-level attentional encoder captures users' interested services in friends' mashups, while a friend-level graph attention network selects informative friends and propagates the friends' social influences. Extensive experiments show that the SGHAN model consistently outperforms the state-of-the-art methods in terms of prediction accuracy for mashup creation.

Hypergraph-based collaborative filtering for recommendations has emerged as an important research topic due to its ability to model complex relations among users and items. However, most existing methods typically construct the hypergraph structures using heuristics (e.g., motifs and jump connections) based on existing graphs (e.g., user-item bipartite graphs and social networks). From a learning perspective, we argue that the fixed heuristic topology of hypergraph may become a limitation and thus potentially compromise the recommendation performance. To tackle this issue, we propose a novel dynamic hypergraph learning framework for collaborative filtering (DHLCF), which learns hypergraph structures and makes recommendations collectively in a unified framework. In the hypergraph learning process, we solve two main challenges, i.e., 1) optimization issue and 2) regularization issue. Firstly, we propose a differentiable hypergraph learner to adaptively learn the optimized hypergraph structures dynamically for the hypergraph convolutions during the training process. Secondly, to better regularize dynamic hypergraph learning, we introduce a novel hypergraph learning objective, which forces the learned hypergraphs to retain the original graph topology. Extensive experiments on public datasets from different domains are provided to show that our proposed model significantly outperforms strong baselines.

Graph convolution networks (GCNs) for recommendations have emerged as an important research topic due to their ability to exploit higher-order neighbors. Despite their success, most of them suffer from the popularity bias brought by a small number of active users and popular items. Also, a real-world user-item bipartite graph contains many noisy interactions, which may hamper the sensitive GCNs. Graph contrastive learning show promising performance for solving the above challenges in recommender systems. Most existing works typically perform graph augmentation to create multiple views of the original graph by randomly dropping edges/nodes or relying on predefined rules, and these augmented views always serve as an auxiliary task by maximizing their correspondence. However, we argue that the graph structures generated from these vanilla approaches may be suboptimal, and maximizing their correspondence will force the representation to capture information irrelevant for the recommendation task. Here, we propose a Contrastive Graph Structure Learning via Information Bottleneck (CGI) for recommendation, which adaptively learns whether to drop an edge or node to obtain optimized graph structures in an end-to-end manner. Moreover, we innovatively introduce the Information Bottleneck into the contrastive learning process to avoid capturing irrelevant information among different views and help enrich the final representation for recommendation. Extensive experiments on public datasets are provided to show that our model significantly outperforms strong baselines.

Users’ social connections have recently shown significant benefits to session-based recommendations, and graph neural networks have demonstrated great success in learning the pattern of information flow among users. However, the current paradigm presumes a given social network, which is not necessarily consistent with the fast-evolving shared interests and is expensive to collect. We propose a novel idea to learn the graph structure among users and make recommendations collectively in a coupled framework. This idea raises two challenges, i.e., scalability and effectiveness. We introduce a novel graph-structure learning framework for session-based recommendations (GSL4Rec) for solving both challenges simultaneously. Our framework has a two-stage strategy, i.e., the coarse neighbor screening and the self-adaptive graph structure learning, to enable the exploration of potential links among all users while maintaining a tractable amount of computation for scalability. We also propose a phased heuristic learning strategy to sequentially and synergistically train the graph learning part and recommendation part of GSL4Rec, thus improving the effectiveness by making the model easier to achieve good local optima. Experiments on five public datasets demonstrate that our proposed model significantly outperforms strong baselines, including state-of-the-art social network-based methods.

Web service recommendation remains a highly demanding yet challenging task in the field of services computing. In recent years, researchers have started to employ side information comprised in a heterogeneous Web service ecosystem to address the issues of data sparsity and cold start in Web service recommendation. Some recent works have exploited the deep learning techniques to learn user/Web service representations accumulating information from multiplex sources. However, we argue that they still struggle to utilize multi-source information in a discriminating, unified and flexible manner. To tackle this problem, this paper presents a novel multi-source information graph-based Web service recommendation framework (MGASR), which can automatically and efficiently extract multifaceted knowledge from the heterogeneous Web service ecosystem. Specifically, different node-type and edge-type dependent parameters are designed to model corresponding types of objects (nodes) and relations (edges) in the Web service ecosystem. We then leverage graph neural networks (GNNs) with an attention mechanism to construct a multi-source information neural network (MIN) layer, for mining diverse significant dependencies among nodes. By stacking multiple MIN layers, each node can be characterized by a highly contextualized representation due to capturing high-order multi-source information. As such, MGASR can generate representations with rich semantic information toward supporting Web service recommendation tasks. Extensive experiments conducted over three real-world Web service datasets demonstrate the superior performance of our proposed MGASR as compared to various baseline methods.

Driven by proliferation of the Service-Oriented Architecture (SOA), the quantity of published software services and their users keeps increasing rapidly in the service ecosystem; thus, personalized service selection and recommendation has remained a hot topic. Recent studies have revealed that users’ social connections may help better model their potential behaviors. Therefore, in this paper, we study how users’ high-order social networks may help improve service recommendation as well as its explainability. Two observations are set forth. First, a user’s service preference may be influenced by his trusted users, whom in turn influenced by their social connections. Second, such chained influences will not remain static and equally-weighted, as a user’s confidence over his social relations may vary confronted with different targeted services. We thus introduce a novel High-order Social Graph Neural Network (HSGNN) to support social-aware service recommendation. The key idea of the model is a graph convolution-based, multi-hop propagation module devised to extract the high-order social similarity signals from users’ local social networks, and encode them into the users’ general representations. Afterwards, a neighbor-level attention module is constructed to adaptively select informative neighbors to model the users’ specific preference. Extensive experiments in a real-world service dataset show that our HSGNN makes service recommendation more accurately, i.e., by 4.71% in terms of normalized discounted cumulative gain (NDCG), than state-of-the-art baseline methods.

With the widespread application of Service-Oriented Architecture, the quantity of web services keeps increasing rapidly over the Internet. Providing personalized service recommendation to users remains to be an important research topic. Recent studies have proved social connections helpful for modeling users' potential preference thus improving the performance of service recommendation. To date, however, one special type of social relation, called high-order social relation, has not been thoroughly studied. In reality, a user's preference may not only be affected by the user's direct neighbors, but also indirect ones. Furthermore, such influences may not remain static in the context of various attentions. To tackle such issues, we have developed a novel neural Attentive network based on High-order Social Graph (A-HSG) toward offering social-aware service recommendation. First, a graph convolution-based, multi-hop propagation module is devised to extract the high-order similarity signals from users' local social networks, and inject them into the users' general representations. Second, a neighbor-level attention module is constructed to adaptively select informative neighbors to model the users' specific preference. Extensive experiments over a real-life service dataset show that A-HSG outperforms baseline methods in terms of prediction accuracy.

Infrared and visible image fusion is an active area in digital image processing. Many methods in spatial or transform domains have been proposed, but there are still several complex challenges. In this paper, we introduce a three-scale image transformation, which possesses multi-scale, translation-invariance and spatial-localization characteristics that are very important for image fusion. The decomposition can be implemented sequentially by shear transform, undecimated framelet transform and average filtering. Furthermore, we propose a multi-direction image fusion method based on the transformation, where the guided filtering based fusion rule and regional energy fusion rule are used. Experimental results show that the proposed method outperforms some representative methods, such as the methods based on Laplacian pyramid, wavelets, guided filtering, framelet, and the combination of nonsub-sampled contourlet transform and nonsubsampled shearlet transform.