Graph self-supervised learning (GSSL) alleviates the graph data labeling bottleneck without supervision, enabling wide application in domains like recommendation systems and social network analysis. High-frequency signals are valuable in GSSL for capturing local structural preferences, thereby enriching graph representations and boosting model performance. However, in practical applications, two critical problems hinder the efficient and robust use of these signals. First, the locality of high-frequency signals limits their full utilization by the model. Second, over-reliance on specific high-frequency signals will affect the model's generalization. To address the above problems, we propose the Frequency-Corrupt Based Graph Self-Supervised Learning (FC-GSSL) algorithm. Specifically, we generate corrupted graphs biased toward high-frequency signals by corrupting nodes and edges according to their low-frequency contributions. These corrupted graphs are fed as input to an autoencoder, with low-frequency and general features serving as the supervision. This compels the model to effectively fuse high- and low-frequency signals, thereby integrating and utilizing more valuable high-frequency information. Additionally, we design multiple sampling strategies and form diverse corrupted graphs based on the intersections and union between the results obtained from these strategies. By aligning the node representations from these views, the model can identify valuable frequency combinations, which helps reduce the negative impact of specific high-frequency components and improve generalization. FC-GSSL optimizes the design of GSSL for web applications, significantly improving model performance on complex web-related graphs, such as social networks and citation networks. This work makes a direct contribution to advancing the ''Graph Algorithms and Modeling for the Web'' research track. Experimental results on 14 datasets across multiple tasks illustrate the superiority of the proposed approach.

Multimodal embeddings of items and users, along with the loss function in the prediction model, are crucial for multimodal recommendation. Previous studies focused on feature extraction and fusion across modalities but lacked collaborative optimization of modal features across different association graphs. While prediction models address data alignment and user preference enhancement, they overlook preference consistency and varying contributions of different modalities to recommendations. In light of these issues, this paper proposes a multimodal recommendation method that combines Modality Shared Encoding with Multidimensional Loss optimization (MSE-ML). We introduce a multi-view feature joint encoding module, leveraging modality sharing mechanism to enable collaborative optimization of item embeddings across the user item graph and multiple modality-aware item association graphs. Additionally, we design a multidimensional loss optimization strategy that simultaneously promotes modality alignment, preserves preference consistency, and enhance the contribution of weak modality. Experimental results on publicly available datasets show that MSE-ML outperforms state-of-the-art approaches in terms of recommendation quality.

Cross-Project Defect Prediction (CPDP) utilizes the existing labeled data in the source project to assist with the prediction of unlabeled projects in the target dataset, which effectively improves the prediction performance and has become a research hotspot in software engineering. At present, CPDP can be categorized into homogeneous CPDP and heterogeneous CPDP (HDP), in which HDP doesn’t require that the source project and the target project have the same feature space, thus, it is more widely used in the actual CPDP. Most of current HDP methods map the original features to the latent feature space and reduce the inter-project variation by transferring domain-independent features, but the transferring process ignores the use of domain-related features, which affects the prediction performance of the model. Moreover, the mapped latent features are not conducive to the model’s interpretability. Based on these, this article proposes a Heterogeneous Defect Prediction method based on Feature Disentanglement (FD-HDP). We disentangle the features using domain-related and domain-independent feature extractors, respectively, to improve the interpretability of the model by maximizing the domain adversarial loss during training and guiding the feature extractors to produce accurate domain-related and domain-independent features. The weighted sum of the prediction results from domain-related and domain-independent predictors is used as the final prediction result of the project during the prediction process, which realizes the combination of domain-independent and domain-related features and effectively improves the prediction performance. In this article, we conducted experiments using four publicly available defect datasets to construct heterogeneous scenarios. The results demonstrate that the FD-HDP model shows significant advantages over state-of-the-art methods in six metrics.

Disentanglement techniques used in collaborative filtering uncover interaction intents between nodes, improving the interpretability of node representations and enhancing recommendation performance. However, existing disentanglement methods still face the following two problems. 1) They focus on local structural features derived from direct node interactions, overlooking the comprehensive graph structure, which limits disentanglement accuracy. 2) The disentanglement process depends on backpropagation signals derived from recommendation tasks, lacking direct supervision, which may lead to biases and overfitting. To address the issues, we propose theIntentPropagationContrastiveCollaborativeFiltering (IPCCF) algorithm. Specifically, we design a double helix message propagation framework to more effectively extract the deep semantic information of nodes, thereby improving the model's understanding of interactions between nodes. An intent message propagation method is also developed that incorporates graph structure information into the disentanglement process, thereby expanding the consideration scope of disentanglement. In addition, contrastive learning techniques are employed to align node representations derived from the structure and intents, providing direct supervision for the disentanglement process, mitigating biases, and enhancing the model's robustness to overfitting. The experiments on three real data graphs illustrate the superiority of the proposed approach.

The purpose of Intrusion Detection Systems (IDS) is to identify security issues in data transmitted by various devices and communication protocols. For domains with sparse data, such as the Internet of Things (IoT), cross-domain models are applied to solve the sparse problem by transfer knowledge from the source domain with rich data to the target domain. However, most of the cross-domain intrusion detection methods map different explicit features in the source and target domains to implicit features in a common implicit space, which weakens the interpretability of these methods. To enhance the interpretability of cross-domain models, we propose a Cross-Domain Intrusion Detection Method Based on Nonlinear Augmented Explicit Features (NAEF). Specifically, we augment the feature space of the source and target domains as the combination of shared features, source domain specific features and target domain specific features. Moreover, we model the nonlinear mapping relationship from shared features to special features in the source and target domains separately. Then, the original features in the source and target domains are mapped to uniform explicit features in the augmented space by migration of the nonlinear mapping relationship. Additionally, a classifier based on ensemble learning and attention mechanism balances the data distribution and selects important features to enhance detection performance. Our experimental results demonstrate the effectiveness of the proposed NAEF method on four public datasets.

Multi-behavior recommendation systems enhance effectiveness by leveraging auxiliary behaviors (such as page views and favorites) to address the limitations of traditional models that depend solely on sparse target behaviors like purchases. Existing approaches to multi-behavior recommendations typically follow one of two strategies: some derive initial node representations from individual behavior subgraphs before integrating them for a comprehensive profile, while others interpret multi-behavior data as a heterogeneous graph, applying graph neural networks to achieve a unified node representation. However, these methods do not adequately explore the intricate patterns of behavior among users and items. To bridge this gap, we introduce a novel algorithm called Behavior Pattern mining-based Multi-behavior Recommendation (BPMR). Our method extensively investigates the diverse interaction patterns between users and items, utilizing these patterns as features for making recommendations. We employ a Bayesian approach to streamline the recommendation process, effectively circumventing the challenges posed by graph neural network algorithms, such as the inability to accurately capture user preferences due to over-smoothing. Our experimental evaluation on three realworld datasets demonstrates that BPMR significantly outperforms existing state-of-the-art algorithms, showing an average improvement of 268.29% in Recall@10 and 248.02% in NDCG@10 metrics. The code of our BPMR is openly accessible for use and further research at https://github.com/rookitkitlee/BPMR.

Bipartite graph representation learning embeds users and items into a low-dimensional latent space based on observed interactions. Previous studies mainly fall into two categories: one reconstructs the structural relations of the graph through the representations of nodes, while the other aggregates neighboring node information using graph neural networks. However, existing methods only explore the local structural information of nodes during the learning process. This makes it difficult to represent the macroscopic structural information and leaves it easily affected by data sparsity and noise. To address this issue, we propose the Intent Distribution based Bipartite graph Representation learning (IDBR) model, which explicitly integrates node intent distribution information into the representation learning process. Specifically, we obtain node intent distributions through clustering and design an intent distribution based graph convolution neural network to generate node representations. Compared to traditional methods, we expand the scope of node representations, enabling us to obtain more comprehensive representations of global intent. When constructing the intent distributions, we effectively alleviated the issues of data sparsity and noise. Additionally, we enrich the representations of nodes by integrating potential neighboring nodes from both structural and semantic dimensions. Experiments on the link prediction and recommendation tasks illustrate that the proposed approach outperforms existing state-of-the-art methods. The code of IDBR is available at https://github.com/rookitkitlee/IDBR.

Disentangled collaborative filtering can explicitly generate embeddings based on users’ interests and help improve the interpretability and robustness of recommendations. However, the existing disentangled graph collaborative filtering methods rely solely on direct interaction constraints between nodes to learn node embeddings, which cannot represent higher-order constraints between nodes and node-type differences, resulting in suboptimal node representations and negatively affecting recommendation performance. To address this problem, we propose a Multi-order Similarity Constraint Disentangled Graph Collaborative Filtering (DGCF-MSC) method, which considers not only direct interaction constraints between nodes but also designs a neighborhood enhancement mechanism based on high-order relationships between homogeneous nodes. We realize the disentanglement of heterogeneous type nodes in different feature spaces in a graph convolutional neural network to make the generated embedding more interpretable and improve the performance of graph collaborative filtering. We conduct extensive experiments with three recommendation system datasets and the results demonstrate that DGCF-MSC outperforms the existing disentangled graph collaborative filtering methods in all performance metrics. Our code is released on https://github.com/lustrelake/DGCF_MSC.

Abstract Generally, open source software (OSS) has a longer bug‐fixing time. If the bug‐fixing time can be predicted accurately as early as possible, it will be beneficial to the efficiency of bug fixing. Traditional bug‐fixing time prediction models are usually based on static features of bug report. It is difficult to go into service due to inappropriate feature extraction of data and low prediction accuracy of models. The HMM prediction model can predict the bug‐fixing time accurately according to earlier fixing activities. However, this method of temporal sequence feature selection results in a large number of inconsistent samples, and the HMM prediction model can only capture the adjacent activity behavior information of one sequence, and hence, it will reduce the performance of bug‐fixing time prediction. By incorporating the activity information and time information of bug activity transfer, the proportion of inconsistent samples is reduced significantly. In this paper, a double‐sequence input LSTM model (LSTM‐DA) is designed to capture both sequences interaction features and long‐distance‐dependent features. The results of the experiments show that the proposed model can improve the F‐measure and accuracy indicators by about 10% compared with the HMM model in all dimensions, which demonstrates the effectiveness of our method.

Recently, recommender systems are applied to provide personalized recomendation for healthcare wearables. However, due to the sparsity problem, traditional recommendation algorithms are difficult to achieve desired performance. Considering that consumers often buy and rate other types of items on E-commerce platforms, we can leverage significant information in the auxiliary domains to improve the recommendation performance of healthcare wearables, which can be regarded as cross-domain recommendation. However, traditional cross-domain recommendation model cannot fully represent user's characteristics and fail to consider the leaks of original auxiliary domain ratings during the information transfer process. To overcome the two shortcomings, this paper proposes a Privacy-Preserving Cross-Domain Healthcare Wearables Recommendation algorithm (PPCDHWRec). Firstly, user's characteristics are divided into domain-dependent features and domain-independent features, which complement each other and fully depict the user's characteristics. Secondly, inspired by the latent factor model, we factorize the original rating information of each auxiliary domain by Funk-SVD and Orthogonal Nonnegative Matrix Tri-Factorization (ONMTF) model, to obtain user's domain-dependent and domain-independent features, respectively. Finally, the Factorization Machine algorithm is used to fuse the obtained user's features with the target domain information to provide the recommendation results. By hiding the item latent factors obtained in the factorization process, PPCDHWRec ensures that the original information cannot be inferred from the transferred user hidden vector. Hence, PPCDHWRec is a privacy-preserving recommendation model. Experiments on two groups of auxiliary domains, having high and low correlations with target domain, show the effectiveness of PPCDHWRec.

Ensemble learning is an effective approach for improving the generalization ability of base classifiers. To generate a set of accurate and diverse base classifiers, different data perturbation schemes have been proposed. For instance, Bagging perturbs the training data via bootstrap sampling. However, when a stable learning algorithm (e.g., KNN, Naive Bayes) is used to train base classifiers, the sole perturbation on the training data may not produce diverse base classifiers. In this paper, by using the attribute reduction technology in rough sets, a multi-modal perturbation-based algorithm (called ‘E _ EARBS’) is proposed for the ensemble of base classifiers. E _ EARBS simultaneously perturbs the feature space, training data and learning parameters, where the relative decision entropy(RDE)-based approximate reducts are used to perturb the feature space, and bootstrap sampling is used to perturb the training data. Experimental results show that E _ EARBS can provide competitive solutions for ensemble learning.

Summary Petri nets are frequently used formal tools to analyze the compatibility of interactive service processes described by Web Services Business Process Execution Language (WS‐BPEL). However, the traditional methods based on Petri nets were with a high computable complexity for state space explosion. To resolve such problem, a logic Petri net–based path detecting method for compatibility analysis of interactive service processes is proposed. From the provided mapping rules, the service process described by WS‐BPEL is modeled as a service net based on logic Petri nets. The evaluation of interactive compatibility of two service processes is converted to analyze whether their service nets can be composed as a non‐blocked synthetic service net. The non‐blocked property is checked by detecting the reachability of the potential connected paths in a service net. To reduce the complexity of computing the connected paths in a service net, we propose a merge‐reduced method to generate the path expression of its skeleton service net. The potential connected paths of a service net can be obtained by unfolding the path expression. Compared with the traditional method based Petri nets, the proposed method is with high efficiency and it can greatly alleviate the problem of state space explosion in analyzing interactive compatibility of service processes.

Reach ability analysis algorithm is the foundation of model checking. State space explosion is accelerated due to the time constraints of timed automata, therefore, the complexity of reach ability analysis algorithm has increased. The existing timed automata reach ability analysis methods can be divided into two types: the first type is the state space traversal algorithm based on equivalent time region map or difference bounded matrices (DBM) using a semi-symbol time constraints structure, the other is a state space traversal method based on BDD structure using full-symbol time constraints. Unfortunately, both types of methods split the symbol correlation of state transition expression and time constraints expression, therefore they are difficult to be applied into a large case study. A full-symbol reach ability analysis method based on BDD structure is proposed in this article, state transition space of timed automata and time constraints are expressed under the same BDD structure to avoid the loss of correlation and to analyze the reach ability of states and satisfiablity of time constraints. The presented method in this paper helps to achieve full-symbol reach ability analysis of timed automata, which would lay a solid foundation for full-symbolic model checking of time system. Finally, a specific case is analyzed by this algorithm.

The traditional relational database model (RDM) is not effective for dealing with imprecise and uncertain data as it deals with precise and unambiguous data. Hence, Beaubouef et al. proposed the rough relational database model (RRDM) for the management of uncertainty in relational databases. Beaubouef et al. defined the corresponding rough relational operators in rough relational databases as in ordinary relational databases. And to give an effective measure of uncertainty in rough relational databases, they defined the rough relation entropy. In this paper, we further discuss the issues of relational operations and uncertainty measure in rough relational databases. We give some new definitions for rough relational operators and rough relation entropy in rough relational databases. Furthermore, we discuss the basic properties of rough relational operators and rough relation entropy, as well as the connections between rough relational operators and rough relation entropy.