With the popularity of smart mobile devices, location-based services (LBS) have been widely applied. Predicting geographical locations from text holds significant value for smart cities and personalized travel. Existing research primarily focuses on the retrieval or prediction of labeled locations, such as cities or points of interest (POIs). However, in scenarios like autonomous driving navigation and autonomous logistics delivery, it is necessary to precisely predict the coordinates of unlabeled locations, for example, 200 meters northwest of a certain location. Consequently, we introduce a new task to infer fine-grained unlabeled locations from text. This task is particularly challenging because of the ambiguous text and the semantic gap between geographic and textual modalities. In this paper, we aim to construct an end-to-end fine-grained location prediction model to accurately predict the unlabeled locations mentioned in texts. First, we encode the geographic coordinates and transform the location prediction problem into a geographic encoding generation problem. Second, we propose a multi-scale cross-modal loss (MCL) to learn the implicit mapping between geographic and textual modalities. Lastly, we design a multi-task prediction model ULP to predict the coordinates of unlabeled locations. We conducted experiments on two real-world datasets, and the results show that our proposed method outperforms existing state-of-the-art retrieval-based methods.

Recently, deep neural networks have revolutionized the field of link prediction, and the state-of-the-art works are typically subgraph-based discriminative methods, which construct features of local subgraphs firstly and predicting potential links via deep learning based binary subgraph classification. However, the discriminative link prediction methods always fail to automatically learn features and perform link prediction, and the performance of them depends on the construction of enclosing subgraphs and the manually-designed features for the subgraphs. To address these issues, we leverage the idea of graph disentangling and propose a novel self-representation-based generative graph neural network framework (GraphLP) for end-to-end link prediction, which learns to extract the latent patterns, i.e., recurring subgraphs, from input graphs via self-supervised learning and reconstruct graphs for link prediction using the subgraphs as structural basis. GraphLP consists of three components: self-representation-based collaborative inference, high-order connectivity computation, and multi-scale pattern fusion. The key idea is to utilize the correlations between the extracted recurring subgraphs on different scales to effectively assist link inference. GraphLP also can effectively exploit the hierarchical organization patterns and incorporate them within the representation procedure, producing robust and accurate results. Compared with traditional methods and state-of-the-art methods, experimental results on public benchmark datasets demonstrate that GraphLP achieves promising performance. Different from the discriminative methods, GraphLP provides a new paradigm for generative neural-network-based link prediction.

The next point of interest (POI) recommendation task recommends POIs to users that they may be interested in next time based on their historical trajectories. This task holds value for both users and businesses. However, it has consistently faced the issue of cold-start caused by sparse user check-in data. Existing research mainly focuses on knowledge transfer among cities within the same data source, but these data are very rare. The abundance of available third-party data presents opportunities to improve cold-start performance, but it is not easy. This third-party data contain numerous entities, such as POIs and users, which have different representations and distributions across different data domains, making knowledge transfer difficult. To address these challenges, we propose the Tree-Based Cross-domain Knowledge Transfer (TCKT) model. First, we construct a multi-granularity Geographical Frequency Tree (GF-Tree), transforming the POI recommendation problem into a path generation problem. Second, we design a pre-training model to mine general user behavior patterns and spatio-temporal features among POIs from large-scale third-party data. Finally, we propose a dual-channel domain adaptation model to facilitate cross-domain knowledge transfer and improve cold-start performance. Experimental results on three public datasets demonstrate that our method outperforms state-of-the-art (SOTA) baseline methods.

Voluntary vaccination is essential to protect oneself from infection and suppress the spread of infectious diseases. Voluntary vaccination behavior is influenced by factors, such as age and interaction patterns. Differences in health consciousness and risk perception based on age result in heterogeneity in vaccination behavior among different age groups. Higher-order interactions among individuals of various ages facilitate the dissemination of vaccine-related information, further influencing vaccination intentions. To investigate the impact of individual age and interaction patterns on vaccination behavior, we propose an epidemic-game coevolution model in which age structure and higher-order interactions are considered. Based on the theoretical framework of epidemic-game coevolution, this work calculates the evolutionarily stable strategies and dynamic equilibrium under imitation dynamics in the well-mixed population. Extensive numerical experiments show that infants and the elderly exhibit conservative attitudes toward vaccination, and the vaccination levels of these two groups have no significant impact on the vaccination behavior of other age groups. The vaccination behavior of children is highly active, while the vaccination behavior of adults depends on the relative cost of vaccination. The increase in vaccination levels among children and adults leads to a decrease in vaccination levels in other groups. Furthermore, the infants exhibit the lowest level of vaccination, while the children have the highest vaccination rate. Higher-order interactions significantly enhance vaccination levels among children and adults.

With the increasing network service content, innovative methods are required for developing optimized network service for e-commerce companies. Accordingly, this study focuses on designing a framework containing personalization, interest group identification, and recommendation mechanisms. The primary contribution of this paper is to propose a recommendation model based on data compensation and dynamic user interest grouping. First, to address the problem of sparse user rating data, homeostasis compensation is performed on native data to more realistically restore the preference relationship between users and items by introducing the advantages of generative adversarial network in learning data distribution and enhancing data samples. Second, to address the problem of user interest generalization, information entropy is introduced to measure the user interest feature space. In addition, the time window marking method is used to further quantify the users’ dynamic interest group around the users’ interest drift. Finally, considering tensor decomposition characteristics in data dimension transformation and data compression, a score prediction model based on the “user-item-interest group” tensor decomposition is constructed. Simultaneously, a time decay function is introduced in the construction of the tensor to dynamically fit the user behavior and further improve prediction accuracy. Experiments show that the proposed framework can effectively improve the recommendation accuracies resulting from both sparse scoring data and dynamic user interest division.

The traditional rumor diffusion model primarily studies the rumor itself and user behavior as the entry points. The complexity of user behavior, multidimensionality of the communication space, imbalance of the data samples, and symbiosis and competition between rumor and anti-rumor are challenges associated with the in-depth study on rumor communication. Given these challenges, this study proposes a group behavior model for rumor and anti-rumor. First, this study considers the diversity and complexity of the rumor propagation feature space and the advantages of representation learning in the feature extraction of data. Further, we adopt the corresponding representation learning methods for their content and structure of the rumor and anti-rumor to reduce the spatial feature dimension of the rumor-spreading data and to uniformly and densely express the full-featured information feature representation. Second, this paper introduces an evolutionary game theory, which is combined with the user-influenced rumor and anti-rumor, to reflect the conflict and symbiotic relationship between rumor and anti-rumor. we obtain a network structural feature expression of the influence degree of users on rumor and anti-rumor when expressing the structural characteristics of group communication relationships. Finally, aiming at the timeliness of rumor topic evolution, the whole model is proposed. Time slice and discretize the life cycle of rumor is used to synthesize the full-featured information feature representation of rumor and anti-rumor. The experiments denote that the model can not only effectively analyze user group behavior regarding rumor but also accurately reflect the competition and symbiotic relation between rumor and anti-rumor diffusion.

The exponential growth of devices connected to the network has resulted in the development of new Internet of Things (IoT) applications and online services, which may have diverse and dynamic requirements on received quality. Although, the emerging software-defined networking (SDN) approach can be leveraged for the IoT environment, to dynamically achieve differentiated quality levels for different IoT tasks in very heterogeneous wireless networking scenarios, the open interfaces in SDN introduces new network attacks, which may make SDN-based IoT malfunctioned. The challenges lies in securely using SDN for IoT systems. To address this challenge, we design a SDN-based data transfer security model middlebox-guard (M-G). M-G aims at reducing network latency, and properly manage dataflow to ensure the network run safely. First, according to different security policies, middleboxes related to the defined secure policies, are placed at the most appropriate locations, using dataflow abstraction and a heuristic algorithm. Next, to avoid any middlebox becoming a hotspot, an offline integer linear program (ILP) pruning algorithm is proposed in M-G, to tackle switch volume constraints. In addition, an online linear program (LP) formulation is come up to handle load balance. Finally, secure mechanisms are proposed to handle different attacks. And network routing is solved flexibly, through dataflow management protocol, which are formulated via combining tunnels and tags. Experimental results demonstrate that this model can improve security performance and manage dataflow effectively in SDN-based IoT system.

In social networks, link establishment among the users is affected by complex factors. In this paper, we try to investigate the internal and external factors that affect the formation of links and propose a three-level hidden Bayesian link prediction model by integrating the user behavior as well as user relationships to link prediction. First, based on the user multiple interest characteristics, a latent Dirichlet allocation (LDA) traditional text modeling method is applied into user behavior modeling. Taking the advantage of LDA topic model in dealing with the problem of polysemy and synonym, we can mine user latent interest distribution and analyze the effects of internal driving factors. Second, owing to the power-law characteristics of user behavior, LDA is improved by Gaussian weighting. In this way, the negative impact of the interest distribution to the high-frequency users can be reduced and the expression ability of interests can be enhanced. Furthermore, taking the impact of common neighbor dependencies in link establishment, the model can be extended with hidden naive Bayesian algorithm. By quantifying the dependencies between common neighbors, we can analyze the effects of external driving factors and combine internal driving factors to link prediction. Experimental results indicate that the model can not only mine user latent interest distribution but also can improve the performance of link prediction effectively.