KHGCN: Knowledge-Enhanced Recommendation with Hierarchical Graph Capsule Network.

Knowledge graphs as external information has become one of the mainstream directions of current recommendation systems. Various knowledge-graph-representation methods have been proposed to promote the development of knowledge graphs in related fields. Knowledge-graph-embedding methods can learn entity information and complex relationships between the entities in knowledge graphs. Furthermore, recently proposed graph neural networks can learn higher-order representations of entities and relationships in knowledge graphs. Therefore, the complete presentation in the knowledge graph enriches the item information and alleviates the cold start of the recommendation process and too-sparse data. However, the knowledge graph's entire entity and relation representation in personalized recommendation tasks will introduce unnecessary noise information for different users. To learn the entity-relationship presentation in the knowledge graph while effectively removing noise information, we innovatively propose a model named knowledge-enhanced hierarchical graph capsule network (KHGCN), which can extract node embeddings in graphs while learning the hierarchical structure of graphs. Our model eliminates noisy entities and relationship representations in the knowledge graph by the entity disentangling for the recommendation and introduces the attentive mechanism to strengthen the knowledge-graph aggregation. Our model learns the presentation of entity relationships by an original graph capsule network. The capsule neural networks represent the structured information between the entities more completely. We validate the proposed model on real-world datasets, and the validation results demonstrate the model's effectiveness.

Wide-field and full-focus optical microscopic imaging system.

As a precision instrument, the microscope is typically used by researchers in criminal investigation, information forensics, biology, metallography, etc. However, the traditional microscope has a dilemma in that if it uses higher magnification, its field of view is smaller and its depth of field is more limited. Hence, it seriously challenges the endurance and brain of the observer to observe an object thoroughly. This paper proposes a wide-field and full-focus imaging method for solving the above problem. First, a high-precision multi-focus image acquisition platform is improved, and its motion displacement is used directly for image calculation, which greatly reduces the amount of calculation. Second, the focus area of each image is segmented by the mask generation algorithm based on a graph cut. Third, a fusion algorithm, whose contrast pyramid is based on the mask region, is proposed, which utilizes the position of the clear area on the mask pyramid to guide the fusion of the contrast pyramid. Finally, a fast and fault-tolerant stitching algorithm based on mechanical and optical parameters is proposed, which effectively eliminates the interference of the cumulative error and successfully completes hundreds of image-stitching tasks. The experimental results demonstrate that the proposed imaging system is obviously superior to the traditional image fusion algorithms and image-stitching approaches. Both the imaging effect and execution time are satisfactory.

RiPPAS: A Ring-Based Privacy-Preserving Aggregation Scheme in Wireless Sensor Networks.

Recently, data privacy in wireless sensor networks (WSNs) has been paid increased attention. The characteristics of WSNs determine that users' queries are mainly aggregation queries. In this paper, the problem of processing aggregation queries in WSNs with data privacy preservation is investigated. A Ring-based Privacy-Preserving Aggregation Scheme (RiPPAS) is proposed. RiPPAS adopts ring structure to perform aggregation. It uses pseudonym mechanism for anonymous communication and uses homomorphic encryption technique to add noise to the data easily to be disclosed. RiPPAS can handle both s u m ( ) queries and m i n ( ) / m a x ( ) queries, while the existing privacy-preserving aggregation methods can only deal with s u m ( ) queries. For processing s u m ( ) queries, compared with the existing methods, RiPPAS has advantages in the aspects of privacy preservation and communication efficiency, which can be proved by theoretical analysis and simulation results. For processing m i n ( ) / m a x ( ) queries, RiPPAS provides effective privacy preservation and has low communication overhead.

Cross Deployment Networking and Systematic Performance Analysis of Underwater Wireless Sensor Networks.

Underwater wireless sensor networks (UWSNs) have become a new hot research area. However, due to the work dynamics and harsh ocean environment, how to obtain an UWSN with the best systematic performance while deploying as few sensor nodes as possible and setting up self-adaptive networking is an urgent problem that needs to be solved. Consequently, sensor deployment, networking, and performance calculation of UWSNs are challenging issues, hence the study in this paper centers on this topic and three relevant methods and models are put forward. Firstly, the normal body-centered cubic lattice to cross body-centered cubic lattice (CBCL) has been improved, and a deployment process and topology generation method are built. Then most importantly, a cross deployment networking method (CDNM) for UWSNs suitable for the underwater environment is proposed. Furthermore, a systematic quar-performance calculation model (SQPCM) is proposed from an integrated perspective, in which the systematic performance of a UWSN includes coverage, connectivity, durability and rapid-reactivity. Besides, measurement models are established based on the relationship between systematic performance and influencing parameters. Finally, the influencing parameters are divided into three types, namely, constraint parameters, device performance and networking parameters. Based on these, a networking parameters adjustment method (NPAM) for optimized systematic performance of UWSNs has been presented. The simulation results demonstrate that the approach proposed in this paper is feasible and efficient in networking and performance calculation of UWSNs.

Data Access Based on a Guide Map of the Underwater Wireless Sensor Network.

Underwater wireless sensor networks (UWSNs) represent an area of increasing research interest, as data storage, discovery, and query of UWSNs are always challenging issues. In this paper, a data access based on a guide map (DAGM) method is proposed for UWSNs. In DAGM, the metadata describes the abstracts of data content and the storage location. The center ring is composed of nodes according to the shortest average data query path in the network in order to store the metadata, and the data guide map organizes, diffuses and synchronizes the metadata in the center ring, providing the most time-saving and energy-efficient data query service for the user. For this method, firstly the data is stored in the UWSN. The storage node is determined, the data is transmitted from the sensor node (data generation source) to the storage node, and the metadata is generated for it. Then, the metadata is sent to the center ring node that is the nearest to the storage node and the data guide map organizes the metadata, diffusing and synchronizing it to the other center ring nodes. Finally, when there is query data in any user node, the data guide map will select a center ring node nearest to the user to process the query sentence, and based on the shortest transmission delay and lowest energy consumption, data transmission routing is generated according to the storage location abstract in the metadata. Hence, specific application data transmission from the storage node to the user is completed. The simulation results demonstrate that DAGM has advantages with respect to data access time and network energy consumption.

Use of customizing kernel sparse representation for hyperspectral image classification.

Sparse representation-based classification (SRC) has attracted increasing attention in remote-sensed hyperspectral communities for its competitive performance with available classification algorithms. Kernel sparse representation-based classification (KSRC) is a nonlinear extension of SRC, which makes pixels from different classes linearly separable. However, KSRC only considers projecting data from original space into feature space with a predefined parameter, without integrating a priori domain knowledge, such as the contribution from different spectral features. In this study, customizing kernel sparse representation-based classification (CKSRC) is proposed by incorporating kth nearest neighbor density as a weighting scheme in traditional kernels. Analyses were conducted on two publicly available data sets. In comparison with other classification algorithms, the proposed CKSRC further increases the overall classification accuracy and presents robust classification results with different selections of training samples.

A Cross Structured Light Sensor and Stripe Segmentation Method for Visual Tracking of a Wall Climbing Robot.

In non-destructive testing (NDT) of metal welds, weld line tracking is usually performed outdoors, where the structured light sources are always disturbed by various noises, such as sunlight, shadows, and reflections from the weld line surface. In this paper, we design a cross structured light (CSL) to detect the weld line and propose a robust laser stripe segmentation algorithm to overcome the noises in structured light images. An adaptive monochromatic space is applied to preprocess the image with ambient noises. In the monochromatic image, the laser stripe obtained is recovered as a multichannel signal by minimum entropy deconvolution. Lastly, the stripe centre points are extracted from the image. In experiments, the CSL sensor and the proposed algorithm are applied to guide a wall climbing robot inspecting the weld line of a wind power tower. The experimental results show that the CSL sensor can capture the 3D information of the welds with high accuracy, and the proposed algorithm contributes to the weld line inspection and the robot navigation.

Feature weighting algorithms for classification of hyperspectral images using a support vector machine.

The support vector machine (SVM) is a widely used approach for high-dimensional data classification. Traditionally, SVMs use features from the spectral bands of hyperspectral images with each feature contributing equally to the classification. In practical applications, although affected by noise, slight contributions can also be obtained from deteriorated bands. Thus, compared with feature reduction or equal assignment of weights to all the features, feature weighting is a trade-off choice. In this study, we examined two approaches to assigning weights to SVM features to increase the overall classification accuracy: (1) "CSC-SVM" refers to a support vector machine with compactness and a separation coefficient feature weighting algorithm, and (2) "SE-SVM" refers to a support vector machine with a similarity entropy feature weighting algorithm. Analyses were conducted on a public data set with nine selected land-cover classes. In comparison with traditional SVMs and other classical feature weighting algorithms, the proposed weighting algorithms increase the overall classification accuracy, and even better results could be obtained with few training samples.

Stability and Hopf Bifurcation Analysis of an Epidemic Model with Time Delay.

Epidemic models are normally used to describe the spread of infectious diseases. In this paper, we will discuss an epidemic model with time delay. Firstly, the existence of the positive fixed point is proven; and then, the stability and Hopf bifurcation are investigated by analyzing the distribution of the roots of the associated characteristic equations. Thirdly, the theory of normal form and manifold is used to drive an explicit algorithm for determining the direction of Hopf bifurcation and the stability of the bifurcation periodic solutions. Finally, some simulation results are carried out to validate our theoretic analysis.

The Quantitative Relationship among the Number of the Pacing Cells Required, the Dimension, and the Diffusion Coefficient.

The purpose of the paper is to derive a formula to describe the quantitative relationship among the number of the pacing cells required (NPR), the dimension i, and the diffusion coefficient D (electrical coupling or gap junction G). The relationship between NPR and G has been investigated in different dimensions, respectively. That is, for each fixed i, there is a formula to describe the relationship between NPR and G; and three formulas are required for the three dimensions. However, there is not a universal expression to describe the relationship among NPR, G, and i together. In the manuscript, surveying and investigating the basic law among the existed data, we speculate the preliminary formula of the relationship among the NPR, i, and G; and then, employing the cftool in MATLAB, the explicit formulas are derived for different cases. In addition, the goodness of fit (R 2) is computed to evaluate the fitting of the formulas. Moreover, the 1D and 2D ventricular tissue models containing biological pacemakers are developed to derive more data to validate the formula. The results suggest that the relationship among the NPR, i, and the G (D) could be described by a universal formula, where the NPR scales with the i (the dimension) power of the product of the square root of G (D) and a constant b which is dependent on the strength of the pacing cells and so on.

A game theory-based trust measurement model for social networks.

BACKGROUND: In social networks, trust is a complex social network. Participants in online social networks want to share information and experiences with as many reliable users as possible. However, the modeling of trust is complicated and application dependent. Modeling trust needs to consider interaction history, recommendation, user behaviors and so on. Therefore, modeling trust is an important focus for online social networks. METHODS: We propose a game theory-based trust measurement model for social networks. The trust degree is calculated from three aspects, service reliability, feedback effectiveness, recommendation credibility, to get more accurate result. In addition, to alleviate the free-riding problem, we propose a game theory-based punishment mechanism for specific trust and global trust, respectively. RESULTS AND CONCLUSIONS: We prove that the proposed trust measurement model is effective. The free-riding problem can be resolved effectively through adding the proposed punishment mechanism.

Collaborative Filtering Recommendation on Users' Interest Sequences.

As an important factor for improving recommendations, time information has been introduced to model users' dynamic preferences in many papers. However, the sequence of users' behaviour is rarely studied in recommender systems. Due to the users' unique behavior evolution patterns and personalized interest transitions among items, users' similarity in sequential dimension should be introduced to further distinguish users' preferences and interests. In this paper, we propose a new collaborative filtering recommendation method based on users' interest sequences (IS) that rank users' ratings or other online behaviors according to the timestamps when they occurred. This method extracts the semantics hidden in the interest sequences by the length of users' longest common sub-IS (LCSIS) and the count of users' total common sub-IS (ACSIS). Then, these semantics are utilized to obtain users' IS-based similarities and, further, to refine the similarities acquired from traditional collaborative filtering approaches. With these updated similarities, transition characteristics and dynamic evolution patterns of users' preferences are considered. Our new proposed method was compared with state-of-the-art time-aware collaborative filtering algorithms on datasets MovieLens, Flixster and Ciao. The experimental results validate that the proposed recommendation method is effective and outperforms several existing algorithms in the accuracy of rating prediction.

A reliability-based track fusion algorithm.

The common track fusion algorithms in multi-sensor systems have some defects, such as serious imbalances between accuracy and computational cost, the same treatment of all the sensor information regardless of their quality, high fusion errors at inflection points. To address these defects, a track fusion algorithm based on the reliability (TFR) is presented in multi-sensor and multi-target environments. To improve the information quality, outliers in the local tracks are eliminated at first. Then the reliability of local tracks is calculated, and the local tracks with high reliability are chosen for the state estimation fusion. In contrast to the existing methods, TFR reduces high fusion errors at the inflection points of system tracks, and obtains a high accuracy with less computational cost. Simulation results verify the effectiveness and the superiority of the algorithm in dense sensor environments.