Search

Non-coding RNAs are gaining prominence in biology and medicine, as they play major roles in cellular homeostasis among which the circRNA-miRNA-mRNA axes are involved in a series of disease-related pathways, such as apoptosis, cell invasion and metastasis. Recently, many computational methods have been developed for the prediction of the relationship between ncRNAs and diseases, which can alleviate the time-consuming and labor-intensive exploration involved with biological experiments. However, these methods handle ncRNAs separately, ignoring the impact of the interactions among ncRNAs on the diseases. In this paper we present a novel approach to discovering disease-related circRNA-miRNA-mRNA axes from the disease-RNA information network. Our method, using graph convolutional network, learns the characteristic representation of each biological entity by propagating and aggregating local neighbor information based on the global structure of the network. The approach is evaluated using the real-world datasets and the results show that it outperforms other state-of-the-art baselines on most of the metrics.

Human pose estimate can be used in action recognition, video surveillance and other fields, which has received a lot of attentions. Since the flexibility of human joints and environmental factors greatly influence pose estimation accuracy, related research is confronted with many challenges. In this paper, we incorporate the pyramid convolution and attention mechanism into the residual block, and introduce a hybrid structure model which synthetically applies the local and global information of the image for the analysis of keypoints detection. In addition, our improved structure model adopts grouped convolution, and the attention module used is lightweight, which will reduce the computational cost of the network. Simulation experiments based on the MS COCO human body keypoints detection data set show that, compared with the Simple Baseline model, our model is similar in parameters and GFLOPs (giga floating-point operations per second), but the performance is better on the detection of accuracy under the multi-person scenes.

The bacterium strain Ideonella sakaiensis 201-F6 is able to hydrolyze low-crystallinity PET films at 30 °C due to two enzymes named PETase and MHETase. Since its discovery, many efforts have been dedicated to elucidating the structure and features of those two enzymes, and various authors have highlighted the necessity to optimize both the substrate binding site and the global structure in order to enhance the stability and catalytic activity of these PET biocatalysts so as to make them more suitable for industrial applications. In this review, the strategies adopted by different research groups to investigate the structure and functionality of both PETase and MHETase in depth are described, emphasizing the advantages provided by the use of computational methods to complement and drive experiments. Subsequently, the modifications implemented with protein engineering are discussed. The versatility of the enzymes secreted by I. sakaiensis enables the prediction that they will find several applications in the disposal of PET debris, encouraging a prioritization of efforts in this prolific research field.

Nonnegative Matrix Factorization (NMF) is a popular technique for machine learning. Its power is that it can decompose a nonnegative matrix into two nonnegative factors whose product well approximates the nonnegative matrix. However, the nonnegative constraint of the data matrix limits its application. Additionally, the representations learned by NMF fail to respect the intrinsic geometric structure of the data. In this paper, we propose a novel unsupervised matrix factorization method, called Robust Structured Convex Nonnegative Matrix Factorization (RSCNMF). RSCNMF not only achieves meaningful factorizations of the mixed-sign data, but also learns a discriminative representation by leveraging local and global structures of the data. Moreover, it introduces the L2,1-norm loss function to deal with noise and outliers, and exploits the L2,1-norm feature regularizer to select discriminative features across all the samples. We develop an alternate iterative scheme to solve such a new model. The convergence of RSCNMF is proven theoretically and verified empirically. The experimental results on eight real-world data sets show that our RSCNMF algorithm matches or outperforms the state-of-the-art methods.

Recently, global structure from motion has successfully gained many followers, mainly because of its computational speed. Most of these global methods take the parameters of relative orientation (ROs ) as input and then perform averaging operations. Therefore, eliminating incorrect ROs is of great significance for improving the robustness of global structure from motion. In this article, we propose a method to eliminate wrong ROs which have resulted from repetitive structure and very short baselines. We present two corresponding criteria that indicate the quality of ROs. Repetitive structure is detected based on counts of conjugate points of the various image pairs, while very short baselines are found by inspecting the intersection angles of corresponding image rays. By analyzing these two criteria, we detect and eliminate incorrect ROs. As correct ROs of image pairs with a longer baseline nearly parallel to both viewing directions can be valuable, a method to identify and keep these ROs is also part of our approach. We demonstrate the new method on various data sets, including public benchmarks as well as close-range images and images from unmanned aerial vehicles, by inserting our refined ROs into a global structure-from-motion pipeline. The experiments show that compared to other methods, we can generate the best results.

Graph neural networks (GNNs) have attracted significant attention from the chemical science community because molecules can be represented as a featured graph. In particular, graph convolutional network (GCN) and its variants have been widely used and have shown a state-of-the-art performance in analyzing molecules, such as molecular label classification, drug discovery, and molecular property prediction. However, in molecular analysis, existing GCNs have two fundamental limitations: (1) information of the molecular scale is distorted and (2) global structures in a molecule are ignored. These limitations can seriously degrade the performance in the machine learning-based molecular analysis because the scale and global structure information of a molecule occasionally have a significant effect on the molecular properties. To overcome the limitations of existing GCNs, we comprehensively analyzed the structure of GCNs and developed a costless solution for the limitations of GCNs. To demonstrate the effectiveness of our solution, extensive experiments were conducted on various benchmark datasets.

Quantitative identification of key nodes in complex networks is of great significance for studying the robustness and vulnerability of complex networks. Although various centralities have been proposed to solve this issue, each approach has its limitations for its own perspective of determining an actor to be “key”. In this paper, we propose a novel method to identify key nodes in complex networks based on global structure. Three aspects including the shortest path length, the number of shortest paths and the number of non-shortest paths are considered, and we establish three corresponding influence matrices. Node efficiency, which can reflect the contribution of one node to the information transmission of the entire network, is selected as the initial value of node's influence on other nodes, and then the comprehensive influence matrix is constructed to reflect the influence among nodes. The proposed method provides a new measure to identify key nodes in complex networks from the perspective of global network structure, and can obtain more accurate identification results. Four experiments are conducted to evaluate the performance of our proposed method based on Susceptible-Infected (SI) model, and the results demonstrate the superiority of our method.

Adaptive voter models (AVMs) are simple mechanistic systems that model the emergence of mesoscopic structure from local networked processes driven by conflict and homophily. AVMs display rich behavior, including a phase transition from a fully-fragmented regime of "echo-chambers" to a regime of persistent disagreement governed by low-dimensional quasistable manifolds. Many extant methods for approximating the behavior of AVMs are either restricted in scope, expensive in computation, or inaccurate in predicting important statistics. In this work, we develop a novel, second-order moment closure approximation method for binary-state rewire-to-random and rewire-to-same model variants. We incorporate a small amount of noise via a random mutation term, which renders the system ergodic. Using ergodicity, we then approximate the voting process, which is non-Markovian in the second moments of the system, with a Markovian term near the phase transition. This approximation exploits an asymmetry between different classes of voting events. The resulting scheme enables us to predict the location of the phase transition and the active edge density in the regime of persistent disagreement, across the entire space of parameters and opinion densities. Numerically, our results are nearly exact for the rewire-to-random model, and competitive with other current approaches for the rewire-to-same model. Moreover, our computations display constant scaling in the mean degree, enabling approximations for denser systems than previously possible. We conclude with suggestions for model refinements and extensions., Comment: 20 Pages, 4 figures. Final revised copy

In recent years, the huge expansion of digital technologies has vastly increased the volume of data to be explored, such that reducing the dimensionality of data is an essential step in data exploration. The integrity of a dimensionality reduction technique relates to the goodness of maintaining the data structure. Dimensionality reduction techniques such as Principal Component Analyses (PCA) and Multidimensional Scaling (MDS) globally preserve the distance ranking at the expense of neglecting small-distance preservation. Conversely, the structure capturing of some other methods such as Isomap, Locally Linear Embedding (LLE), Laplacian Eigenmaps ${t}$ -Stochastic Neighbour Embedding ( t -SNE), Uniform Manifold Approximation and Projection (UMAP), and TriMap rely on the number of neighbours considered. This paper presents a dimensionality reduction technique, Same Degree Distribution (SDD) that does not rely on the number of neighbours, thanks to using degree-distributions in both high and low dimensional spaces. Degree-distribution is similar to Student- t distribution and is less expensive than Gaussian distribution. As such, it enables better global data preservation in less processing time. Moreover, to improve the data structure capturing, SDD has been extended to Multi-SDD s (MSDD), which employs various degree-distributions on top of SDD. The proposed approach and its extension demonstrated a greater performance compared with eight other benchmark methods, tested in several popular synthetics and real datasets such as Iris, Breast Cancer, Swiss Roll, MNIST, and Make Blob evaluated by the co-ranking matrix and Kendall’s Tau coefficient. For further work, we aim to approximate the number of distributions and their degrees in relation to the given dataset. Reducing the computational complexity is another objective for further work.

In many fields of machine learning such as classifying and clustering, neighborhood construction algorithms are used to model local relationships between data samples and to build global structure from local information. In finding connection among the data points, neighborhood is undeniably useful for data processing. Therefore, a very major issue is to find a novel approach to locating neighborhood among data points. If the geometric relationships existing between the data points in the neighborhood area are accurately explored, it will be feasible to observe the behavioral rules as well as the similarities among the data. This will also help identify indirect and direct neighborhood ranges. This study aims to find neighborhood accurately by means of the Apollonius circle zones. The experimental validation against well-known k-nearest neighbor and e-neighborhood is also an indication of the robustness of the method in real data sets.

The global structure of the chamber graph of certain rank [math] geometries that are almost buildings is determined. Computer files containing extensive details of these graphs accompany this paper.

Substructuring methods possess many merits in model updating and damage identification of large-scale structures. With substructuring methods, a global structure is divided into a number of independent substructures. Only the substructures are repeatedly analyzed and the re-analysis of the global structure is thereby avoided. This article reviews widely used dynamic substructuring methods for model updating and damage identification of large-scale structures. These methods can be categorized into forward and inverse substructuring approaches. The former is a conventional process that assembles the vibration properties of each substructure to obtain the vibration properties of the global structure. The latter, on the contrary, disassembles the vibration properties of the global structure into those of the substructures. In each category, both frequency and time domain methods have been developed and will be reviewed.

This paper develops locality preserving difference component analysis in which the intrinsic and global structure of data is exploited, and the model we propose also provides the flexibility to adapt some characteristics of data by applying the Lq norm. In order to solve the proposed model that is non-convex or non-smooth, we resort to the proximal alternating linearized optimization approach where each subproblem has good optimization properties. It is observed that the objective function in the proposed model is a semi-algebraic function. This allows us to give the convergence analysis of algorithms in terms of the Kurdyka–Lojasiewicz property. To be specific, the sequence of iterations generated by the proposed approach converges to a critical point of the objective function. The experiments on several data sets have been conducted to demonstrate the effectiveness of the proposed approach.

Recently, extreme events have highlighted their potentially tragic effects on structural and infrastructure systems. Resilience of the Community to these extreme vents is an important issue of increasing more concern for developing design methods. Such extreme events scenarios involve many uncertainties, such as the intensity, location, and period. The extreme events may include those caused by various natural or manmade hazards, such as earthquake, strong winds, fire, blast, etc. Compared to other events, earthquake and wind are particularly critical due to their significant threats to the global structure performance and more challenges for design. Researchers have recognized that proper evaluation, modeling, and assessment of the effects of extreme events are fundamental to ensure the desired performance of structures. Therefore, the concern for developing appropriate methodologies to evaluate and design structures that can withstand the effects of extreme events has become a very active field of research in recent years. Improvement of building codes and development of new strategies are needed to mitigate the disastrous effects of extreme events. This paper presents a comprehensive review of literature surrounding designing building structures for extreme events. First, a general overview of the extreme events design and different objectives of approaches is conducted. Furthermore, a review related literature surrounding designing for earthquake resistance guidelines is presented, also highlights Performance-Based Seismic Design objectives. The available literature includes many studies for the provisions included in different design codes (China, United States and Europe). A review of literature related to wind resistance design with an overview of Performance Based Wind Design of building design method for the control of winds impacting on building structures is also presented.

Expert networks are formed by a group of expert-professionals with different specialties to collaboratively resolve specific queries posted to the network. In such networks, when a query reaches an expert who does not have sufficient expertise, this query needs to be routed to other experts for further processing until it is completely solved; therefore, query answering efficiency is sensitive to the underlying query routing mechanism being used. Among all possible query routing mechanisms, decentralized search, operating purely on each expert’s local information without any knowledge of network global structure, represents the most basic and scalable routing mechanism, which is applicable to any network scenarios even in dynamic networks. However, there is still a lack of fundamental understanding of the efficiency of decentralized search in expert networks. In this regard, we investigate decentralized search by quantifying its performance under a variety of network settings. Our key findings reveal the existence of network conditions, under which decentralized search can achieve significantly short query routing paths (i.e., between O (log n ) and O (log 2 n ) hops, n : total number of experts in the network). Based on such theoretical foundation, we further study how the unique properties of decentralized search in expert networks are related to the anecdotal small-world phenomenon. In addition, we demonstrate that decentralized search is robust against estimation errors introduced by misinterpreting the required expertise levels. The developed performance bounds, confirmed by real datasets, are able to assist in predicting network performance and designing complex expert networks.

The ostentatiously steep increase in the number of malwares a year has given rise to a growing demand for a decision support system which would help against the same. The reason for such a steep increase in malware is because of slight mutation added to them to avoid detection. Thus, the malware from the same family with slight mutation is proliferated across the internet causing harm. We have come up with a Convolutional Neural Network (CNN) (Ciresan DC, Meier U, Masci J, Maria Gambardella L, Schmidhuber J, IJCAI Proceedings-international joint conference on artificial intelligence, vol 22, p 1237 (2011)) approach to battle a quarter of a million malwares a day, which heavily relies on the classification of the malwares based on the visual resemblance of malware samples of the same family. Binaries of malwares are represented visually as it has the ability to retain changes while keeping the global structure intact to help in detecting variations. The feasibility and performance of the proposed approach have been tested on the Malimg dataset which has shown avant-garde performance results. 98.97% precision was obtained by the proposed process.

Accurate identification of influential nodes facilitates the control of rumor propagation and interrupts the spread of computer viruses. Many classical approaches have been proposed by researchers regarding different aspects. To explore the impact of location information in depth, this paper proposes an improved global structure model to characterize the influence of nodes. The method considers both the node’s self-information and the role of the location information of neighboring nodes. First, degree centrality of each node is calculated, and then degree value of each node is used to represent self-influence, and degree values of the neighbor layer nodes are divided by the power of the path length, which is path attenuation used to represent global influence. Finally, an extended improved global structure model that considers the nearest neighbor information after combining self-influence and global influence is proposed to identify influential nodes. In this paper, the propagation process of a real network is obtained by simulation with the SIR model, and the effectiveness of the proposed method is verified from two aspects of discrimination and accuracy. The experimental results show that the proposed method is more accurate in identifying influential nodes than other comparative methods with multiple networks.

The aim of the work is to substantiate and predict measures to ensure the increase in the values of the target indicators of the university's activity in the international institutional rating QS to the values necessary to increase the rating in subsequent years. Mathematical modeling of cognitive maps is used in conjunction with selective dendritic networks of neurons. The possibility of increasing the efficiency of using cognitive maps for predicting the development of a university using selective dendritic networks of neurons is shown. The proposed structure of hybrid cognitive maps allows you to naturally simplify the structure by removing non-working network connections of the global structure. The universal matrix description of the mathematical model of the network allows you to effectively form a computational algorithm and software for predicting the indicators of the development of the university. The results obtained allowed us to form a scenario plan for the necessary step-by-step increase in the values of target indicators, taking into account the hidden factors affecting them. The possibility of forming a cognitive map based on a multilayer dendritic network of direct propagation in the presence of unidirectional content links is shown.

XRP is a modern crypto-asset (crypto-currency) developed by Ripple Labs, which has been increasing its financial presence. We study its transaction history available as ledger data. An analysis of its basic statistics, correlations, and network properties are presented. Motivated by the behavior of some nodes with histories of large transactions, we propose a new index: the ``Flow Index.'' The Flow Index is a pair of indices suitable for characterizing transaction frequencies as a source and destination of a node. Using this Flow Index, we study the global structure of the XRP network and construct bow-tie/walnut structure., Comment: 15 pages, 15 Figures, for "Blockchain Kyoto 2011" JPS Conference Proceedings

One of the most significant challenges for machine learning nowadays is the discovery of causal relationships from data. This causal discovery is commonly performed using Bayesian like algorithms. However, more recently, more and more causal discovery algorithms have appeared that do not fall into this category. In this paper, we present a new algorithm that explores global causal association rules with Uncertainty Coefficient. Our algorithm, CRPA-UC, is a global structure discovery approach that combines the advantages of association mining with causal discovery and can be applied to binary and non-binary discrete data. This approach was compared to the PC algorithm using several well-known data sets, using several metrics.

Map construction is the problem of reconstructing a travel network based on trajectory data of entities travelling on the network. While many map construction algorithms reconstruct the global structure of a network well, local features such as location of crossings and turns are generally harder to reconstruct correctly, in particular for noisy and irregularly sampled data. We demonstrate how subtrajectory clustering can be used to construct maps that capture both the global structure and local features well.

Contextualized neural language models have gained much attention in Information Retrieval (IR) with its ability to achieve better text understanding by capturing contextual structure. However, to achieve better document understanding, it is necessary to involve global structure of a document. In this paper, we take the advantage of Graph Convolutional Networks (GCN) to model global word-relation structure of a document to improve context-aware document ranking. We propose to build a graph for a document to model the global structure. The nodes and edges of the graph are constructed from contextual embeddings. Then we apply graph convolution on the graph to learning a new representation, and this representation covers both contextual and global structure information. The experimental results show that our method outperforms the state-of-the-art contextual language models, which demonstrate that incorporating global structure is useful for improving document ranking and GCN is an effective way to achieve it.

Entity alignment (EA) aims to identify entities located in different knowledge graphs (KGs) that refer to the same real-world object. To learn the entity representations, most EA approaches rely on either translation-based methods which capture the local relation semantics of entities or graph convolutional networks (GCNs), which exploit the global KG structure. Afterward, the aligned entities are identified based on their distances. In this paper, we propose to jointly leverage the global KG structure and entity-specific relational triples for better entity alignment. Specifically, a global structure and local semantics preserving network is proposed to learn entity representations in a coarse-to-fine manner. Experiments on several real-world datasets show that our method significantly outperforms other entity alignment approaches and achieves the new state-of-the-art performance.

We discuss Manchak (2009a)'s result that there are locally (but not globally) isometric universes observationally indistinguishable from our own. This theorem makes the epistemic predicament of modern cosmology particularly problematic and the prospects of ever gaining knowledge of the global structure of the universe rather unlikely in the context of general relativity. We argue however that this conclusion is too quick; indeed, Manchak's theorem deploys spacetimes which are not physically reasonable, since they have features which are not the product of any physical process. This ultimately rests on the fact that local isometry between two spacetimes is not sufficient to guarantee that they are both physically reasonable. We propose an additional condition to properly define when a spacetime is physically reasonable, and we show that Manchak's spacetimes do not satisfy this further demand.

Fine-grained vehicle recognition is a challenging problem due to the subtle intra-category appearance variation, which requires the recognition model can capture discriminative features from distinguishing regions. The structure is an important characteristic of vehicles which can help to find substantial parts and learn distinguishing representations. In this paper, we propose an approach that introduces the structure graph into consideration to learn distinguishing representations for vehicle recognition. Our proposed method first constructs a global structure graph from the features generated by the convolutional network and then it applies the graph as the guidance to produce effective representations of vehicles. The results of extensive experiments demonstrate that our proposed method can produce more promising results than other state-of-the-art methods. The results of the visualization illustrate that our approach can construct a suitable structure graph and the global structure information facilitates learning discriminative representations at crucial parts of vehicles.

Although deep learning-based approaches have been very effective in solving problems with plenty of labeled data, they suffer in tackling problems for which labeled data are scarce. In few-shot classification, the objective is to train a classifier from only a handful of labeled examples in a support set. In this paper, we propose a few-shot learning framework based on structured margin loss which takes into account the global structure of the support set in order to generate a highly discriminative feature space where the features from distinct classes are well separated in clusters. Moreover, in our meta-learning-based framework, we propose a context-aware query embedding encoder for incorporating support set context into query embedding and generating more discriminative and task-dependent query embeddings. The task-dependent features help the metalearner to learn a distribution over tasks more effectively. Extensive experiments based on few-shot, zero-shot and semi-supervised learning on three benchmarks show the advantages of the proposed model compared to state-of-the- art.

Unaffordable memory requirement becomes a bottleneck for the cell mapping method when global analysis is carried out in a chosen domain of the state space with a specific computer in order to depict complex invariant sets with high resolution, and/or in high dimensions. This may also be often encountered even though the subdivision technique is adopted. Thus a subdomain synthesis method for cell mapping is proposed in this paper in order to release this difficulty. The proposed method first divides the chosen domain into smaller subdomains with affordable memory requirements, then carries the cell mapping analysis out on each subdomain independently meanwhile records the message on inter-subdomain mappings, finally recovers the global structure in the complete chosen domain based on the dynamical information revealed in the subdomains. The method can effectively avoid the loss of the invariant sets in the chosen domain that are unavoidably sliced into pieces by the state space partition without prior knowledge on them. Furthermore, both the subdivision technique and the parallel computing based on graphical processing units can also be incorporated into the proposed method that enhances the computation efficiency. Three examples of application on global analysis of nonlinear dynamical systems are presented to demonstrate the basic idea and the potential of the proposed method.

This paper proposes a new unsupervised spectral feature selection method to preserve both the local and global structure of the features as well as the samples. Specifically, our method uses the self-expressiveness of the features to represent each feature by other features for preserving the local structure of features, and a low-rank constraint on the weight matrix to preserve the global structure among samples as well as features. Our method also proposes to learn the graph matrix measuring the similarity of samples for preserving the local structure among samples. Furthermore, we propose a new optimization algorithm to the resulting objective function, which iteratively updates the graph matrix and the intrinsic space so that collaboratively improving each of them. Experimental analysis on 12 benchmark datasets showed that the proposed method outperformed the state-of-the-art feature selection methods in terms of classification performance.

An alternative formulation of the theory of generalized resonance structures (Clar covers) of single zigzag chains N(n), based on the novel concept of interfaces, is presented. The global structure of every Clar cover can be conveniently expressed as a unique sequence of n interfaces. The complete set of Clar covers is then robustly constructed as the complete set of walks of length $$n+1$$ on the associated interface connectivity graph. The presented algorithm is readily generalizable to an arbitrary class of benzenoid structures.

Multiple kernel clustering (MKC) based on global structure alignment (GSA) has unified many existing MKC algorithms, and shown outstanding clustering performance. However, we observe that most of existing GSA-based MKC algorithms only maximally align global structure of data with an ideal similarity matrix, while ignoring the local geometrical structure hidden in data, which is regarded to be important in improving the clustering performance. To address this issue, we propose a global and local structure alignment framework for MKC (GLSAMKC) which well considers both the alignment between the global structure and local structure of data with the same ideal similarity matrix. To illustrate the effectiveness of the proposed framework, we instantiate two specific GLSAMKC-based algorithms by exploiting the local structure with local linear embedding and locality preserving projection, respectively. A two-step alternate iterative and convergent optimization algorithm is developed to implement the resultant optimization problem. Extensive experimental results on five benchmark data sets demonstrate the superiority of proposed algorithms compared with the many state-of-the-art MKC algorithms, indicating the effectiveness of the proposed framework.

Graph plays crucially important roles in graph-based semi-supervised learning (SSL). Most SSL methods construct a single graph over all instances to explore the manifold structure of instances, and then enforce the smoothness constraint over such graph. However, instances in the real world are not always evenly distributed. Some instances from different classes but close to decision boundary may be close to each other, and thus they are easy to be misclassified. To mitigate this issue, we propose an approach called semi-supervised classification based on clustering adjusted similarity (SSC-CAS). SSC-CAS firstly takes advantage of clustering on both labeled and unlabeled instances to explore the global structure and discrimination of instances, and then quantifies the similarity between pairwise cluster centers. Second, it adjusts the similarity between pairwise instances by multiplying the similarity between centers of clusters they belong to. In this way, if two instances are from different clust...

Drug targets are biomacromolecules or biomolecular structures that bind to specific drugs and produce therapeutic effects. Therefore, the prediction of drug-target interactions (DTIs) is important for disease therapy. Incorporating multiple similarity measures for drugs and targets is of essence for improving the accuracy of prediction of DTIs. However, existing studies with multiple similarity measures ignored the global structure information of similarity measures, and required manual extraction features of drug-target pairs, ignoring the non-linear relationship among features. In this paper, we proposed a novel approach MDADTI for DTIs prediction based on MDA. MDADTI applied random walk with restart method and positive pointwise mutual information to calculate the topological similarity matrices of drugs and targets, capturing the global structure information of similarity measures. Then, MDADTI applied multimodal deep autoencoder to fuse multiple topological similarity matrices of drugs and targets, automatically learned the low-dimensional features of drugs and targets, and applied deep neural network to predict DTIs. The results of 5-repeats of 10-fold cross-validation under three different cross-validation settings indicated that MDADTI is superior to the other four baseline methods. In addition, we validated the predictions of the MDADTI in six drug-target interactions reference databases, and the results showed that MDADTI can effectively identify unknown DTIs.

In the field of complex networks, how to identify influential nodes is a significant issue in analyzing the structure of a network. In the existing method proposed to identify influential nodes based on the local dimension, the global structure information in complex networks is not taken into consideration. In this paper, a node information dimension is proposed by synthesizing the local dimensions at different topological distance scales. A case study of the Netscience network is used to illustrate the efficiency and practicability of the proposed method.

Pooling layers are crucial components for efficient deep representation learning. As to graph data, however, it’s not trivial to decide which nodes to retain in order to represent the high-level structure of a graph. Recently many different graph pooling methods have been proposed. However, they all rely on local features to conduct global pooling over all nodes, which contradicts poolings in CNNs that only use local features to conduct local pooling. We analyze why this may hinder the performance of graph pooling, then propose a novel graph pooling method called Bottom-Up and Top-Down graph POOLing (BUTDPool). BUTDPool aims to learn a more fine-grained pooling criterion based on coarse global structure information produced by a bottom-up pooling layer, and can enhance local features with global features. Specifically, we propose to use one or multiple pooling layers with a relatively high retain ratio to produce a coarse high-level graph. Injecting the high-level information back into low-level representation, BUTDPool enhances learning a better pooling criterion. Experiments demonstrate the superior performance of the proposed method over compared methods.

Local optima networks (LONs) are a useful tool to analyse and visualise the global structure of fitness landscapes. The main goal of our study is to use LONs to contrast the global structure of synthetic benchmark functions against those of real-world continuous optimisation problems of similar dimensions. We selected two real-world problems, namely, an engineering design problem and a machine learning problem. Our results indicate striking differences in the global structure of synthetic vs real-world problems. The real-world problems studied were easier to solve than the synthetic ones, and our analysis reveals why; they have easier to traverse global structures with fewer nodes and edges, no sub-optimal funnels, higher neutrality and multiple global optima with shorter trajectories towards them.

Traditional supervised keyphrase extraction models depend on the features of labeled keyphrases while prevailing unsupervised models mainly rely on global structure of the word graph, with nodes representing candidate words and edges/links capturing the co-occurrence between words. However, the local context information of the word graph can not be exploited in existing unsupervised graph-based keyphrase extraction methods and integrating different types of information into a unified model is relatively unexplored. In this paper, we propose a new word embedding model specially for keyphrase extraction task, which can capture local context information and incorporate them with other types of crucial information into the low-dimensional word vector to help better extract keyphrases. Experimental results show that our method consistently outperforms 7 state-of-the-art unsupervised methods on three real datasets in Computer Science area for keyphrase extraction.

Emerging self-ensembling methods have achieved promising semi-supervised segmentation performances on medical images through forcing consistent predictions of unannotated data under different perturbations. However, the consistency only penalizes on independent pixel-level predictions, making structure-level information of predictions not exploited in the learning procedure. In view of this, we propose a novel structure-aware entropy regularized mean teacher model to address the above limitation. Specifically, we firstly introduce the entropy minimization principle to the student network, thereby adjusting itself to produce high-confident predictions of unannotated images. Based on this, we design a local structural consistency loss to encourage the consistency of inter-voxel similarities within the same local region of predictions from teacher and student networks. To further capture local structural dependencies, we enforce the global structural consistency by matching the weighted self-information maps between two networks. In this way, our model can minimize the prediction uncertainty of unannotated images, and more importantly that it can capture local and global structural information and their complementarity. We evaluate the proposed method on a publicly available 3D left atrium MR image dataset. Experimental results demonstrate that our method achieves outstanding segmentation performances than the state-of-the-art approaches in scenes with limited annotated images.

One of the most ubiquitous analysis tools employed in single-cell transcriptomics and cytometry is t-distributed stochastic neighbor embedding (t-SNE) [1], used to visualize individual cells as points on a 2D scatter plot such that similar cells are positioned close together. Recently, a related algorithm, called uniform manifold approximation and projection (UMAP) [2] has attracted substantial attention in the single-cell community. In Nature Biotechnology, Becht et al. [3] argued that UMAP is preferable to t-SNE because it better preserves the global structure of the data and is more consistent across runs. Here we show that this alleged superiority of UMAP can be entirely attributed to different choices of initialization in the implementations used by Becht et al.: t-SNE implementations by default used random initialization, while the UMAP implementation used a technique called Laplacian eigenmaps [4] to initialize the embedding. We show that UMAP with random initialization preserves global structure as poorly as t-SNE with random initialization, while t-SNE with informative initialization performs as well as UMAP with informative initialization. Hence, contrary to the claims of Becht et al., their experiments do not demonstrate any advantage of the UMAP algorithm per se, but rather warn against using random initialization.

A wide variety of deep neural network models for graph-structured data have been proposed to solve tasks like node/graph classification and link prediction. By effectively learning low-dimensional embeddings of graph nodes, they have shown state-of-the-art performance. However, most existing models learn node embeddings by exploring flat information propagation across the edges within the local neighborhood of each node. We argue that incorporating hierarchical node embeddings can capture the inherently hierarchical topological features of many realistic graphs such as social networks, biological network and World Wide Web. In this paper we propose GRAHIES, a general framework for graph neural networks to learn node representations that preserve hierarchical graph information at higher-orders. GRAHIES adaptively learns a multi-level hierarchical structure of the input graph, which consists of successively coarser (smaller) graphs that preserve the global structure of the original graphs at different levels. By combining the graph representations from different levels of the graph hierarchy, the final node representation captures the inherent global hierarchical structure of the original graph. Our experiments show that applying GRAHIES's hierarchical paradigm yields improved accuracy for existing graph neural networks on the node classification tasks.

Accumulating evidence has demonstrated that RNAs play an important role in identifying various complex human diseases. However, the number of known disease related RNAs is still small and many biological experiments are time-consuming and labor-intensive. Therefore, researchers have focused on developing useful computational algorithms to predict associations between diseases and RNAs. It is useful for people to identify complex human diseases at molecular level, especially in diseases diagnosis, therapy, prognosis and monitoring. In this paper, we propose a novel framework Graph Convolutional Attention Network(GCAN) to predict potential disease-RNAs associations. Facing thousands of associations, GCAN benefits from the efficiency of deep learning model. Compared to other disease-RNAs association prediction methods, GCAN operates the computation process from global structure of disease-RNAs network with graph convolution networks(GCN) and can also integrate local neighborhoods with the attention mechanism. What is more, GCAN is at the first attempt to utilize GCN to discover the feature representation of the latent nodes in disease-RNAs network. In order to evaluate the performance of GCAN, we conduct experiments on two different disease-RNAs networks: disease-miRNA and disease-lncRNA. Comparisons of several state-of-the-art methods using disease-RNAs networks show that our novel frameworks outperform baselines by a wide margin in potential disease-RNAs associations.

We propose a Bayesian nonparametric model based on Markov Chain Monte Carlo methods for unveiling the structure of the invariant global stable manifold from observed time-series data. The underlying unknown dynamical process could have been contaminated by additive noise. We introduce the Stable Manifold Geometric Stick Breaking Reconstruction model with which we reconstruct the unknown dynamic equations, while at the same time, we estimate the global structure of the perturbed stable manifold. Our method works for noninvertible maps without modifications. The stable manifold estimation procedure is demonstrated specifically in the case of polynomial maps. Simulations based on synthetic time-series are presented.We propose a Bayesian nonparametric model based on Markov Chain Monte Carlo methods for unveiling the structure of the invariant global stable manifold from observed time-series data. The underlying unknown dynamical process could have been contaminated by additive noise. We introduce the Stable Manifold Geometric Stick Breaking Reconstruction model with which we reconstruct the unknown dynamic equations, while at the same time, we estimate the global structure of the perturbed stable manifold. Our method works for noninvertible maps without modifications. The stable manifold estimation procedure is demonstrated specifically in the case of polynomial maps. Simulations based on synthetic time-series are presented.

The article examines the problem of cognitive model of the micro text. Here the complex sentence is learned as a micro text in the basis of hierarchical relations of its components. In the process of the analysis of language material local and global structure of the components of the microtext is taken into account, which is important for the definition of its cognitive model. In general, the micro text is defined as a global structure, while its separate components are studied as local structures. It is known, that each syntactic structure has a definite situation beyond the microtext, and its components also have a situation. This is also a basic definition of general and private cognitive model of the microtext.

It is still a crucial issue to identify influential nodes effectively in the study of complex networks. As for the existing efficiency centrality (EffC), it cannot be applied to a weighted network. In this paper, a modified efficiency centrality (EffC $$^\mathrm{m}$$ ) is proposed by extending EffC into weighted networks. The proposed measure trades off the node degree and global structure in a weighted network. The influence of both the sum of the average degree of nodes in the whole network and the average distance of the network is taken into account. Numerical examples are used to illustrate the efficiency of the proposed method.

We have developed a software package, namely, PASP (Property Analysis and Simulation Package for materials), to analyze the structural, electronic, magnetic, and thermodynamic properties of complex condensed matter systems. Our package integrates several functionalities including symmetry analysis, global structure searching methods, effective Hamiltonian methods, and Monte Carlo simulation methods. In conjunction with first-principles calculations, PASP has been successfully applied to diverse physical systems. In this paper, we give a brief introduction to its main features and underlying theoretical formulism. Some typical applications are provided to demonstrate the usefulness, high efficiency, and reliability of PASP. We expect that further developments will make PASP a general-purpose tool for material simulation and property calculation of condensed matters.

The article describes how the radio station presents itself in the news service and elaborates on the position of self-presentation in the global structure of the news service and in single news programme. The purpose of the article is to discuss the roles, function and qualities of the radio stations as well as the language that is being used. The paper is based on findings from more than 100 news services broadcast by 15 radio stations located in Warsaw, Cracow and Poznan.

A novel high-order three-scale reduced homogenization (HTRH) approach is introduced for analyzing the nonlinear heterogeneous materials with multiple periodic microstructure. The first-order, second-order and high-order local cell solutions at microscale and mescoscale gotten by solving the distinct multiscale cell functions are derived at first. Then, two kinds of homogenized parameters are calculated, and the nonlinear homogenization equations defined on global structure are evaluated, successively. Further, the displacement and stress fields are established as high-order multiscale approximate solutions by assembling the various unit cell solutions and homogenization solutions. The significant characteristics of the presented approach are an efficient reduced model form for solving high-order nonlinear local cell problems, and hence reducing the computational cost in comparison to direct computational homogenization. Besides, the new asymptotic high-order nonlinear homogenization does not need higher order continuities of the coarse-scale (or macroscale) solutions. Finally, by some representative examples, the efficiency and accuracy of the presented algorithms are verified. The numerical results clearly illustrate that the HTRH approach reported in this work is effective and accurate to predict the macroscopic nonlinear properties, and can capture the microscale and mesoscale behavior of the composites accurately.

Graph Neural Networks (GNNs) have received wide acclaim in recent times due to their performance on inference tasks for unstructured data. Typically, GNNs operate by exploiting local structural information in graphs and disregarding their global structure. This is influenced by assumptions of homophily and unbiased class distributions. As a result, this could impede model performance on noisy real-world graphs such as spatial graphs where these assumptions may not be sufficiently held. In this article, we study the problem of graph learning on spatial graphs. Particularly, we focus on transductive learning methods for the imbalanced case. Given the nature of these graphs, we hypothesize that taking the global structure of the graph into account when aggregating local information would be beneficial especially with respect to generalisability. Thus, we propose a novel approach to training GNNs for these type of graphs. We achieve this through a sampling technique: Structure-Aware Sampling (SAS), which leverages the intra-class and global-geodesic distances between nodes. We model the problem as a node classification one for street networks with high variance between class sizes. We evaluate our approach using large real-world graphs against state-of-the-art methods. In the majority of cases, our approach outperforms traditional methods by up to a mean F1-score of 20%.

Indexing exploits assumptions on the inner structures of a dataset to make the nearest neighbor queries cheaper to resolve. Datasets are generally indexed at once into a unique index for similarity search. By indexing a given dataset as a whole, one faces the parameters of its global structure, which may be adverse. A typical well-studied example is a high global dimensionality of the dataset, making any indexing strategy inefficient due to the curse of dimensionality. We conjecture that a dataset may be partitioned into subsets of variable indexability. The strategy is, therefore, to define a procedure to extract parts of the dataset with predictable indexability and to adapt the index structure to this parameter. In this paper, we define and discuss indexability related to the curse of dimensionality and propose a related heuristic to partition the dataset into low-dimensional parts. Each data object is ranked according to its degree centrality, under a connected sparse graph, the Half-Space Proximal Graph (HSP). We postulate centrality measures are good predictors of dimensionality and indexability. In view of validation, we conducted an experiment using the degree centrality of the HSP graph as unique dimensionality/indexability measure. We ranked the data objects by their respective centrality degree under the HSP graph, then extracted the lower dimensional subsets, recomputed the HSP and repeated. Subsets were then indexed with an exact method in increasing, decreasing and random order. We measured the complexity of a fixed set of queries for each of the three arrangements. For each set we used a fixed dataset with 250 queries. The above single experiment demonstrated that the heuristic can extract low dimensional subsets, and also that those subsets are easier to index. This initial results demonstrate the validity of our conjecture and motivate the need for exploring further the notion of indexability and related dataset partitioning strategies.