Search

Improving the capability of models at limited computational cost is an urgent demand in many vision-based Internet-of-Things applications. Recent progress on deep convolutional neural network (CNN) has largely accelerated the development of image classification. Although the hierarchical structure of CNN naturally helps to extract image features in different scales and locations progressively, conventional convolution can only handle contexts of one scale and on a limited area of a single location in a specific layer, limiting the utilization of multiscale and multilocation information. In this work, we present a cyclic CNN framework, which enables sufficient utilization of multiscale and multilocation contexts in a single layer of convolution. The cyclic CNN is an extremely simple but effective improvement upon conventional convolution, which occupies no additional parameter and negligible computation (even less than 0.1%). Moreover, cyclic CNN can be easily plugged into many existing CNN pipelines, e.g., the ResNet family, obtaining extremely low-cost performance gain upon them. Extensive experiments on both small-scale (CIFAR10 and CIFAR100) and large-scale (ILSVRC2012) image classification benchmarks demonstrate that a consistent performance promotion is obtained with the help of cyclic CNN.

Real-world data is often organized at different levels of granularity with specified concept hierarchies. Multi-scale information tables represent such a type of data set in a hierarchical form. In this paper, we aim to focus on rule extraction in multi-scale decision tables. Unlike the case of attribute reduct, we consider a type of generalization reduct in multi-scale decision tables. This approach requires both the least number of attributes and the coarsest level of scales according to different reduct standards. The present study is conducted at the three different levels, that is, generalization reducts for objects, decision rules and multi-scale decision tables, respectively. The construction of granularity trees and the selection of cuts play a crucial role. At each level, we present various types of generalization reducts according to different requirements. The relationship between them is also investigated. Moreover, a comparative study between generalization reducts and attribute reducts is also performed. Based on the notion of generalization reducts, the procedure to extract the set of optimal decision rule in multi-scale decision tables is provided and an illustrative example is also given to show the proposed approach.

The increasing availability of single-cell data revolutionizes the understanding of biological mechanisms at cellular resolution. For differential expression analysis in multi-subject single-cell data, negative binomial mixed models account for both subject-level and cell-level overdispersions, but are computationally demanding. Here, we propose an efficient NEgative Binomial mixed model Using a Large-sample Approximation (NEBULA). The speed gain is achieved by analytically solving high-dimensional integrals instead of using the Laplace approximation. We demonstrate that NEBULA is orders of magnitude faster than existing tools and controls false-positive errors in marker gene identification and co-expression analysis. Using NEBULA in Alzheimer’s disease cohort data sets, we found that the cell-level expression of APOE correlated with that of other genetic risk factors (including CLU, CST3, TREM2, C1q, and ITM2B) in a cell-type-specific pattern and an isoform-dependent manner in microglia. NEBULA opens up a new avenue for the broad application of mixed models to large-scale multi-subject single-cell data., The application of negative binomial mixed models (NBMMs) to single-cell data is computationally demanding. To address this issue, Liang He et al. have developed NEBULA, an efficient algorithm that can analyze differential gene expression or co-expression networks in multi-subject single-cell data sets, and validate it on snRNA-seq and scRNA-seq data sets comprising ~200k cells from cohorts of Alzheimer’s disease and multiple sclerosis patients.

The main purpose of the presented paper is to obtain some time scale inequalities for different divergences and distances by using weighted time scales Jensen’s inequality. These results offer new inequalities in h-discrete calculus and quantum calculus and extend some known results in the literature. The lower bounds of some divergence measures are also presented. Moreover, the obtained discrete results are given in the light of the Zipf–Mandelbrot law and the Zipf law.

This article presents a multiple objective fractional fixed-charge transportation problem (MFFTP) in a rough decision-making framework. A transformation procedure is modified to convert non-linear multi-objective transportation problem to its linear version. The parameters of the designed model are considered to be fuzzy. We employ separate kinds of fuzzy scale, i.e., possibility, credibility and necessity measures, to deal with the fuzzy parameters. Using the fuzzy chance-constrained rough approximation (FCRA) technique, we extract the more preferable optimal solution from our suggested MFFTP. The initial result is compared with that of the robust ranking (RR) technique. We also use the theory of rough sets for expanding as well as dividing the feasible domain of the MFFTP to accommodate more information by considering two approximations. Employing these approximations, we introduce two variants, namely, the lower approximation (LA) and the upper approximation (UA), of the suggested MFFTP. Finally, by using these models, we provide the optimal solutions for our proposedproblem. We also associate our MFFTP with a real-world example to showcase its applicability as well as performance. Our core concept of this article is that it tackles an MFFTP using two separate kinds of uncertainty and expands its feasible domain for optimal solutions. Optimal solutions of the designed model (obtained from FCRA technique) belong to two separate regions, namely, “surely region” and “possible region”. The optimal solution which belongs to the “surely region” is better (as these are minimum values) than the one in the “possible region” and other cases. An interpretation of our approach along with offers about the intended future research work are provided at last.

Large-scale group decision-making (LSGDM) deals with complex decision- making problems which involve a large number of decision makers (DMs). Such a complex scenario leads to uncertain contexts in which DMs elicit their knowledge using linguistic information that can be modelled using different representations. However, current processes for solving LSGDM problems commonly neglect a key concept in many real-world decision-making problems, such as DMs' regret aversion psychological behavior. Therefore, this paper introduces a novel consensus based linguistic distribution LSGDM (CLDLSGDM) approach based on a statistical inference principle that considers DMs' regret aversion psychological characteristics using regret theory and which aims at obtaining agreed solutions. Specifically, the CLDLSGDM approach applies the statistical inference principle to the consensual information obtained in the consensus process, in order to derive the weights of DMs and attributes using the consensus matrix and adjusted decision-making matrices to solve the decision-making problem. Afterwards, by using regret theory, the comprehensive perceived utility values of alternatives are derived and their ranking determined. Finally, a performance evaluation of public hospitals in China is given as an example in order to illustrate the implementation of the designed method. The stability and advantages of the designed method are analyzed by a sensitivity and a comparative analysis.

We introduce and study bi-Lipschitz-invariant dimensions that range between the box and Assouad dimensions. The quasi-Assouad dimensions and $\theta$-spectrum are other special examples of these intermediate dimensions. These dimensions are localized, like Assouad dimensions, but vary in the depth of scale which is considered, thus they provide very refined geometric information. We investigate the relationship between these and the familiar dimensions. We construct a Cantor set with a non-trivial interval of dimensions, the endpoints of this interval being given by the quasi-Assouad and Assouad dimensions of the set. We study continuity-like properties of the dimensions. In contrast with the Assouad-type dimensions, we see that decreasing sets in $\mathbb{R}$ with decreasing gaps need not have dimension $0$ or $1$. Formulas are given for the dimensions of Cantor-like sets and these are used in some of our constructions. We also show that, as is the case for Hausdorff and Assouad dimensions, the Cantor set and the decreasing set have the extreme dimensions among all compact sets in $\mathbb{R}$ whose complementary set consists of open intervals of the same lengths., Comment: Updated reference

Katz centrality is a fundamental concept to measure the influence of a vertex in a social network. However, existing approaches to calculating Katz centrality in a large-scale network are unpractical and computationally expensive. In this article, we propose a novel method to estimate Katz centrality based on graph sampling techniques, which object to achieve comparable estimation accuracy of the state-of-the-arts with much lower computational complexity. Specifically, we develop a Horvitz–Thompson estimate for Katz centrality by using a multi-round sampling approach and deriving an unbiased mean value estimator. We further propose SAKE , a S ampling-based A lgorithm for fast K atz centrality E stimation. We prove that the estimator calculated by SAKE is probabilistically guaranteed to be within an additive error from the exact value. Extensive evaluation experiments based on four real-world networks show that the proposed algorithm can estimate Katz centralities for partial vertices with low sampling rate, low computation time, and it works well in identifying high influence vertices in social networks.

Streaming graph partitioning methods have recently gained attention due to their ability to scale to very large graphs with limited resources. However, many such methods do not consider workload and graph characteristics. This may degrade the performance of queries by increasing inter-node communication and computational load imbalance. Moreover, existing workload-aware methods cannot consistently provide good performance as they do not consider dynamic workloads that keep emerging in graph applications. We address these issues by proposing a novel workload-adaptive streaming partitioner named WASP, that aims to achieve low-latency and high-throughput online graph queries. As each workload typically contains frequent query patterns, WASP exploits the existing workload to capture active vertices and edges which are frequently visited and traversed, respectively. This information is used to heuristically improve the quality of partitions either by avoiding the concentration of active vertices in a few partitions proportional to their visit frequencies or by reducing the probability of the cut of active edges proportional to their traversal frequencies. In order to assess the impact of WASP on a graph store and to show how easily the approach can be plugged on top of the system, we exploit it in a distributed graph-based RDF store. Our experiments over three synthetic and real-world graph datasets and the corresponding static and dynamic query workloads show that WASP achieves a better query performance against state-of-the-art graph partitioners, especially in dynamic query workloads.

Quasi-cliques are dense incomplete subgraphs of a graph that generalize the notion of cliques. Enumerating quasi-cliques from a graph is a robust way to detect densely connected structures with applications in bioinformatics and social network analysis. However, enumerating quasi-cliques in a graph is a challenging problem, even harder than the problem of enumerating cliques. We consider the enumeration of top- k degree-based quasi-cliques and make the following contributions: (1) we show that even the problem of detecting whether a given quasi-clique is maximal (i.e., not contained within another quasi-clique) is NP-hard. (2) We present a novel heuristic algorithm K ernel QC to enumerate the k largest quasi-cliques in a graph. Our method is based on identifying kernels of extremely dense subgraphs within a graph, followed by growing subgraphs around these kernels, to arrive at quasi-cliques with the required densities. (3) Experimental results show that our algorithm accurately enumerates quasi-cliques from a graph, is much faster than current state-of-the-art methods for quasi-clique enumeration (often more than three orders of magnitude faster), and can scale to larger graphs than current methods.

We introduce a Gaussian process latent factor model for multi-label classification that can capture correlations among class labels by using a small set of latent Gaussian process functions. To address computational challenges, when the number of training instances is very large, we introduce several techniques based on variational sparse Gaussian process approximations and stochastic optimization. Specifically, we apply doubly stochastic variational inference that sub-samples data instances and classes which allows us to cope with Big Data. Furthermore, we show it is possible and beneficial to optimize over inducing points, using gradient-based methods, even in very high dimensional input spaces involving up to hundreds of thousands of dimensions. We demonstrate the usefulness of our approach on several real-world large-scale multi-label learning problems.

Nowadays, the scale of various graphs soars rapidly, which imposes a serious challenge to develop processing and analytic algorithms. Among them, graph pattern matching is the one of the most primitive tasks that find a wide spectrum of applications, the performance of which is yet often affected by the size and dynamicity of graphs. In order to handle large dynamic graphs, incremental pattern matching is proposed to avoid re-computing matches of patterns over the entire data graph, hence reducing the matching time and improving the overall execution performance. Due to the complexity of the problem, little work has been reported so far to solve the problem, and most of them only solve the graph pattern matching problem under the scenario of the data graph varying alone. In this article, we are devoted to a more complicated but very practical graph pattern matching problem, continuous matching of evolving patterns over dynamic graph data, and the investigation presents a novel algorithm for continuously pattern matching along with changes of both pattern graph and data graph. Specifically, we propose a concise representation of partial matching solutions, which can help to avoid re-computing matches of the pattern and speed up subsequent matching process. In order to enable the updates of data graph and pattern graph, we propose an incremental maintenance strategy, to efficiently maintain the intermediate results. Moreover, we conceive an effective model for estimating step-wise cost of pattern evaluation to drive the matching process. Extensive experiments verify the superiority of .

We present an overview of fine-scale features in the Sun’s atmosphere, with a focus on spicules and jets. We consider older and newer observations and theories for chromospheric spicules and coronal jets. We also consider the connection between these features and some other solar atmospheric phenomena. We then discuss the possibility that there is a continuum of jet-like features ranging from spicules to large-scale CME-producing eruptions, all driven by similar magnetic processes operating on differing corresponding size scales. Future observational and theoretical studies will help clarify further the nature of these solar events, and elucidate possible connections between them.

Community detection aims to partition a set of nodes with more similarities in the set than out of it based on different criteria like neighborhood similarity or vertex connectivity. Most present day community detection methods principally concentrate on the topological structure, largely ignoring the heterogeneous properties of the vertex. This paper proposes a new community detection model, based on the possibilistic c-means model, by using structural as well as attribute similarities in a large scale in social networks. In the majority of real social networks, different clusters share nodes, resulting in the formation of overlapping communities. The proposed model, on the basis of structural and attribute similarity (PCMSA), serves as a fuzzy community detection model addressing the overlapping community detection problem, and detecting communities in a way that each community has a densely connected sub-graph with homogeneous attribute values. The function of the proposed model is assessed by a trade-off between intra-cluster and inter-cluster density and homogeneity. Therefore, to validate the proposed community detection algorithm (PCMSA) and its results, an index, compatible with the proposed model, is defined; and to assess the efficiency of the proposed fuzzy community detection, several experimental results in variety sizes from very small to very large sizes of real social networks are given, and the results are contrasted with other community detection models like FCAN, CODICIL, SA-cluster, K-SNAP and PCM. The experimental findings reveal the superiority of this novel model and its promising scalability and computational complexity over others.

The top eigenpairs at the title mean the maximal, the submaximal, or a few of the subsequent eigenpairs of an Hermitizable matrix. Restricting on top ones is to handle with the matrices having large scale, for which only little is known up to now. This is different from some mature algorithms, that are clearly limited only to medium-sized matrix for calculating full spectrum. It is hoped that a combination of this paper with the earlier works, to be seen soon, may provide some effective algorithms for computing the spectrum in practice, especially for matrix mechanics.

Complex networks in the real world are often with heterogeneous degree distributions. The structure and function of nodes can vary significantly, with vital nodes playing a crucial role in information spread and other spreading phenomena. Identifying and taking action on vital nodes enables change to the network’s structure and function more efficiently. Previous work either redefines metrics used to measure the nodes’ importance or focuses on developing algorithms to efficiently find vital nodes. These approaches typically rely on global knowledge of the network and assume that the structure of the network does not change over time, both of which are difficult to achieve in the real world. In this paper, we propose a localized strategy that can find vital nodes without global knowledge of the network. Our joint nomination (JN) strategy selects a random set of nodes along with a set of nodes connected to those nodes, and together they nominate the vital node set. Experiments are conducted on 12 network datasets that include synthetic and real-world networks, and undirected and directed networks. Results show that average degree of the identified node set is about 3–8 times higher than that of the full node set, and higher-degree nodes take larger proportions in the degree distribution of the identified vital node set. Removal of vital nodes increases the average shortest path length by 20–70% over the original network, or about 8–15% longer than the other decentralized strategies. Immunization based on JN is more efficient than other strategies, consuming around 12–40% less immunization resources to raise the epidemic threshold to $$\tau \sim 0.1$$ τ ∼ 0.1 . Susceptible-infected-recovered simulations on networks with 30% vital nodes removed using JN delays the arrival time of infection peak significantly and reduce the total infection scale to 15%. The proposed strategy can effectively identify vital nodes using only local information and is feasible to implement in the real world to cope with time-critical scenarios such as the sudden outbreak of COVID-19.

Numerous Bayesian Network (BN) structure learning algorithms have been proposed in the literature over the past few decades. Each publication makes an empirical or theoretical case for the algorithm proposed in that publication and results across studies are often inconsistent in their claims about which algorithm is ‘best’. This is partly because there is no agreed evaluation approach to determine their effectiveness. Moreover, each algorithm is based on a set of assumptions, such as complete data and causal sufficiency, and tend to be evaluated with data that conforms to these assumptions, however unrealistic these assumptions may be in the real world. As a result, it is widely accepted that synthetic performance overestimates real performance, although to what degree this may happen remains unknown. This paper investigates the performance of 15 state-of-the-art, well-established, or recent promising structure learning algorithms. We propose a methodology that applies the algorithms to data that incorporates synthetic noise, in an effort to better understand the performance of structure learning algorithms when applied to real data. Each algorithm is tested over multiple case studies, sample sizes, types of noise, and assessed with multiple evaluation criteria. This work involved learning approximately 10,000 graphs with a total structure learning runtime of seven months. In investigating the impact of data noise, we provide the first large scale empirical comparison of BN structure learning algorithms under different assumptions of data noise. The results suggest that traditional synthetic performance may overestimate real-world performance by anywhere between 10% and more than 50%. They also show that while score-based learning is generally superior to constraint-based learning, a higher fitting score does not necessarily imply a more accurate causal graph. The comparisons extend to other outcomes of interest, such as runtime, reliability, and resilience to noise, assessed over both small and large networks, and with both limited and big data. To facilitate comparisons with future studies, we have made all data, raw results, graphs and BN models freely available online.

The aim of this study is to investigate the applicability of reliability theory on surface square/rectangular footing against bearing capacity failure using fuzzy set theory in conjunction with the finite element method. Soil is modeled as a three-dimensional spatially varying medium, where its parameters (cohesion, friction angle, unit weight, etc.) are considered as fuzzy variables that maintain some membership functions. Soil is idealized as an elastic-perfectly plastic material obeying the Mohr-Coulomb failure criterion, where both associated and non-associated flow rules are considered in estimating the ultimate bearing capacity of the footing. The spatial variability of the soil is incorporated for both isotropic and anisotropic fields, which are determined by the values of scales of fluctuation in both the horizontal and vertical directions. A new parameter namely, limiting applied pressure at zero failure probability is proposed, and it indirectly predicts the failure probability of the footing. The effect of the coefficient of variation of the friction angle of the soil on the probability of failure is analyzed, and it is observed that the effect is significant. Furthermore, the effect of the scale of fluctuation on the probability of failure is investigated, and the necessity for considering spatial variability in the reliability analysis is well proven.

As the size of available networks is continuously increasing (even with millions of nodes), large-scale complex networks are receiving significant attention. While existing overlapping-community detection algorithms are quite effective in analyzing complex networks, most of these algorithms suffer from scalability issues when applied to large-scale complex networks, which can have more than 1,000,000 nodes. To address this problem, we propose an efficient local-expansion-based overlapping-community detection algorithm using local-neighborhood information (OCLN). During the iterative expansion process, only neighbors of nodes added in the last iteration (rather than all neighbors) are considered to determine whether they can join the community. This significantly reduces the computational cost and enhances the scalability for community detection in large-scale networks. A belonging coefficient is also proposed in OCLN to filter out incorrectly identified nodes. Theoretical analysis demonstrates that the computational complexity of the proposed OCLN is linear with respect to the size of the network to be detected. Experiments on large-scale LFR benchmark and real-world networks indicate the effectiveness of OCLN for overlapping-community detection in large-scale networks, in terms of both computational efficiency and detected-community quality.

Motivation Molecular property prediction is a hot topic in recent years. Existing graph-based models ignore the hierarchical structures of molecules. According to the knowledge of chemistry and pharmacy, the functional groups of molecules are closely related to its physio-chemical properties and binding affinities. So, it should be helpful to represent molecular graphs by fragments that contain functional groups for molecular property prediction. Results In this article, to boost the performance of molecule property prediction, we first propose a definition of molecule graph fragments that may be or contain functional groups, which are relevant to molecular properties, then develop a fragment-oriented multi-scale graph attention network for molecular property prediction, which is called FraGAT. Experiments on several widely used benchmarks are conducted to evaluate FraGAT. Experimental results show that FraGAT achieves state-of-the-art predictive performance in most cases. Furthermore, our case studies show that when the fragments used to represent the molecule graphs contain functional groups, the model can make better predictions. This conforms to our expectation and demonstrates the interpretability of the proposed model. Availability and implementation The code and data underlying this work are available in GitHub, at https://github.com/ZiqiaoZhang/FraGAT. Supplementary information Supplementary data are available at Bioinformatics online.

The machine learning of small sample size is one of the most challenging problems in face recognition. Multiscale patch cooperative representation for face recognition provides multiple patch scales to a sample set. In each patch scale, all samples in the training set and test set are segmented into patches of the same size, followed by a collaborative representation classification. A sample belongs to the category in which the number of patch blocks is the highest. Under a single scale, however, the patch blocks of a sample often belong to different categories. If the decision value of the sample is defined as the category with the largest number of patches, the possibility of the sample belonging to other categories is often disregarded; this will then significantly reduce the recognition accuracy. Therefore, a multiscale patched collaborative representation based on fuzzy decision is proposed herein. In a single scale, among all the patch blocks in a sample, the proportion of patch blocks belonging to a category to the total patch blocks is used to represent the degree of a sample belonging to the category. Hence, a fuzzy decision matrix can be obtained for a sample set in each scale. The elements of the decision matrix represent the possibility of samples belonging to categories, thereby solving the absolute problem of classification. Corresponding weights are applied to the patch scales, and multiscale outputs can be integrated by regularizing boundary distribution optimization. It is experimentally demonstrated that the proposed method exhibits high recognition accuracy and is superior to some existing algorithms.

In this paper, we extend the Zipf distribution by means of the Randomly Stopped Extreme mechanism; we establish the conditions under which the maximum and minimum families of distributions intersect in the original family; and we demonstrate how to generate data from the extended family using any Zipf random number generator. We study in detail the particular cases of geometric and positive Poisson stopping distributions, showing that, in log-log scale, the extended models allow for top-concavity (top-convexity) while maintaining linearity in the tail. We prove the suitability of the models presented, by fitting the degree sequences in a collaboration and a protein-protein interaction networks. The proposed models not only give a good fit, but they also allow for extracting interesting insights related to the data generation mechanism.

The connection between pharmacokinetic models and system theory has been established for a long time. In this approach, the drug concentration is seen as the output of a system whose input is the drug administered at different times. In this article we further explore this connection. We show that system theory can be used to easily accommodate any therapeutic regime, no matter its complexity, allowing the identification of the pharmacokinetic parameters by means of a non-linear regression analysis. We illustrate how to exploit the properties of linear systems to identify non-linearities in the pharmacokinetic data. We also explore the use of bootstrapping as a way to compare populations of pharmacokinetic parameters and how to handle the common situation of using multiple hypothesis tests as a way to distinguish two different populations. Finally, we demonstrate how the bootstrap values can be used to estimate the distribution of derived parameters, as can be the allometric scale factors.

In recent years, more and more scholars devoted themselves to the research of the target detection algorithm due to the continuous development of deep learning. Among them, the detection and recognition of small and complex targets are still a problem to be solved. The authors of this article have understood the shortcomings of the deep learning detection algorithm in detecting small and complex defect targets and would like to share a new improved target detection algorithm in steel surface defect detection. The steel surface defects will affect the quality of steel seriously. We find that most of the current detection algorithms for NEU-DET dataset detection accuracy are low, so we choose to verify a steel surface defect detection algorithm based on machine vision on this dataset for the problem of defect detection in steel production. A series of improvement measures are carried out in the traditional Faster R-CNN algorithm, such as reconstructing the network structure of Faster R-CNN. Based on the small features of the target, we train the network with multiscale fusion. For the complex features of the target, we replace part of the conventional convolution network with a deformable convolution network. The experimental results show that the deep learning network model trained by the proposed method has good detection performance, and the mean average precision is 0.752, which is 0.128 higher than the original algorithm. Among them, the average precision of crazing, inclusion, patches, pitted surface, rolled in scale and scratches is 0.501, 0.791, 0.792, 0.874, 0.649, and 0.905, respectively. The detection method is able to identify small target defects on the steel surface effectively, which can provide a reference for the automatic detection of steel defects.

SOS is a global optimization algorithm, based on nature, and is utilized to execute the various complex hard optimization problems. Be that as it may, some basic highlights of SOS, for example, pitfall among neighborhood optima and weaker convergence zone should be upgraded to discover better answers for progressively intricate, nonlinear, many optimum solution type problems. To diminish these deficiencies, as of late, numerous analysts increase the exhibition of the SOS by designing up a few changed form of the SOS. This paper suggests an improved form of the SOS to build up an increasingly steady balance between discovery and activity cores. This technique uses three unique procedures called adjusted benefit factor, altered parasitism stage, and random weighted number-based search. The technique is referred to as mISOS and tested in a popular series of twenty classic benchmarks. The dimension of these problems is considered to be hundred to monitor the impact of the suggested technique on the versatility of the test problems. Also, some real-life optimization problems are solved with the help of the proposed mISOS. The results investigated based on three different way and theses are statistical measures, convergence, and statistical analyses. The comparison of results of the mISOS with the standard SOS, SOS variants, and certain other cutting-edge algorithms shows its improved search performance.

Due to the importance of correlation coefficient in data analysis, researchers have shown interest in the concept of correlation coefficient for the extensions of fuzzy sets, in particular, for a recent established extension known as hesitant fuzzy set (HFS). Most of the existing correlation coefficients for fuzzy sets return the value one to reflect a perfect linear relationship between objects, but they do not give us sufficient information about the scale with that the objects are correlated. This can be regarded as a disadvantage from the decision-making viewpoint. Moreover, some of the existing correlation coefficients were undefined in the case that the two objects are the same. To overcome such drawbacks, we propose an approach for deriving the modified correlation coefficients of HFSs based on hesitancy degree and then extend the approach to that of interval-valued hesitant fuzzy sets. Furthermore, we put forward a number of new correlation coefficients for HFSs based on Jaccard’s and Dice’s similarity criteria. Then, we give a practical example to illustrate the application of the improved correlation coefficients for HFSs in medical diagnosis. However, comparing the results of the proposed correlation coefficients with those of the other existing definitions shows that the diagnosis results may be quite different, and this is due to their orientation.

A detailed optical and infrared photometric analysis of the open star cluster IC 361 is given in the present paper. On studying the radial density profile, radial extent of the cluster is found to be $$8.0\pm 0.5$$ arcmin. The basic physical parameters of the cluster such as $$E(B-V) = 0.56\pm 0.10$$ mag, $$E(V-K)=1.72\pm0.12$$ mag, $$\hbox {log(Age)}=9.10\pm 0.05$$ , and $$(m-M)_{0} = 12.54\pm 0.05$$ mag are obtained using the colour–colour and colour–magnitude diagrams. IC 361 is found to be located at a distance of $$3.22\pm 0.07$$ kpc. Using the archival proper motion catalogues, we estimate the mean proper motions of IC 361 as $$4.97\pm 0.17$$ mas yr $$^{-1}$$ and $${-}\,5.80\pm 0.18$$ mas yr $$^{-1}$$ in the direction of RA and Dec, respectively. A uniform distribution of mass between small scale and large scale is confirmed by MF slope of IC 361. However, the mass function slope of cluster region is found to be $$-1.06\pm 0.09$$ , which is too low compared to the Salpeter value. Our study is further showing a dynamical relaxation behaviour of the cluster IC 361.

Macroseismic intensity provides a qualitative description of seismic damage. It can be associated with Ground Motion Parameters (GMPs), which are extracted in near real-time from instrumental recordings during an earthquake. Several formulations of this empirical association exist in literature for Italy, mainly focusing on the relationship between intensity expressed on the Mercalli-Cancani-Sieberg (MCS) scale and peak ground acceleration or velocity. They are usually in the form of Ground Motion to Intensity Conversion Equations (GMICEs), which treat intensity as a continuous quantity. We propose an alternative approach, the Gaussian Naïve Bayes (GNB) classifiers, which allows to correctly treat intensity according to its ordinal definition. As a comparison, we also implement a modified version of the standard GMICE approach. We expand the existing database of GMP/MCS-intensity points with new, high-quality accelerometric data recorded in Italy in the period from 2002 to 2016 and resample the database by treating the intermediate intensities with half integer values (which are not meaningful in the MCS description) as both belong to the above and below full integer classes with an assigned weight. As a result, we estimate a new set of regression relations and GNB probability distributions between integer MCS intensity classes and eight GMPs (peak acceleration, velocity, displacement, Arias and Housner intensities, spectral acceleration at 0.3, 1.0 and 3.0 s). Results based on PGA and PGV are the most stable on the whole intensity scale. GNB models score better than GMICEs in terms of performance on unseen data and classification scores.

In this study, we present a metric of consensus for Likert-type scales. The statistic provides the level of agreement for any given number of response options as the percentage of consensus among respondents. With this aim, we use a geometric framework that allows us to analytically derive a positional indicator. The statistic is obtained as the relative weight of the distance from the point containing the proportions of observations that fall in each category to the centre of a regular polygon with as many vertices as categories, which corresponds to the point of maximum dissent. The polygon can be regarded as the area that encompasses all possible answering combinations. In order to assess the performance of the proposed metric of consensus, we conduct an iterated forecasting experiment to test whether the inclusion of the degree of agreement in households' expectations improves out-of-sample forecast accuracy of the unemployment rate in seven European countries and the Euro Area. We find evidence that the level of consensus among households contains useful information to predict unemployment rates in all cases. This result shows the potential of agreement metrics to track the evolution of economic variables. Finally, we design a simulation experiment in which we compare the sampling distribution of the proposed metric for three- and five-response alternatives, finding that the distribution of the former shows a higher level of granularity and dispersion.

We consider the optimal prediction problem of stopping a spectrally negative Lévy process as close as possible to a given distance $b \geq 0$ from its ultimate supremum, under a squared-error penalty function. Under some mild conditions, the solution is fully and explicitly characterised in terms of scale functions. We find that the solution has an interesting non-trivial structure: if b is larger than a certain threshold then it is optimal to stop as soon as the difference between the running supremum and the position of the process exceeds a certain level (less than b), while if b is smaller than this threshold then it is optimal to stop immediately (independent of the running supremum and position of the process). We also present some examples.

In statistical parameter estimation problems, how well the parameters are estimated largely depends on the sampling design used. In the current paper, a modification of ranked set sampling (RSS) called moving extremes RSS (MERSS) is considered for the estimation of the scale and shape parameters for the log-logistic distribution. Several traditional estimators and ad hoc estimators will be studied under MERSS. The estimators under MERSS are compared to the corresponding ones under SRS. The simulation results show that the estimators under MERSS are significantly more efficient than the ones under SRS.

On the basis of interpolation theory, several new inequalities are established for both general orthonormal systems and various specific classes of orthonormal systems including the Haar and Franklin systems and wavelets. The solution of the problem of characterizing the Fourier coefficients of continuous functions for general orthonormal systems is completed. For every complete orthonormal system, a continuous function is constructed that generates a universal singularity similar to the one appearing in Carleman’s theorem. This result significantly strengthens Olevskii’s theorem and turns into Orlicz’s theorem at the other end of the power scale. It is proved that the results obtained are, in a sense, final.

The new platforms for Earth observation from space are characterized by measurements made at great spatial and temporal resolutions. While this abundance of information makes it possible to detect and study localized phenomena, it may be difficult to manage this large amount of data for the study of global and large-scale phenomena. A particularly significant example is the use by assimilation systems of Level 2 products that represent gas profiles in the atmosphere. The models on which assimilation systems are based are discretized on spatial grids with horizontal dimensions of the order of tens of kilometres in which tens or hundreds of measurements may fall in the future. A simple procedure to overcome this problem is to extract a subset of the original measurements, but this involves a loss of information. Another option is the use of simple averages of the profiles, but this approach also has some limitations that we will discuss in the paper. A more advanced solution is to resort to the so-called fusion algorithms, capable of compressing the size of the dataset while limiting the information loss. A novel data fusion method, the Complete Data Fusion algorithm, was recently developed to merge a set of retrieved products in a single product a posteriori. In the present paper, we apply the Complete Data Fusion method to ozone profile measurements simulated in the thermal infrared and ultraviolet bands in a realistic scenario. Following this, the fused products are compared with the input profiles; comparisons show that the output products of data fusion have smaller total errors and higher information contents in general. The comparisons of the fused products with the fusing products are presented both at single fusion grid box scale and with a statistical analysis of the results obtained on large sets of fusion grid boxes of the same size. We also evaluate the grid box size impact, showing that the Complete Data Fusion method can be used with different grid box sizes even if this possibility is connected to the natural variability of the considered atmospheric molecule.

In this work, we continue our study of string theory in the background that interpolates between AdS3 in the IR to flat spacetime with a linear dilaton in the UV. The boundary dual theory interpolates between a CFT2 in the IR to a certain two-dimensional Little String Theory (LST) in the UV. In particular, we study computational complexity of such a theory through the lens of holography and investigate the signature of non-locality in the short distance behavior of complexity. When the cutoff UV scale is much smaller than the non-locality (Hagedorn) scale, we find exotic quadratic and logarithmic divergences (for both volume and action complexity) which are not expected in a local quantum field theory. We also generalize our computation to include the effects of finite temperature. Up to second order in finite temperature correction, we do not any find newer exotic UV-divergences compared to the zero temperature case.

Intersections appear where one street crosses another, acting as fundamental nodes embedded in the network of urban public space. Though discussed in configurational and perceptive studies, limited attention has been paid to the morphological aspects of intersections. This study proposes and tested a new approach to modeling the urban intersection form. First, effective cylinder was introduced to define the related space for each intersection based on its scale. Second, a set of morphological indicators were presented to measure the physical properties of the intersection form. Third, the affinity propagation algorithm was used to examine the patterns of intersection form. Using the Old City of Nanjing, China as the study area, 844 intersections were analyzed to test this new method. As a result, we were able to classify the intersections into eight types. This study shows that the intersection can be modelled as a volumetric and integral spatial unit of urban form, which may demand more attention from urban designers and architects in the future for shaping the built environment. The quantitative nature of our method could also open more possibilities for intersection-based studies.

Recently popularized Graph Neural Network (GNN) has been attaching great attention along with its successful industry applications. This paper focuses on two challenges traditional GNN frameworks face: (i) most of them are transductive and mainly concentrate on homogeneous networks considering single typed nodes and edges; (ii) they are difficult to handle the real-time changing network structures as well as scale to big graph data. To address these issues, a novel attention-based Heterogeneous Multi-view Graph Neural Network (aHMGNN) solution is introduced. aHMGNN models a more intricate heterogeneous multi-view network, where various node and edge types co-exist and each of these objects also contain specific attributes. It is end-to-end, and two stages are designed for node embeddings learning and multi-typed node and edge representations fusion, respectively. Experimental studies on large-scale spam detection and link prediction tasks clearly verify the efficiency and effectiveness of our proposed aHMGNN. Furthermore, we have implemented our approach in one of the largest e-commerce platforms which further verifies that aHMGNN is arguably promising and scalable in real-world applications.

Accurate performance modeling of utility-scale PV plants is crucial for reducing the risk of investment, as well as improving operations and maintenance. In spite of much research, the precise rate at which PV plants degrade is still not well known. This article uses a well-known methodology for calculating degradation rates to assess the impact of climate on long-term performance. The RdTools analytic approach is used to perform a comparative case study of two utility-scale PV plants with the same setup and almost identical technology yet located in distinct climates. A significant difference is observed between the degradation rates of each park. The median of the nominal system degradation rate measured at the dc-side of the inverters in the hotter and more humid climate is $ - 1.4{{\%}}/{\rm{yr}}$ , with 95% of the inverters between $ - 1.5{{\%}}/{\rm{yr}}$ and $ - 1.3{{\%}}/{\rm{yr}}$ , whereas in the temperate climate the median rate is $ - 0.94{{\%}}/{\rm{yr}}$ , with 95% of the inverters between $ - 1.0{{\%}}/{\rm{yr}}$ and $ - 0.9{{\%}}/{\rm{yr}}$ . These results suggest that climate zones must be included when evaluating the long-term performance of PV plants.

TOD is designed for public transport passengers and pedestrians to provide comfort and safety in creating a healthy lifestyle. Green Building Council Indonesia (GBCI) is needed as a reference and assessment to optimize building design. The purpose of this paper is to determine the extent of the green building concept applied, calculate the met and unmet Greenship criteria, also measure the GBCI Greenship predicate obtained. The data collection method used is project data and data analysis with the index scale method on Greenship New Building version 1.2. The results showed that in Appropriate Land Use categories credit are 11 points of 17 maximum points with percentage 64,71%, the Efficiency and Energy Conservation categories credit is 1 point of 26 maximum points with percentage 3,85%, the Water Conservation categories credit are 9 points of 21 maximum points with percentage 42,86%, the Source and Material Cycle categories credit are 2 points of 2 maximum points with percentage 100%, the Health and Safety Space categories credit are 4 points of 5 maximum points with percentage 80%, and the Building Environmental Management categories credit are 4 points of 6 maximum points with percentage 66,67%. The final results of TOD Pondok Cina Project are 31 points of 77 maximum points with percentage 40,26% which categorized as Building with Bronze predicate.

Quantum embedding theories are powerful tools for approximately solving large-scale strongly correlated quantum many-body problems. The main idea of quantum embedding is to glue together a highly accurate quantum theory at the local scale and a less accurate quantum theory at the global scale. We introduce the first quantum embedding theory that is also variational, in that it is guaranteed to provide a one-sided bound for the exact ground-state energy. Our method, which we call the variational embedding method, provides a lower bound for this quantity. The method relaxes the representability conditions for quantum marginals to a set of linear and semidefinite constraints that operate at both local and global scales, resulting in a semidefinite program (SDP) to be solved numerically. The accuracy of the method can be systematically improved. The method is versatile and can be applied, in particular, to quantum many-body problems for both quantum spin systems and fermionic systems, such as those arising from electronic structure calculations. We describe how the proper notion of quantum marginal, sufficiently general to accommodate both of these settings, should be phrased in terms of certain algebras of operators. We also investigate the duality theory for our SDPs, which offers valuable perspective on our method as an embedding theory. As a byproduct of this investigation, we describe a formulation for efficiently implementing the variational embedding method via a partial dualization procedure and the solution of quantum analogs of the Kantorovich problem from optimal transport theory.

Found in varied contexts from neurons to ants to fish, binary decision-making is one of the simplest forms of collective computation. In this process, information collected by individuals about an uncertain environment is accumulated to guide behavior at the aggregate scale. We study binary decision-making dynamics in networks responding to inputs with small signal-to-noise ratios, looking for quantitative measures of collectivity that control decision-making performance. We find that decision accuracy is controlled largely by three factors: the leading eigenvalue of the network adjacency matrix, the corresponding eigenvector's participation ratio, and distance from the corresponding symmetry-breaking bifurcation. This allows us to predict how decision-making performance scales in large networks based on their spectral properties. Specifically, we explore the effects of localization caused by the hierarchical assortative structure of a "rich club" topology. This gives insight into the tradeoffs involved in the higher-order structure found in living networks performing collective computations., Comment: 16 pages, 4 figures

This work presentsObject Landmarks, a new type of visual feature designed for visual localization over major changes in distance and scale. AnObject Landmarkconsists of a bounding box$${\mathbf {b}}$$bdefining an object, a descriptor$${\mathbf {q}}$$qof that object produced by a Convolutional Neural Network, and a set of classical point features within$${\mathbf {b}}$$b. We evaluateObject Landmarkson visual odometry and place-recognition tasks, and compare them against several modern approaches. We find thatObject Landmarksenable superior localization over major scale changes, reducing error by as much as 18% and increasing robustness to failure by as much as 80% versus the state-of-the-art. They allow localization under scale change factors up to 6, where state-of-the-art approaches break down at factors of 3 or more.

In Release 001 and 002 of the ICESat-2 sea ice products, candidate height segments used to estimate the reference sea surface height for freeboard calculations included two surface types: specular and smooth dark leads. We found that the uncorrected photon rates, used as proxies of surface reflectance, are attenuated due to clouds resulting in the potential misclassification of sea ice as dark leads, biasing the reference sea surface height relative to those derived from the more reliable specular returns. This results in higher reference sea surface heights and lower estimated ice freeboards. The resolution of available cloud flags from the ICESat-2 atmosphere data product is too coarse to provide useful filtering at the lead segment scale. In Release 003, we have modified the surface-reference-finding algorithm so that only specular leads are used. The consequence of this change can be seen in the composites of mean freeboard of the Arctic and Southern oceans. Broadly, coverages have decreased by ∼10–20 % because there are fewer leads (by excluding the dark leads), and the composite means have increased by 0–4 cm because of the use of more consistent specular leads.

Influence maximization aims at identifying a limited set of key individuals in a (social) network which spreads information based on some propagation model and maximizes the number of individuals reached. We show that influence maximization based on the probabilistic independent cascade model can be modeled as a stochastic maximal covering location problem. A reformulation based on Benders decomposition is proposed and a relation between obtained Benders optimality cuts and submodular cuts for correspondingly defined subsets is established. We introduce preprocessing tests, which allow us to remove variables from the model and develop efficient algorithms for the separation of Benders cuts. Both aspects are shown to be crucial ingredients of the developed branch-and-cut algorithm since real-life social network instances may be very large. In a computational study, the considered variants of this branch-and-cut algorithm outperform the state-of-the-art approach for influence maximization by orders of magnitude.

We introduce a method for generating a continuous, mass-conserving and high-order differentiable density field from a discrete point distribution such as particles or halos from an N-body simulation or galaxies from a spectroscopic survey. The method consists on generating an ensemble of point realizations by perturbing the original point set following the geometric constraints imposed by the Delaunay tessellation in the vicinity of each point in the set. By computing the mean field of the ensemble we are able to significantly reduce artifacts arising from the Delaunay tessellation in poorly sampled regions while conserving the features in the point distribution. Our implementation is based on the Delaunay Tessellation Field Estimation (DTFE) method, however other tessellation techniques are possible. The method presented here shares the same advantages of the DTFE method such as self-adaptive scale, mass conservation and continuity, while being able to reconstruct even the faintest structures of the point distribution usually dominated by artifacts in Delaunay-based methods. Additionally, we also present preliminary results of an application of this method to image denoising and artifact removal, highlighting the broad applicability of the technique introduced here., Submitted to MNRAS

Pedestrian detection in traffic environment requires high accuracy and speed of the algorithm. Traditional methods can meet the speed requirement, but there is a long gap in accuracy. Traditional methods based on convolution neural network have higher accuracy, but the amount of calculation is enormous. Aiming at the problems of background confusion, there are pedestrian ambiguity and pedestrian multi-scale in pedestrian detection, this paper constructs a pedestrian detection method with higher accuracy and faster speed based on Faster Regions Convolutional Neural Network (Faster RCNN) method, which is the leading method in the field of target detection. We have mainly improved three aspects: (1) The design criteria based on the summary and the scale characteristics of pedestrians; (2) The setting of anchor windows and the way of creating regional networks have been adjusted, and the pooling layer of environmental regions has been added; (3) Fusion of different levels of features acquires more comprehensive features. Then, based on the open-source deep learning framework, the network is implemented on the California Institute of Technology pedestrian data set. The experimental results show that our method improves the detection accuracy by 2.9% and the detection speed by 10.1 frames/second compared with the original Faster RCNN on the same data set.

We characterize the Schauder and the unconditional basis properties for the Haar system in the Triebel–Lizorkin spaces $$F^s_{p,q}({{\mathbb {R}}}^d)$$ , at the endpoint cases $$s=1$$ , $$s=d/p-d$$ , and $$p=\infty $$ . Together with the earlier results in Garrigos et al. (J Fourier Anal Appl 24(5):1319–1339, 2018) and Seeger and Ullrich (Math Z 285:91–119, 2017), this completes the picture for such properties in the Triebel–Lizorkin scale, and complements a similar study for the Besov spaces given in Garrigos et al. (Basis properties of the Haar system in limiting Besov spaces. In: Geometric aspects of harmonic analysis: a conference in honour of Fulvio Ricci, Springer-INdAM series, arXiv.1901.09117 ).

Security X-ray baggage scanners provide images based on the different levels of radiation absorption by different materials. Images captured by such scanners are inspected by a human operator, which can slow down the verification process. To speed up inspection time, computer vision and machine learning methods are increasingly being used. While object recognition has been the subject of a huge number of articles, the problem of material recognition in X-ray images still requires some work to achieve equivalent accuracy. This paper focuses on the problem of discrimination of materials into several classes, such as organic substances or metals, in images obtained from dual-energy X-ray security scanners. We propose a new multi-scale convolutional neural network (CNN) for predicting the material class, in which five different sizes of patches are implemented parallelly to balance the trade-off between the increase in the receptive field and the loss of detail. We analyze some regularization methods and activation functions and their impact on the effectiveness of our architecture. The results were compared with other popular CNN architectures and demonstrate the superiority of our solution.

This paper uses data from the 2004 pre-election survey of the American National Election Study to test empirically different ways of incorporating a valence parameter into a Downsian utility function. We call particular attention to the problem of interpersonal incomparability of responses to the liberal-conservative scale, and use Aldrich–McKelvey’s pathbreaking method to obtain accurate distances between respondents and candidates, the key regressors. We find that the utility function the most supported by the empirical evidence, the intensity valence utility function, is the one which permits to make the better predictions for the 2004 presidential election. We also consider counterfactual analyses wherein we test if Bush, the candidate with the highest intensity valence, has dominant strategies which would have insured him to obtain a majority of the popular vote. According to the theory, it is known that the candidate with the highest intensity valence does not have such dominant strategies if the distribution of voters in the policy space is too heterogenous. Nevertheless, we show that the distribution of voters in 2004 is sufficiently homogenous for Bush to have dominant strategies.

As the scale of datasets used for big data applications expands rapidly, there have been increased efforts to develop faster algorithms. This paper addresses big data summarisation problems using the submodular maximisation approach and proposes an efficient algorithm for maximising general non-negative submodular objective functions subject tok-extendible system constraints. Leveraging a random sampling process and a decreasing threshold strategy, this work proposes an algorithm, named Sample Decreasing Threshold Greedy (SDTG). The proposed algorithm obtains an expected approximation guarantee of$$\frac{1}{1+k}-\epsilon $$11+k-ϵfor maximising monotone submodular functions and of$$\frac{k}{(1+k)^2}-\epsilon $$k(1+k)2-ϵin non-monotone cases with expected computational complexity of$$O\left(\frac{n}{(1+k)\epsilon }\ln \frac{r}{\epsilon }\right)$$On(1+k)ϵlnrϵ. Here,ris the largest size of feasible solutions, and$$\epsilon \in \left(0, \frac{1}{1+k}\right)$$ϵ∈0,11+kis an adjustable designing parameter for the trade-off between the approximation ratio and the computational complexity. The performance of the proposed algorithm is validated and compared with that of benchmark algorithms through experiments with a movie recommendation system based on a real database.

The k-medoids (k-median) problem is one of the best known unsupervised clustering problems. Due to its complexity, finding high-quality solutions for huge-scale datasets remains extremely challenging. The application of many approaches finding optimal or quality solutions is limited to only small and medium-size instances. On the other hand, many parallel, distributed algorithms that can handle huge-scale datasets usually provide very poor solutions. In this paper, we develop a first parallel, distributed primal–dual heuristic algorithm for the k-medoids problem. Its main component is a very efficient parallel subgradient column generation that solves a Lagrangian dual problem and finds a tight bound on solution quality. High-quality solutions are then produced by a parallel core selection technique. We considerably reduce computational burden and memory load by employing a nearest neighbor strategy to approximate the dissimilarity matrix. We demonstrate that our algorithm finds very close to optimal solutions, confirmed by the tightness of dual bounds, of instances that are much larger than those considered in the literature to date. Our experiments include clustering large-scale collections of face images into several thousand of clusters. We show that our approach outperforms parallel improved versions of the most popular k-medoids clustering algorithms, achieving nearly linear parallel speedup.