In recent years, deep neural nets have triumphed over many computer vision problems, including semantic segmentation, which is a critical task in emerging autonomous driving and medical image diagnostics applications. general, training requires humongous amount of labeled data, laborious costly to collect annotate. Recent advances graphics shed light on utilizing photo-realistic synthetic data with generated annotations train nets. Nevertheless, the domain mismatch between real images ones...

Abstract The flapping motion of the current sheet is a common dynamic phenomenon in planetary magnetosphere and plays an important role transportation energy disturbances. Based on measurements from Cassini spacecraft, we investigate short‐period motions characterized by periods significantly much smaller than rotation cycle Saturn's magnetosphere. Employing Minimum Variance Analysis (MVA) method, new technique developed to distinguish propagation radial azimuthal directions. A total 105...

One of the key challenges machine learning-based anomaly detection relies on difficulty obtaining data for training, which is usually rare, diversely distributed, and difficult to collect. To address this challenge, we formulate as a Positive Unlabeled (PU) learning problem where only labeled positive (normal) unlabeled (normal anomaly) are required an detector. As semi-supervised method, it does not require providing thus easily deployed various applications. can be extremely unbalanced,...

Images are transmitted or stored in their compressed form and most of the AI tasks performed from re-constructed domain. Convolutional neural network (CNN)-based image compression reconstruction is growing rapidly it achieves surpasses state-of-the-art heuristic methods, such as JPEG BPG. A major limitation application CNN-based on computation complexity during reconstruction. Therefore, learning domain desirable to avoid latency caused by In this paper, we show that can achieve comparative...

Abstract Electron vortices are usually embedded within different magnetic structures in space plasmas. The effects, including the nonideal electric field, energy dissipation and of electron on these still unclear. Utilizing unprecedented high‐resolution data from Magnetospheric Multiscale mission terrestrial magnetosheath, we statistically investigate effects structures. Both fields have no obvious correlations with scales vortices. However, compared to scales, stronger found between...

Many methods have been proposed to accelerate 2D ConvNets by removing redundant parameters. However, few efforts are devoted the problem of accelerating 3D Convolutional Networks. The ConvNets, which mainly designed for extracting spatiotemporal features, widely used in many video analytics tasks, such as action recognition and scene analysis. In this paper, we focus on two motivations: (1) Fast processing techniques dire need due explosive growth data; (2) Compared with individual images,...

There is increased pressure by governments and industry to develop national surveillance programmes evaluate antimicrobial usage (AMU) in animals.This article presents a methodological approach cost-effectiveness analysis of such programmes.Seven objectives are proposed for AMU animals: quantifying use, finding trends, detecting hotspots, identifying risk factors, encouraging research, evaluating the impact policies diseases, demonstrating compliance with regulations.Achieving these would...

Photovoltaic (PV) power shows different fluctuation characteristics under weather types as well strong randomness and uncertainty in changeable such sunny to cloudy, cloudy rain, so on, resulting low forecasting accuracy. For the type of weather, an ultra-short-term photovoltaic method is proposed based on affinity propagation (AP) clustering, complete ensemble empirical mode decomposition with adaptive noise algorithm (CEEMDAN), bi-directional long short-term memory network (BiLSTM). First,...

Anomaly detection is of great interest to big data applications, and both supervised unsupervised learning have been applied for anomaly detection. However, it still remains a challenging problem because: (1) learning, difficult acquire training samples; while (2) the performance may not be satisfactory due lack data. To address limitations, we propose hybrid solution by using normal (positive) unlabeled (could positive or negative) semi-supervised Particularly, introduce new framework based...

Most of current visual search systems focus on image-to-image (point-to-point) such as image and object retrieval. Nevertheless, fast image-to-video (point-to-set) is much less exploited. This paper tackles instance in videos, where efficient point-to-set matching essential. Through jointly optimizing vector quantization hashing, we propose compressive method to compressM proposals extracted from each video into only k binary codes, ≪ M. Then the similarity between query whole can be...

Abstract In this Paper, we carry out a new model-independent cosmological test for the cosmic distance–duality relation (CDDR) by combining latest five baryon acoustic oscillation (BAO) measurements and Pantheon type Ia supernova (SNIa) sample. Particularly, BAO measurement from extended Baryon Oscillation Spectroscopic Survey data release 16 quasar sample at effective redshift z = 1.48 is used, two methods, i.e., compressed form of artificial neural network combined with binning SNIa...

Brain–computer interface (BCI) has developed rapidly over the past two decades, mainly due to advancements in machine learning. Subjects must learn modulate their brain activities ensure a successful BCI. Feedback training is practical approach this learning process; however, commonly used classifier-dependent approaches have inherent limitations such as need for calibration and lack of continuous feedback long periods time. This paper proposes an online data visualization protocol that...

Motivated by the recent success of supervised and weakly common object discovery, in this paper, we move forward one step further to tackle discovery a fully unsupervised way. Generally, co-localization aims at simultaneously localizing objects same class across group images. Traditional localization/detection usually trains specific detectors which require bounding box annotations instances, or least image-level labels indicate presence/absence an image. Given collection images without any...

PDF HTML阅读 XML下载 导出引用 引用提醒 一种基于可变多簇结构的动态概率粒子群优化算法 DOI: 作者: 作者单位: 作者简介: 通讯作者: 中图分类号: 基金项目: Supported by the National Natural Science Foundation of China under Grant No.90412014 (国家自然科学基金) Dynamic Probabilistic Particle Swarm Optimization Based on Varying Multi-Cluster Structure Author: Affiliation: Fund Project: 摘要 | 图/表 访问统计 参考文献 相似文献 引证文献 资源附件 文章评论...

Explicit feature mapping is an appealing way to linearize additive kernels, such as χ <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> kernel for training large-scale support vector machines (SVMs). Although accurate in approximation, could pose computational challenges high-dimensional settings it expands the original features a higher dimensional space. To handle this issue context of SVMs learning, we introduce simple yet efficient...

As the software project has been more and complex than ever, test case automatic generation is of great importance for testing. In this paper, we aim to propose a novel technology promote quality The main innovation paper lies in that introduce PSO algorithm generate case. Particularly, order enhance performance PSO, an improved algorithm, which position velocity each particle represented as vectors multiple dimensional binary solution space. Afterwards, based on proposed. Experimental...

The detection of rail fastener defects is the key to ensure running safety high-speed trains.Traditional method usually be detected rely on train workers who walk along railway lines find out potential risks.The by artificial maintenance slowly, costly, and dangerous.As solve problem, an automatic detect based machine vision proposed for all kinds defects.background subtraction algorithm used achieve accurate positioning fasteners in this paper.The video sequence processing image.First all,...

Crowdfunding, as a new financing model, has drawn more and attention in the rapid development of Internet financing. This paper will select investors awardbased crowdfunding study object, apply grounded theory to explore their participation motivations crowdfunding, including 4 major categories: internal motivation, external internalized motivation behavior 17 sub-categories, further discuss relationship between build model for analysis.

In this paper, we present a concise framework to approximately construct feature maps for nonlinear additive kernels such as the Intersection, Hellinger's, and X <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> kernels. The core idea is each individual set of anchor points assign every query map its nearest neighbor or weighted combination those k-nearest neighbors in anchors. resultant can be compactly stored by group (binary) indication...

Abstract The objective of this study was to investigate the effects concomitantly increasing supplementation Ca and phytase on growth performance, balance P, bone mineralization in nursery pigs. There were eight experimental diets. positive control (PC) one two formulated contain 0.64% 0.85% total Ca, respectively, whereas dietary concentrations other nutrients identical adequate. negative (NC) deficient (0.48%) P (0.41%). Five combinations incremental levels (0.48% 1,750 units [FYT]/kg,...

Abstract The objective of this study was to investigate the effects two dietary total Ca/P ratios on available P release by phytase, measured using growth performance and bone mineralization with 528 barrows gilts according a randomized complete block design. Three were 11 diets in factorial 2 4 plus 3, including 3 reference consisting 0.25% (control), 0.70%, or 1.15% monocalcium phosphate (MCP) 8 from combining phytase doses (500, 1,000, 2,000, 3,000 FYT/kg) MCP (1.05 1.20). Each diet fed 6...

Drug-target association plays an important role in drug discovery, repositioning, synergy prediction, etc. Currently, a lot of drug-related databases, such as DrugBank and BindingDB, have emerged. However, these databases are separate, incomplete non-uniform with different criteria. Here, we integrated eight databases; collected, filtered supplemented drugs, target genes experimentally validated (highly confident) associations built highly confident drug-target (HCDT:...