Main Dataset Research Articles

Multi-Scale and Single-Scale Fully Convolutional Networks for Sound Event Detection

Among various Sound Event Detection (SED) systems, Recurrent Neural Networks (RNN), such as long short-term memory unit and gated recurrent unit, is used to capture temporal dependencies, but it is confined in its length of temporal dependencies, resulting in a failure to model sound events with long duration. What’s more, RNN is incapable to process datasets in parallel, leading to low efficiency and low industrial value. Given these shortcomings, we propose to use dilated convolution (and causal dilated convolution) to capture temporal dependencies, as its great ability to ensure high time resolution and obtain longer temporal dependencies under the filter size and the network depth unchanged. In addition, dilated convolution can be parallelized, so it has higher efficiency and industrial value. Based on this, we propose Single-Scale Fully Convolutional Networks (SS-FCN) composed of convolutional neural networks and dilated convolutional networks, with the former to provide frequency invariance and the later to capture temporal dependencies. With the help of dilated convolution to control the length of temporal dependencies, we observe SS-FCN modeling a single length of temporal dependencies achieves superior detection performance for finite kinds of events. For better performance, we propose Multi-Scale Fully Convolutional Networks (MS-FCN), in which the feature fusion module is introduced to capture long short-term dependencies by fusing features with different length of temporal dependencies. The proposed methods achieve competitive performance on three main datasets with higher efficiency. The results show that SED systems based on Fully Convolutional Networks have further research value and potential.

Development of BDMS Utilization Review Technology (BURT): An Artificial Intelligence Tool Using Thai Natural Language Processing to Assess Appropriateness of Hospitalization

OBJECTIVES: To develop an effective artificial intelligence (AI) driven platform to optimize the process of assessing appropriateness of hospitalization. MATERIALS AND METHODS: Anonymized data of 22,020 insured-patient admissions in a BDMS network hospital were included to build a prediction model based on a comprehensive guideline for appropriate hospitalization. To develop Thai Natural Language Processing (NLP) model, 77,707 sentences from medical records were used and separated into two datasets, 80% for training and 20% for testing. A combined NLP and rule-based algorithms formed an AI engine and outputs were displayed using a web-based application. An expert panel of five Utilization Management (UM) physicians had several collaborative discussions to fine tune the NLP model, application of clinical criteria, and classification engine. Eventually, NLP model in the latest version (BURT1.1), had satisfactory features with overall higher than 99% accuracy, precision, recall, and F1.  RESULTS: Performance of BURT1.1 was assessed using 300 cases randomly selected from the main dataset, against other methods, including concurrent review by UM nurses at the participating hospital, and UM nurses at Bangkok Hospital Headquarters (BHQ). Agreement upon UM Physician Panel consensus was set as one of the performance indicators, and BURT1.1 showed a favorable outcome with the highest rate of agreement (86%) among all the methods. The precision rate was 99% as compared to insurance claim approval status. Additionally, dramatic time savings were achieved with 0.59 second of processing time as compared to 10-15 minutes per case by conventional manual review. CONCLUSION: BURT1.1 should be effectively implemented as an automatic daily tool to screen inappropriate hospitalization. It can immediately identify patients at high risk of inappropriate hospitalization that require further assessment by UM nurse, thus providing feedback to attending physicians on the completeness and quality of documentation, with parallel notification to UM physicians. Ultimately, BURT1.1 can contribute to increase UM efficiency, speeding up the claim process, reducing health care costs due to unnecessary hospitalization, and reduction of claim denials.

Variable Magellanic HMXB Sources versus Variable ULX Sources: Nothing to Brag about the ULX Sources

We carry out a meta-analysis of ultraluminous X-ray (ULX) sources that show large variabilities (by factors of >10) between their highest and lowest emission states in the X-ray energy range of 0.3–10 keV. We are guided by a recent stringent compilation of 25 such X-ray sources by Song et al. We examine the relation of logN versus logSmax, where N is the number of sources radiating above the maximum-flux level Smax. We find a strong deviation from all previously determined slopes in various high-mass X-ray binary (HMXB) samples. In fact, the ULX data clearly show a slope of −0.91. Thus, ULX sources do not appear to be uniform and isotropic in our Universe. We compare the ULX results against the local X-ray luminosity function of HMXBs in the Small Magellanic Cloud (SMC) constructed from our latest library that includes 41 Chandra 0.3–8 keV sources and 56 XMM-Newton 0.2–12 keV sources. The ULX data are not drawn from the same continuous distribution as the SMC data (the ULX data peak at the low tails of the SMC distributions), and none of our data sets is drawn from a normal distribution or from a log-normal distribution (they all show marked excesses at both tails). At a significance level of α=0.05 (2σ), the two-sample p-value of the Kolmogorov–Smirnov (KS) test gives p=4.7×10−3<α for the ULX versus the small Chandra sample and p=1.1×10−5<<α for the ULX versus the larger XMM-Newton sample, respectively. This adds to the evidence that ULX sources are not simply the higher end of the known local Be/X-ray pulsar distribution, but they represent a class of X-ray sources different from the young sources found in the SMC and in individual starburst galaxies. On the other hand, our two main SMC data sets are found to be statistically consistent, as they are drawn from the same continuous parent distribution (null hypothesis H0): at the α=0.05 significance level, the two-sample KS test shows an asymptotic p-value of 0.308>α, which tells us to accept H0.

Warwick electron microscopy datasets

Large, carefully partitioned datasets are essential to train neural networks and standardize performance benchmarks. As a result, we have set up new repositories to make our electron microscopy datasets available to the wider community. There are three main datasets containing 19769 scanning transmission electron micrographs, 17266 transmission electron micrographs, and 98340 simulated exit wavefunctions, and multiple variants of each dataset for different applications. To visualize image datasets, we trained variational autoencoders to encode data as 64-dimensional multivariate normal distributions, which we cluster in two dimensions by t-distributed stochastic neighbor embedding. In addition, we have improved dataset visualization with variational autoencoders by introducing encoding normalization and regularization, adding an image gradient loss, and extending t-distributed stochastic neighbor embedding to account for encoded standard deviations. Our datasets, source code, pretrained models, and interactive visualizations are openly available at https://github.com/Jeffrey-Ede/datasets.

Effects of Union Certification on Workplace-Safety Enforcement: Regression-Discontinuity Evidence

The authors study how union certification affects the enforcement of workplace-safety laws. To generate credible causal estimates, a regression discontinuity design compares outcomes in establishments in which unions barely won representation elections to outcomes in establishments in which unions barely lost such elections. The study combines two main data sets: the census of National Labor Relations Board (NLRB) representation elections and the Occupational Safety and Health Administration’s (OSHA) enforcement database since 1985. Evidence shows positive effects of union certification on establishment’s rate of OSHA inspection, the share of inspections carried out in the presence of a union representative, violations cited, and penalties assessed.

A novel dataset of permanent plots in extremely species-rich temperate grasslands

The extreme species richness of some temperate grasslands is a globally relevant yet understudied phenomenon. Until now, few records from thoroughly sampled, though not permanently fixed, vegetation plots were available. We therefore established a network of 40 permanent 4 × 4-m2 plots in species-rich grasslands of the White Carpathians (Czechia), Ukraine, Hungary and Romania. Thirty-one of them form the main dataset comprising plots from discrete forest-steppe grasslands (the Brachypodio-Molinietum association of the Cirsio-Brachypodion pinnati alliance), while nine of them forms the local dataset from a single locality in the White Carpathians, whose vegetation lacks most forest-steppe species and is transitional between the Bromion erecti and Arrhenatherion elatioris alliances (the Anthoxantho-Agrostietum and Anthyllido vulnerariae-Trifolietum montani associations). The permanent plots contain a nested subplot of 1 m2, except for five plots from Ukraine and Romania. Most plots harboured at least 80 vascular plant species per 16 m2, with the mean richness of 16-m2 plots was 95.3 species in the main dataset and 84 species in the local dataset. Mean richness of the 1-m2 subplots, presumably reflecting local environmental conditions and competitive relationships more than that of 16 m2 plots, did not vary between the two datasets, with a general mean of 57.6 species. Species counts of 16-m2 plots and 1-m2 subplots correlated, albeit only moderately, only in the main dataset. The presented datasets may be used for comparative analyses as well as for a future revisiting of the plots and may bring novel insights into the dynamics of extreme species richness.

1670O Prospective data of first 1,797 hospitalised patients with cancer and COVID-19 derived from the COVID-19 Clinical Information Network and international Severe Acute Respiratory and emerging Infections Consortium, WHO Coronavirus Clinical Characterisation Consortium

The SARS-CoV-2 pandemic in the UK triggered a national characterisation protocol and information on co-morbidities including malignant neoplasm is recorded. A lack of prospective data regarding cancer patients with COVID-19 hampers the development of an evidence based approach in this population. The Clinical Characterisation Protocol-CANCER-UK is a UK multi-disciplinary project aimed at characterising the presentation and course of COVID-19 in cancer patients with the aim of informing practice. The international Severe Acute Respiratory and emerging Infections Consortium (ISARIC)-4C COVID-19 Clinical Information Network (CO-CIN) collects data on hospital inpatients with proven/high likelihood of COVID-19. Data was collected in 166 UK sites using a questionnaire adopted by the WHO. Data on patients with malignant neoplasm was extracted from the main dataset. We chose a priori to restrict any analysis of outcome to patients who were admitted more than 14 days before data extraction (13th May 2020). As of 13th May 2020 1797 of 16160 participants had malignant neoplasm (8.6% of all cases). Age<50 62 (3.5%), 50-60 378 (21%), 70-79 558 (31%), 80+ 1002 (42%). Male 1147 (64%); Female 645 (36%). Commonest comorbidities chromic pulmonary disease (22%), chronic kidney disease (21%), uncomplicated diabetes (19%) and dementia (14%). Outcomes 35% discharged alive, 30% care ongoing & 35% died. Admiited to ICU: 150 cases (25% discharged alive,31% care ongoing & 45% died). Receiving invasive ventiation: 67 cases (18% discharged alive, 25% care ongoing:25% & 57% died). HR mortality for malignancy (adjusted for age, sex, other comorbidity): 1.13 (1.02-1.24, p=0.017). Data on presentation will be presented. Europe’s largest prospective COVID-19 dataset demonstrates that cancer is independently associated with mortality in patients admitted with COVID-19. Data collection is on-going and updated data will be presented including a comparison of cancer vs. non-cancer cohort with regard to presentation, comorbidity and otucomes.

Target quality review of internal risk models and how to inspect them

The European Central Bank performs a targeted review of internal models with the objective of reducing the variability in risk-weighted assets (RWAs). This will be accomplished by harmonising practices and checking the compliance of Pillar 1 internal models of credit risk (CR), market risk (MR) and counterparty credit risk (CCR) with regular requirements. The paper will help assessment teams with guidance on which situations should trigger findings, covering a selection of mandatory key variables and allowing for additional tests to be performed. Specific attention will be paid to the different core banking systems and data sources that are used through to the (historical) risk database identified in the inspection. Although the fact that the information technology architecture and infrastructure for the credit rating systems are mode-specific, a simplified process will be outlined throughout the paper to illustrate the main process steps, systems and datasets. Targets are the retail and corporate small medium enterprises (SMEs) portfolios, including information based on personal experiences of the author. For the future, more sophisticated methods like artificial intelligence and machine learning, as described in literature, need to be found and applied.

STA-GCN: two-stream graph convolutional network with spatial–temporal attention for hand gesture recognition

Skeleton-based hand gesture recognition is an active research topic in computer graphics and computer vision and has a wide range of applications in VR/AR and robotics. Although the spatial–temporal graph convolutional network has been successfully used in skeleton-based hand gesture recognition, these works often use a fixed spatial graph according to the hand skeleton tree or use a fixed graph on the temporal dimension, which may not be optimal for hand gesture recognition. In this paper, we propose a two-stream graph attention convolutional network with spatial–temporal attention for hand gesture recognition. We adopt pose stream and motion stream as the two input streams for our network. In pose stream, we use the joint in each frame as the input; In motion stream, we use the joint offsets between neighboring frames as the input. We propose a new temporal graph attention module to model the temporal dependency and also use a spatial graph attention module to construct dynamic skeleton graph. For each stream, we adopt graph convolutional network with spatial–temporal attention to extract the features. Then, we concatenate the feature of the pose stream and motion stream for gesture recognition. We achieve the competitive performance on the main hand gesture recognition benchmark datasets, which demonstrates the effectiveness of our method.

Most Cephalaspidea have a shell, but transcriptomes can provide them with a backbone (Gastropoda: Heterobranchia)

Cephalaspidea is an order of marine gastropods found worldwide, often in sandy or muddy habitats, which has a convoluted taxonomic history based on convergent or ill-defined morphological characters. The cephalaspidean shell—which can be external and robust, internal, or altogether absent in the adult—is of particular interest in this group, and a well-resolved phylogeny can give us greater insight into the evolution of this character. Molecular data have clarified many relationships within Cephalaspidea, but studies involving few Sanger sequenced phylogenetic markers remain limited in the resolution they provide. Here we take a phylogenomic approach, the first to address internal cephalaspidean relationships, sequencing and assembling transcriptomes de novo from 22 ingroup taxa—representing the five currently accepted superfamilies, 10 of the 21 currently recognized families, and 21 genera—and analyzing these along with publicly available data. We generated two main datasets varying by a minimum taxon occupancy threshold (50% and 75%), and analyzed these using maximum likelihood, Bayesian inference and a coalescence-based method. We find a consistent, well-supported topology, with full support across most nodes including at the family and genus level, which also appears to be robust to the effect of compositional heterogeneity among amino acids in the dataset. Our analyses find Newnesioidea as the sister group to the rest of Cephalaspidea. Within the rest of the order, Philinoidea is the sister group to a clade that comprises (Bulloidea (Haminoeoidea, Cylichnoidea)). There is strong support for several previously suggested, but tenuously supported relationships such as the genus Odontoglaja nesting within the family Aglajidae, and a sister group relationship between Gastropteridae and Colpodaspididae, with Philinoglossidae as their sister group. We discuss these results and their implications in the context of current cephalaspidean taxonomy and evolution. Genomic-scale data give a backbone to this group of snails and slugs, and hold promise for a completely resolved Cephalaspidea.

Predicting Hot Spots Using a Deep Neural Network Approach.

Targeting protein-protein interactions is a challenge and crucial task of the drug discovery process. A good starting point for rational drug design is the identification of hot spots (HS) at protein-protein interfaces, typically conserved residues that contribute most significantly to the binding. In this chapter, we depict point-by-point an in-house pipeline used for HS prediction using only sequence-based features from the well-known SpotOn dataset of soluble proteins (Moreira et al., Sci Rep 7:8007, 2017), through the implementation of a deep neural network. The presented pipeline is divided into three steps: (1) feature extraction, (2) deep learning classification, and (3) model evaluation. We present all the available resources, including code snippets, the main dataset, and the free and open-source modules/packages necessary for full replication of the protocol. The users should be able to develop an HS prediction model with accuracy, precision, recall, and AUROC of 0.96, 0.93, 0.91, and 0.86, respectively.

AntiCP 2.0: an updated model for predicting anticancer peptides.

Increasing use of therapeutic peptides for treating cancer has received considerable attention of the scientific community in the recent years. The present study describes the in silico model developed for predicting and designing anticancer peptides (ACPs). ACPs residue composition analysis show the preference of A, F, K, L and W. Positional preference analysis revealed that residues A, F and K are favored at N-terminus and residues L and K are preferred at C-terminus. Motif analysis revealed the presence of motifs like LAKLA, AKLAK, FAKL and LAKL in ACPs. Machine learning models were developed using various input features and implementing different machine learning classifiers on two datasets main and alternate dataset. In the case of main dataset, dipeptide composition based ETree classifier model achieved maximum Matthews correlation coefficient (MCC) of 0.51 and 0.83 area under receiver operating characteristics (AUROC) on the training dataset. In the case of alternate dataset, amino acid composition based ETree classifier performed best and achieved the highest MCC of 0.80 and AUROC of 0.97 on the training dataset. Five-fold cross-validation technique was implemented for model training and testing, and their performance was also evaluated on the validation dataset. Best models were implemented in the webserver AntiCP 2.0, which is freely available at https://webs.iiitd.edu.in/raghava/anticp2/. The webserver is compatible with multiple screens such as iPhone, iPad, laptop and android phones. The standalone version of the software is available at GitHub; docker-based container also developed.

Evaluation of CYGNSS Observations for Flood Detection and Mapping during Sistan and Baluchestan Torrential Rain in 2020

Flood detection and produced maps play essential roles in policymaking, planning, and implementing flood management options. Remote sensing is commonly accepted as a maximum cost-effective technology to obtain detailed information over large areas of lands and oceans. We used remote sensing observations from Global Navigation Satellite System-Reflectometry (GNSS-R) to study the potential of this technique for the retrieval of flood maps over the regions affected by the recent flood in the southeastern part of Iran. The evaluation was made using spaceborne GNSS-R measurements over the Sistan and Baluchestan provinces during torrential rain in January 2020. This area has been at a high risk of flood in recent years and needs to be continuously monitored by means of timely observations. The main dataset was acquired from the level-1 data product of the Cyclone Global Navigation Satellite System (CYGNSS) spaceborne mission. The mission consisted of a constellation of eight microsatellites with GNSS-R sensors onboard to receive forward-scattered GNSS signals from the ocean and land. We first focused on data preparation and eliminating the outliers. Afterward, the reflectivity of the surface was calculated using the bistatic radar equations formula. The flooded areas were then detected based on the analysis of the derived reflectivity. Images from Moderate-Resolution Imaging Spectroradiometer (MODIS) were used for evaluation of the results. The analysis estimated the inundated area of approximately 19,644 km2 (including Jaz-Murian depression) to be affected by the flood in the south and middle parts of the Sistan and Baluchestan province. Although the main mission of CYGNSS was to measure the ocean wind speed in hurricanes and tropical cyclones, we showed the capability of detecting floods in the study area. The sensitivity of the spaceborne GNSS-R observations, together with the relatively short revisit time, highlight the potential of this technique to be used in flood detection. Future GNSS-R missions capable of collecting the reflected signals from all available multi-GNSS constellations would offer even more detailed information from the flood-affected areas.

Investigation and Prediction of Human Interactome Based on Quantitative Features.

Protein is one of the most significant components of all living creatures. All significant and essential biological structures and functions relies on proteins and their respective biological functions. However, proteins cannot perform their unique biological significance independently. They have to interact with each other to realize the complicated biological processes in all living creatures including human beings. In other words, proteins depend on interactions (protein-protein interactions) to realize their significant effects. Thus, the significance comparison and quantitative contribution of candidate PPI features must be determined urgently. According to previous studies, 258 physical and chemical characteristics of proteins have been reported and confirmed to definitively affect the interaction efficiency of the related proteins. Among such features, essential physiochemical features of proteins like stoichiometric balance, protein abundance, molecular weight and charge distribution have been validated to be quite significant and irreplaceable for protein-protein interactions (PPIs). Therefore, in this study, we, on one hand, presented a novel computational framework to identify the key factors affecting PPIs with Boruta feature selection (BFS), Monte Carlo feature selection (MCFS), incremental feature selection (IFS), and on the other hand, built a quantitative decision-rule system to evaluate the potential PPIs under real conditions with random forest (RF) and RIPPER algorithms, thereby supplying several new insights into the detailed biological mechanisms of complicated PPIs. The main datasets and codes can be downloaded at https://github.com/xypan1232/Mass-PPI.

Integrating Remote Sensing and GIS to Model Forest Fire Rik in Virunga Massif, Central - Eastern Africa

This study aimed at developing a forest fire risk model using a combination of GIS and Remote sensing techniques, which helped to identify the level of forest fire vulnerability in Virunga Massif, located at the edge of central and eastern Africa. The Analytic Hierarchical Process (AHP) approach was employed to rank and weigh the key variables and combine them into different fire risk input factors which were later integrated into the main forest fire risk model. The main input datasets, which were linked with a potential source of a forest fire, include the land cover (specifically vegetation type data generated through the Landsat 8 image classification); topographic variables such as slope, elevation and aspect retrieved from the existing Digital Elevation Model (DEM) of Rwanda; the concentration of illegal activities and proximity to beehives sites; as well as visibility from the road and human settlements. Input factor maps were generated, assigned weights and combined in a GIS environment to produce a Virunga massif fire risk model map, which was validated using the existing burnt areas map, and ground truth points recorded using GPS. The study found that the ignition factors are the most forest fire triggering factors in Virunga massif, followed by topographic factors which play a major role in the fire spreading across the ecosystem. The high forest fire risk areas were found in steep slope location around the peaks of the volcanoes, whereas areas with the lowest risk of forest fire were found inside the forest at gentle slopes. The model was validated at 75% accuracy using ground truth data. The study proposes measure to halt the ignition factors through prevention of illegal activities in the Virunga massif for the successful prevention of the forest fire risk in the ecosystem, with much effort invested during the dry season, along with the relocation of beehives to a farther distance from the ecosystem’s edge.&#x0D; Keywords: Forest Fire Risk Modelling, Biodiversity, Illegal Activities, Ignition Factors, Topographic Factors, Analytic Hierarchy Process

Microvasculature Segmentation and Intercapillary Area Quantification of the Deep Vascular Complex Using Transfer Learning.

PurposeOptical coherence tomography angiography (OCT-A) permits visualization of the changes to the retinal circulation due to diabetic retinopathy (DR), a microvascular complication of diabetes. We demonstrate accurate segmentation of the vascular morphology for the superficial capillary plexus (SCP) and deep vascular complex (DVC) using a convolutional neural network (CNN) for quantitative analysis.MethodsThe main CNN training dataset consisted of retinal OCT-A with a 6 × 6-mm field of view (FOV), acquired using a Zeiss PlexElite. Multiple-volume acquisition and averaging enhanced the vasculature contrast used for constructing the ground truth for neural network training. We used transfer learning from a CNN trained on smaller FOVs of the SCP acquired using different OCT instruments. Quantitative analysis of perfusion was performed on the resulting automated vasculature segmentations in representative patients with DR.ResultsThe automated segmentations of the OCT-A images maintained the distinct morphologies of the SCP and DVC. The network segmented the SCP with an accuracy and Dice index of 0.8599 and 0.8618, respectively, and 0.7986 and 0.8139, respectively, for the DVC. The inter-rater comparisons for the SCP had an accuracy and Dice index of 0.8300 and 0.6700, respectively, and 0.6874 and 0.7416, respectively, for the DVC.ConclusionsTransfer learning reduces the amount of manually annotated images required while producing high-quality automatic segmentations of the SCP and DVC that exceed inter-rater comparisons. The resulting intercapillary area quantification provides a tool for in-depth clinical analysis of retinal perfusion.Translational RelevanceAccurate retinal microvasculature segmentation with the CNN results in improved perfusion analysis in diabetic retinopathy.

Enabling Accurate Indoor Localization Using a Machine Learning Algorithm

In this paper, fingerprint referencing methods based on wireless fidelity Wi-Fi received signal strength (RSS) have used for indoor positioning. More precisely, Naïve Bayes, decision tree (DT), and support vector machine (SVM) one-to-one multi-classes and error-correcting-output-codes classifier are to enable accurate indoor positioning. Then, normalization is used to reduce positioning error by reducing the fluctuation and diverse distribution of the RSS values. Different devices are used in this experiment; the training dataset is not included in the main dataset. Nonetheless, the learned model by the SVM algorithm cannot be affected by the elimination of train datasets of the test device. The efficiency of DT is lower than the other machine learning algorithms, because it performs by Boolean function, and it provides the low accuracy of prediction for dataset than the algorithms. Naïve Bayes technique based on Bayes Theorem is better than DT and close to SVM for positioning approves that 1–1.5 m positioning accuracy for indoor environments can be achieved by the proposed approach which is an excellent result than traditional protocol.

Validation of Consumer Styles Inventory for consumer decision making styles

AbstractThe Consumer Styles Inventory (CSI) developed by Sproles and Kendall, Journal of Consumer Affairs, 1986, 20: 267–279, is one of the most widely used instruments to classify consumers by their decision making styles. With changes in consumption culture attributable to the coexistence of global and local consumer culture in the past few decades, it is essential to reaffirm that the CSI measurement instrument is valid for contemporary consumers and can be applied in various contexts. This study, therefore, aimed to present a set of CSI validation procedures using contemporaneous participants and targeting two groups of participants in the United States. Two sets of exploratory factor analysis using the main data set (n = 390) were performed and confirmatory factor analysis was applied to the main data set and the cross‐validation data set (n = 172). Results from both the exploratory factor and confirmatory factor analyses seem to validate the eight underlying consumer characteristics of CSI.

Job Instability and Wages for Young Adult Men

Debate continues over whether job instability is rising in the United States, with data and measurement complicating the search for definitive answers. In this paper we compare two cohorts of young white men from the National Longitudinal Surveys (NLS), construct a rigorous measure of job change, and confirm earlier findings of a significant increase in job instability in recent years. Further validation of this increase is found when we benchmark the NLS against the other main datasets in the field and conduct a thorough attrition analysis. Extending the analysis to wages, we find that the wage returns to job changing have both declined and become more unequal for young men, mirroring the trends in their long-term wage growth, and in the overall cross-sectional trends in earnings.

Overview and methods of correlation filter algorithms in object tracking

An important area of computer vision is real-time object tracking, which is now widely used in intelligent transportation and smart industry technologies. Although the correlation filter object tracking methods have a good real-time tracking effect, it still faces many challenges such as scale variation, occlusion, and boundary effects. Many scholars have continuously improved existing methods for better efficiency and tracking performance in some aspects. To provide a comprehensive understanding of the background, key technologies and algorithms of single object tracking, this article focuses on the correlation filter-based object tracking algorithms. Specifically, the background and current advancement of the object tracking methodologies, as well as the presentation of the main datasets are introduced. All kinds of methods are summarized to present tracking results in various vision problems, and a visual tracking method based on reliability is observed.