Search

Learning from complementary labels (CLs) is a useful learning paradigm, where the CL specifies the classes that the instance does not belong to, instead of providing the ground truth as in the ordinary supervised learning scenario. In general, although it is less laborious and more efficient to collect CLs compared with ordinary labels, the less informative signal in the complementary supervision is less helpful to learn competent feature representation. Consequently, the final classifier's performance greatly deteriorates. In this article, we leverage generative adversarial networks (GANs) to derive an algorithm GAN-CL to effectively learn from CLs. In addition to the role in original GAN, the discriminator also serves as a normal classifier in GAN-CL, with the objective constructed partly with the complementary information. To further prove the effectiveness of our schema, we study the global optimality of both generator and discriminator for the GAN-CL under mild assumptions. We conduct extensive experiments on benchmark image datasets using deep models, to demonstrate the compelling improvements, compared with state-of-the-art CL learning approaches.

In the context of big data, organizations and individuals can often benefit from the data mining techniques, such as classification. However, decision-makers must quickly react to insights over time under dynamic environments. In this paper, we present a novel perspective, named concept-cognitive computing system (C3S), to achieve dynamic classification learning over the partially labeled data and labeled data. More specifically, to store and consolidate knowledge, a concept falling space is first employed as a basic knowledge memory mechanism in C3S. Then, we design a new concept-cognitive process by means of simulating human learning processes, which can incorporate new information into the old knowledge. Finally, a strategy of constructing two different concept spaces is considered in our system when faced with the scenario of a partially labeled dynamic data. Although there exist significant differences between C3S and the conventional incremental learning methods in the learning paradigm, our proposed C3S still performs strong performance for dynamic classification in comparison with several state-of-the-art incremental learning approaches. In addition, the experiments on various datasets have demonstrated that our system can obtain a good performance on the partially labeled data and labeled data simultaneously in dynamic environments.

In human concept learning, people can naturally combine a handful of labeled data with abundant unlabeled data when they make classification decisions, which is also known as semi-supervised learning (SSL) in machine learning. Especially, human concept learning not only is a static process in human cognition but also can vary gradually with dynamic environments. Nevertheless, the classical SSL algorithms must be redesigned to accommodate newly input data. In this sense, concept-cognitive learning may be a good choice, as it can implement dynamic processes by imitating human cognitive processes. Meanwhile, numerous SSL methods were designed based on the feature vector information of instances, while ignoring concept structural information. Based on this idea, a novel SSL method, named semi-supervised concept learning method (S2CL), is proposed for dynamic SSL by employing concept lattices. Moreover, to make full use of the global and local conceptual information, we further propose an extended version of S2CL for concept learning. More specifically, we first show some new related theories for S2CL based on a regular formal decision context; then a novel SSL framework and its corresponding algorithms are designed. Finally, we conduct some experiments on various datasets to demonstrate the effectiveness of our methods.

ObjectiveThe aim of the study was to investigate the role and regulatory mechanisms of fibroblast-like synoviocytes (FLSs) and their senescence in the progression of osteoarthritis (OA).MethodsSynovial tissues from normal patients and patients with OA were collected. Synovium FLS senescence was analysed by immunofluorescence and western blotting. The role of methyltransferase-like 3 (METTL3) in autophagy regulation was explored using N6-methyladenosine (m6A)-methylated RNA and RNA immunoprecipitation assays. Mice subjected to destabilisation of the medial meniscus (DMM) surgery were intra-articularly injected with or without pAAV9 loaded with small interfering RNA (siRNA) targeting METTL3. Histological analysis was performed to determine cartilage damage.ResultsSenescent FLSs were markedly increased with the progression of OA in patients and mouse models. We determined that impaired autophagy occurred in OA-FLS, resulting in the upregulation of senescence-associated secretory phenotype (SASP). Re-establishment of autophagy reversed the senescent phenotype by suppressing GATA4. Further, we observed for the first time that excessive m6A modification negatively regulated autophagy in OA-FLS. Mechanistically, METTL3-mediated m6A modification decreased the expression of autophagy-related 7, an E-1 enzyme crucial for the formation of autophagosomes, by attenuating its RNA stability. Silencing METTL3 enhanced autophagic flux and inhibited SASP expression in OA-FLS. Intra-articular injection of synovium-targeted METTL3 siRNA suppressed cellular senescence propagation in joints and ameliorated DMM-induced cartilage destruction.ConclusionsOur study revealed the important role of FLS senescence in OA progression. Targeted METTL3 inhibition could alleviate the senescence of FLS and limit OA development in experimental animal models, providing a potential strategy for OA therapy.

Feature processing is an important step for modeling and can improve the accuracy of machine learning models. Feature extraction methods can effectively extract features from high-dimensional data sets and enhance the accuracy of tasks. However, the performance of feature extraction methods is not stable in low-dimensional data sets. This article extends the broad learning system (BLS) to a framework for constructing robust representations in low-dimensional and small data sets. First, the BLS changed from a supervised prediction method to an ensemble feature extraction method. Second, feature extraction methods instead of random mapping are used to generate mapped features. Third, deep representations, called enhancement features, are learned from the ensemble mapped features. Fourth, data for generating mapped features and enhancement features can be randomly selected. The ensemble of mapped features and enhancement features can provide robust representations to enhance the performance of downstream tasks. A label-based autoencoder (LA) is embedded in the BLS framework as an example to show the effectiveness of the framework. A random LA (RLA) is presented to generate more different features. The experimental results show that the BLS framework can construct robust representations and significantly promote the performance of machine learning models.

The Three Gorges reservoir area in the Yangtze River economic belt has frequent geological disasters. To assist decision-makers make effective emergency decisions based on their experience, this paper proposes a dynamic geological disaster emergency plan selection scheme based on case-based reasoning and prospect theory. Geological information and data of disaster sites in the Three Gorges reservoir area are collected from the China Geological Environment Monitoring Institute and China Geological Survey of the Ministry of Natural Resources. First, historical geological disasters are analyzed to extract the gene structure characteristics and develop a gene library, from which geological and economic disaster indicators are selected. The indicator values are then divided into several intervals according to their corresponding distributions, which are then processed with a pruning method. Prospect theory is applied to compensate for the differences in decision-making, for which comprehensive foreground values and rank schemes are calculated. A comparison is made between the proposed method and other methods, besides, the Zigui County landslide in Hubei Province is taken as an example to verify the feasibility and effectiveness of the proposed method.

Cardiovascular disease (CVD) is one of the main causes of death in primary Sjogren' s syndrome (pSS) patients, while hypertension is considered an independent risk factor for CVD in pSS patients and its prevalence is increased compared with the general population. In order to identify risk factors of hypertension in patients with pSS, this study included 293 pSS patients admitted in the Third People's Hospital of Chengdu from April 2011 to August 2020 and 195 pSS patients admitted in Sichuan Provincial People's Hospital from January 2011 to June 2018. The multiple logistic regression was used to screen risk factors and nomogram was drawn based on regression coefficients. Then C-index and calibration plot were used to estimate its discrimination and calibration, respectively. Compared with pSS patients without hypertension, those with hypertension were older, higher in blood pressure, longer in duration of pSS, and had higher rates of smoking, xerophthalmia, xerostomia, previous use of botanicals, NSAIDs, and statins. Moreover, they were more likely to have pulmonary interstitial fibrosis and had higher levels of serum lipids and renal function indicators (all P

Broad Learning System (BLS) has been proven to be one of the most important techniques for classification and regression in machine learning and data mining. BLS directly collects all the features from feature and enhancement nodes as input of the output layer, which neglects vast amounts of redundant information. It usually leads to be inefficient and overfitting. To resolve this issue, we propose sparse regularization-based compact broad learning system (CBLS) framework, which can simultaneously remove redundant nodes and weights. To be more specific, we use group sparse regularization based on [Formula: see text] norm to promote the competition between different nodes and then remove redundant nodes, and a class of nonconvex sparsity regularization to promote the competition between different weights and then remove redundant weights. To optimize the resulting problem of the proposed CBLS, we exploit an efficient alternative optimization algorithm based on proximal gradient method together with computational complexity. Finally, extensive experiments on the classification task are conducted on public benchmark datasets to verify the effectiveness and superiority of the proposed CBLS.

Most existing algorithms of anomaly detection are suitable for static data where all data are available during detection but are incapable of handling dynamic data streams. In this study, we proposed an improved iLOF (incremental local outlier factor) algorithm based on the landmark window model, which provides an efficient method for anomaly detection in data streams and outperforms conventional methods. What is more, data windows as updating units are introduced to reduce the false alarm rate, and multiple tests are taken here to identify candidate anomalies and real anomalies. The improved iLOF shows its obvious advantage with its false positive rate. Furthermore, the proposed algorithm instantly deletes data points of identified real anomalies. We analyzed the performance of the improved algorithm and the sensitivity of certain parameters via empirical experiments using synthetic and real data sets. The experimental results demonstrate that the proposed improved algorithm achieved better performance on the higher detection rate and the lower false alarm rate compared with the original iLOF algorithm and its improvements.

We aimed to construct and validate a risk assessment model to identify risk factors for heart failure (HF) in patients with Takayasu’s arteritis (TAK). Three hundred sixty-five patients with TAK were recruited in the East China Takayasu Arteritis Cohort from January 2012 to December 2019. Patients were assigned into training and validation sets following a 2:1 ratio according to the date of enrollment. Clinical characteristics were compared between heart failure (HF) and non-HF subgroups in the training set, and a risk assessment model for HF and its scoring algorithm was established based on logistic regression, which was tested in the validation set. Among total of 74 (20.27%) TAK patients exhibited HF, and 55 cases (74.32%) were in the training set. The risk factors for HF of TAK patients included onset age >38 years, serum tumor necrosis factor (TNF)-α concentration >10 pg/ml, aortic valve involvement, coronary artery involvement, and pulmonary hypertension. We constructed the model without TNF-α (Model 1) and with TNF-α (Model 2). Patients in the training set with the score ≥ 3 appeared to be associated with an increased risk of HF with an area under curve (AUC) of 0.88 and 0.90 in Model 1 and Model 2 respectively. The AUC reached to 0.88 and 0.89 in the validation set that proved the accuracy of the model. We presented a risk assessment model of HF in TAK, which may help clinicians alert the complication of HF in the patients with specifically cardiac impairments.

Concept-cognitive learning (CCL) is an emerging field of concerning incremental concept learning and dynamic knowledge processing in the context of dynamic environments. Although CCL has been widely researched in theory, the existing studies of CCL have one problem: the concepts obtained by CCL systems do not have generalization ability. In the meantime, the existing incremental algorithms still face some challenges that: 1) classifiers have to adapt gradually and 2) the previously acquired knowledge should be efficiently utilized. To address these problems, based on the advantage that CCL can naturally integrate new data into itself for enhancing flexibility of concept learning, we first propose a new CCL model (CCLM) to extend the classical methods of CCL, which is not only a new classifier but also good at incremental learning. Unlike the existing CCL systems, the theory of CCLM is mainly based on a formal decision context rather than a formal context. In learning concepts from dynamic environments, we show that CCLM can naturally incorporate new data into itself with a sufficient theoretical guarantee for incremental learning. For classification task and knowledge storage, our results on various data sets demonstrate that CCLM can simultaneously: 1) achieve the state-of-the-art static and dynamic classification task and 2) directly accomplish preservation of previously acquired knowledge (or concepts) under dynamic environments.

Automatic detecting anomalous regions in images of objects or textures without priors of the anomalies is challenging, especially when the anomalies appear in very small areas of the images, making difficult-to-detect visual variations, such as defects on manufacturing products. This paper proposes an effective unsupervised anomaly segmentation approach that can detect and segment out the anomalies in small and confined regions of images. Concretely, we develop a multi-scale regional feature generator which can generate multiple spatial context-aware representations from pre-trained deep convolutional networks for every subregion of an image. The regional representations not only describe the local characteristics of corresponding regions but also encode their multiple spatial context information, making them discriminative and very beneficial for anomaly detection. Leveraging these descriptive regional features, we then design a deep yet efficient convolutional autoencoder and detect anomalous regions within images via fast feature reconstruction. Our method is simple yet effective and efficient. It advances the state-of-the-art performances on several benchmark datasets and shows great potential for real applications.

This paper presents a novel design of micro solar sails for emerging lightweight chip scale spacecraft based on flexible electronic circuits. To acquire large deformation, the micro solar sails were designed to be bilayer beams that were able to be electro-thermally actuated by Joule heating. The concept design of the solar sail with high area-to-mass ratios allowed the solar sailing system, named as ChipSail, for efficient orbital transfer and attitude adjustment. The principle of solar sailing with ChipSail was illustrated, and the thickness of the two metals for the bilayer sails should be no more than 1 µm, so as to achieve the efficient solar sailing. Then, the fabrication and characterization of such bilayer microstructures for solar sails were introduced briefly. After that, the electro-thermal analysis of such solar sails deployed on the low earth orbit was carried out, and it was found that the balanced temperature of the sails under the effect of solar radiation and thermal reemission of the sails was 315.31 K, followed by electro-thermal modelling of the sails under the Joule heating. A nonlinear second order differential equation was derived, which allowed rapid prediction of the thermal distribution across the sail. Equivalence of the bilayer solar sail to a width-changing 980 µm long bilayer beam was proposed and validated by finite element analysis. Finally, the thermo-mechanical model on the bilayer sail was then established and solved numerically. Results showed that the maximum bending angle could reach to 94.05o by applying a voltage of 0.05 V across the sail. The electro-thermo-mechanical model laid a solid foundation for dynamic control of the configuration of the ChipSail for efficient orbital transfer and attitude adjustment in space.

Single Image Super-Resolution (SISR) is a fundamental and important low-level computer vision (CV) task, yet its performance on real-world applications is not always satisfactory. Different from the previous SISR research, we focus on a specific but realistic SR issue: How can we obtain satisfied SR results from compressed JPG (C-JPG) images, which is a ubiquitous image format to greatly release storage space while missing fine details. the JPG SR task is deeply analyzed to discover the connotation. Then, we propose an effective two-step model structure named RGSR, involving two specifically designed components, i.e., JPG recovering and SR generation, instead of the perspective of noise elimination in traditional SR approaches. Besides, we further integrate the cycle loss to build a hybrid objective across scales for better SR generation. Experimental results on both of the standard test data sets and real images show that our approach achieves outstanding results and succeed in applying to practical C-JPG SR tasks.

High-performance DC interfaces play a pivotal role in microgrids with distributed energy resources (DERs), which would significantly improve the utilization of DERs and the flexibility of the prosumers. For a candidate DC interface converter, how to maintain the performance over a wide operation range becomes the biggest challenge. This paper proposes a flexible structure soft-switching three-level (TL) dc-dc converter, which adds four high-speed MOSFETs and one low-speed electronic relay to achieve overall optimum performance over a wide range. Under high output or low input voltage conditions, the relay and the extra MOSFETs on the primary side are ON permanently. The proposed converter is a zero-voltage switching (ZVS) converter, which guarantees ZVS for all primary switches even under 0 load currents. Under low output or high input voltage conditions, the relay and the extra MOSFETs in the secondary side are OFF permanently. The proposed converter is a zero-voltage and zero-current switching (ZVZCS) converter, and two added MOSFETs provide ZVZCS operation not only for the main primary switches but also themselves over a wide range. Other advantages of the presented converter include reduced filter size, no primary side circulating currents, and reduced current stress of the clamping capacitor. Besides, the switching loss caused by the added power devices is low because of the full range of soft-switching operation and low on-state resistance of these devices. This paper discusses the circuit configuration, operation principle, soft switching characteristics, technical comparison, and experimental results from a 1-kW prototype prove the rightness of the presented converter.

BNT162b2 Vaccine and Emerging SARS-CoV-2 Variants A total of 20 serum samples from 15 persons who received the BNT162b2 vaccine showed strong neutralization activity against recombinant viruses eng...

A novel zero-voltage and zero-current switching (ZVZCS) full-bridge (FB) three-level (TL) dc–dc converter is proposed, which can satisfy most requirements of wide operation range high-input dc–dc industrial applications. OFF-voltage on the primary switches can be maintained to $V_{\mathrm {in}}$ /2 by the input capacitors, and no extra primary clamping device is needed. The primary switches can obtain zero-voltage switching (ZVS) or zero-current switching (ZCS) easily. Furthermore, the operation mode can be altered under different output and input conditions, which ensures high efficiency over a wide operation range. The extra MOSFETs have low ON-state resistance and cause less switching loss; thus, the power loss caused by these MOSFETs can be omitted. Furthermore, these components are not inserted into the circuit loop of the input capacitors and primary switches, which may reduce the voltage spike on the primary switches at turn-off instants. The VA rating of the primary switches and input capacitors is identical to that of conventional FB TL dc–dc converters, and the utilization of the dc-link voltage is close to 100%. To evaluate the proposed converter, an industrial application example and some experimental results are also provided.

The past few decades have seen many meaningful trials on development of solar sail spacecraft with high area-to-mass ratios and thus allowing efficient, propellant-free photon propulsion. However, the restriction to bending stiffness of thin solar films has proven to be an obstacle for building solar sails with high area-to-mass ratios. This paper presents a novel design of micro solar sails for emerging lightweight chip scale spacecraft based on flexible electronics. A representative concept design of the chip scale solar sail spacecraft with area-to-mass ratios over 100 m2/kg is proposed, which enables efficient orbital transfer and attitude adjustment. To acquire large deformation for adjustment of forces, the micro solar sails consisting of bilayer beams are actuated by Joule heating. Electro-thermal analysis on the sails after deployed into space should be the prioritized and fundamental problem for following evaluation of deflection and thus change of solar radiation pressure on the reflective sails. Solar radiation, thermal reemission, and Joule heating are incorporated into the electro-thermal model for such solar sails deployed in geospace. A nonlinear second order differential equation is derived, which is solved numerically for temperature distribution across the sail. The proposed electro-thermal model is validated by finite element analysis and lays a solid foundation for the thermo-mechanical modelling, as well as flight modelling, of the chip scale spacecraft with such micro solar sails.

An effective vaccine is needed to halt the spread of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) pandemic. Recently, we reported safety, tolerability and antibody response data from an ongoing placebo-controlled, observer-blinded phase I/II coronavirus disease 2019 (COVID-19) vaccine trial with BNT162b1, a lipid nanoparticle-formulated nucleoside-modified mRNA that encodes the receptor binding domain (RBD) of the SARS-CoV-2 spike protein1. Here we present antibody and T cell responses after vaccination with BNT162b1 from a second, non-randomized open-label phase I/II trial in healthy adults, 18–55 years of age. Two doses of 1–50 μg of BNT162b1 elicited robust CD4+ and CD8+ T cell responses and strong antibody responses, with RBD-binding IgG concentrations clearly above those seen in serum from a cohort of individuals who had recovered from COVID-19. Geometric mean titres of SARS-CoV-2 serum-neutralizing antibodies on day 43 were 0.7-fold (1-μg dose) to 3.5-fold (50-μg dose) those of the recovered individuals. Immune sera broadly neutralized pseudoviruses with diverse SARS-CoV-2 spike variants. Most participants had T helper type 1 (TH1)-skewed T cell immune responses with RBD-specific CD8+ and CD4+ T cell expansion. Interferon-γ was produced by a large fraction of RBD-specific CD8+ and CD4+ T cells. The robust RBD-specific antibody, T cell and favourable cytokine responses induced by the BNT162b1 mRNA vaccine suggest that it has the potential to protect against COVID-19 through multiple beneficial mechanisms. In a phase I/II dose-escalation clinical trial, the mRNA COVID-19 vaccine BNT162b1 elicits specific T cell and antibody responses that suggest it has protective potential.

In March 2020, the World Health Organization (WHO) declared coronavirus disease 2019 (COVID-19), which is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)1, a pandemic. With rapidly accumulating numbers of cases and deaths reported globally2, a vaccine is urgently needed. Here we report the available safety, tolerability and immunogenicity data from an ongoing placebo-controlled, observer-blinded dose-escalation study (ClinicalTrials.gov identifier NCT04368728) among 45 healthy adults (18–55 years of age), who were randomized to receive 2 doses—separated by 21 days—of 10 μg, 30 μg or 100 μg of BNT162b1. BNT162b1 is a lipid-nanoparticle-formulated, nucleoside-modified mRNA vaccine that encodes the trimerized receptor-binding domain (RBD) of the spike glycoprotein of SARS-CoV-2. Local reactions and systemic events were dose-dependent, generally mild to moderate, and transient. A second vaccination with 100 μg was not administered because of the increased reactogenicity and a lack of meaningfully increased immunogenicity after a single dose compared with the 30-μg dose. RBD-binding IgG concentrations and SARS-CoV-2 neutralizing titres in sera increased with dose level and after a second dose. Geometric mean neutralizing titres reached 1.9–4.6-fold that of a panel of COVID-19 convalescent human sera, which were obtained at least 14 days after a positive SARS-CoV-2 PCR. These results support further evaluation of this mRNA vaccine candidate. In a dose-escalation study of the COVID-19 RNA vaccine BNT162b1 in 45 healthy adults, RBD-binding IgG concentrations and SARS-CoV-2 neutralizing titres in sera increased with dose level and after a second vaccine dose.

Renewable energy resources and technologies, especially solar thermal and photovoltaic systems, can provide advantageous and clean options for meeting electrical demands. The location and sizing of photovoltaic systems affect the system losses significantly. Here, a new hybrid framework for the optimal location and size for an off-grid solar/hydrogen energy system is proposed using an improved heuristic approach: improved harmony search and geographic information system (GIS). To determine the appropriate capacity and location for a grid-independent photovoltaic scheme in rural areas, a GIS module is used subject to several criteria: social, economic, technical, and environmental. The life cycle cost, the main objective function, is minimized and employed to assess the viability of the stand-alone photovoltaic scheme with hydrogen storage accounting for the reliability of the scheme. The system reliability is measured with the parameter loss of power supply probability. The proposed framework is useful, and is applied here to an actual case study in Iran for illustrative purposes. To validate the effectiveness of the proposed algorithm, the results obtained by proposed algorithm are compared with those of harmony search algorithm-based geographic information systems. The results of the simulation show the advantages of utilizing the proposed framework, based on improved harmony search and geographic information systems, for a grid-independent photovoltaic scheme considering both power quality and cost in remote and island areas.

The inquiry letters of annual reports play an increasingly important role in performing the first-line supervision duties of Stock Exchange. Based on 2015–2017 annual report inquiry letters, this p...

In many disciplines, the evaluation of algorithms for processing massive data is a challenging research issue. However, different algorithms can produce different or even conflicting evaluation performance, and this phenomenon has not been fully investigated. The motivation of this paper aims to propose a solution scheme for the evaluation of clustering algorithms to reconcile different or even conflicting evaluation performance. The goal of this research is to propose and develop a model, called decision-making support for evaluation of clustering algorithms (DMSECA), to evaluate clustering algorithms by merging expert wisdom in order to reconcile differences in their evaluation performance for information fusion during a complex decision-making process. The proposed model is tested and verified by an experimental study using six clustering algorithms, nine external measures, and four MCDM methods on 20 UCI data sets, including a total of 18,310 instances and 313 attributes. The proposed model can generate a list of algorithm priorities to produce an optimal ranking scheme, which can satisfy the decision preferences of all the participants. The results indicate our developed model is an effective tool for selecting the most appropriate clustering algorithms for given data sets. Furthermore, our proposed model can reconcile different or even conflicting evaluation performance to reach a group agreement in a complex decision-making environment.

Nowadays, datasets are always dynamic and patterns in them are changing. Instances with different labels are intertwined and often linearly inseparable, which bring new challenges to traditional learning algorithms. This paper proposes adaptive hyper-sphere (AdaHS), an adaptive incremental classifier, and its kernelized version: Nys-AdaHS. The classifier incorporates competitive training with a border zone. With adaptive hidden layer and tunable radii of hyper-spheres, AdaHS has strong capability of local learning like instance-based algorithms, but free from slow searching speed and excessive memory consumption. The experiments showed that AdaHS is robust, adaptive, and highly accurate. It is especially suitable for dynamic data in which patterns are changing, decision borders are complicated, and instances with the same label can be spherically clustered.

To make better use of the cloud computing technology, and to overcome the computing and storage requirements which increase rapidly with the number of training samples, in this paper, a new parallel algorithm is proposed—Parallel Regularized Multiple-Criteria Linear Programming (PRMCLP) algorithm—The RMCLP model is converted into a unconstrained optimization problem, and then, in the parallel version, it is split into several tasks, where each part is mapped and computed on a separate processor. This approach enables us to obtain efficiently the final optimization solution of the whole classification problem. At last, we apply this algorithm to Medical data classification. All experiments show that our method and approach greatly increases the training speed of RMCLP in the parallel case.

In this paper, a novel symmetrical topology for the small power high frequency bidirectional AC-DC converter has been proposed, which is applied to the bi-directional wireless power transmission system based on the magnetic coupling resonance. It has dual functions of the inverter and rectifier. The research is mainly reflected in the following aspects: First, based on analyzing the principle of the wireless power bidirectional transmission technology, the operation mechanism of the topology circuit is analyzed emphatically. Second, the feasibility of the designed converter is verified by the simulation of the PSPICE software. Third, the hardware system has been operated at 200 kHz in the modes of the constant current, the constant voltage and the constant resistance, respectively. The results show the maximal output power of the designed circuit is 7.8 W, meeting the design requirement. This circuit has the advantages of the simple circuit structure and simple control. It's suitable for the bidirectional wireless power charging of the small power portable devices.

Development of optimal SARS-CoV-2 vaccines to induce potent, long-lasting immunity and provide cross-reactive protection against emerging variants remains a high priority. Here, we report that a modified porous silicon microparticle (mPSM)-adjuvanted SARS-CoV-2 receptor-binding domain (RBD) vaccine activated dendritic cells and generated more potent and durable SARS-CoV-2-specific systemic humoral and type 1 helper T (Th) cell-mediated immune responses than alum-formulated RBD following parenteral vaccination, and protected mice from SARS-CoV-2 and Beta variant infection. mPSM facilitated the uptake of SARS-CoV-2 RBD antigens by nasal and airway epithelial cells. Parenteral and intranasal prime and boost vaccinations with mPSM-RBD elicited potent systemic and lung resident memory T and B cells and SARS-CoV-2 specific IgA responses, and markedly diminished viral loads and inflammation in the lung following SARS-CoV-2 Delta variant infection. Our results suggest that mPSM can serve as potent adjuvant for SARS-CoV-2 subunit vaccine which is effective for systemic and mucosal vaccination.

In this study, We design a trajectory recording and analysis system to record and analysis the changes in the movement behavior of the cockroach robot after stimulation. The external hardware of this system is an infrared touchpad as the experimental platform for the cockroach robot to crawl freely, and the infrared matrices densely distributed in the X and Y directions of the infrared touchpad are used to detect and locate the position of the cockroach robot. The cockroach robot’s movement trajectory is displayed visually through the projector’s interface projection on the infrared touchpad. The system software consists of three main parts: the electrical signal parameter setting module, the movement trajectory recording module, and the data analysis module. The electrical signal parameter setting module sets the stimulation parameters and configures the corresponding serial port to independently stimulate the left and right antenna and cercus of the cockroach; the trajectory recording module is used to record the trajectory of the cockroach robot through the coordinate positioning method. The data analysis module explores the change of motion behavior of the cockroach robot with time after receiving the stimulus by using the stage analysis method, and explores the change of motion of the cockroach robot with different voltage stimulus by using the module analysis method. The system is tested in experiments and the results demonstrated its applicability to the recording and analysis of the cockroach robot’s trajectories.

Hybrid islanded microgrids with diesel generators (DGEs) and renewable energy systems have wide applications in remote areas and islands where no electricity can be accessed. Compared with grid-tied microgrids, a more complex impedance interaction feature caused by the randomness and fluctuation of renewable energy sources (RES) and loads brings about a greater challenge on the analysis of small-signal stability for HIMG. In this paper, the impedance aggregation model of the HIMG is derived to analyze the parameter stability region of the HIMG, and the equivalent damping ratio is extracted to investigate the main factors that affect HIMG’s stability. Based on the impedance aggregation model and the equivalent damping ratio, the parameters stability region for the typical operation scenarios is obtained. The HIMG simulation model with multiple DGEs, multiple photovoltaic inverters, and constant power loads is modeled in the PSCAD to validate the parameters stability region by the proposed theoretical analysis method, which can be used as the guidance for the selection of the equipment parameters and the formulation for the operation strategy of HIMG.

Objective: Accidental pharyngeal fishbone ingestion is a common complaint in ear, nose, and throat clinics. Approximately two-thirds of the accidentally ingested fishbones can be removed using tongue depressors and indirect laryngoscopy. However, the remaining third is challenging to identify and remove using these methods. These difficult fishbones require identification and removal via more advanced approaches. Video-guided laryngoscope is used to deal with difficult fishbones in our center. This study aimed to explore the risk factors for difficult fishbones. Methods: A prospective study was performed at a teaching hospital on 2080 patients. Univariate and multivariate analyses were performed to identify the risk factors. Results: The common fishbone locations were the tonsils (39.8%; defined as STEP-I), tongue base (37.1%), vallecula (13.3%; STEP-II), and hypopharynx (9.8%; STEP-III). With increasing STEP level, the ratio of difficult fishbones correspondingly increased (Z = 13.919, P < .001), and the proportions were 21.1%, 41.9%, and 70% in STEP-I, II, and III, respectively. In particular, fishbones in STEP-III (vs STEP-I) had a higher risk of difficult fishbones (odds ratio [OR]: 11.573, 95% CI: 7.987-16.769). Complaints of neck pain (yes vs no), foreign body sensation (yes vs no), and shorter length of fishbones always had a lower risk of difficult fishbones (OR: 0.455, 95% CI: 0.367-0.564; OR: 0.284, 95% CI: 0.191-0.422; OR: 0.727, 95% CI: 0.622-0.85). Missing teeth (yes vs no), swallowing behavior after fishbone ingestion (yes vs no), and male patients (vs female) had a higher risk of difficult fishbones (OR: 1.9, 95% CI: 1.47-2.456; OR: 1.631, 95% CI: 1.293-2.059; OR: 1.278, 95% CI: 1.047-1.56). Conclusions: Neck pain, foreign body sensation, fishbone length, patient age and sex, tooth status, and swallowing behavior after fishbone ingestion are independent risk factors for difficult fishbones.

Background Although uncommon during hepatobiliary-pancreatic (HBP) surgery, iatrogenic intraoperative injury to a major artery requires prompt and appropriate repair. Here, we outline our surgical experience with the repair of this injury, and compare our experience to findings garnered from a selective review of the literature. Methods We retrospectively analyzed the clinical diagnoses, surgical methods, sites of arterial injury, operative repair, intra-operative blood loss, blood transfusion requirements, post-operative management, and outcome of 17 consecutive patients who sustained iatrogenic intra-operative injuries to major arteries during HBP surgery between January 2008 and December 2013. Results Depending on the location and extent of injury, suture repair, primary end-to-end anastomosis, artery transposition, interposition grafting, or arterio-portal shunting were used. Postoperative morbidity occurred in three cases and there was only one case of in-hospital mortality (5.9%). No arterial thrombosis or other repairrelated complications were found after the operation with a follow-up duration of 6 months. Conclusions The use of an optimal repair method for injured arteries based on their location and extent resulted in a satisfactory outcome.

Dynamic electrocardiogram and multi-spiral computed tomography (MSCT) myocardial perfusion imaging are commonly used in clinical examination methods, which can provide important reference for the diagnosis of myocardial ischemia in coronary heart disease. To analyze and compare the application value of dynamic electrocardiogram and MSCT myocardial perfusion imaging in the diagnosis of myocardial ischemia in coronary heart disease.The clinical data of 79 patients diagnosed with coronary heart disease admitted in our hospital from October 2017 to October 2019 were retrospectively collected. Exercise load/rest myocardial perfusion imaging (MPI) confirmed the presence of myocardial ischemia, and the patients were all examined by dynamic electrocardiogram and MSCT. The diagnosis of exercise load/resting MPI was used as a benchmark to compare the sensitivity, specificity, and accuracy of the different tests in the diagnosis of myocardial ischemia in coronary heart disease.As confirmed by the exercise load/rest MPI, the 79 included patients had myocardial ischemia. Fifty-five cases of myocardial ischemia were detected by dynamic electrocardiogram, of which 33 cases showed horizontal depression of the ST segment. Also, 15 cases exhibited downward oblique depression, and seven cases showed upward oblique depression. MSCT examination showed that 78 patients had myocardial ischemia, of which 74 showed abnormal myocardial perfusion when contracted. Of these patients, 47 showed reversible perfusion filling, 20 exhibited partially reversible perfusion filling, and seven showed fixed perfusion defects. The sensitivity, specificity, and accuracy of dynamic electrocardiogram in the diagnosis of myocardial ischemia in coronary heart disease were 60.76%, 72.15%, 69.62%, respectively, while those of MSCT were 91.14%, 75.94%, and 86.07%, respectively. The sensitivity and accuracy of MSCT examination were significantly higher than those of dynamic electrocardiogram examination (P0.05), and the specificity was slightly higher than that of dynamic electrocardiogram examination, but this difference was not statistically significant (P0.05).Compared to dynamic electrocardiogram examination, MSCT myocardial perfusion imaging has superior sensitivity and accuracy in the diagnosis of myocardial ischemia in coronary heart disease, and can provide a reference for clinical diagnosis and treatment.

Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) is the causative agent of the COVID-19 global pandemic. Vaccines are needed to control the disease and bring an end to the pandemic. SARS-CoV-2 is an enveloped RNA virus that relies on its trimeric surface glycoprotein, spike, for entry into host cells. Here we describe the COVID-19 vaccine candidate MV-014-212, a live attenuated, recombinant human respiratory syncytial virus (RSV) expressing a chimeric SARS-CoV-2 spike as the only viral envelope protein. MV-014-212 was attenuated and immunogenic in African green monkeys (AGMs). One mucosal administration of MV-014-212 in AGMs protected against SARS-CoV-2 challenge, reducing the peak shedding of SARS-CoV-2 in the nose by more than 200-fold. MV-014-212 elicited mucosal immunity in the nose and neutralizing antibodies in serum that exhibited cross neutralization against two virus variants of concern. Intranasally delivered, live attenuated vaccines such as MV-014-212 entail low-cost manufacturing suitable for global deployment. MV-014-212 is currently in phase I clinical trials as a single-dose intranasal COVID-19 vaccine.

A candidate multigenic SARS-CoV-2 vaccine based on an MVA vector expressing both viral N and S proteins (MVA-S + N) was immunogenic, and induced T-cell responses and binding antibodies to both antigens but in the absence of detectable neutralizing antibodies. Intranasal immunization with the vaccine diminished viral loads and lung inflammation in mice after SARS-CoV-2 challenge, which correlated with the T-cell response induced by the vaccine in the lung, indicating that T-cell immunity is also likely critical for protection against SARS-CoV-2 infection in addition to neutralizing antibodies.

The original version of this Article contained an error in the Acknowledgements, which incorrectly listed the NIH grant ‘AI142759’. This has been corrected in both the PDF and HTML versions of the Article. © The Author(s) 2021.

Network intrusion detection (NID) is an important cyber security scheme to identify attacks in network traffic. Recent years, a large amount of studies try to improve the accuracy of the NID by kinds of deep learning approaches. However, these models always require a lot of calculation and space, which constitutes a major hurdle to practical implementation of DL-based models. Thus, lightweight model is imperative, but there are very few applications of DL-based lightweight algorithms in NID models. In this paper, we propose a lightweight knowledge distillation (LKD) model for NID using the idea of knowledge distillation and separable convolution. To the best of our knowledge, it is the first system to use the knowledge distillation approach for NID. The experiment results show that the accuracy of the proposed approach reaches 91.46% and 94.30% on the KDD-CUP99 and UNSW-NB15 datasets respectively. The performance of our model is superior to some approaches based on deep neural network or some machine learning methods. Moreover, both the computational cost and model size of our model are reduced by about 99% compared to the original model.

SARS-CoV-2 mutations may diminish vaccine-induced protective immune responses, and the durability of such responses has not been previously reported. Here, we present a comprehensive assessment of the impact of variants B.1.1.7, B.1.351, P.1, B.1.429, and B.1.526 on binding, neutralizing, and ACE2-blocking antibodies elicited by the vaccine mRNA-1273 over seven months. Cross-reactive neutralizing responses were rare after a single dose of mRNA-1273. At the peak of response to the second dose, all subjects had robust responses to all variants. Binding and functional antibodies against variants persisted in most subjects, albeit at low levels, for 6 months after the primary series of mRNA-1273. Across all assays, B.1.351 had the greatest impact on antibody recognition, and B.1.1.7 the least. These data complement ongoing studies of clinical protection to inform the potential need for additional boost vaccinations., One-Sentence Summary: Most mRNA-1273 vaccinated individuals maintained binding and functional antibodies against SARS-CoV-2 variants for 6 months.

In stock trading markets, trade duration (i. e., inter-arrival times of trades) usually exhibits high uncertainty and excessive zero values. To forecast conditional distribution of trade duration, this study proposes a hybrid model called “DL-ZIACD” for short, which addresses the problem of excessive zero values by a zero-inflated distribution. Meanwhile, dynamics of the distribution time-varying parameters are captured by a specially designed deep learning (DL) architecture in which the behavioral patterns of large traders and small individual traders are represented separately by different blocks. The proposed hybrid model takes advantage of the strong fitting ability of deep learning methods while allowing for providing a probabilistic output. This paper empirically applied the established model to a large-scale dataset, containing 9,900,000 transactions of the Chinese Shenzhen Stock Exchange 100 Index (SZSE 100) constituents. To the best of our knowledge, no previous studies have applied conditional duration models to a dataset of such a large scale. For both the central location forecasting and the extreme quantile forecasting, our proposed model exhibited significant superiority over the benchmark models, which indicates that our DL-ZIACD model can provide accurate forecasts in conditional duration distribution.

SARS-CoV-2 variants that attenuate antibody neutralization could jeopardize vaccine efficacy and the end of the COVID-19 pandemic. We recently reported the protective activity of a single-dose intranasally-administered spike protein-based chimpanzee adenovirus-vectored vaccine (ChAd-SARS-CoV-2-S) in animals, which has advanced to human trials. Here, we assessed its durability, dose-response, and cross-protective activity in mice. A single intranasal dose of ChAd-SARS-CoV-2-S induced durably high neutralizing and Fc effector antibody responses in serum and S-specific IgG and IgA secreting long-lived plasma cells in the bone marrow. Protection against a historical SARS-CoV-2 strain was observed across a 100-fold vaccine dose range and over a 200-day period. At 6 weeks or 9 months after vaccination, serum antibodies neutralized SARS-CoV-2 strains with B.1.351 and B.1.1.28 spike proteins and conferred almost complete protection in the upper and lower respiratory tracts after challenge. Thus, in mice, intranasal immunization with ChAd-SARS-CoV-2-S provides durable protection against historical and emerging SARS-CoV-2 strains.

Edge computing or Internet of Things (IoT) comprises a set of devices that are interconnected ranging from our daily used objects to advanced networked equipment. It is constantly evolving as the number of devices owned by users is increasing at a rapid speed. These devices are used for various scenarios such as health care, monitoring, autonomous vehicles etc. However, as the edges perform more complex operations and IoTs carry increasing heavy workloads, they demand more energy to perform such tasks. In this paper, we perform a comprehensive study of state-of-the-art energy saving on edge platforms. Specifically, energy efficiency of the devices that run on the edges as well as corresponding solutions including hardware, software, algorithms, etc. will be thoroughly analyzed and we also presented the strengths and weakness of various researches in each area.

Containers, an emerging service to manage and deploy applications into isolated boxes, are quickly increasing in popularity in the cloud and edge computing. In order to provide connectivity among multiple hosts, cloud providers adopt overlay networks, which not only impose significant overhead in throughput and latency in containerized applications, but also consume more CPU resources of the system. Through profiling and code analysis, this paper reveals that the overwhelming interrupts, as well as its load imbalance in the kernel processing contribute to the inefficiency of the container overlay networks. Specifically, every packet in container networks might raise multiple software interrupts compared to that in VM networks. Our results indicate that the container network throughput drops 2/3 and the tail latency increases more than 37 times if the interrupt overhead is not well optimized.

At present, flexible die medium represented by fluid is mainly used in tube hydropiercing, but it is relatively rare in sheet forming field because of the high requirement of sealing. For this reason, magnetorheological fluid, one of the representatives of intelligent materials, is used as a force transfer medium to provide backpressure for the piercing process of sheet for the first time in this paper, so as to reduce the collapse and improve the quality of fracture surface. The results show that with the increase of internal pressure, the collapse of the area near the cutting edge of punch decreases, the length of burnish zone increases, the aperture is closer to the nominal diameter, and the quality of the hole is significantly improved. The theoretical mechanics model of a single straight chain is established through theoretical analysis, and the magnitude and loading mode of the force exerted on the sheet by magnetic medium are deduced. Through the balanced relationship between punching force and backpressure in the deformed area of sheet, the key parameters affecting the collapse of piercing of sheet are obtained, which has important guiding significance for improving the quality of sheet metal piercing products with magnetic medium.

Concept-cognitive learning (CCL) concerns the manner in which to incorporate new information into itself by mimicking human cognitive processes. Numerous CCL systems have been proposed to meet different requirements based on formal concept analysis. However, most of these can only operate effectively on small-scale datasets and lack classification ability. To overcome these two challenges, based on a regular formal decision context and multi-thread technique, a new concurrent CCL model is proposed in this study as an extension of classical CCL. More precisely, to enhance computational efficiency, a new concurrent learning framework is designed and its corresponding learning algorithms at the initial concept construction and CCL stages are developed. We propose a concurrent incremental learning technique by continuously accommodating newly added data to meet the requirements of classification tasks. Finally, experiments on various datasets, including real-world and synthetic data, demonstrate that the proposed concurrent method can achieve comparable classification effectiveness and significantly improve the concept learning performance simultaneously.

To address the variabilities of the number and position of extracted feature points for the traditional scale-invariant feature transform (SIFT) method, an improved SIFT algorithm is proposed for robust emotion recognition. Specifically, shape decomposition is first performed on the detected facial images by defining a weight vector. Then, a feature point constraint algorithm is developed to determine the optimum position of the feature points that can effectively represent the expression change regions. On this basis, the SIFT descriptors are applied to extract the regional gradient information as feature parameters. Finally, the support vector machine classifier combined with the principal component analysis method is used to reduce the feature dimensions and facial expression recognition. Experiments have been performed under different conditions, i.e., varied illuminations, face poses and facial moisture levels, using 15 participants. In the cases of frontal face and 5-degree face rotation views, the average recognition accuracies are 98.52% and 94.47% (no additional light sources), as well as 96.97% and 95.40% (two additional light sources), respectively. In addition, as an effective supplement to the problem of changes in illumination, the average recognition ratios are 96.23% and 96.20% under dry and wet face conditions, respectively. The experimental results reveal the robust performance of the proposed method in facial expression recognition.

Traffic prediction is a complex, nonlinear spatiotemporal relationship modeling task with the randomness of traffic demand, the spatial and temporal dependency between traffic flows, and other recurrent and nonrecurrent factors. Based on the ability to learn generic features from history information, deep learning approaches have been recently applied to traffic prediction. Convolutional neural network (CNN) methods that learn traffic as images can improve the predictive accuracy by leveraging the implicit correlations among nearby links. Traffic prediction based on CNN is still in its initial stage without making full use of spatiotemporal traffic information. In this paper, we improve the predictive accuracy by directly capturing the relationship between the input sequence and the predicted value. We propose the new local receptive fields for spatiotemporal traffic information to provide the constraints in the task domain for CNN which is different from traditionally learning traffic as images. We explore a max-pooled CNN followed by a fully connected layer with a nonlinear activation function to convolute the new local receptive fields. The higher global-level features are fed into a predictor to generate the predicted output. Based on the dataset provided by Highways England, we validate the assumption that there exists direct relationship between the input sequence and the predicted value. We train the proposed method by using the backpropagation approach, and we employ the AdaGrad method to update the parameters of the proposed method. The experimental results show that the proposed method can improve the predictive accuracy.