Search

In this paper, a novel hybrid Neural Network Algorithm-Differential Evolution (NNA-DE) optimizer which integrates both NNA and DE is proposed. The proposed hybrid NNA-DE optimizer demonstrates better performance compared to the standard NNA and the other state-of-the-art optimization algorithms. The proposed hybrid NNA-DE algorithm is then employed for optimizing an observer-based interval type-2 fuzzy PID (OB-IT2FPID) controller applied to the proton exchange membrane fuel cell (PEMFC) air feeding system. The whole design parameters of the OB-IT2FPID controller including the scaling factors, interval type-2 membership function parameters and the footprint-of-uncertainty (FOU) are optimized using the proposed hybrid NNA-DE algorithm. The results show that the proposed NNA-DE optimized OB-IT2FPID controller achieves better performance in terms of set-point tracking, disturbance rejection and the time-domain performance indices. The robustness of the proposed NNA-DE optimized OB-IT2FPID controller against parametric uncertainty in the PEMFC air-feeding system is tested. The proposed controller demonstrated better robustness against parameter uncertainty in the system. Processor-in-the-Loop (PIL) approach is adopted to validate the performance of the proposed controller on embedded control hardware.

Brain-derived extracellular vesicles (EVs) play an active role in Alzheimer's disease (AD), relaying important physiological information about their host tissues. Circulating EVs are protected from degradation, making them attractive AD biomarkers. However, it is unclear how circulating EVs relate to EVs isolated from disease-vulnerable brain regions. We developed a novel method for collecting EVs from the hippocampal interstitial fluid (ISF) of live mice. EVs (EVISF) were isolated via ultracentrifugation and characterized by nanoparticle tracking analysis, immunogold labeling, and flow cytometry. Mass spectrometry and proteomic analyses were performed on EVISF cargo. EVISF were 40-150 nm in size and expressed CD63, CD9, and CD81. Using a model of cerebral amyloidosis (e.g. APPswe,PSEN1dE9 mice), we found protein concentration increased but protein diversity decreased with A deposition. Genotype, age, and Aβ deposition modulated proteostasis- and immunometabolic-related pathways. Changes in the microglial EVISF proteome were sexually dimorphic and associated with a differential response of plaque associated microglia. We found that female APP/PS1 mice have more amyloid plaques, less plaque associated microglia, and a less robust- and diverse- EVISF microglial proteome. Thus, in vivo microdialysis is a novel technique for collecting EVISF and offers a unique opportunity to explore the role of EVs in AD.

Background. Lung metastasis of malignant tumor signifies worse prognosis and immensely deteriorates patients’ life quality. Spatholobi Caulis (SC) has been reported to reduce lung metastasis, but the mechanism remains elusive. Methods. The active components and corresponding targets of SC were obtained from the Traditional Chinese Medicine Database and Analysis Platform (TCMSP) database and the SwissTargetPrediction database. The disease targets were acquired from DisGeNET and GeneCards databases. Venn map was composed to figure out intersection targets by using R. The PPI network was constructed through STRING and Cytoscape, and MCODE plug-in was used to sift hub targets. Gene Ontology (GO)-Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was carried out by utilizing clusterProfiler package (R3.6.1) with adjusted P value

Recognition of obstacle type based on visual sensors is important for navigation by unmanned surface vehicles (USV), including path planning, obstacle avoidance, and reactive control. Conventional detection techniques may fail to distinguish obstacles that are similar in visual appearance in a cluttered environment. This work proposes a novel obstacle type recognition approach that combines a dilated operator with the deep-level features map of ResNet50 for autonomous navigation. First, visual images are collected and annotated from various different scenarios for USV test navigation. Second, the deep learning model, based on a dilated convolutional neural network, is set and trained. Dilated convolution allows the whole network to learn deep features with increased receptive field and further improves the performance of obstacle type recognition. Third, a series of evaluation parameters are utilised to evaluate the obtained model, such as the mean average precision (mAP), missing rate and detection speed. Finally, some experiments are designed to verify the accuracy of the proposed approach using visual images in a cluttered environment. Experimental results demonstrate that the dilated convolutional neural network obtains better recognition performance than the other methods, with an mAP of 88%.

This paper investigates the dynamic positioning (DP) control problem of surface vessels with input saturation, external disturbances and error constraints, and a finite-time disturbance observer-based DP control law with prescribed performance is proposed. First, a smooth saturation approximation model is constructed to avoid input saturation, which integrates a stable auxiliary system to transform the design problem of the DP control law into the design problem of an auxiliary control signal. Next, by introducing an auxiliary dynamic, a finite-time disturbance observer is designed to estimate the external disturbance, and it is proved that the estimation error can converge to zero in finite-time. Specially, an adaptive update law is constructed to estimate the upper bound of the derivative of external disturbances, which releases the requirement for the prior knowledge of the disturbance in the observer design. Then, combined with finite-time performance functions and error transformation functions, a prescribed performance dynamic surface control (DSC) strategy is designed to realize the surface vessel's prescribed performance control with guaranteed transient and steady-state performance in finite-time. Moreover, all signals in the closed-loop control system are proved to be ultimately uniformly bounded. Finally, the numerical simulations of a surface vessel illustrate the effectiveness of the proposed DP control law.

Because it is very expensive to collect a large number of labeled samples to train deep neural networks in certain fields, semi-supervised learning (SSL) researcher has become increasingly important in recent years. There are many consistency regularization-based methods for solving SSL tasks, such as the $$\Pi $$ model and mean teacher. In this paper, we first show through an experiment that the traditional consistency-based methods exist the following two problems: (1) as the size of unlabeled samples increases, the accuracy of these methods increases very slowly, which means they cannot make full use of unlabeled samples. (2) When the number of labeled samples is vary small, the performance of these methods will be very low. Based on these two findings, we propose two methods, metric learning clustering (MLC) and auxiliary fake samples, to alleviate these problems. The proposed methods achieve state-of-the-art results on SSL benchmarks. The error rates are 10.20%, 38.44% and 4.24% for CIFAR-10 with 4000 labels, CIFAR-100 with 10,000 labels and SVHN with 1000 labels by using MLC. For MNIST, the auxiliary fake samples method shows great results in cases with the very few labels.

With the advantages of flexible control ability, unmanned surface vehicle (USV) has been widely applied in civil and military fields. A number of researchers have been working on the development of intelligent path planning algorithms to plan a high-quality and collision-free path which is applied to guide USV through cluttered environments. The conventional algorithms may either have issues with trapping into a local optimal solution or face a slow convergence problem. This paper presents a novel multi-subpopulation bacterial foraging optimisation (MS-BFO) algorithm for USV path planning that enhances the searching performance, especially, in a complex environment. This method constructs the deletion and immigration strategies (DIS), which guarantees the elite optimised individual of each subpopulation to be inherited by others, thus to consequently lead to fast convergence speed. The experimental results show that the proposed method is able to suggest an optimised path within the shortest length of time, compared with other optimisation algorithms.

This paper proposes a double-layer model predictive control (MPC) algorithm for the integrated path planning and trajectory tracking of autonomous vehicles on roads. The upper module is responsible for generating collision-free lane trajectories, while the lower module is responsible for tracking this trajectory. A simplified vehicle model based on the friction cone is proposed to reduce the computation time for trajectory planning in the upper layer module. To achieve dynamic and accurate collision avoidance, a polygonal distance-based dynamic obstacle avoidance method is proposed. A vertical load calculation method for the tires is introduced to design the anti-rollover constraint in the lower layer module. Numerical simulations, with static and dynamic obstacle scenarios, are conducted on the MATLAB platform and compared with two state-of-the-art MPC algorithms. The results demonstrate that the proposed algorithm outperforms the other two algorithms regarding computation time and collision avoidance efficiency.

The packet loss during transmission of load control commands can lead to regulation errors in the smart grid and increase the cost of utility agencies due to the purchase of additional automatic generation control (AGC) services. In this paper, a two-layer cooperative communication network between the utility company and relays is presented. The utility company rents the relay to assist with the downlink transmission to improve the reliability of communication and reduce the data transmission cost due to packet loss. Furthermore, the uncertainty of channel gain is considered, and a two-tier game model is established. A distributed robust power control algorithm based on the continuous convex approximation method is proposed to obtain the optimal relay power allocation and price. Through the simulation analysis of the proposed scheme and the two comparison schemes, the cost of the utility company was reduced by 6% and 21%, and the standard deviation of income value between the relays was reduced by 40% and 48%, respectively.

Objective To investigate the effect of artesunate (ART) on T lymphocyte immune function in patients with lung cancer. Methods Fifteen healthy people (NC group) and twenty-one lung cancer patients (LC group) were randomly selected to collect their clinical information and isolate peripheral blood mononuclear cells (PBMCs). After 24 hour-treatment of PBMCs with ART, the median lethal concentration (LC

Recent positron emission tomography (PET) studies of kappa opioid receptors (KOR) in humans reported significant relationships between KOR availability and social status, as well as cocaine choice. In monkey models, social status influences physiology, receptor pharmacology and behavior; these variables have been associated vulnerability to cocaine abuse. The present study utilized PET imaging to examine KOR availability in socially housed, cocaine-naïve female and male monkeys, and peripheral measures of KORs with neuron-derived extracellular vesicles (NDE). KOR availability was assessed in dominant and subordinate female and male cynomolgus macaques (N = 4/rank/sex), using PET imaging with the KOR selective agonist [

Deep learning-based methods for fast target segmentation of magnetic resonance imaging (MRI) have become increasingly popular in recent years. Generally, the success of deep learning methods in medical image segmentation tasks relies on a large amount of labeled data. The time-consuming and labor-intensive problem of data annotation is a major challenge in medical image segmentation tasks. The aim of this work is to enhance the segmentation of MR images using a semi-supervised learning-based method using a small amount of labeled data and a large amount of unlabeled data.To utilize the effective information of the unlabeled data, we designed the method of guiding the Student segmentation model simultaneously by the Dual-Teacher structure of CNN and transformer forming the subject network. Both Teacher A and Student models are CNNs, and the TA-S module they form is a mean teacher structure with added data noise. In the TB-S module formed by the combination of Student and Teacher B models, their backbone networks CNN and transformer capture the local and global information of the image at the same time, respectively, to create pseudo labels for each other and perform cross-supervision. The Dual-Teacher guides the Student through synchronous training and performs knowledge rectification and communication with each other through consistent regular constraints, which better utilizes the valid information in the unlabeled data. In addition, the segmentation predictions of Teacher A and Student and Teacher A and Teacher B are screened for uncertainty assessment during the training process to enhance the prediction accuracy and generalization of the model. This method uses the mechanism of simultaneous training of the synthetic structure composed of TA-S and TB-S modules to jointly guide the optimization of the Student model to obtain better segmentation ability.We evaluated the proposed method on a publicly available MRI dataset from a cardiac segmentation competition organized by MICCAI in 2017. Compared with several existing state-of-the-art semi-supervised segmentation methods, the method achieves better segmentation results in terms of Dice coefficient and HD distance evaluation metrics of 0.878 and 4.9 mm and 0.886 and 5.0 mm, respectively, using a training set containing only 10% and 20% of labeled data.This method fuses CNN and transformer to design a new Teacher-Student semi-supervised learning optimization strategy, which greatly improves the utilization of a large number of unlabeled medical images and the effectiveness of model segmentation results.

In the practical application of large-scale photovoltaic module monitoring, adopting wireless sensor network (WSN) technology is a method worth researching. With increasing nodes in the wireless sensor network, widely existing clock skew, increased geometrically, is bringing about greater energy consumption. Due to the random distribution of nodes, in order to improve the transmission efficiency and reduce the computational load of the coordinator, the node processor needs to the use edge computing for preliminary analysis. This paper puts forward an improved energy-efficient reference broadcast synchronization algorithm (ERBS). This algorithm firstly calculates the average phase offset of nonadjacent nodes in the network after receiving a message. It then uses the least square method to solve the clock skew to achieve high-precision synchronization of the whole network. Simulation results show that compared with RBS, the time synchronization precision of ERBS is greatly improved and synchronization times are greatly reduced, decreasing energy consumption significantly.

As we acknowledge that vehicle-to-grid (V2G) network holds obvious predominance in economic benefits, it attracts wide attentions of power system and electric vehicle (EV) users. However, lots of security issues hinder its further development. For example, many existing V2G schemes use the bilinear pairing to implement authentication protocol, resulting in high computational costs. In addition, the system master key in the existing V2G scheme is generated independently by a trusted third party, so the system master key is vulnerable to leakage if the third party is compromised by the adversary. In this article, we propose a novel privacy-preserving authentication scheme for V2G networks. Specifically, this scheme employs the nonsupersingular elliptic curve to construct a lightweight authentication protocol for EV and smart grid. Meanwhile, system master key in the proposed scheme is generated by the key agreement protocol, i.e., a protocol that is constructed using a secure two-party computation between the third party authority and the dispatching center. Security analysis shows that the proposed scheme is provably secure under the random oracle model. Compared with related schemes, the proposed scheme can not only improve the authentication efficiency of V2G networks, but also prevent internal attacker. Experimental results indicate that our scheme is of practical significance for V2G networks.

The bacterial foraging optimisation (BFO) algorithm is a commonly adopted bio-inspired optimisation algorithm. However, BFO is not a proper choice in coping with continuous global path planning in the context of unmanned surface vehicles (USVs). In this paper, a grid partition-based BFO algorithm, named AS-BFO, is proposed to address this issue in which the enhancement is contributed by the involvement of the A* algorithm. The chemotaxis operation is redesigned in AS-BFO. Through repeated simulations, the relative optimal parameter combination of the proposed algorithm is obtained and the most influential parameters are identified by sensitivity analysis. The performance of AS-BFO is evaluated via five size grid maps and the results show that AS-BFO has advantages in USV global path planning.

To solve the long-tail problem and improve the testing efficiency for autonomous navigation systems of unmanned surface vehicles (USVs), a visual image-based navigation scene complexity perception method is proposed. In this paper, we intend to accurately construct a mathematical model between navigation scene complexity and visual features from the analysis and processing of image textures. First, the typical complex elements are summarized, and the navigation scenes are divided into four levels according to whether they contain these typical elements. Second, the textural features are extracted using the gray level cogeneration matrix (GLCM) and Tamura coarseness, which are applied to construct the feature vectors of the navigation scenes. Furthermore, a novel paired bare bone particle swarm clustering (PBBPSC) method is proposed to classify the levels of complexity, and the exact value of the navigation scene complexity is calculated using the clustering result and an interval mapping method. By comparing different methods on the classical and self-collected datasets, the experimental results show that our proposed complexity perception method can not only better describe the level of complexity of navigation scenes but also obtain more accurate complexity values.

Extracellular vesicles have emerged as a less-invasive nano-tool for discovering biomarkers of Alzheimer’s disease and related dementia. Here, we analysed different neuron-enriched extracellular vesicles from plasma to predict response and molecular mechanisms of ketogenic diet’s efficacy in mild cognitive impairment participants. The study was a randomized crossover design in which cognitively normal and mild cognitive impairment participants consumed a modified Mediterranean-ketogenic diet or American Heart Association diet for 6 weeks, followed by other diet after washout. L1 cell adhesion molecule, synaptophysin and neural cell adhesion molecule surface markers were used to enrich for neuron-secreted small extracellular vesicles (sEVL1CAM, sEVSYP and sEVNCAM). For the first time, we have presented multiple evidences, including immunogold labelling/transmission electron microscopy, clusters of differentiation 63-ELISA-based assay, confocal microscopy fluorescent images and flow cytometry data confirming the presence of L1 cell adhesion molecule on the surface of sEVL1CAM, validating purity and relative abundance of sEVL1CAM in the plasma. Cargo analysis of sEVL1CAM showed that modified Mediterranean-ketogenic diet intervention reduces amyloid beta 1–42 (50.3%, P = 0.011), p181-tau (34.9%, P = 0.033) and neurofilament light (54.2%, P = 0.020) in mild cognitive impairment participants. Moreover, sEVL1CAMshowed better sensitivity compared with CSF in analysing increased glutamate (6-folds, P < 0.0001) from mild cognitive impairment participants following modified Mediterranean-ketogenic diet intervention. sEVL1CAM characterization also suggested that modified Mediterranean-ketogenic diet differentially targets the expression of various glutamate receptors—glutamate receptor ionotropic NMDA1, glutamate receptor ionotropic NMDA2A, glutamate receptor ionotropic NMDA2B and glutamate receptor ionotropic AMPA type subunit 1. Importantly, these sEVL1CAM measures strongly correlated with corresponding clinical CSF biomarkers (neurogranin, amyloid beta 1–42, neurofilament light and tau). Furthermore, sEVL1CAM were loaded with less advanced glycation endproducts and exhibited anti-inflammatory activity following modified Mediterranean-ketogenic diet intervention. Most importantly, the expression of monocarboxylate transporter 2 on the surface of sEVL1CAM predicted the amyloid beta 1–42 response to modified Mediterranean-ketogenic diet intervention (area under the curve = 0.87, P = 0.0044) and offered a novel screening tool to identify participants responsive to this dietary intervention. Finally, sEVL1CAM, sEVSYP and sEVNCAM showed significantly high concordance in analysing amyloid beta 1–42 (Pearson correlation coefficient ≥ 0.63, P < 0.01) and neurofilament light (Pearson correlation coefficient ≥ 0.49, P < 0.05). Together, small extracellular vesicles in plasma offers promise in assessing the efficacy of dietary/therapeutic intervention against mild cognitive impairment/Alzheimer’s disease.

The development of novel imaging technologies allows the analysis of the expression and spatial distribution of multiple markers simultaneously, providing necessary information about a cellular identity and the surrounding microenvironment. This chapter describes the utilization of immunofluorescence to identify such biomarkers in fixed tissue from prostate cancer (PCa) xenografts. One such marker detectable by immunofluorescence is pimonidazole, which has been utilized to locate areas of low oxygen (hypoxia). Pimonidazole, in combination with other biomarkers, could be utilized to identify “niches” in the microenvironment harboring more aggressive cells both within and outside hypoxic areas. Specifically, we describe the method to use pimonidazole for the identification of hypoxic regions in PCa xenograft tumors along with CPT1A (carnitine palmitoyltransferase 1A) expression, an indicator of β-oxidation. This approach could be useful to characterize various biomarkers in the complex hypoxic tumor microenvironment.

Exosomes are an intriguing class of nanosized vesicles (~30-150 nm in diameter) released by all cell types for intercellular communication and also for cellular metabolic waste removal to maintain cellular homeostasis. Exosomes secreted by cancer cells play an important role in supporting tumor growth and metastasis by communicating with other cells in the tumor microenvironment and distant sites. Several studies have reported that the exosomes secreted by cancer cells show distinct characteristics, including size, cargo, and surface proteins from the normal cells, and can be used as important biomarkers for diagnosis and prognosis for various cancer types. Exosomes represent many distinct biochemical and morphological characteristics than other -extracellular vesicles (EVs), including their size and surface proteins. Understanding the functional role of exosomes requires specific methods for their characterization to distinguish them from other EV and non-EV structures. Transmission electron microscopy with the immunogold labeling method allows direct detection of exosomes based on their size and specific surface protein. In this chapter, we outlined the required materials and detailed method for immunogold labeling for exosomal surface proteins and size characterization.

Detecting beneficial feature interactions is essential in recommender systems, and existing approaches achieve this by examining all the possible feature interactions. However, the cost of examining all the possible higher-order feature interactions is prohibitive (exponentially growing with the order increasing). Hence existing approaches only detect limited order (e.g., combinations of up to four features) beneficial feature interactions, which may miss beneficial feature interactions with orders higher than the limitation. In this paper, we propose a hypergraph neural network based model named HIRS. HIRS is the first work that directly generates beneficial feature interactions of arbitrary orders and makes recommendation predictions accordingly. The number of generated feature interactions can be specified to be much smaller than the number of all the possible interactions and hence, our model admits a much lower running time. To achieve an effective algorithm, we exploit three properties of beneficial feature interactions, and propose deep-infomax-based methods to guide the interaction generation. Our experimental results show that HIRS outperforms state-of-the-art algorithms by up to 5% in terms of recommendation accuracy., Comment: KDD 2022, 11 pages, 12 figures, 5 tables

Short-term wind speed prediction plays a significant role in the management of large-scale wind power plants. However, wind speed prediction is extremely complex and difficult due to the volatility and non-linearity of wind. For this purpose, a broad learning system (BLS) with ensemble and classification named BLS-EC is proposed to predict multi-step-ahead wind speed. The proposed method is based on a new neural network termed the BLS, which could work out the complex non-linear relation by learning model while ensuring the computational efficiency. To overcome the randomness and instability of a single BLS, this paper proposes the BLS ensemble method to improve the generalization and stability of the network. In order to improve the accuracy of prediction, a method called classification-guided regression is proposed to distinguish different variation patterns of initial predicted wind speed. According to the classification result, different pattern sequences are re-predicted to obtain the final prediction result. Applying this thinking and method into research of three real-time wind speed datasets which were taken from Sotavento Galicia SA (SG), Alberta (ALB), and Newfoundland (NFL), the validity and practical value of this method can be demonstrated. Results obtained clearly show that BLS is better than existing methods ARIMA and RBF. Moreover, the BLS-EC method improved generalization performance and the predicting precision of a single BLS. In this study, the BLS-EC was proposed and successfully applied to wind speed prediction.

Magnetic bearing is widely used in helium-turbine circle of the high temperature gas-cooled reactor and many other highspeed rotating machinery because of its unique advantages in vibration and noise reduction. However, the power consumption of magnetic bearing increases its cost of use. Moreover, the design of magnetic bearing controller relies on accurate system modeling. All these restrict the industrial application of magnetic bearings. Based on the structure of the eight-pole magnetic bearing and its corresponding traditional decentralized differential PID control strategy, this paper proposes a magnetic bearing control framework including expected bearing force realization control strategy and centralized control strategy. Under this framework, a nonlinear low bias current control method for magnetic bearing system is given. Afterwards, an active disturbance rejection controller based on low(zero) bias current is proposed to compensate the gyroscopic disturbance and modeling uncertainty of the system. The controller can keep small loss of magnetic bearing and have good stability. It has a frame of active disturbance rejection control (ADRC) and its compensation performance is analyzed. In order to verify the effectiveness of the controller, a corresponding experimental verification is carried out on the test rig. The results show that the control strategy is effective.

User and item attributes are essential side-information; their interactions (i.e., their co-occurrence in the sample data) can significantly enhance prediction accuracy in various recommender systems. We identify two different types of attribute interactions, inner interactions and cross interactions: inner interactions are those between only user attributes or those between only item attributes; cross interactions are those between user attributes and item attributes. Existing models do not distinguish these two types of attribute interactions, which may not be the most effective way to exploit the information carried by the interactions. To address this drawback, we propose a neural Graph Matching based Collaborative Filtering model (GMCF), which effectively captures the two types of attribute interactions through modeling and aggregating attribute interactions in a graph matching structure for recommendation. In our model, the two essential recommendation procedures, characteristic learning and preference matching, are explicitly conducted through graph learning (based on inner interactions) and node matching (based on cross interactions), respectively. Experimental results show that our model outperforms state-of-the-art models. Further studies verify the effectiveness of GMCF in improving the accuracy of recommendation., Comment: 10 pages, 6 figures, 4 tables, SIGIR 2021