Journal articles on the topic 'Matrix capsule networks'

An optimized matrix capsule network is proposed for the problem of contraband in parcels with different poses, different sizes, random occlusion and sample imbalance. The network improves the recognition accuracy with the help of the matrix capsule network's recognition ability for object poses, and is mainly composed of a multi-branch feature extraction network and a side branch matrix capsule network, which is used to extract large and small targets; the side branch matrix capsule network uses a larger capsule convolution kernel, which is capable of detecting larger targets, and uses the operation of randomly discarding the capsules in the side branch, which makes the parameter amount to be reduced while enhancing the learning ability of the network. The region of interest of the network is obtained by using heat map approach with the help of weight back propagation mapping back to the original map to localize the contraband. Through a large number of experiments on the SIXray dataset, it is proved that the network in this paper improves the detection accuracy by 9.43% and the processing speed of the model by about 1/3 compared with the original capsule network.

We investigate a latent variable model for multinomial classification inspired by recent capsule architectures for visual object recognition (Sabour, Frosst, & Hinton, 2017 ). Capsule architectures use vectors of hidden unit activities to encode the pose of visual objects in an image, and they use the lengths of these vectors to encode the probabilities that objects are present. Probabilities from different capsules can also be propagated through deep multilayer networks to model the part-whole relationships of more complex objects. Notwithstanding the promise of these networks, there still remains much to understand about capsules as primitive computing elements in their own right. In this letter, we study the problem of capsule regression—a higher-dimensional analog of logistic, probit, and softmax regression in which class probabilities are derived from vectors of competing magnitude. To start, we propose a simple capsule architecture for multinomial classification: the architecture has one capsule per class, and each capsule uses a weight matrix to compute the vector of hidden unit activities for patterns it seeks to recognize. Next, we show how to model these hidden unit activities as latent variables, and we use a squashing nonlinearity to convert their magnitudes as vectors into normalized probabilities for multinomial classification. When different capsules compete to recognize the same pattern, the squashing nonlinearity induces nongaussian terms in the posterior distribution over their latent variables. Nevertheless, we show that exact inference remains tractable and use an expectation-maximization procedure to derive least-squares updates for each capsule's weight matrix. We also present experimental results to demonstrate how these ideas work in practice.

The squash function in capsule networks (CapsNets) dynamic routing is less capable of performing discrimination of non-informative capsules which leads to abnormal activation value distribution of capsules. In this paper, we propose vertical squash (VSquash) to improve the original squash by preventing the activation values of capsules in the primary capsule layer to shrink non-informative capsules, promote discriminative capsules and avoid high information sensitivity. Furthermore, a new neural network, (i) skip-connected convolutional capsule (S-CCCapsule), (ii) Integrated skip-connected convolutional capsules (ISCC) and (iii) Ensemble skip-connected convolutional capsules (ESCC) based on CapsNets are presented where the VSquash is applied in the dynamic routing. In order to achieve uniform distribution of coupling coefficient of probabilities between capsules, we use the Sigmoid function rather than Softmax function. Experiments on Guangzhou Women and Children’s Medical Center (GWCMC), Radiological Society of North America (RSNA) and Mendeley CXR Pneumonia datasets were performed to validate the effectiveness of our proposed methods. We found that our proposed methods produce better accuracy compared to other methods based on model evaluation metrics such as confusion matrix, sensitivity, specificity and Area under the curve (AUC). Our method for pneumonia detection performs better than practicing radiologists. It minimizes human error and reduces diagnosis time.

Aiming at the problem of low detection accuracy caused by the different postures, different sizes, complex backgrounds and overlapping occlusions of contraband in the security checking process, a Matrix capsule network based on attention mechanism (MCAM) is designed by introducing attention mechanism into the capsule network. Firstly, a multi-feature-extraction (MFE) module is designed to solve the difficulty of detecting contraband with different sizes and complex backgrounds; then the Conv2d with Attention for ConvCap (CACC) is constructed in the convolutional layer of the capsule, and the weight information is computed in the channel dimensions of the feature map to give the contraband regions higher coefficients to enhance the contraband detection ability when the poses are different and the occlusion is severe; finally, a new capsule detection layer is designed by utilizing the pose matrix of the capsule to give full play to the detection ability of the matrix capsule network. The mAPs of the proposed model on SIXray, SIXray10, and SIXray100 are 85.10%, 64.05%, and 53.67%, respectively, which are 42.79%, 19.73%, and 9.41% higher than those of the original network, higher than those of the current mainstream detection networks, and the number of parameters of the network and the amount of computation are also lower.

Medical image classification plays an essential role in disease diagnosis and clinical treatment. More and more research efforts have been dedicated to the design of effective methods for medical image classification. As an effective framework, the capsule network (CapsNet) can realize translation equivariance. Lots of current research applies capsule networks in medical image analysis. In this paper, we propose an attentive octave convolutional capsule network (AOC-Caps) for medical image classification. In AOC-Caps, an AOC module is used to replace the traditional convolution operation. The purpose of the AOC module is to process and fuse the high- and low-frequency information in the input image simultaneously, and weigh the important parts automatically. Following the AOC module, a matrix capsule is used and the expectation maximization (EM) algorithm is applied to update the routing weights. The proposed AOC-Caps and comparative methods are tested on seven datasets, including PathMNIST, DermaMNIST, OCTMNIST, PneumoniaMNIST, OrganMNIST_Axial, OrganMNIST_Coronal, and OrganMNIST_Sagittal, which are from MedMNIST. In the experiments, baselines include the traditional CNN models, automated machine learning (AutoML) methods, and related capsule network methods. The experimental results demonstrate that the proposed AOC-Caps achieves better performance on most of the seven medical image datasets.

According to the functional matrix hypothesis, changes in size and shape and localization of facial bones during individual ontogeny are influenced by periosteal and capsular matrices. However, the interaction of the functional matrices with the distribution of areas of bone remodeling has not been extensively studied yet. Here we evaluate the changes in the volume of the paranasal sinuses and orbital capsule with age and their association with facial growth arising from bone remodeling patterns of the upper and middle face in a sample of prehistoric human populations from South America. We found an association between capsule size and bone cell proportions, however the trajectories of variation are ambiguous across bones. The frontal and maxillary sinuses had a significant increase from 4.5 up to 14.5 years old, while the orbital capsule had an increase in volume even in adult stages. In turn, the volume of the frontal sinus increases while the bone formation remains relatively stable in subadults and decreases in adults, while the maxilla and the zygomatic bones display a lower proportion of formation when the bones are growing. Our study contributes with information concerning the covariation between bone growth remodeling and the increments of the capsular matrices.

This paper reports a study that aims to solve the problem of the weak adaptability to angle transformation of current monocular depth estimation algorithms. These algorithms are based on convolutional neural networks (CNNs) but produce results lacking in estimation accuracy and robustness. The paper proposes a lightweight network based on convolution and capsule feature fusion (CNNapsule). First, the paper introduces a fusion block module that integrates CNN features and matrix capsule features to improve the adaptability of the network to perspective transformations. The fusion and deconvolution features are fused through skip connections to generate a depth image. In addition, the corresponding loss function is designed according to the long-tail distribution, gradient similarity, and structural similarity of the datasets. Finally, the results are compared with the methods applied to the NYU Depth V2 and KITTI datasets and show that our proposed method has better accuracy on the C1 and C2 indices and a better visual effect than traditional methods and deep learning methods without transfer learning. The number of trainable parameters required by this method is 65% lower than that required by methods presented in the literature. The generalization of this method is verified via the comparative testing of the data collected from the internet and mobile phones.

With the widespread success of deep learning in the two-dimensional field, how to apply deep learning methods from two-dimensional to three-dimensional field has become a current research hotspot. Among them, the polygon mesh structure in the three-dimensional representation as a complex data structure provides an effective shape approximate representation for the three-dimensional object. Although the traditional method can extract the characteristics of the three-dimensional object through the graphical method, it cannot be applied to more complex objects. However, due to the complexity and irregularity of the mesh data, it is difficult to directly apply convolutional neural networks to 3D mesh data processing. Considering this problem, we propose a deep learning method based on a capsule network to effectively classify mesh data. We first design a polynomial convolution template. Through a sliding operation similar to a two-dimensional image convolution window, we directly sample on the grid surface, and use the window sampling surface as the minimum unit of calculation. Because a high-order polynomial can effectively represent a surface, we fit the approximate shape of the surface through the polynomial, use the polynomial parameter as the shape feature of the surface, and add the center point coordinates and normal vector of the surface as the pose feature of the surface. The feature is used as the feature vector of the surface. At the same time, to solve the problem of the introduction of a large number of pooling layers in traditional convolutional neural networks, the capsule network is introduced. For the problem of nonuniform size of the input grid model, the capsule network attitude parameter learning method is improved by sharing the weight of the attitude matrix. The amount of model parameters is reduced, and the training efficiency of the 3D mesh model is further improved. The experiment is compared with the traditional method and the latest two methods on the SHREC15 data set. Compared with the MeshNet and MeshCNN, the average recognition accuracy in the original test set is improved by 3.4% and 2.1%, and the average after fusion of features the accuracy reaches 93.8%. At the same time, under the premise of short training time, this method can also achieve considerable recognition results through experimental verification. The three-dimensional mesh classification method proposed in this paper combines the advantages of graphics and deep learning methods, and effectively improves the classification effect of 3D mesh model.

Wireless Capsule Endoscopy is a technology used to examine and view the gastro intestinal tract. Here we propose a methodology for the detection of bleeding and non-bleeding regions. The edge regions are first detected and then removed before identifying the bleeding regions. The edge and the bleeding regions have the same hue value and also the bleeding and non-bleeding regions have same luminance. The canny edge detection algorithm is used to detect edges since it have the ability to detect more edge pixels and preserves more bleeding regions. After the edge detection the regions are segmented by using super-pixel segmentation. Here Statistical features and texture features are extracted from Gray Level Co-occurrence Matrix. Finally the bleeding and non-bleeding regions are classified by using the Artificial Neural Networks

Lactic acid bacteria (LAB) constitute the microbial group most used as probiotics; however, many strains reduce their viability during their transit through the body. The objective of this study was to evaluate the effect of two microencapsulation techniques, as well as the incorporation of lactulose as a prebiotic and the use of chitosan coating on the microcapsules, on the viability of the Lactobacillus sp. strain FM4.C1.2. LAB were microencapsulated by extrusion or emulsion, using 2% sodium alginate as encapsulating matrix and lactulose (2 or 4%) as the prebiotic. The encapsulation efficiency was evaluated, and the capsules were measured for moisture and size. The encapsulation efficiency ranged between 80.64 and 99.32% for both techniques, with capsule sizes between 140.64 and 1465.65 µm and moisture contents from 88.23 to 98.04%. The microcapsules of some selected treatments (five) were later coated with chitosan and LAB survival was evaluated both in coated and uncoated microcapsules, through tolerance to pH 2.5, bile salts and storage for 15 days at 4 °C. The highest survival of the probiotic strain under the conditions of pH 2.5 (96.78–99.2%), bile salts (95.54%) and storage for 15 days (84.26%), was found in the microcapsules obtained by emulsion containing 4% lactulose and coated with chitosan. These results demonstrate the possible interaction of lactulose with alginate to form better encapsulating networks, beyond its sole probiotic effect. Additional research may shed more light on this hypothesis.

When using a fiber optic gyroscope as the core measurement element in an inertial navigation system, its work stability and reliability directly affect the accuracy of the navigation system. The modeling and fault diagnosis of the gyroscope is of great significance in ensuring the high accuracy and long endurance of the inertial system. Traditional diagnostic models often encounter challenges in terms of reliability and accuracy, for example, difficulties in feature extraction, high computational cost, and long training time. To address these challenges, this paper proposes a new fault diagnostic model that performs a fault diagnosis of gyroscopes using the enhanced capsule neural network (iCaps NN) optimized by the improved gray wolf algorithm (NIGWO). The wavelet packet transform (WPT) is used to construct a two-dimensional feature vector matrix, and the deep feature extraction module (DFE) is added to extract deep-level information to maximize the fault features. Then, an improved gray wolf algorithm combined with the adaptive algorithm (Adam) is proposed to determine the optimal values of the model parameters, which improves the optimization performance. The dynamic routing mechanism is utilized to greatly reduce the model training time. In this paper, effectiveness experiments were carried out on the simulation dataset and real dataset, respectively; the diagnostic accuracy of the fault diagnosis method in this paper reached 99.41% on the simulation dataset; the loss value in the real dataset converged to 0.005 with the increase in the number of iterations; and the average diagnostic accuracy converged to 95.42%. The results show that the diagnostic accuracy of the NIGWO-iCaps NN model proposed in this paper is improved by 13.51% compared with the traditional diagnostic methods. It effectively confirms that the method in this paper is capable of efficient and accurate fault diagnosis of FOG and has strong generalization ability.

Extracting robust features against geometric transformations and adversarial perturbations remains a critical challenge in deep learning. Although capsule networks exhibit resilience through vector-encapsulated features and dynamic routing, they suffer from computational inefficiency due to iterative routing, dense matrix operations, and extra activation scalars. To address these limitations, we propose a method that integrates (1) compact vector-grouped neurons to eliminate activation scalars, (2) a non-iterative voting algorithm that preserves spatial relationships with reduced computation, and (3) efficient weight-sharing strategies that balance computational efficiency with generalizability. Our approach outperforms existing methods in image classification on CIFAR-10 and SVHN, achieving up to a 0.31% increase in accuracy with fewer parameters and lower FLOPs. Evaluations demonstrate superior performance over competing methods, with 0.31% accuracy gains on CIFAR-10/SVHN (with reduced parameters and FLOPs) and 1.93%/1.09% improvements in novel-view recognition on smallNORB. Under FGSM and BIM attacks, our method reduces attack success rates by 47.7% on CIFAR-10 and 32.4% on SVHN, confirming its enhanced robustness and efficiency. Future work will extend vexel representations to MLPs and RNNs and explore applications in computer graphics, natural language processing, and reinforcement learning.

Transfer learning is the act of using the data or knowledge in a problem to help solve different but related problems. In a brain computer interface (BCI), it is important to deal with individual differences between topics and/or tasks. A kind of capsule decision neural network (CDNN) based on transfer learning is proposed. In order to solve the problem of feature distortion caused by EEG feature extraction algorithm, a deep capsule decision network was constructed. The architecture includes multiple primary capsules to form a hidden layer, and the connection between the advanced capsule and the primary capsule is determined by the neural decision routing algorithm. Unlike the dynamic routing algorithm that iteratively calculates the similarity between primary capsules and advanced capsules, the neural decision network computes the relationship between each capsule in the deep and shallow hidden layers in a probabilistic manner. At the same time, the distribution of the EEG covariance matrix is aligned in Riemann space, and the regional adaptive method is further introduced to improve the independent decoding ability of the capsule decision neural network for the subject’s EEG signals. Experiments on two motor imagery EEG datasets show that CDNN outperforms several of the most advanced transfer learning methods.

Abstract Axons from deep heterotopia do not extend through U-fibers, except transmantle dysplasias. Keratan sulfate (KS) in fetal spinal cord/brainstem median septum selectively repels glutamatergic axons while enabling GABAergic commissural axons. Immunocytochemical demonstration of KS in neocortical resections and forebrain at autopsy was studied in 12 fetuses and neonates 9–41 weeks gestational age (GA), 9 infants, children, and adolescents and 5 patients with focal cortical dysplasias (FCD1a). From 9 to 15 weeks GA, no KS is seen in the cortical plate; 19-week GA reactivity is detected in the molecular zone. By 28 weeks GA, patchy granulofilamentous reactivity appears in extracellular matrix and adheres to neuronal somata with increasing intensity in deep cortex and U-fibers at term. Perifascicular KS surrounds axonal bundles of both limbs of the internal capsule and within basal ganglia from 9 weeks GA. Thalamus and globus pallidus exhibit intense astrocytic reactivity from 9 weeks GA. In FCD1a, U-fiber reactivity is normal, discontinuous or radial. Ultrastructural correlates were not demonstrated; KS is not electron-dense. Proteoglycan barrier of the U-fiber layer impedes participation of deep heterotopia in cortical epileptic networks. Perifascicular KS prevents aberrant axonal exit from or entry into long and short tracts. KS adhesion to neuronal somatic membranes may explain inhibitory axosomatic synapses.

Background Joint local context that is primarily processed by pre-trained models has emerged as a prevailing technique for text classification. Nevertheless, there are relatively few classification applications on small sample of industrial text datasets. Methods In this study, an approach of employing global enhanced context representation of the pre-trained model to classify industrial domain text is proposed. To achieve the application of the proposed technique, we extract primary text representations and local context information as embeddings by leveraging the BERT pre-trained model. Moreover, we create a text information entropy matrix through statistical computation, which fuses features to construct the matrix. Subsequently, we adopt BERT embedding and hyper variational graph to guide the updating of the existing text information entropy matrix. This process is subjected to iteration three times. It produces a hypergraph primary text representation that includes global context information. Additionally, we feed the primary BERT text feature representation into capsule networks for purification and expansion as well. Finally, the above two representations are fused to obtain the final text representation and apply it to text classification through feature fusion module. Results The effectiveness of this method is validated through experiments on multiple datasets. Specifically, on the CHIP-CTC dataset, it achieves an accuracy of 86.82% and an F1 score of 82.87%. On the CLUEEmotion2020 dataset, the proposed model obtains an accuracy of 61.22% and an F1 score of 51.56%. On the N15News dataset, the accuracy and F1 score are 72.21% and 69.06% respectively. Furthermore, when applied to an industrial patent dataset, the model produced promising results with an accuracy of 91.84% and F1 score of 79.71%. All four datasets are significantly improved by using the proposed model compared to the baselines. The evaluation result of the four dataset indicates that our proposed model effectively solves the classification problem.

Decorin is an archetypal member of the small leucine-rich proteoglycan gene family and is involved in various biological functions and many signaling networks, interacting with extra-cellular matrix (ECM) components, growth factors, and receptor tyrosine kinases. Decorin also modulates the growth factors, cell proliferation, migration, and angiogenesis. It has been reported to be involved in many ischemic and fibrotic eye diseases, such as congenital stromal dystrophy of the cornea, anterior subcapsular fibrosis of the lens, proliferative vitreoretinopathy, et al. Furthermore, recent evidence supports its role in secondary posterior capsule opacification (PCO) after cataract surgery. The expression of decorin mRNA in lens epithelial cells in vitro was found to decrease upon transforming growth factor (TGF)-β-2 addition and increase upon fibroblast growth factor (FGF)-2 addition. Wound healing of the injured lens in mice transgenic for lens-specific human decorin was promoted by inhibiting myofibroblastic changes. Decorin may be associated with epithelial–mesenchymal transition and PCO development in the lens. Gene therapy and decorin administration have the potential to serve as excellent therapeutic approaches for modifying impaired wound healing, PCO, and other eye diseases related to fibrosis and angiogenesis. In this review, we present findings regarding the roles of decorin in the lens and ocular diseases.

Electroencephalogram (EEG) and functional near-infrared spectroscopy (fNIRS) can objectively reflect a person’s emotional state and have been widely studied in emotion recognition. However, the effective feature fusion and discriminative feature learning from EEG–fNIRS data is challenging. In order to improve the accuracy of emotion recognition, a graph convolution and capsule attention network model (GCN-CA-CapsNet) is proposed. Firstly, EEG–fNIRS signals are collected from 50 subjects induced by emotional video clips. And then, the features of the EEG and fNIRS are extracted; the EEG–fNIRS features are fused to generate higher-quality primary capsules by graph convolution with the Pearson correlation adjacency matrix. Finally, the capsule attention module is introduced to assign different weights to the primary capsules, and higher-quality primary capsules are selected to generate better classification capsules in the dynamic routing mechanism. We validate the efficacy of the proposed method on our emotional EEG–fNIRS dataset with an ablation study. Extensive experiments demonstrate that the proposed GCN-CA-CapsNet method achieves a more satisfactory performance against the state-of-the-art methods, and the average accuracy can increase by 3–11%.

The pericellular microenvironment around articular cartilage chondrocytes must play a key role in regulating the interaction between the cell and its extracellular matrix. The potential contribution of type VI collagen to this interaction was investigated in this study using isolated canine tibial chondrons embedded in agarose monolayers. The immunohistochemical distribution of an anti-type VI collagen antibody was assessed in these preparations using fluorescence, peroxidase and gold particle probes in combination with light, confocal and transmission electron microscopy. Light and confocal microscopy both showed type VI collagen concentrated in the pericellular capsule and matrix around the chondrocyte with reduced staining in the tail region and the interconnecting segments between adjacent chondrons. Minimal staining was recorded in the territorial and interterritorial matrices. At higher resolution, type VI collagen appeared both as microfibrils and as amorphous deposits that accumulated at the junction of intersecting capsular fibres and microfibrils. Electron microscopy also showed type VI collagen anchored to the chondrocyte membrane at the articular pole of the pericellular capsule and tethered to the radial collagen network through the tail at the basal pole of the capsule. We suggest that type VI collagen plays a dual role in the maintenance of chondron integrity. First, it could bind to the radial collagen network and stabilise the collagens, proteoglycans and glycoproteins of the pericellular microenvironment. Secondly, specific cell surface receptors exist, which could mediate the interaction between the chondrocyte and type VI collagen, providing firm anchorage and signalling potentials between the pericellular matrix and the cell nucleus. In this way type VI collagen could provide a close functional interrelationship between the chondrocyte, its pericellular microenvironment and the load bearing extracellular matrix of adult articular cartilage.

Smartphone apps are closely integrated with our daily lives, and mobile malware has brought about serious security issues. However, the features used in existing traffic-based malware detection techniques have a large amount of redundancy and useless information, wasting the computational resources of training detection models. To overcome this drawback, we propose a feature selection method; the core of the method involves choosing selected features based on high irrelevance, thereby removing redundant features. Furthermore, artificial intelligence has implemented malware detection and achieved outstanding detection ability. However, almost all malware detection models in deep learning include pooling operations, which lead to the loss of some local information and affect the robustness of the model. We also propose designing a malware detection model for malicious traffic identification based on a capsule network. The main difference between the capsule network and the neural network is that the neuron outputs a scalar, while the capsule outputs a vector. It is more conducive to saving local information. To verify the effectiveness of our method, we verify it from three aspects. First, we use four popular machine learning algorithms to prove the effectiveness of the proposed feature selection method. Second, we compare the capsule network with the convolutional neural network to prove the superiority of the capsule network. Finally, we compare our proposed method with another state-of-the-art malware detection technique; our accuracy and recall increased by 9.71% and 20.18%, respectively.

Verification of font style followed in a file is a difficult task to classify. An artificial intelligence based algorithm network can effectively perform this task in reduced time. Capsule network is one among such algorithm and an emerging technique implemented for so many classification process with limited datasets. The proposed font style classification algorithm is enforced with Capsule Network (CapsNet) algorithm for executing the font style classification task. The proposed method is confirmed by classifying times new roman, Arial black and Algerian font style in English letters along with the performance evaluation in terms of accuracy and confusion matrix parameters. The proposed network structure is also compared with the existing Naive Bayes (NB), Decision Tree (DT) and K nearest neighbor (KNN) algorithms for comparative study and the evaluation result indicates that the proposed font style classification model based on CapsNet is classifying the images with better accuracy, F1 score and Gmean.

Self-healing cementitious materials containing microcapsules filled with healing agents can autonomously seal cracks and restore structural integrity. However, optimising the microcapsule mechanical properties to survive concrete mixing whilst still rupturing at the cracked interface to release the healing agent remains challenging. This study develops an integrated numerical modelling and machine learning approach for tailoring acrylate-based microcapsules for triggering within cementitious matrices. Microfluidics is first utilised to produce microcapsules with systematically varied shell thickness, strength, and cement compatibility. The capsules are characterised and simulated using a continuum damage mechanics model that is able to simulate cracking. A parametric study investigates the key microcapsule and interfacial properties governing shell rupture versus matrix failure. The simulation results are used to train an artificial neural network to rapidly predict the triggering behaviour based on capsule properties. The machine learning model produces design curves relating the microcapsule strength, toughness, and interfacial bond to its propensity for fracture. By combining advanced simulations and data science, the framework connects tailored microcapsule properties to their intended performance in complex cementitious environments for more robust self-healing concrete systems.

Biomaterials can display complex spatial patterns of cellular responses to external forces. Revealing and predicting the role of these patterns in material failure require an understanding of the statistical dependencies between spatially distributed changes in a cell's local biomechanical environment, including altered collagen fibre kinematics in the extracellular matrix. Here, we develop and apply a novel extension of network science methods to investigate how excessive tensile stretch of the human cervical facet capsular ligament (FCL), a common source of chronic neck pain, affects the local reorganization of collagen fibres. We define collagen alignment networks based on similarity in fibre alignment angles measured by quantitative polarized light imaging. We quantify the reorganization of these networks following macroscopic loading by describing the dynamic reconfiguration of network communities, regions of the material that display similar fibre alignment angles. Alterations in community structure occur smoothly over time, indicating coordinated adaptation of fibres to loading. Moreover, flexibility, a measure of network reconfiguration, tracks the loss of FCL's mechanical integrity at the onset of anomalous realignment (AR) and regions of AR display altered community structure. These findings use novel network-based techniques to explain abnormal collagen fibre reorganization, a dynamic and coordinated multivariate process underlying tissue failure.

Classification of crops using time-series vegetation index (VI) curves requires appropriate modeling of phenological events and their characteristics. The current study explores the use of capsules, a group of neurons having an activation vector, to learn the characteristic features of the phenological curves. In addition, joint optimization of denoising and classification is adopted to improve the generalizability of the approach and to make it resilient to noise. The proposed approach employs reconstruction loss as a regularizer for classification, whereas the crop-type label is used as prior information for denoising. The activity vector of the class capsule is applied to sample the latent space conditioned on the cell state of a Long Short-Term Memory (LSTM) that integrates the sequences of the phenological events. Learning of significant phenological characteristics is facilitated by adversarial variational encoding in conjunction with constraints to regulate latent representations and embed label information. The proposed architecture, called the variational capsule network (VCapsNet), significantly improves the classification and denoising results. The performance of VCapsNet can be attributed to the suitable modeling of phenological events and the resilience to outliers and noise. The maxpooling-based capsule implementation yields better results, particularly with limited training samples, compared to the conventional implementations. In addition to the confusion matrix-based accuracy measures, this study illustrates the use of interpretability-based evaluation measures. Moreover, the proposed approach is less sensitive to noise and yields good results, even at shallower depths, compared to the main existing approaches. The performance of VCapsNet in accurately classifying wheat and barley crops indicates that the approach addresses the issues in crop-type classification. The approach is generic and effectively models the crop-specific phenological features and events. The interpretability-based evaluation measures further indicate that the approach successfully identifies the crop transitions, in addition to the planting, heading, and harvesting dates. Due to its effectiveness in crop-type classification, the proposed approach is applicable to acreage estimation and other applications in different scales.

Emotion recognition based on multi-channel electroencephalograph (EEG) signals is becoming increasingly attractive. However, the conventional methods ignore the spatial characteristics of EEG signals, which also contain salient information related to emotion states. In this paper, a deep learning framework based on a multiband feature matrix (MFM) and a capsule network (CapsNet) is proposed. In the framework, the frequency domain, spatial characteristics, and frequency band characteristics of the multi-channel EEG signals are combined to construct the MFM. Then, the CapsNet model is introduced to recognize emotion states according to the input MFM. Experiments conducted on the dataset for emotion analysis using EEG, physiological, and video signals (DEAP) indicate that the proposed method outperforms most of the common models. The experimental results demonstrate that the three characteristics contained in the MFM were complementary and the capsule network was more suitable for mining and utilizing the three correlation characteristics.

Abstract Introduction: Defibrotide (DF), now in clinical trials in Veno-Occlussive Disease (VOD), and Multi-Organ Failure (MOF) as an anti-thrombotic agent is also recognized as a modulator of phosphorilated nucleotides of ATP, ADP, NAD/NADH and cAMP..Thus far the mechanism of its polypharmacology, and the.identity of its biologically active components are not known.The precursor molecule is a double stranded mammalian DNA. And the final commercial product is a mixture of nucleotidic fractions which have formed by the ability of the molecule to form intra-molecular bonding., with random sequences such as:P1-5, (dAp)12–24, (dGp) 10–20, (dTp)13–26, (dCp)10–20, Wherein, P=phosphoric radical; dAp= deoxyadenylic monomer; dGp= deoxyguanylic monomer; dTp= deoxythymidylic monomer; dCp= deoxycytidilic monomer.(US Patent # 4,985,552,Crinos). This report aims to investigate for the first time the sequence spesificity, and the reproducability of the nucleotidic fractions of DF. and to define a structural base for DF as a global modulator of the signalling network. Method:.Materials from two commercial forms of DF, Prociclide and Noravid in solution and capsule forms, were separated in each case by HPLC on a Vydac, C8 column using 0.1% trifluoroacetate in water. 513 mass spectra peaks were collected, freeze dried, preserved on ice and subjected to Voyager Elite Biospectrometry Workstation using the technique of delayed extraction matrix assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry..A blank run representing DNA matrix was done in each case on MALDI-TOF to detect the low MW particles that were not related to the sample. The peaks were superimposed on the HPLC profiles of mono-, di-, tri-phosphates of T,C,G, A via crossmatching the MW, together with retention times as described above.260,000 crossmatches were done. Statistics: The min, max, mean and median mass values were stratified into four respective batches of DF, capsule and liquid forms, and two commercial labels The standard deviations (SD) were calculated in each of the above stratifications with the “n-1” method. Results: HPLC profiles contained more than one molecular species with MWs ranging from <300 to <1500 Da. and were not superimposable. A representative panel would include deoxyadenylic and deoxycytidylic monomers, guanine and other nucleotides of 3–25 bases, such as dC,dA, G, dGMP, AMP, oligonucleotides with 3–5 bases such as UTP, dTTP, CTP, ATP, dGTP, cTMP, CMP, cGMP, and dAMP. Since the study was only directed at bases, deoxy-bases and mono-and di-, and tri-nucleotides of </= 560 Da (the calculated MW of alkali sodium salt of ATP), statistical calculations were carried out only for these fractions. When compared to capsules, liquid DF showed double number of peaks of ATP, ADP, AMP, GTP, GDP, and GMP. The standard deviations (STD) on the nucleotides in the various liquid forms of Defibrotide were much narrower, ranging between (+ / −) 0–1.4, capsule forms showing less consistency with STD’s of (+ / −)0–2.1 Conclusion: The sequence analysis of the nucleotidic fractions of <560 D MW displayed presence of cAMP, ATP, ADP, GTP, GDP,,GMP, CTP,CMP,CDP, TTP,TDP,TMP reproducably between batches. DF maybe a direct and indirect supplier of messenger nucleotides. This structural analysis may support this hypothesis generating report presenting DF as the prototype global modulator of the signalling network as a mechanism of its polypharmacology.

Unsupervised feature extraction is crucial in machine learning and data mining for handling high-dimensional and unlabeled data. However, existing methods often ignore feature relationaships, resulting in suboptimal feature subsets. This paper reviews the current state of unsupervised feature extraction methods, discussing the limitations of traditional methods such as Principal Component Analysis (PCA) and Independent Component Analysis (ICA), particularly in terms of interpretability, sensitivity to outliers, and computational resource challenges. In recent years, improvement strategies such as information theory, sparse learning, and deep learning (e.g., deep autoencoders and generative adversarial networks) have significantly progressed in feature extraction. This paper analyzes the practical applications of these methods in image processing, gene analysis, text mining, and network security. For example, in image processing, deep autoencoder-based methods such as Matrix Capsules with EM Routing can effectively extract key features from complex images. In text mining, unsupervised feature selection methods combined with generative adversarial networks significantly improve the efficiency of processing high-dimensional text data. Additionally, this paper explores future research directions such as multimodal data processing, improving real-time processing capabilities, and integration with other machine learning techniques (e.g., reinforcement learning, transfer learning), providing insights for further development of unsupervised feature extraction technologies.

Point cloud registration is the foundation and key step for many vital applications, such as digital city, autonomous driving, passive positioning, and navigation. The difference of spatial objects and the structure complexity of object surfaces are the main challenges for the registration problem. In this paper, we propose a graph attention capsule model (named as GACM) for the efficient registration of terrestrial laser scanning (TLS) point cloud in the urban scene, which fuses graph attention convolution and a three-dimensional (3D) capsule network to extract local point cloud features and obtain 3D feature descriptors. These descriptors can take into account the differences of spatial structure and point density in objects and make the spatial features of ground objects more prominent. During the training progress, we used both matched points and non-matched points to train the model. In the test process of the registration, the points in the neighborhood of each keypoint were sent to the trained network, in order to obtain feature descriptors and calculate the rotation and translation matrix after constructing a K-dimensional (KD) tree and random sample consensus (RANSAC) algorithm. Experiments show that the proposed method achieves more efficient registration results and higher robustness than other frontier registration methods in the pairwise registration of point clouds.

Metal cabinets such as switch cabinets and ring network cabinets have the advantages of small footprints and good protection for equipment and offer neat and orderly protection. They are widely used in power systems. In a hot and humid environment, condensation can easily cause equipment to age and even cause insulation failure. Therefore, research on reliable anticondensation methods is of great significance for the safe operation of power equipment. In this study, phase change capsules with phase transition temperatures of 22 and 32 °C were used as fillers and liquid silicone rubber was used as a matrix to prepare liquid silicone rubber composites filled with phase change capsules for a temperature-control coating. Studies have shown that liquid silicone rubber coatings containing phase change capsules can significantly enhance the anticondensation performance of metal cabinets. By using a temperature-control coating on the surface where the cabinet experiences condensation, the temperature difference between the surface and the dew point is reduced, thereby slowing down the condensation rate and even preventing condensation.

Background: Endometriosis (EMs) is a common chronic inflammatory gynecological disease that belongs to the classification of Traditional Chinese Medicine Syndromes “Zheng Jia,” and the classic Chinese formula Guizhi Fuling (GZFL) demonstrates significant clinical efficacy in the treatment of this condition. This study aims to investigate GZFL’s effect and potential mechanism in EMs. Methods: The search reviewed randomized controlled trials in 7 databases from inception to 2024, assessed quality with the Cochrane tool, and analyzed data with STATA 15 by 2 reviewers. In the network pharmacology study, we searched and screened the components and targets of GZFL, subsequently compared these targets to EMs targets, and used bioinformatics techniques to analyze and explore their potential interactions. Results: Nine randomized controlled trials involving 897 participants were analyzed. Meta-analysis showed that GZFL combined with dienogest significantly enhanced the clinical effectiveness rate (odds ratio = 2.404, 95% confidence intervals [CI], 1.868 to 3.093; P < .001). Specifically, combination therapy with GZFL reduced serum carbohydrate antigen 125 (standardized mean differences [SMD] = −1.65, 95% CI = −2.13 to −1.17, P < .001), estradiol (SMD = −1.54, 95% CI = −1.89 to −1.19, P = .003), matrix metalloproteinases (SMD = −2.636, 95% CI = −2.993 to −2.279, P < .001), pain scores (SMD = −0.88, 95% CI = −1.11 to −0.67, P < .001) and the diameter of ectopic cysts (SMD = −1.7, 95% CI = −2.42 to −0.98, P < .001). Network pharmacology analysis identified 136 components and 145 common targets, focusing on interleukin-6, cellular tumor antigen p53, epidermal growth factor receptor, estrogen receptor alpha, Cyclooxygenase-2, and matrix metalloproteinases-9. Gene ontology and Kyoto Encyclopedia of Genes and Genomes analyses suggested GZFL modulates hormone receptors and inflammatory responses in EMs treatment. Conclusion: In conclusion, GZFL combination treatment could increase the clinical effectiveness rate of EMs patients, and reduce the serum level of carbohydrate antigen 125, estradiol, matrix metalloproteinases, pain scores, and the diameter of the ectopic cyst. The potential mechanism might be linked to the modulation of hormone receptors and inflammation.

Endometriosis is a chronic condition in which the lining of the uterus grows outside the uterus, causing pain, swelling, and fertility problems. It usually affects the ovaries, fallopian tubes, and pelvic lining, leading to severe menstrual cramps and other complications. Traditional methods for diagnosing endometriosis, such as laparoscopy and ultrasound, are often invasive, time-consuming, and can lead to delayed diagnosis. Relying on symptom-based assessment lacks accuracy, making early and affective treatment challenging. To solve the problem a novel Hyper capsule Resnet50-CNN algorithm is introduced for classifying the ovarian cysts by utilizing the ultrasound images processed datasets and applied to the image processing technique. Initially, Butterworth Filter preprocessing enhanced the details of the input data set and gave a clear view of the input dataset. Modified Watershed Segmentation algorithm (MWSA) separates follicles or cysts that specifically differentiate for selection features. An improved Recursive Bee Colony (RBC) Feature Selection algorithm is trained to identify biologically significant markers, ensuring accurate feature extraction without errors. ResNet50 with CNN architecture is a deep learning approach to extract complex features methodically hyper capsule ResNet50 contains 50 layers of network operation, which is a disappearing gradient issue that is frequently observed. RESNET 50 classification identifies ovarian cysts into three stages based on the condition: regular nodule, ovarian growth, and polycystic ovary. The accuracy is 94.15%, sensitivity 95.82%, Specificity 94.54%, FI– Score 94.89% and RMSE 84.25% measure parameters are analyzed, and the performance Matrix obtains results.

Self-healing of concrete is the process in which the material regenerates itself repairing inner cracks. This process can be produced by autogenous or autonomous healing. Autogenous healing is a natural process, produced by carbonation and/or continuing hydration. Autonomous healing is based on the use of specific agents to produce self-healing, which can be added directly to the concrete matrix, embedded in capsules or introduced through vascular networks. Some examples are superabsorbent polymers, crystalline admixtures, microencapsulated sodium silicate, and bacteria. This review is structured into two parts. The first part is an overview of self-healing concrete that summarises the basic concepts and the main advances produced in the last years. The second part is a critical discussion on the feasibility of self-healing concrete, its possibilities, current weaknesses, and challenges that need to be addressed in the coming years.

Natural rubber with 50 mol % epoxidation (ENR-50) was filled with carbon nanotubes (CNTs) and conductive carbon black (CCB) hybrid fillers with various CCB loadings of 2.5, 5.0, 7.0, 10.0 and 15.0 phr, and the compounds were mixed with ferric ion (Fe3+) as a crosslinking agent. The ENRs filled exclusively with CNTs, and CNT–CCB hybrid fillers exhibited typical curing curves at different CCB loadings, i.e., increasing torque with time and thus crosslinked networks. Furthermore, the incorporation of CNT–CCB hybrid fillers and increasing CCB loadings caused an enhancement of tensile properties (modulus and tensile strength) and crosslink densities, which are indicated by the increasing torque difference and the crosslink densities. The crosslink densities are determined by swelling and temperature scanning stress relaxation (TSSR). Increasing CCB loadings also caused a significant improvement in bound rubber content, filler–rubber interactions, thermal resistance, glass transition temperature (Tg) and electrical conductivity. A combination of 7 phr CNT and CCB with loading higher than 2.5 phr gave superior properties to ENR vulcanizates. Furthermore, the secondary CCB filler contributes to the improvement of CNT dispersion in the ENR matrix by networking the CNT capsules and forming CNT–CCB–CNT pathways and thus strong CNT–CCB networks, indicating the improvement in the tensile properties, bound rubber content and dynamic properties of the ENR composites. Moreover, higher electrical conductivity with a comparatively low percolation threshold of the hybrid composites was found as compared to the ENR filled with CNTs without CCB composite. The superior mechanical and other properties are due to the finer dispersion and even distribution of CNT–CCB hybrid fillers in the ENR matrix.

Excessive deformation of nerve fibres (axons) in the spinal facet capsular ligaments (FCLs) can be a cause of pain. The axons are embedded in the fibrous extracellular matrix (ECM) of FCLs, so understanding how local fibre organization and micromechanics modulate their mechanical behaviour is essential. We constructed a computational discrete-fibre model of an axon embedded in a collagen fibre network attached to the axon by distinct fibre–axon connections. This model was used to relate the axonal deformation to the fibre alignment and collagen volume concentration of the surrounding network during transverse, axial and shear deformations. Our results showed that fibre alignment affects axonal deformation only during transverse and axial loading, but higher collagen volume concentration results in larger overall axonal strains for all loading cases. Furthermore, axial loading leads to the largest stretch of axonal microtubules and induces the largest forces on axon's surface in most cases. Comparison between this model and a multiscale continuum model for a representative case showed that although both models predicted similar averaged axonal strains, strain was more heterogeneous in the discrete-fibre model.

Among many applications, elevated-temperature cured epoxy resins are widely used for high-performance applications especially for structural adhesive and as a matrix for structural composites. This is due to their superior chemical and mechanical properties. The thermosetting nature of epoxy produces a highly cross-linked polymer network during the curing process where the resulting material exhibited excellent properties. However, due to this cross-linked molecular structure, epoxies are also known to be brittle, and once a crack initiated in the material, it is difficult to arrest the crack propagation. Earlier research found that the inclusion of encapsulated healing agents is able to introduce self-healing ability to the room-temperature cured epoxies. The current study investigated the self-healing behaviour of an elevated-temperature cured epoxy, which incorporated the dual-capsule system loaded with diglycidyl-ether of bisphenol-A (DGEBA) resin and mercaptan. The microcapsules were prepared by the in-situ polymerisation method while the fracture toughness and the self-healing capability of the tapered-double-cantilever-beam (TDCB) epoxy specimens were measured under Mode-I fracture toughness testing. We investigated the effect of temperature on viscosity of the healing agents and how these values influence the formation of uniform healing on the fracture surfaces. It was found that incorporation of the dual-capsule self-healing system onto an elevated-temperature cured epoxy slightly changed the fracture toughness of the epoxy as indicated by the Mode-I testing. In the case of thermal healing at 70°C, the self-healing epoxy exhibited a recovery of up to 111% of its original fracture toughness, where a uniform spreading of the healant was observed. The excellent healing behaviour is attributed to the lower viscosity of the healant at higher temperature and the higher glass transition temperature ( Tg) of the produced healant film. The DSC analysis confirmed that the healing process was not contributed by the post-curing of the host epoxy.

Purpose. To study the cause and mechanism of joint degeneration in osteoarthritis, through histopathological and ultrastructural-histochemical experiments on the articular cartilage of the knees of the C57 black mouse. Methods. 192 C57 black mice and a control group of 64 C57BL/6J mice were used in this study. The left and right knee articular capsules of the joints were removed and stained. Each articular cartilage sample was examined and osteoarthritic changes were assessed using a transmission electron microscope. The severity of osteoarthritis in the knee joint cartilage of C57 black mice was histologically assessed using a classification system described by Okabe, based on Maier's system. Results. The incidence and the severity of osteoarthritis gradually increased with age; the incidence increased from 20% at 2 months to 80% at 16 months. Irreversible changes appeared at an advanced stage, and the process of degeneration was quite similar to that in human osteoarthritis. Through transmission electron microscopy, we observed poorly developed Golgi apparatus, markedly increased intracellular microfilaments, decreased proteoglycan granules, and broken collagen networks in all stages of osteoarthritis. By contrast, Golgi apparatus and other organelles were well developed in histologically normal mice of all ages. Proteoglycan granules, which mainly consisted of keratan sulphate, were observed; collagen networks were maintained. Conclusion. Disturbed protein transport and sugar synthesis in chondrocytes, caused by the deficient development of the Golgi apparatus, could result in degenerative changes in articular cartilage. The structure and function of the matrix were maintained mainly because of the continued presence of keratan sulphate.

An appropriately designed sustained release dosage form is opted to be a major goal in solving the problems which arises regarding the targeting of a drug to a specific organ or tissue and for controlling its rate of delivery to the target site. The development of oral sustained release system has proven to be a major challenge to formulation scientist due to their inability to restrain as well as localize the system at targeted areas of the gastrointestinal tract. Therefore the development of matrix type drug delivery system is promising option regarding the development of an oral sustained release system. There is availability of wide variety of polymers which helps the formulation scientist to develop sustained/controlled release products. The attractiveness of these dosage forms is increasing because of their awareness towards toxicity and ineffectiveness when administered by oral route in the form of tablets and capsules. Numerous advantages are provided by sustained release products over conventional dosage forms through optimizing various bio-pharmaceutics, pharmacokinetic and pharmacodynamics properties of drugs and finally leads to reduction in dosing frequency to such an extent that only once daily dose is required for therapeutic management with maximum utility of drug with reduction in both local as well as systemic side effects. They can cure or control diseased condition in shortest possible time with smallest quantity of drug to assure greater patient compliance. Polymer swelling, drug dissolution and its diffusion are the known mechanisms for drug release through polymer network. Keywords: Oral drug delivery system, sustained release dosage form, matrix system, polymer swelling, drug diffusion.

Mesenchymal stromal cells (MSC) can promote tissue protection following injury, in part by modulating inflammatory cell responses. The aim of this study was to investigate the potential tissue protective properties of encapsulated MSCs (eMSC) in an organotypic injury model induced by fibronectin culture. MSC were encapsulated in alginate beads containing a network of nanopores, which segregate the cells from the extracapsular milieu, while still permitting diffusion into and out of the capsule. An increase in blood brain barrier permeability during pathological conditions permits the influx of blood plasma constituents that can be quite harmful to surrounding tissues. In particular, increased concentrations of fibronectin have been shown in a number of diseases and CNS traumas, co-localizing in areas of activated microglia. We observed over a 14-day period, a consistent increase in OHC degradation in the presence of fibronectin measured by a significant decrease in slice area, the breakdown in OHC pyramidal layers, and consistent cell death over the culture period. Microglial ionized calcium-binding adapter molecule 1 (IBA-1) expression remained elevated throughout the culture period with the majority found within the pyramidal layers. When eMSC were added to the cultures, a significant decrease in OHC degradation was observed as reflected by a reduction in OHC area shrinkage and in the amount of cell death. In the presence of eMSC, pyramidal layer structure was maintained and axonal extension from the periphery of the OHCs was observed. Therefore, MSC, delivered in a nanoporous alginate matrix, can modulate responses to injury by reversing fibronectin-induced OHC degradation.

In this paper, the application of 3-dimensional (3D) functional magnetic resonance imaging (FMRI) in the diagnosis of the 5th lumbar (L5) nerve root compression and brain functional areas in patients with lumbar disc herniation (LDH) was analyzed. The traditional fast independent component analysis (Fast ICA) algorithm was optimized based on the modified whitening matrix to establish a new type of Modified-Fast ICA (M-Fast ICA) algorithm that was compared with the introduced traditional Fast ICA and ICA. M-Fast ICA was applied to the 3D FMRI diffusion tensor imaging (DTI) evaluation of 65 patients with L5 nerve root pain due to LDH (group A) and 50 healthy volunteers (group B). The values of fractional anisotropy (FA) and apparent diffusion coefficient (ADC) in the lumbar nerve roots (L3, L4, L5, and the 1st sacral vertebra (S1)) were recorded among subjects from the two groups. Besides, the score of edema degree in the lumbar nerve roots (L5 and S1) and activity of brain functional areas were also recorded among all subjects of the two groups. The results showed that the mean square error of M-Fast ICA was smaller than that of traditional Fast ICA and ICA, while its signal-to-noise ratio (SNR) was greater than that of Fast ICA and ICA ( P < 0.05 ). The FA of L5 and S1 nerve roots in patients of group A was sharply lower than the values of group B, while the ADC of patients in group A was greater than that of the control group ( P < 0.05 ). Besides, the score of edema in L5 and S1 nerve roots of patients in group A increased in contrast to group B ( P < 0.05 ). The brain areas were activated after surgery including bilateral temporal lobe, left thalamus, splenium of corpus callosum, and right internal capsule. In conclusion, the 3D image denoising performance of M-Fast ICA optimized and constructed in this study was superior to that of the traditional Fast ICA and ICA. The FA of patients with L5 nerve root pain due to LDH decreased steeply, while the ADC increased dramatically. L5 nerve root pain caused by LDH resulted in changes in brain functional areas of the patients to inhibit the resting state default network activity, and the corresponding brain functional areas could be activated through treatment.

This article provides the review of the medical use of pH- and temperature-sensitive polymer hydrogels. Such polymers are characterised by their thermal and pH sensitivity in aqueous solutions at the functioning temperature of living organisms and can react to the slightest changes in environmental conditions. Due to these properties, they are called stimuli-sensitive polymers. This response to an external stimulus occurs due to the amphiphilicity (diphilicity) of these (co)polymers. The term hydrogels includes several concepts of macrogels and microgels. Microgels, unlike macrogels, are polymer particles dispersed in a liquid and are nano- or micro-objects. The review presents studies reflecting the main methods of obtainingsuch polymeric materials, including precipitation polymerisation, as the main, simplest, and most accessible method for mini-emulsion polymerisation, microfluidics, and layer-by-layer adsorption of polyelectrolytes. Such systems will undoubtedly be promising for use in biotechnology and medicine due to the fact that they are liquid-swollen particles capable of binding and carrying various low to high molecular weight substances. It is also important that slight heating and cooling or a slight change in the pH of the medium shifts the system from a homogeneous to a heterogeneous state and vice versa. This providesthe opportunity to use these polymers as a means of targeted drug delivery, thereby reducing the negative effect of toxic substances used for treatment on the entire body and directing the action to a specific point. In addition, such polymers can be used to create smart coatings of implanted materials, as well as an artificial matrix for cell and tissue regeneration, contributing to a significant increase in the survival rate and regeneration rate of cells and tissues.
 
 
 
 
 References
 1. Gisser K. R. C., Geselbracht M. J., Cappellari A.,Hunsberger L., Ellis A. B., Perepezko J., et al. Nickeltitaniummemory metal: A "Smart" material exhibitinga solid-state phase change and superelasticity. Journalof Chemical Education. 1994;71(4): 334. DOI: https://doi.org/10.1021/ed071p3342. Erman B., Flory P.J.. Critical phenomena andtransitions in swollen polymer networks and in linearmacromolecules. Macromolecules. 1986;19(9): 2342–2353. DOI: https://doi.org/10.1021/ma00163a0033. Tanaka T., Fillmore D., Sun S.-T., Nishio I.,Swislow G., Shah A. Phase transitions in ionic gels.Physical Review Letters. 1980;45(20): 1636–1639. DOI:https://doi.org/10.1103/physrevlett.45.16364. Polymer Gels. DeRossi D., Kajiwara K., Osada Y.,Yamauchi A. (eds.). Boston, MA: Springer US; 1991.354 p. DOI: https://doi.org/10.1007/978-1-4684-5892‑35. Ilmain F., Tanaka T., Kokufuta E. Volumetransition in a gel driven by hydrogen bonding. Nature.1991;349(6308): 400–401. DOI: https://doi.org/10.1038/349400a06. Kuhn W., Hargitay B., Katchalsky A., EisenbergH. Reversible dilation and contraction by changing thestate of ionization of high-polymer acid networks.Nature. 1950;165(4196): 514–516. DOI: https://doi.org/10.1038/165514a07. Steinberg I. Z., Oplatka A., Katchalsky A.Mechanochemical engines. Nature. 1966;210(5036):568-571. DOI: https://doi.org/10.1038/210568a08. Tian H., Tang Z., Zhuang X., Chen X., Jing X.Biodegradable synthetic polymers: Preparation,functionalization and biomedical application. Progressin Polymer Science. 2012;37(2): 237–280. DOI: https://doi.org/10.1016/j.progpolymsci.2011.06.0049. Gonçalves C., Pereira P., Gama M. Self-Assembledhydrogel nanoparticles for drug delivery applications.Materials. 2010;3(2): 1420–1460. DOI: https://doi.org/10.3390/ma302142010. Pangburn T. O., Petersen M. A., Waybrant B.,Adil M. M., Kokkoli E. Peptide- and Aptamer-functionalizednanovectors for targeted delivery of therapeutics.Journal of Biomechanical Engineering. 2009;131(7):074005. DOI: https://doi.org/10.1115/1.316076311. Caldorera-Moore M. E., Liechty W. B., PeppasN. A. Responsive theranostic systems: integrationof diagnostic imaging agents and responsive controlledrelease drug delivery carriers. Accounts of ChemicalResearch. 2011;44(10): 1061–1070. DOI: https://doi.org/10.1021/ar200177712. Das M., Sanson N., Fava D., Kumacheva E.Microgels loaded with gold nanorods: photothermallytriggered volume transitions under physiologicalconditions†’. Langmuir. 2007;23(1): 196–201. DOI:https://doi.org/10.1021/la061596s13. Oh J. K., Lee D. I., Park J. M. Biopolymer-basedmicrogels/nanogels for drug delivery applications.Progress in Polymer Science. 2009;34(12): 1261–1282.DOI: https://doi.org/10.1016/j.progpolymsci.2009.08.00114. Oh J. K., Drumright R., Siegwart D. J.,Matyjaszewski K. The development of microgels/nanogels for drug delivery applications. Progress inPolymer Science. 2008;33(4): 448–477. DOI: https://doi.org/10.1016/j.progpolymsci.2008.01.00215. Talelli M., Hennink W. E. Thermosensitivepolymeric micelles for targeted drug delivery.Nanomedicine. 2011;6(7): 1245–1255. DOI: https://doi.org/10.2217/nnm.11.9116. Bromberg L., Temchenko M., Hatton T. A. Smartmicrogel studies. Polyelectrolyte and drug-absorbingproperties of microgels from polyether-modifiedpoly(acrylic acid). Langmuir. 2003;19(21): 8675–8684.DOI: https://doi.org/10.1021/la030187i17. Vinogradov S. V. Polymeric nanogel formulationsof nucleoside analogs. Expert Opinion on Drug Delivery.2007;4(1): 5–17. DOI: https://doi.org/10.1517/17425247.4.1.518. Vinogradov S. V. Colloidal microgels in drugdelivery applications. Current Pharmaceutical Design.2006;12(36): 4703–4712. DOI: https://doi.org/10.2174/13816120677902625419. Kabanov A. V., Vinogradov S. V. Nanogels aspharmaceutical carriers: finite networks of infinitecapabilities. Angewandte Chemie International Edition.2009;48(30): 5418–5429. DOI: https://doi.org/10.1002/anie.20090044120. Lee E. S., Gao Z., Bae Y. H. Recent progress intumor pH targeting nanotechnology. Journal ofControlled Release. 2008;132(3): 164–170. DOI: https://doi.org/10.1016/j.jconrel.2008.05.00321. Dong H., Mantha V., Matyjaszewski K.Thermally responsive PM(EO)2MA magnetic microgelsvia activators generated by electron transfer atomtransfer radical polymerization in miniemulsion.Chemistry of Materials. 2009;21(17): 3965–3972. DOI:https://doi.org/10.1021/cm901143e22. Nayak S., Lyon L. A. Soft nanotechnology withsoft nanoparticles. Angewandte Chemie InternationalEdition. 2005;44(47): 7686–7708. DOI: https://doi.org/10.1002/anie.20050132123. Hennink W. E., van Nostrum C. F. Novelcrosslinking methods to design hydrogels. AdvancedDrug Delivery Reviews. 2012;64: 223–236. DOI: https://doi.org/10.1016/j.addr.2012.09.00924. Motornov M., Roiter Y., Tokarev I., Minko S.Stimuli-responsive nanoparticles, nanogels and capsulesfor integrated multifunctional intelligent systems.Progress in Polymer Science. 2010;35(1-2): 174–211. DOI:https://doi.org/10.1016/j.progpolymsci.2009.10.00425. Saunders B. R., Laajam N., Daly E., Teow S.,Hu X., Stepto R. Microgels: From responsive polymercolloids to biomaterials. Advances in Colloid andInterface Science. 2009;147-148: 251–262. DOI: https://doi.org/10.1016/j.cis.2008.08.00826. Landfester K. Miniemulsion polymerizationand the structure of polymer and hybrid nanoparticles.chemInform. 2009;40(33). DOI: https://doi.org/10.1002/chin.20093327927. Seo M., Nie Z., Xu S., Mok M., Lewis P.C.,Graham R., et al. Continuous microfluidic reactors forpolymer particles. Langmuir. 2005;21(25): 11614–11622. DOI: https://doi.org/10.1021/la050519e28. Nie Z., Li W., Seo M., Xu S., Kumacheva E. Janusand ternary particles generated by microfluidicsynthesis: design, synthesis, and self-assembly. Journalof the American Chemical Society. 2006;128(29): 9408–9412. DOI: https://doi.org/10.1021/ja060882n29. Seiffert S., Thiele J., Abate A. R., Weitz D. A.Smart microgel capsules from macromolecularprecursors. Journal of the American Chemical Society.2010;132(18): 6606–6609. DOI: https://doi.org/10.1021/ja102156h30. Rossow T., Heyman J. A., Ehrlicher A. J.,Langhoff A., Weitz D. A., Haag R., et al. Controlledsynthesis of cell-Laden Microgels by Radical-FreeGelation in Droplet Microfluidics. Journal of theAmerican Chemical Society. 2012;134(10): 4983–4989.DOI: https://doi.org/10.1021/ja300460p31. Perry J. L., Herlihy K. P., Napier M. E.,DeSimone J. M. PRINT: A novel platform toward shapeand size specific nanoparticle theranostics. Accountsof Chemical Research. 2011;44(10): 990–998. DOI:https://doi.org/10.1021/ar200031532. Caruso F., Sukhorukov G. Coated Colloids:Preparation, characterization, assembly and utilization.In: Decher G., Schlenoff J. B., editors. MultilayerThin Films. Weinheim, FRG: Wiley-VCH Verlag GmbH& Co. KGaA; 2002. p. 331-362.33. Sauzedde F., Elaïssari A., Pichot C. Hydrophilicmagnetic polymer latexes. 2. Encapsulation ofadsorbed iron oxide nanoparticles. Colloid & PolymerScience. 1999;277(11): 1041–1050. DOI: https://doi.org/10.1007/s00396005048834. Sauzedde F., Elaïssari A., Pichot C. Hydrophilicmagnetic polymer latexes. 1. Adsorption of magneticiron oxide nanoparticles onto various cationic latexes.Colloid & Polymer Science. 1999;277(9): 846–855. DOI:https://doi.org/10.1007/s00396005046135. Pich A., Richtering W. Microgels by PrecipitationPolymerization: Synthesis, Characterization, andFunctionalization. In: Pich A., Richtering W. (eds.)Chemical Design of Responsive Microgels. SpringerHeidelberg Dordrecht London New York; 2011. p. 1–37.DOI: https://doi.org/10.1007/978-3-642-16379-136. Yamada N., Okano T., Sakai H., Karikusa F.,Sawasaki Y., Sakurai Y. Thermo-responsive polymericsurfaces; control of attachment and detachment ofcultured cells. Die Makromolekulare Chemie, RapidCommunications. 1990;11(11): 571–576. DOI: https://doi.org/10.1002/marc.1990.03011110937. Kushida A., Yamato M., Konno C., Kikuchi A.,Sakurai Y., Okano T. Decrease in culture temperaturereleases monolayer endothelial cell sheets togetherwith deposited fibronectin matrix from temperatureresponsiveculture surfaces. Journal of BiomedicalMaterials Research. 1999;45(4): 355–362. DOI: https://doi.org/10.1002/(sici)1097-4636(19990615)45:4<355::aid-jbm10>3.0.co;2-738. Sekine H., Shimizu T., Dobashi I., Matsuura K.,Hagiwara N., Takahashi M., et al. Cardiac cell sheettransplantation improves damaged heart function viasuperior cell survival in comparison with dissociatedcell injection. Tissue Engineering Part A. 2011;17(23-24): 2973–2980. DOI: https://doi.org/10.1089/ten.tea.2010.065939. Nishida K., Yamato M., Hayashida Y.,Watanabe K., Yamamoto K., Adachi E., et al. Corneal reconstruction with tissue-engineered cell sheetscomposed of autologous oral mucosal epithelium. TheNew England Journal of Medicine. 2004;351(12): 1187–1196. DOI: https://doi.org/10.1056/nejmoa04045540. Kanzaki M., Yamato M., Yang J., Sekine H.,Kohno C., Takagi R., et al. Dynamic sealing of lungair leaks by the transplantation of tissue engineeredcell sheets. Biomaterials. 2007;28(29): 4294–4302.DOI: https://doi.org/10.1016/j.biomaterials.2007.06.00941. Iwata T., Yamato M., Tsuchioka H., Takagi R.,Mukobata S., Washio K., et al. Periodontal regenerationwith multi-layered periodontal ligament-derived cellsheets in a canine model. Biomaterials. 2009;30(14):2716–2723. DOI: https://doi.org/10.1016/j.biomaterials.2009.01.03242. Sawa Y., Miyagawa S., Sakaguchi T., Fujita T.,Matsuyama A., Saito A., et al. Tissue engineeredmyoblast sheets improved cardiac function sufficientlyto discontinue LVAS in a patient with DCM: report ofa case. Surgery Today. 2012;42(2): 181–184. DOI:https://doi.org/10.1007/s00595-011-0106-443. Ohki T., Yamato M., Ota M., Takagi R.,Murakami D., Kondo M., et al. Prevention of esophagealstricture after endoscopic submucosal dissection usingtissue-engineered cell sheets. Gastroenterology.2012;143(3): 582–588. DOI: https://doi.org/10.1053/j.gastro.2012.04.05044. Ebihara G., Sato M., Yamato M., Mitani G.,Kutsuna T., Nagai T., et al. Cartilage repair intransplanted scaffold-free chondrocyte sheets usinga minipig model. Biomaterials. 2012;33(15): 3846–3851. DOI: https://doi.org/10.1016/j.biomaterials.2012.01.05645. Sato M., Yamato M., Hamahashi K., Okano T.,Mochida J. Articular cartilage regeneration using cellsheet technology. The Anatomical Record. 2014;297(1):36–43. DOI: https://doi.org/10.1002/ar.2282946. Kuramoto G., Takagi S., Ishitani K., Shimizu T.,Okano T., Matsui H. Preventive effect of oral mucosalepithelial cell sheets on intrauterine adhesions. HumanReproduction. 2014;30(2): 406–416. DOI: https://doi.org/10.1093/humrep/deu32647. Yamamoto K., Yamato M., Morino T.,Sugiyama H., Takagi R., Yaguchi Y., et al. Middle earmucosal regeneration by tissue-engineered cell sheettransplantation. NPJ Regenerative Medicine. 2017;2(1):6. DOI: https://doi.org/10.1038/s41536-017-0010-748. Gan D., Lyon L. A. Synthesis and Proteinadsorption resistance of PEG-modified poly(Nisopropylacrylamide) core/shell microgels.Macromolecules. 2002;35(26): 9634–9639. DOI: https://doi.org/10.1021/ma021186k49. Veronese F. M., Mero A. The impact ofPEGylation on biological therapies. BioDrugs.2008;22(5): 315–329. DOI: https://doi.org/10.2165/00063030-200822050-0000450. Sahay G., Alakhova D. Y., Kabanov A. V.Endocytosis of nanomedicines. Journal of ControlledRelease. 2010;145(3): 182–195. DOI: https://doi.org/10.1016/j.jconrel.2010.01.03651. Nolan C. M., Reyes C. D., Debord J. D.,García A. J., Lyon L. A. Phase transition behavior,protein adsorption, and cell adhesion resistance ofpoly(ethylene glycol) cross-linked microgel particles.Biomacromolecules. 2005;6(4): 2032–2039. DOI:https://doi.org/10.1021/bm050008752. Scott E. A., Nichols M. D., Cordova L. H., GeorgeB. J., Jun Y.-S., Elbert D. L. Protein adsorption and celladhesion on nanoscale bioactive coatings formed frompoly(ethylene glycol) and albumin microgels.Biomaterials. 2008;29(34): 4481–4493. DOI: https://doi.org/10.1016/j.biomaterials.2008.08.00353. South A. B., Whitmire R. E., García A. J.,Lyon L. A. Centrifugal deposition of microgels for therapid assembly of nonfouling thin films. ACS AppliedMaterials & Interfaces. 2009;1(12): 2747–2754. DOI:https://doi.org/10.1021/am900543554. Wang Q., Uzunoglu E., Wu Y., Libera M. Selfassembledpoly(ethylene glycol)-co-acrylic acidmicrogels to inhibit bacterial colonization of syntheticsurfaces. ACS Applied Materials & Interfaces. 2012;4(5):2498–2506. DOI: https://doi.org/10.1021/am300197m55. Wang Q., Libera M. Microgel-modified surfacesenhance short-term osteoblast response. Colloids andSurfaces B: Biointerfaces. 2014;118: 202–209. DOI:https://doi.org/10.1016/j.colsurfb.2014.04.00256. Tsai H.-Y., Vats K., Yates M. Z., Benoit D. S. W.Two-dimensional patterns of poly(N-isopropylacrylamide)microgels to spatially control fibroblastadhesion and temperature-responsive detachment.Langmuir. 2013;29(39): 12183–12193. DOI: https://doi.org/10.1021/la400971g57. Lynch I. , Miller I. , Gallagher W. M. ,Dawson K. A. Novel method to prepare morphologicallyrich polymeric surfaces for biomedical applicationsvia phase separation and arrest of microgel particles.The Journal of Physical Chemistry B. 2006;110(30):14581–14589. DOI: https://doi.org/10.1021/jp061166a58. Li Y., Chen P., Wang Y., Yan S., Feng X., Du W.,et al. Rapid assembly of heterogeneous 3D cellmicroenvironments in a microgel array. AdvancedMaterials. 2016;28(18): 3543–3548. DOI: https://doi.org/10.1002/adma.20160024759. Bridges A. W., Singh N., Burns K. L., BabenseeJ. E., Andrew Lyon L., García A. J. Reduced acuteinflammatory responses to microgel conformalcoatings. Biomaterials. 2008;29(35): 4605–4615. DOI:https://doi.org/10.1016/j.biomaterials.2008.08.01560. Bridges A. W., Whitmire R. E., Singh N.,Templeman K. L., Babensee J. E., Lyon L. A., et al.Chronic inflammatory responses to microgel-basedimplant coatings. Journal of Biomedical Materials Research Part A. 2010;94A(1): 252–258. DOI: https://doi.org/10.1002/jbm.a.3266961. Gutowski S. M., Templeman K. L., South A. B.,Gaulding J. C., Shoemaker J. T., LaPlaca M. C., et al.Host response to microgel coatings on neuralelectrodes implanted in the brain. Journal of BiomedicalMaterials Research Part A. 2014;102(5): 1486–1499.DOI: https://doi.org/10.1002/jbm.a.3479962. da Silva R. M. P., Mano J. F., Reis R. L. Smartthermoresponsive coatings and surfaces for tissueengineering: switching cell-material boundaries.Trends in Biotechnology. 2007;25(12): 577–583. DOI:https://doi.org/10.1016/j.tibtech.2007.08.01463. Schmidt S., Zeiser M., Hellweg T., Duschl C.,Fery A., Möhwald H. Adhesion and mechanicalproperties of PNIPAM microgel films and theirpotential use as switchable cell culture substrates.Advanced Functional Materials. 2010;20(19): 3235–3243. DOI: https://doi.org/10.1002/adfm.20100073064. Uhlig K., Wegener T., He J., Zeiser M., BookholdJ., Dewald I., et al. Patterned thermoresponsivemicrogel coatings for noninvasive processing ofadherent cells. Biomacromolecules. 2016;17(3): 1110–1116. DOI: https://doi.org/10.1021/acs.biomac.5b01728

Aiming at the problem of low detection accuracy caused by the different postures, different sizes, complex backgrounds and overlapping occlusions of contraband in the security checking process, a Matrix capsule network based on attention mechanism (MCAM) is designed by introducing attention mechanism into the capsule network. Firstly, a multi-feature-extraction (MFE) module is designed to solve the difficulty of detecting contraband with different sizes and complex backgrounds; then the Conv2d with Attention for ConvCap (CACC) is constructed in the convolutional layer of the capsule, and the weight information is computed in the channel dimensions of the feature map to give the contraband regions higher coefficients to enhance the contraband detection ability when the poses are different and the occlusion is severe; finally, a new capsule detection layer is designed by utilizing the pose matrix of the capsule to give full play to the detection ability of the matrix capsule network. The mAPs of the proposed model on SIXray, SIXray10, and SIXray100 are 85.10%, 64.05%, and 53.67%, respectively, which are 42.79%, 19.73%, and 9.41% higher than those of the original network, higher than those of the current mainstream detection networks, and the number of parameters of the network and the amount of computation are also lower.

Abstract Background Text Matching (TM) is a fundamental task of natural language processing widely used in many application systems such as information retrieval, automatic question answering, machine translation, dialogue system, reading comprehension, etc. In recent years, a large number of deep learning neural networks have been applied to TM, and have refreshed benchmarks of TM repeatedly. Among the deep learning neural networks, convolutional neural network (CNN) is one of the most popular networks, which suffers from difficulties in dealing with small samples and keeping relative structures of features. In this paper, we propose a novel deep learning architecture based on capsule network for TM, called CapsTM, where capsule network is a new type of neural network architecture proposed to address some of the short comings of CNN and shows great potential in many tasks. Methods CapsTM is a five-layer neural network, including an input layer, a representation layer, an aggregation layer, a capsule layer and a prediction layer. In CapsTM, two pieces of text are first individually converted into sequences of embeddings and are further transformed by a highway network in the input layer. Then, Bidirectional Long Short-Term Memory (BiLSTM) is used to represent each piece of text and attention-based interaction matrix is used to represent interactive information of the two pieces of text in the representation layer. Subsequently, the two kinds of representations are fused together by BiLSTM in the aggregation layer, and are further represented with capsules (vectors) in the capsule layer. Finally, the prediction layer is a connected network used for classification. CapsTM is an extension of ESIM by adding a capsule layer before the prediction layer. Results We construct a corpus of Chinese medical question matching, which contains 36,360 question pairs. This corpus is randomly split into three parts: a training set of 32,360 question pairs, a development set of 2000 question pairs and a test set of 2000 question pairs. On this corpus, we conduct a series of experiments to evaluate the proposed CapsTM and compare it with other state-of-the-art methods. CapsTM achieves the highest F-score of 0.8666. Conclusion The experimental results demonstrate that CapsTM is effective for Chinese medical question matching and outperforms other state-of-the-art methods for comparison.

Abstract Intrusion detection has become a research focus in internet information security, with deep learning algorithms playing a crucial role in its development. Typically, intrusion detection data are transformed into a two-dimensional matrix by segmenting, stacking and padding them with zeros for input into deep learning models. However, this method consumes computational resources and fails to consider the correlation between features. In this paper, we transform the data into images through visualization operations and propose an information entropy weighted scheme to optimize the collision element problem during the transformation process. This method enhances the correlation between pixel frame features, leading to approximately 2% improvement in accuracy of the classification model when using the generated image samples for detection in experiments. To address the issues of insensitivity to target feature locations and incomplete feature extraction in traditional neural networks, this paper introduces a new network model called CBAM-CapsNet, which combines the advantages of the lightweight Convolutional Block Attention Module and capsule networks. Experimental results on the UNSW-NB15 and IDS-2017 datasets demonstrate that the proposed model achieves accuracies of 92.94% and 99.72%, respectively. The F1 scores obtained are 91.83% and 99.56%, indicating a high level of detection.

Abstract Background Since protein-DNA interactions are highly essential to diverse biological events, accurately positioning the location of the DNA-binding residues is necessary. This biological issue, however, is currently a challenging task in the age of post-genomic where data on protein sequences have expanded very fast. In this study, we propose iProDNA-CapsNet – a new prediction model identifying protein-DNA binding residues using an ensemble of capsule neural networks (CapsNets) on position specific scoring matrix (PSMM) profiles. The use of CapsNets promises an innovative approach to determine the location of DNA-binding residues. In this study, the benchmark datasets introduced by Hu et al. (2017), i.e., PDNA-543 and PDNA-TEST, were used to train and evaluate the model, respectively. To fairly assess the model performance, comparative analysis between iProDNA-CapsNet and existing state-of-the-art methods was done. Results Under the decision threshold corresponding to false positive rate (FPR) ≈ 5%, the accuracy, sensitivity, precision, and Matthews’s correlation coefficient (MCC) of our model is increased by about 2.0%, 2.0%, 14.0%, and 5.0% with respect to TargetDNA (Hu et al., 2017) and 1.0%, 75.0%, 45.0%, and 77.0% with respect to BindN+ (Wang et al., 2010), respectively. With regards to other methods not reporting their threshold settings, iProDNA-CapsNet also shows a significant improvement in performance based on most of the evaluation metrics. Even with different patterns of change among the models, iProDNA-CapsNets remains to be the best model having top performance in most of the metrics, especially MCC which is boosted from about 8.0% to 220.0%. Conclusions According to all evaluation metrics under various decision thresholds, iProDNA-CapsNet shows better performance compared to the two current best models (BindN and TargetDNA). Our proposed approach also shows that CapsNet can potentially be used and adopted in other biological applications.

Self-Attention based Generative Adversarial Capsule Network optimized with Atomic orbital search algorithm based Sentiment Analysis is proposed in this manuscript for Online Product Recommendation (SFA-AGCN-AOSA-SA-OPR). Here, Collaborative filtering (CF) and product-product (P-P) similarity method is utilized for designing the new recommendation system. CF is employed for predicting the best shops and P-P similarity method is employed to predict the better product. Initially, the datas are gathered via Amazon Product recommendation dataset. After that, the datas are given to pre-processing. During pre-processing, Markov chain random field (MCRF) co-simulation is used to remove the unwanted content and filtering relevant text. The preprocessing output is fed to feature extraction. The features, like manufacturing date, Manufacturing price, discounts, offers, quality ratings, and suggestions or reviews are extracted using Gray level co-occurrence matrix (GLCM) window adaptive algorithm based feature extraction method. Finally, Self-Attention based Generative Adversarial Capsule Network (SFA-AGCN) categorizes the product recommendation as excellent, good, very good, bad, very bad. Atomic orbital search algorithm optimizes the SFA-AGCN weight parameters. The performance metrics, like accuracy, precision, sensitivity, recall, F-measure, mean absolute error (MAE), mean squared error (MSE), mean absolute percentage error (MAPE) is examined. The efficiency of the proposed method provides higher mean absolute percentage error 98.23%, 88.34%, 90.35% and 78.96% and lower Mean squared error 92.15%, 90.25%, 89.64% and 92.48% compared to the existing methods, such as sentiment analysis of online product reviews using DLMNN and future prediction of online product using IANFIS (DLMNN-IANFIS-SA-OPR), intelligent sentiment analysis approach using edge computing based deep learning technique (DCNN-SA-OPR), sentiment analysis for online product reviews in Chinese depending on sentiment lexicon and deep learning (CNN-BiGRU-SA-OPR) and sentiment analysis on product reviews depending on weighted word embedding and deep neural networks (CNN-LSTM-SA-OPR) respectively.

Abnormal loading of a joint's ligamentous capsule causes pain by activating the capsule's nociceptive afferent fibers, which reside in the capsule's collagenous matrix alongside fibroblast-like synoviocytes (FLS) and transmit pain to the dorsal root ganglia (DRG). This study integrated FLS into a DRG-collagen gel model to better mimic the anatomy and physiology of human joint capsules; using this new model, the effect of FLS on multiscale biomechanics and cell physiology under load was investigated. Primary FLS cells were co-cultured with DRGs at low or high concentrations, to simulate variable anatomical FLS densities, and failed in tension. Given their roles in collagen degradation and nociception, matrix-metalloproteinase (MMP-1) and neuronal expression of the neurotransmitter substance P were probed after gel failure. The amount of FLS did not alter (p > 0.3) the gel failure force, displacement, or stiffness. FLS doubled regional strains at both low (p < 0.01) and high (p = 0.01) concentrations. For high FLS, the collagen network showed more reorganization at failure (p < 0.01). Although total MMP-1 and neuronal substance P were the same regardless of FLS concentration before loading, protein expression of both increased after failure, but only in low FLS gels (p ≤ 0.02). The concentration-dependent effect of FLS on microstructure and cellular responses implies that capsule regions with different FLS densities experience variable microenvironments. This study presents a novel DRG-FLS co-culture collagen gel system that provides a platform for investigating the complex biomechanics and physiology of human joint capsules, and is the first relating DRG and FLS interactions between each other and their surrounding collagen network.

AbstractWith the rise of deep learning, spoken language understanding (SLU) for command-and-control applications such as a voice-controlled virtual assistant can offer reliable hands-free operation to physically disabled individuals. However, due to data scarcity, it is still a challenge to process dysarthric speech. Pre-training (part of) the SLU model with supervised automatic speech recognition (ASR) targets or with self-supervised learning (SSL) may help to overcome a lack of data, but no research has shown which pre-training strategy performs better for SLU on dysarthric speech and to which extent the SLU task benefits from knowledge transfer from pre-training with dysarthric acoustic tasks. This work aims to compare different mono- or cross-lingual pre-training (supervised and unsupervised) methodologies and quantitatively investigates the benefits of pre-training for SLU tasks on Dutch dysarthric speech. The designed SLU systems consist of a pre-trained speech representations encoder and a SLU decoder to map encoded features to intents. Four types of pre-trained encoders, a mono-lingual time-delay neural network (TDNN) acoustic model, a mono-lingual transformer acoustic model, a cross-lingual transformer acoustic model (Whisper), and a cross-lingual SSL Wav2Vec2.0 model (XLSR-53), are evaluated complemented with three types of SLU decoders: non-negative matrix factorization (NMF), capsule networks, and long short-term memory (LSTM) networks. The acoustic analysis of the four pre-trained encoders are tested on Dutch dysarthric home-automation data with word error rate (WER) results to investigate the correlations of the dysarthric acoustic task (ASR) and the semantic task (SLU). By introducing the intelligibility score (IS) as a metric of the impairment severity, this paper further quantitatively analyzes dysarthria-severity-dependent models for SLU tasks.

The eye is an immune-privileged site, with both vasculature and lymphatics absent from the central light path. Unique adaptations have made it possible for immune cells to be recruited to this region of the eye in response to ocular injuries and pathogenic insults. The induction of such immune responses is typically activated by tissue resident immune cells, considered the sentinels of the immune system. We discovered that, despite the absence of an embedded vasculature, the embryonic lens becomes populated by resident immune cells. The paths by which they travel to the lens during development were not known. However, our previous studies show that in response to corneal wounding immune cells travel to the lens from the vascular-rich ciliary body across the zonules that link these two tissues. We now examined whether the zonule fibers provide a path for immune cells to the embryonic lens, and the zonule-associated matrix molecules that could promote immune cell migration. The vitreous also was examined as a potential source of lens resident immune cells. This matrix-rich site in the posterior of the eye harbors hyalocytes, an immune cell type with macrophage-like properties. We found that both the zonules and the vitreous of the embryonic eye contained fibrillin-2-based networks and that migration-promoting matrix proteins like fibronectin and tenascin-C were linked to these fibrils. Immune cells were seen emerging from the ciliary body, migrating along the ciliary zonules to the lens, and invading through the lens capsule at its equator. This is just adjacent to where immune cells take up residence in the embryonic lens. In contrast, the immune cells of the vitreous were not detected in the region of the lens. These results strongly suggest that the ciliary zonules are a primary path of immune cell delivery to the developing lens.

ABSTRACT The marine bacterium and human pathogen Vibrio parahaemolyticus rapidly colonizes surfaces by using swarming motility and forming robust biofilms. Entering one of the two colonization programs, swarming motility or sessility, involves differential regulation of many genes, resulting in a dramatic shift in physiology and behavior. V. parahaemolyticus has evolved complex regulation to control these two processes that have opposing outcomes. One mechanism relies on the balance of the second messenger c-di-GMP, where high c-di-GMP favors biofilm formation. V. parahaemolyticus possesses four homologous regulators, the Scr transcription factors, that belong in a Vibrio-specific family of W[F/L/M][T/S]R motif transcriptional regulators, some members of which have been demonstrated to bind c-di-GMP. In this work, we explore the role of these Scr regulators in biofilm development. We show that each protein binds c-di-GMP, that this binding requires a critical R in the binding motif, and that the biofilm-relevant activities of CpsQ, CpsS, and ScrO but not ScrP are dependent upon second messenger binding. ScrO and CpsQ are the primary drivers of biofilm formation, as biofilms are eliminated when both of these regulators are absent. ScrO is most important for capsule expression. CpsQ is most important for RTX-matrix protein expression, although it contributes to capsule expression when c-di-GMP levels are high. Both regulators contribute to O-antigen ligase expression. ScrP works oppositely in a minor role to repress the ligase gene. CpsS plays a regulatory checkpointing role by negatively modulating expression of these biofilm-pertinent genes under fluctuating c-di-GMP conditions. Our work further elucidates the multifactorial network that contributes to biofilm development in V. parahaemolyticus. IMPORTANCE Vibrio parahaemolyticus can inhabit open ocean, chitinous shells, and the human gut. Such varied habitats and the transitions between them require adaptable regulatory networks controlling energetically expensive behaviors, including swarming motility and biofilm formation, which are promoted by low and high concentrations of the signaling molecule c-di-GMP, respectively. Here, we describe four homologous c-di-GMP-binding Scr transcription factors in V. parahaemolyticus. Members of this family of regulators are present in many vibrios, yet their numbers and the natures of their activities differ across species. Our work highlights the distinctive roles that these transcription factors play in dynamically controlling biofilm formation and architecture in V. parahaemolyticus and serves as a powerful example of regulatory network evolution and diversification.