Search

Abstract Urban greening plays a crucial role in maintaining environmental sustainability and enhancing people’s well-being. However, limited by the shortcomings of traditional methods, studying the heterogeneity and nonlinearity between environmental factors and green view index (GVI) still faces many challenges. To address the concerns of nonlinearity, spatial heterogeneity, and interpretability, an interpretable spatial machine learning framework incorporating the Geographically Weighted Random Forest (GWRF) model and the SHapley Additive exPlanation (Shap) model is proposed in this paper. In this paper, we combine multi-source big data, such as Baidu Street View data and remote sensing images, and utilize semantic segmentation models and geographic data processing techniques to study the global and local interpretation of the Beijing region with GVI as the key indicator. Our research results show that: (1) Within the Sixth Ring Road of Beijing, GVI shows significant spatial clustering phenomenon and positive correlation linkage, and at the same time exhibits significant spatial differences; (2) Among many environmental variables, the increase of green coverage rate has the most significant positive effect on GVI, while the increase of building density shows a strong negative correlation with GVI; (3) The performance of the GWRF model in predicting GVI is excellent and far exceeds that of comparison models.; (4) Whether it is the green coverage rate, urban built environment or socioeconomic factors, their influence on GVI shows non-linear characteristics and a certain threshold effect. With the help of these non-linear influences and explicit threshold effects, quantitative analyses of greening are provided, which can help to assist urban planners in making more scientific and rational decisions when allocating greening resources.

Abstract Background Breast milk contains various crucial nutrients and biologically active substances and is ideal for newborns. This study aimed to analyze the composition of breast milk from mothers of premature and full-term infants and its influences on the growth of infants. Methods Infant-mother dyads examined at our Hospital (March 2016 to May 2017) were included. Milk was collected at 0–1 month, 2–3 months, and 5–6 months and analyzed using a MIRIS human milk analyzer. Z-scores of weight-for-length (WLZ), weight-for-age (WAZ), and length-for-age (LAZ) were calculated. Results This study included full-term (> 37 weeks of gestation, n = 177) and premature (

A compact and robust optical excitation photoacoustic sensor with a self-integrated laser module excitation and an optimized differential resonator was developed to achieve high sensitivity and full linear range detection of carbon dioxide (CO2) based on dual modes of wavelength modulated photoacoustic spectroscopy (WMPAS) and resonant frequency tracking (RFT). The integrated laser module equipped with three lasers (a quantum cascade laser (QCL), a distributed feedback laser (DFB) and a He-Ne laser) working in a time-division multiplexing mode was used as an integrated set of spectroscopic sources for detection of the designated concentration levels of CO2. With the absorption photoacoustic mode, the WMPAS detection with the QCL and DFB sources was capable of CO2 detection at concentrations below 20 %, yielding a noise equivalent concentration (NEC) as low as 240 ppt and a normalized noise equivalent absorption coefficient (NNEA) of 4.755 × 10−10 W cm−1/√Hz, and dynamic range as great as 11 orders of magnitude. Higher concentration detection ranges (20 %-100 %) of CO2 were investigated using the RFT mode with an amplitude-stabilized He-Ne laser and a mechanical chopper. With the dual modes of WMPAS and RFT, the optical excitation sensor achieved full-range CO2 detection, with an R² ≥ 0.9993 and a response time of 5 seconds. The compact and full-range CO2 sensor combines the advantages of WMPAS and RFT and offers a solution for high sensitivity, linearity and full-range CO2 detection.

Crack propagation is an important cause of damage to rock bodies. In this study, uniaxial compression tests were conducted on specimens with rock-like mass containing fissures with different inclination angles to study the effect of crack angle on the crack evolution and fracture characteristics of rock bodies. The specimen surface deformation and internal response characteristics during fracture were analyzed via digital image correlation (DIC) and acoustic emission (AE) techniques. The results indicated that the AE characteristics of the fractured specimens exhibited a high degree of activity during the pore compaction and crack propagation stages. The prefabricated fissure configuration affected the stress state at the fissure tip, leading to differences in the crack evolution paths and rupture modes of fissure specimens with different angles. Under the uniaxial peak intensity, the relative position of the normalized global strain curve peak point gradually shifted from the specimen tip to the middle of the specimen as the crack angle increased, which corresponded to the shear damage-tension-shear mixed damage-tension damage modes of the specimen. The findings of this study indicate that normalized global strain curves can reflect the characteristics of crack evolution and provide a basis for the discrimination of fissured rock mass damage modes.

This paper addresses the problem of multi-source survey data sharing in power system engineering by proposing two improved methods: a survey data sharing method combined with differential privacy and a permission change method based on attribute encryption. The survey data sharing method integrated with differential privacy achieves effective cross-professional and cross-departmental data sharing while ensuring data security by introducing multi-discriminator architecture and dynamic noise adjustment. To reduce the computational and communication overhead when user permissions change during survey data sharing, the attribute encryption-based permission change method supports dynamic changes in user permissions. The effectiveness of the proposed methods has been validated through targeted experiments in different scenarios. The work in this thesis provides a new solution for dynamic sharing of survey data in power network engineering. It contributes to the digital transformation of power network projects.

RGB-T salient object detection (SOD) has made significant progress in recent years. However, most existing works are based on heavy models, which are not applicable to mobile devices. Additionally, there is still room for improvement in the design of cross-modal feature fusion and cross-level feature fusion. To address these issues, we propose a lightweight cross-modal information mutual reinforcement network for RGB-T SOD. Our network consists of a lightweight encoder, the cross-modal information mutual reinforcement (CMIMR) module, and the semantic-information-guided fusion (SIGF) module. To reduce the computational cost and the number of parameters, we employ the lightweight module in both the encoder and decoder. Furthermore, to fuse the complementary information between two-modal features, we design the CMIMR module to enhance the two-modal features. This module effectively refines the two-modal features by absorbing previous-level semantic information and inter-modal complementary information. In addition, to fuse the cross-level feature and detect multiscale salient objects, we design the SIGF module, which effectively suppresses the background noisy information in low-level features and extracts multiscale information. We conduct extensive experiments on three RGB-T datasets, and our method achieves competitive performance compared to the other 15 state-of-the-art methods.

Recent research on recommender systems has proved that by leveraging social network information, the quality of recommendations can be evidently improved. Traditional social recommendation models typically linearly combine social network information. For instance, matrix factorization based models linearly combine latent factors of relevant users and items. However, in practice, the multifaceted social relations are so complex that simple linear combination may not be able to reasonably organize such information for accurate social recommendation. On the other hand, existing deep learning based non-linear methods lack systematic modeling of user-item-friend relations. To handle these issues, we propose a novel, non-linear latent factor model for social recommendations leveraging Gaussian process. By introducing a social-aware covariance function, we organize individual users’ past feedback, as well as the associated social information (e.g., friends’ feedback to the same items) into a covariance matrix, which non-linearly and systematically learns the complex interactions among users, their interacted items and their friends’ opinions. A stochastic gradient descent based optimization algorithm is developed to fit the model. Extensive experiments conducted on three real-world datasets demonstrate that the proposed model outperforms the state-of-the-art social recommendation models and Gaussian process based models.

Abstract Cross modal retrieval can retrieve images through a text query and vice versa. In recent years, cross modal retrieval has attracted extensive attention. The purpose of most now available cross modal retrieval methods is to find a common subspace and maximize the different modal correlation. To generate specific representations consistent with cross modal tasks, this paper proposes a novel cross modal retrieval framework, which integrates feature learning and latent space embedding. In detail, we proposed a deep CNN and a shallow CNN to extract the feature of the samples. The deep CNN is used to extract the representation of images, and the shallow CNN uses a multi‐dimensional kernel to extract multi‐level semantic representation of text. Meanwhile, we enhance the semantic manifold by constructing cross modal ranking and within‐modal discriminant loss to improve the division of semantic representation. Moreover, the most representative samples are selected by using online sampling strategy, so that the approach can be implemented on a large‐scale data. This approach not only increases the discriminative ability among different categories, but also maximizes the relativity between different modalities. Experiments on three real word datasets show that the proposed method is superior to the popular methods.

This work presents a security analysis of the QUIC handshake protocol based on symbolic model checking. As a newly proposed secure transport protocol, the purpose of QUIC is to improve the transport performance of HTTPS traffic and enable rapid deployment and evolution of transport mechanisms. QUIC is currently in the IETF standardization process and will potentially carry a significant portion of Internet traffic in the emerging future. For a better understanding of the essential security properties, we have developed a formal model of the QUIC handshake protocol and perform a comprehensive formal security analysis by using two state-of-the-art model checking tools for cryptographic protocols, i.e., ProVeirf and Verifpal. Our analysis shows that ProVerif is generally more powerful than Verifpal in terms of verifying authentication properties. Moreover, both tools produce a counterexample to some security properties, which reveal a design flaw in the current protocol specification. Last but not least, we analyze the root causes of this counterexample and suggest a possible fix.

The performance of the salient object detection of strip surface defects has been promoted largely by deep learning based models. However, due to the complexity of strip surface defects, the existing models perform poorly in the challenging scenes such as noise disturbance, and low contrast between defect regions and background. Meanwhile, the detection results of existing models often suffer from coarse boundary details. Therefore, we propose a novel saliency model, namely an Edge-aware Multi-level Interactive Network, to detect the defects from the strip steel surface. Concretely, our model adopts the U-shape architecture where the two crucial points are the interactive feature integration and the edge-guided saliency fusion. Firstly, except the skip connection that combines the same stage of encoder and decoder, we deploy another connection, where the features from adjacent levels of encoder are transferred to the same stage of decoder. By this way, we are able to provide an effective fusion of multi-level deep features, yielding a well depiction for defects. Secondly, to give well-defined boundaries for prediction results, we add the edge extraction branch after each decoder block, where the progressive feature aggregation endows the edge with precise details and complete object cues. Meanwhile, together with the edge extraction branches, we deploy the saliency prediction branch at each decoder stage. After that, coupled with the fine edge information, we fuse all outputs of saliency prediction branches into the final saliency map, where the edge cue steers the saliency result to pay more attention to the boundary details. Following this way, we can provide a high-quality saliency map which can accurately locate and segment the defects. Extensive experiments are performed on the public dataset, and the results prove the effectiveness and robustness of our model which consistently outperforms the state-of-the-art models.

Alzheimer's disease (AD) has raised extensive concern in healthcare and academia as one of the most prevalent health threats to the elderly. Due to the irreversible nature of AD, early and accurate diagnoses are significant for effective prevention and treatment. However, diverse clinical symptoms and limited neuroimaging accuracy make diagnoses challenging. In this article, we built a brain network for each subject, which assembles several commonly used neuroimaging data simply and reasonably, including structural magnetic resonance imaging (MRI), diffusion-weighted imaging (DWI), and amyloid positron emission tomography (PET). Based on some existing research results, we applied statistical methods to analyze (i) the distinct affinity of AD burden on each brain region, (ii) the topological lateralization between left and right hemispheric sub-networks, and (iii) the asymmetry of the AD attacks on the left and right hemispheres. In the light of advances in graph convolutional networks for graph classifications and summarized characteristics of brain networks and AD pathologies, we proposed a regional brain fusion-graph convolutional network (RBF-GCN), which is constructed with an RBF framework mainly, including three sub-modules, namely, hemispheric network generation module, multichannel GCN module, and feature fusion module. In the multichannel GCN module, the improved GCN by our proposed adaptive native node attribute (ANNA) unit embeds within each channel independently. We not only fully verified the effectiveness of the RBF framework and ANNA unit but also achieved competitive results in multiple sets of AD stages' classification tasks using hundreds of experiments over the ADNI clinical dataset.

In this paper, a content adaptive coding method for plenoptic video is proposed to reduce both of the spatial and temporal redundancy. Based on the spatial correlations among subapertures, the plenoptic video is divided into two subaperture groups: the central view videos and the residual videos. Coding methods, multiview coding method based on spatial correlation(S-MVC) and multiview coding method based on temporal correlation(T-MVC), are adopted for residual videos according to correlation and residual energy analysis, while the central view videos are always encoded by S-MVC. Correlation analysis is performed first, and residual energy analysis is further performed if necessary. Coding the central view videos and the residual videos separately reduces bitstream, choosing different encoding methods for residual videos makes full use of the characteristics of different video content. Both the central view videos and the residual videos are compressed by MV-HEVC. The experimental results demonstrate that the proposed method outperforms a pseudo-sequence-based MVC method for plenoptic video by an average of 25.06% bitrate saving.

Recently, deep neural networks are widely used in recommendation systems, but most of them are used to process auxiliary information of recommendation systems such as items' descriptions and images. When it comes to how to learn a better interaction function to model the relation between user latent features and item latent features, which is the most critical step in a recommendation task, most works employ matrix factorization together with the inner product. However, it is sub-optimal because of ignoring many correlations between latent factors. As deep neural networks perform well in building more complex non-linear models, we employ deep neural networks to improve the collaborative filtering algorithm, solving the problem of implicit feedback which is the most common scene in real applications. Some recent work has contributed to finding better interaction function, but these functions are not exact enough to model comprehensive correlations among latent features. In this work, we propose the Convolutional Neural Networks based Deep Collaborative Filtering model (CNN-DCF) to solve the key problem in the recommendation system. Based on the outer product and deep neural networks, we develop a correlation extraction module that can learn high-order correlations between item latent features and user latent features. Extensive experiments on the public implicit feedback dataset Yelp show that the proposed CNN-DCF model brings significant improvements over the state-of-the-art methods.

Graph clustering is a fundamental task in data analysis and has attracted considerable attention in recommendation systems, mapping knowledge domain, and biological science. Because graph convolution is very effective in combining the feature information and topology information of graph data, some graph clustering methods based on graph convolution have achieved superior performance. However, current methods lack the consideration of structured information and the process of graph convolution. Specifically, most of existing methods ignore the implicit interaction between topology information and feature information, and the stacking of a small number of graph convolutional layers leads to insufficient learning of complex information. Inspired by graph convolutional network and auto-encoder, we propose a deep graph structured clustering network that applies a deep clustering method to graph structured data processing. Deep graph convolution is employed in the backbone network, and evaluates the result of each iteration with node feature and topology information. In order to optimize the network without supervision, a triple self-supervised module is designed to help update parameters for overall network. In our model, we exploit all information of the graph structured data and perform self-supervised learning. Furthermore, improved graph convolution layers significantly alleviate the problem of clustering performance degradation caused by over-smoothing. Our model is designed to perform on representative and indirect graph datasets, and experimental results demonstrate that our model achieves superior performance over state-of-the-art models.

With improvements in lubrication and material strength, the power transmitted by plastic gears has increased significantly. To develop high-performance transmission systems, it is necessary to gain deep insights into the dynamic characteristics of plastic gears. However, because plastics are viscoelastic materials, they do not obey Hooke’s law, which is the basis of traditional gear dynamic models. In this study, a refined dynamic model for an epoxy gear pair considering material viscoelasticity and extended tooth contact is established, and the differences in the dynamic responses between an epoxy and a steel gear pair are compared with respect to the dynamic meshing force and dynamic transmission error. The results show that: (1) the plastic gear can restrain the meshing impact, it has a generally lower dynamic meshing force than steel gear pair; (2) the position accuracy is the weak point of plastic gears, and this is significantly affected by the rotation speed; (3) the way to indirectly evaluate the dynamic meshing force by measuring the dynamic transmission error, which is often used for metal gears and is less effective for plastic gears.

Abstract Background The diagnosis of cow’s milk protein allergy(CMPA) may be easily misdiagnosed due to its lack of specific symptoms. Thus, experts have proposed the use of Cow’s milk-related symptom scores (CoMiSS) to predict CMPA. There has been no relevant report on the clinical application value of CoMiSS in Chinese children. This study aimed to evaluate the effect of CoMiSS in early identification of CMPA in Chinese infants. Methods We calculated CoMiSS for 38 infants with suspected CMPA diagnosed in the pediatric gastroenterologic clinic in our hospital. After 4 weeks of dietary elimination and symptomatic improvement, these infants returned to our hospital to undergo oral food challenge (OFC). The ROC curve was used to determine the sensitivity and specificity of CoMiSS and evaluate the effect of CoMiSS in early identification of CMPA in Chinese infants. We didn’t determine the CoMiSS of presumed healthy infants as control group. Results Of 38 infants who underwent OFC testing, the average CoMiSS of infants with positive OFC testing was 7.4 ± 2.3, while the average CoMiSS of infants with negative OFC testing was 4.1 ± 1.6, and there was a significant difference between two groups(F = 2.13, P

Online sequential extreme learning machine (OS-ELM) and its variants provide a promising way to solve data stream problems, but most of them do not take the timeliness of the problems into account, which may degrade the performance of the model. The main reason is that these algorithms are unable to adapt to the latest data accordingly when the distribution of the data stream changes. To mitigate this limitation, the forgetting factor is introduced into the relevant models, which is used to balance the relative importance of past data and new data when necessary. However, there is no efficient way to set the forgetting factor properly so far. In this paper, we have developed a novel updating strategy for setting the forgetting factor and proposed a dynamic forgetting factor based OS-ELM algorithm (DOS-ELM). In the sequential learning phase of DOS-ELM, the forgetting factor can be adjusted dynamically according to the change degree of the model accuracy in each learning epoch. This updating process does not require setting any parameters artificially and thus greatly improves the flexibility of the model. The experimental results on ten classification problems, five regression problems, one time-series problem show that DOS-ELM can deal well with both stationary and non-stationary data stream problems. In addition, we have extended DOS-ELM to an online deep model named ML-DOS-ELM, which can handle more complex tasks such as the face recognition problem and the handwritten digit recognition problem. Our experimental evaluations show that both DOS-ELM and ML-DOS-ELM can achieve higher prediction accuracy compared to the other similar algorithms.

Ultrasound is one of the most commonly used methods in the diagnosis and therapy of diseases due to its safety, deep penetration into tissue, and non-invasive nature. In the drug/gene delivery systems, ultrasound shows many advantages in terms of site-specific delivery and spatial release control of drugs/genes and attracts increasing attention. Microbubbles are the most well-known ultrasound-responsive delivery materials. Recently, nanobubbles, droplets, micelles, and nanoliposomes have been developed as novel carriers in this field. Herein, we review advances of novel ultrasound-responsive materials (nanobubbles, droplets, micelles and nanoliposomes) and discuss the challenges of ultrasound-responsive materials in delivery systems to boost the development of ultrasound-responsive materials as delivery carriers.

Idiopathic central precocious puberty (ICPP) is a relatively common condition in preadolescent girls, and its pathogenesis remains to be uncovered. A variety of studies have highlighted the association of gut microbiota (GM) with endocrine diseases, such as obesity, which is commonly associated with ICPP. However, the relationship between GM and ICPP remains unexplored. Feces samples were collected from 25 girls with ICPP (ICPP group) and 23 healthy girls (Control group). We applied 16S rDNA sequencing to compare the GM between two groups. The ICPP group had higher GM diversity and was enriched for several GM species, including Ruminococcus gnavus, Ruminococcus callidus, Ruminococcus bromii, Roseburia inulinivorans, Coprococcus eutactus, Clostridium leptum, and Clostridium lactatifermentans, which are known to be associated with obesity and are related to the production of short-chain fatty acids. Additionally, 36 candidate GM biomarkers for patients with ICPP screening were identified with high accuracy (AUC = 0.95, 95% CI 0.88 to 1). We observed that the GM of the ICPP group was enriched for the microbial functions of cell motility, signal transduction, and environmental adaptation. Positive correlations were also detected between Fusobacterium and follicle-stimulating hormone, and Gemmiger and luteinizing hormone. This study documents relationships between GM and ICPP, and the implication of these findings remains to be determined.

The quality of detected lane lines has a great influence on the driving decisions of unmanned vehicles. However, during the process of unmanned vehicle driving, the changes in the driving scene cause much trouble for lane detection algorithms. The unclear and occluded lane lines cannot be clearly detected by most existing lane detection models in many complex driving scenes, such as crowded scene, poor light condition, etc. In view of this, we propose a robust lane detection model using vertical spatial features and contextual driving information in complex driving scenes. The more effective use of contextual information and vertical spatial features enables the proposed model more robust detect unclear and occluded lane lines by two designed blocks: feature merging block and information exchange block. The feature merging block can provide increased contextual information to pass to the subsequent network, which enables the network to learn more feature details to help detect unclear lane lines. The information exchange block is a novel block that combines the advantages of spatial convolution and dilated convolution to enhance the process of information transfer between pixels. The addition of spatial information allows the network to better detect occluded lane lines. Experimental results show that our proposed model can detect lane lines more robustly and precisely than state-of-the-art models in a variety of complex driving scenarios.

The rapidly growing location-based social network (LBSN) has become a promising platform for studying users’ mobility patterns. Many online applications can be built based on such studies, among which, recommending locations is of particular interest. Previous studies have shown the importance of spatial and temporal influences on location recommendation; however, most existing approaches build a universal spatial–temporal model for all users despite the fact that users always demonstrate heterogeneous check-in behavior patterns. In order to realize truly personalized location recommendations, we propose a Gaussian process based model for each user to systematically and non-linearly combine temporal and spatial information to predict the user’s displacement from their currently checked-in location to the next one. The locations whose distances to the user’s current checked-in location are the closest to the predicted displacement are recommended. We also propose an enhancement to take into account category information of locations for semantic-aware recommendation. A unified recommendation framework called spatial–temporal–semantic (STS) is introduced to combine displacement prediction and the semantic-aware enhancement to provide final top-N recommendation. Extensive experiments over real datasets show that the proposed STS framework significantly outperforms the state-of-the-art location recommendation models in terms of precision and mean reciprocal rank (MRR).

We have previously shown that increases in blood–brain barrier permeability represent an important component of ischemia–reperfusion related brain injury in the fetus. Pro-inflammatory cytokines could contribute to these abnormalities in blood–brain barrier function. We have generated pharmacological quantities of mouse anti-ovine interleukin-1β monoclonal antibody and shown that this antibody has very high sensitivity and specificity for interleukin-1β protein. This antibody also neutralizes the effects of interleukin-1β protein in vitro. In the current study, we hypothesized that the neutralizing anti-interleukin-1β monoclonal antibody attenuates ischemia–reperfusion related fetal blood–brain barrier dysfunction. Instrumented ovine fetuses at 127 days of gestation were studied after 30 min of carotid occlusion and 24 h of reperfusion. Groups were sham operated placebo-control- (n = 5), ischemia-placebo- (n = 6), ischemia-anti-IL-1β antibody- (n = 7), and sham-control antibody- (n = 2) treated animals. Systemic infusions of placebo (0.154 M NaCl) or anti-interleukin-1β monoclonal antibody (5.1 ± 0.6 mg/kg) were given intravenously to the same sham or ischemic group of fetuses at 15 min and 4 h after ischemia. Concentrations of interleukin-1β protein and anti-interleukin-1β monoclonal antibody were measured by ELISA in fetal plasma, cerebrospinal fluid, and parietal cerebral cortex. Blood–brain barrier permeability was quantified using the blood-to-brain transfer constant (Ki) with α-aminoisobutyric acid in multiple brain regions. Interleukin-1β protein was also measured in parietal cerebral cortices and tight junction proteins in multiple brain regions by Western immunoblot. Cerebral cortical interleukin-1β protein increased (P

This paper proposed a multi-keyword ciphertext search, based on an improved-quality hierarchical clustering (MCS-IQHC) method. MCS-IQHC is a novel technique, which is tailored to work with encrypted data. It has improved search accuracy and can self-adapt when performing multi-keyword ciphertext searches on privacy-protected sensor network cloud platforms. Document vectors are first generated by combining the term frequency-inverse document frequency (TF-IDF) weight factor and the vector space model (VSM). The improved quality hierarchical clustering (IQHC) algorithm then generates document vectors, document indices, and cluster indices, which are encrypted via the k-nearest neighbor algorithm (KNN). MCS-IQHC then returns the top-k search result. A series of experiments proved that the proposed method had better searching efficiency and accuracy in high-privacy sensor cloud network environments, compared to other state-of-the-art methods.