Search

Graph Convolutional Networks (GCNs) are widely applied in classification tasks by aggregating the neighborhood information of each sample to output robust node embedding. However, conventional GCN methods do not update the graph during the training process so that their effectiveness is always influenced by the quality of the input graph. Moreover, previous GCN methods lack the interpretability to limit their real applications. In this paper, a novel personalized diagnosis technique is proposed for early Alzheimer’s Disease (AD) diagnosis via coupling interpretable feature learning with dynamic graph learning into the GCN architecture. Specifically, the module of interpretable feature learning selects informative features to provide interpretability for disease diagnosis and abandons redundant features to capture inherent correlation of data points. The module of dynamic graph learning adjusts the neighborhood relationship of every data point to output robust node embedding as well as the correlations of all data points to refine the classifier. The GCN module outputs diagnosis results based on the learned inherent graph structure. All three modules are jointly optimized to perform reliable disease diagnosis at an individual level. Experiments demonstrate that our method outputs competitive diagnosis performance as well as provide interpretability for personalized disease diagnosis.

With the trend of people expressing opinions and emotions via images online, increasing attention has been paid to affective analysis of visual content. Traditional image affective analysis mainly focuses on single-label classification, but an image usually evokes multiple emotions. To this end, emotion distribution learning is proposed to describe emotions more explicitly. However, most current studies ignore the ambiguity included in emotions and the elusive correlations with complex visual features. Considering that emotions evoked by images are delivered through various visual features, and each feature in the image may have multiple emotion attributes, this paper develops a novel model that extracts multiple features and proposes an enhanced fuzzy k-nearest neighbor (EFKNN) to calculate the fuzzy emotional memberships. Specifically, the multiple visual features are converted into fuzzy emotional memberships of each feature belonging to emotion classes, which can be regarded as an intermediate representation to bridge the affective gap. Then, the fuzzy emotional memberships are fed into a fully connected neural network to learn the relationships between the fuzzy memberships and image emotion distributions. To obtain the fuzzy memberships of test images, a novel sparse learning method is introduced by learning the combination coefficients of test images and training images. Extensive experimental results on several datasets verify the superiority of our proposed approach for emotion distribution learning of images.

The functional connectomic profile is one of the non-invasive imaging biomarkers in the computer-assisted diagnostic system for many neuro-diseases. However, the diagnostic power of functional connectivity is challenged by mixed frequency-specific neuronal oscillations in the brain, which makes the single Functional Connectivity Network (FCN) often underpowered to capture the disease-related functional patterns. To address this challenge, we propose a novel functional connectivity analysis framework to conduct joint feature learning and personalized disease diagnosis, in a semi-supervised manner, aiming at focusing on putative multi-band functional connectivity biomarkers from functional neuroimaging data. Specifically, we first decompose the Blood Oxygenation Level Dependent (BOLD) signals into multiple frequency bands by the discrete wavelet transform, and then cast the alignment of all fully-connected FCNs derived from multiple frequency bands into a parameter-free multi-band fusion model. The proposed fusion model fuses all fully-connected FCNs to obtain a sparsely-connected FCN (sparse FCN for short) for each individual subject, as well as lets each sparse FCN be close to its neighbored sparse FCNs and be far away from its furthest sparse FCNs. Furthermore, we employ the $\ell _{{1}}$ -SVM to conduct joint brain region selection and disease diagnosis. Finally, we evaluate the effectiveness of our proposed framework on various neuro-diseases, i.e., Fronto-Temporal Dementia (FTD), Obsessive-Compulsive Disorder (OCD), and Alzheimer’s Disease (AD), and the experimental results demonstrate that our framework shows more reasonable results, compared to state-of-the-art methods, in terms of classification performance and the selected brain regions. The source code can be visited by the url https://github.com/reynard-hu/mbbna .

The cartoon animation industry has developed into a huge industrial chain with a large potential market involving games, digital entertainment, and other industries. However, due to the coarse-grained classification of cartoon materials, cartoon animators can hardly find relevant materials during the process of creation. The polar emotions of cartoon materials are an important reference for creators as they can help them easily obtain the pictures they need. Some methods for obtaining the emotions of cartoon pictures have been proposed, but most of these focus on expression recognition. Meanwhile, other emotion recognition methods are not ideal for use as cartoon materials. We propose a deep learning-based method to classify the polar emotions of the cartoon pictures of the "Moe" drawing style. According to the expression feature of the cartoon characters of this drawing style, we recognize the facial expressions of cartoon characters and extract the scene and facial features of the cartoon images. Then, we correct the emotions of the pictures obtained by the expression recognition according to the scene features. Finally, we can obtain the polar emotions of corresponding picture. We designed a dataset and performed verification tests on it, achieving 81.9% experimental accuracy. The experimental results prove that our method is competitive.

The increasing tendency of people expressing opinions via images online has motivated the development of automatic assessment of sentiment from visual contents. Based on the observation that visual sentiment is conveyed through many visual elements in images, we put forward to tackle visual sentiment analysis under multiple instance learning (MIL) formulation. We propose a deep multiple clustered instance learning formulation, under which a deep multiple clustered instance learning network (DMCILN) is constructed for visual sentiment analysis. Specifically, the input image is converted into a bag of instances through visual instance generation module, which is composed of a pre-trained convolutional neural network (CNN) and two adaptation layers. Then, a fuzzy c-means routing algorithm is introduced for generating clustered instances as semantic mid-level representation to bridge the instance-to-bag gap. To explore the relationships between clustered instances and bags, we construct an attention based MIL pooling layer for representing bag features. A multi-head mechanism is integrated to form MIL ensembles, which enables to weigh the contribution of each clustered instance in different subspaces for generating more robust bag representation. Finally, we conduct extensive experiments on several datasets, and the experimental results verify the feasibility of our proposed approach for visual sentiment analysis.

Conventional spectral clustering is generally divided into two main steps, 1) constructing a reliable spectral representation by similarity matrix or Laplacian matrix; 2) performing K-means clustering on the spectral representation. In this paper, we propose a novel spectral clustering algorithm to improve clustering performance by separately improving these two steps in a same learning framework. Specifically, we first utilize two dimensionality reduction methods to learn the robust low-dimensional spectral representation, and a hypergraph structure to make the spectral representation keep the high-order relation of original data. Furthermore, a spectral rotation clustering method is embedded into the spectral representation learning model to conduct one-step clustering, which effectively reduces the deviation of clustering and obtains a reliable clustering performance. Besides, an effective optimization algorithm is further proposed to solve the objective problem to have a fast convergence. Experimental analysis on real data sets showed that our proposed clustering method outperformed the classical and state-of-the-art spectral clustering methods in terms of four frequently-used clustering metrics.

Previous feature selection methods equivalently consider the samples to select important features. However, the samples are often diverse. For example, the outliers should have small or even zero weights while the important samples should have large weights. In this paper, we add a self-paced regularization in the sparse feature selection model to reduce the impact of outliers for conducting feature selection. Specifically, the proposed method automatically selects a sample subset which includes the most important samples to build an initial feature selection model, whose generalization ability is then improved by involving other important samples until a robust and generalized feature selection model has been established or all the samples have been used. Experimental results on eight real datasets show that the proposed method outperforms the comparison methods.

In social media, images and texts are used to convey individuals’ attitudes and feelings; thus, social media has become an indispensable part of people’s lives. To understand social behavior and provide better recommendations, sentiment analysis on social media is helpful. One sentiment analysis task is polarity prediction. Although current research on visual or textual sentiment analysis has achieved quite good progress, multimodal and cross-modal analysis combining visual and textual correlation is still in the exploration stage. To capture a semantic connection between images and captions, this paper proposes a cross-modal approach that considers both images and captions in classifying image sentiment polarity. This method transfers the correlation between textual content to images. First, the image and its corresponding caption are sent into an inner-class mapping model, where they are transformed into vectors in Hilbert space to obtain their labels by calculating the inner-class maximum mean discrepancy (MMD). Then, a class-aware sentence representation (CASR) model assigns the distributed representation to the labels with a class-aware attention-based gated recurrent unit (GRU). Finally, an inner-class dependency LSTM (IDLSTM) classifies the sentiment polarity. Experiments carried out on the Getty Images dataset and Twitter 1269 dataset demonstrate the effectiveness of our approach. Moreover, extensive experimental results show that our model outperforms baseline solutions.

Plants fine-tune the growth-defense trade-off to survive when facing pathogens. Meanwhile, plant-associated microbes, such as the endophytes inside plant tissues, can benefit plant growth and stress resilience. However, the mechanisms for the beneficial microbes to increase stress resistance with little yield penalty in host plants remain poorly understood. In the present study, we report that endophytic Streptomyces hygroscopicus OsiSh-2 can form a sophisticated interaction with host rice, maintaining cellular homeostasis under pathogen-infection stress, and optimize plant growth and disease resistance in rice. Four-year field trials consistently showed that OsiSh-2 could boost host resistance to rice blast pathogen Magnaporthe oryzae while still maintaining a high yield. The integration of the proteomic, physiological, and transcriptional profiling analysis revealed that OsiSh-2 induced rice defense priming and controlled the expression of energy-consuming defense-related proteins, thus increasing the defense capability with the minimized costs of plant immunity. Meanwhile, OsiSh-2 improved the chloroplast development and optimally maintained the expression of proteins related to plant growth under pathogen stress, thus promoting the crop yield. Our results provided a representative example of an endophyte-mediated modulation of disease resistance and fitness in the host plant. The multilayer effects of OsiSh-2 implicate a promising future of using endophytic actinobacteria for disease control and crop yield promotion. IMPORTANCE Under disease stress, activation of defense response in plants often comes with the cost of a reduction in growth and yield, which is referred as the growth-defense trade-off. The microorganisms which can be recruited by plants to mitigate the growth-defense trade-off are of great value in crop breeding. Here, we reported a rice endophytic actinomycetes Streptomyces hygroscopicus OsiSh-2, which can improve host performances on resistance to rice blast while still sustaining high yield in the 4-year field trials. The proteomic, physiological, and transcriptional profiling data offer insights into the molecular basis underlying the balancing between defense and growth in OsiSh-2-rice symbiont. The findings provide an example for the endophyte-mediated modulation of growth-defense trade-offs in plants and indicated the promising application of endophytic actinobacterial strains in agriculture to breed “microbe-optimized crops.”

Support Vector Machine (SVM) has been widely applied in real application due to its efficient performance in the classification task so that a large number of SVM methods have been proposed. In this paper, we present a novel SVM method by taking the dynamic graph learning and the self-paced learning into account. To do this, we propose utilizing self-paced learning to assign important samples with large weights, learning a transformation matrix for conducting feature selection to remove redundant features, and learning a graph matrix from the low-dimensional data of original data to preserve the data structure. As a consequence, both the important samples and the useful features are used to select support vectors in the SVM framework. Experimental analysis on four synthetic and sixteen benchmark data sets demonstrated that our method outperformed state-of-the-art methods in terms of both binary classification and multi-class classification tasks.

Server workload in the form of cloud-end clusters is a key factor in server maintenance and task scheduling. How to balance and optimize hardware resources and computation resources should thus receive more attention. However, we have observed that the disordered execution of running application and batching seriously cuts down the efficiency of the server. To improve the workload prediction accuracy, this paper proposes an approach using the long short-term memory (LSTM) encoder-decoder network with attention mechanism. First, the approach extracts the sequential and contextual features of the historical workload data through the encoder network. Second, the model integrates the attention mechanism into the decoder network, through which the prediction for batch workloads can be carried out. Third, experiments carried out on Alibaba and Dinda workload traces dataset demonstrate that our method achieves state-of-the-art performance in mixed workload prediction in cloud computing environment. Furthermore, we also propose a scroll prediction method, which splits a long prediction sequence into several small sequences to monitor and control prediction accuracy. This work helps to dynamically guide the configuration for workload balancing.

The hardware-in-the-loop (HIL) employs the real controller and true signal transmission, which is efficient to test control algorithms for the situation that is inconvenient to conduct experiments....

The sentiment analysis of microblog text has always been a challenging research field due to the limited and complex contextual information. However, most of the existing sentiment analysis methods for microblogs focus on classifying the polarity of emotional keywords while ignoring the transition or progressive impact of words in different positions in the Chinese syntactic structure on global sentiment, as well as the utilization of emojis. To this end, we propose the emotion-semantic-enhanced bidirectional long short-term memory (BiLSTM) network with the multi-head attention mechanism model (EBILSTM-MH) for sentiment analysis. This model uses BiLSTM to learn feature representation of input texts, given the word embedding. Subsequently, the attention mechanism is used to assign the attentive weights of each words to the sentiment analysis based on the impact of emojis. The attentive weights can be combined with the output of the hidden layer to obtain the feature representation of posts. Finally, the sentiment polarity of microblog can be obtained through the dense connection layer. The experimental results show the feasibility of our proposed model on microblog sentiment analysis when compared with other baseline models.

Spectral clustering is an important component of clustering method, via tightly relying on the affinity matrix. However, conventional spectral clustering methods 1). equally treat each data point, so that easily affected by the outliers; 2). are sensitive to the initialization; 3). need to specify the number of cluster. To conquer these problems, we have proposed a novel spectral clustering algorithm, via employing an affinity matrix learning to learn an intrinsic affinity matrix, using the local PCA to resolve the intersections; and further taking advantage of a robust clustering that is insensitive to initialization to automatically generate clusters without an input of number of cluster. Experimental results on both artificial and real high-dimensional datasets have exhibited our proposed method outperforms the clustering methods under comparison in term of four clustering metrics.

Previous Spectral Feature Selection (SFS) methods output promising feature selection results in many real-world applications, which deeply depend on the preservation of the local or global structures of the data via learning a graph matrix. However, current SFS methods 1) learn the graph matrix in the original data which may contain a number of noise to affect the results of feature selection, 2) conduct the learning of a low-dimensional feature space and the graph matrix individually, thus hard achieve the optimal results of feature selection even though both of these two steps achieve their individual optimization, and 3) consider either the local or global structure of data to difficult provide complementary information for feature selection. To address the above issues, this paper proposes a novel supervised feature selection algorithm to simultaneously preserve the local structure (via adaptive structure learning in a low-dimensional feature space of the original data) and the global structure (via a low-rank constraint) of the data. Moreover, we also propose a new optimization method to fast optimize the resulting objective function. We finally verify the proposed method on eight real-word and benchmark datasets, by comparing with the state-of-the-art feature selection methods, and experimental results show that our proposed method achieves competitive results in term of classification performance.

Purpose To evaluate the biodistribution, metabolism, and pharmacokinetics of a new type I collagen-targeted magnetic resonance (MR) probe, CM-101, and to assess its ability to help quantify liver fibrosis in animal models. Materials and Methods Biodistribution, pharmacokinetics, and stability of CM-101 in rats were measured with mass spectrometry. Bile duct-ligated (BDL) and sham-treated rats were imaged 19 days after the procedure by using a 1.5-T clinical MR imaging unit. Mice were treated with carbon tetrachloride (CCl

This paper proposes a new unsupervised spectral feature selection method to preserve both the local and global structure of the features as well as the samples. Specifically, our method uses the self-expressiveness of the features to represent each feature by other features for preserving the local structure of features, and a low-rank constraint on the weight matrix to preserve the global structure among samples as well as features. Our method also proposes to learn the graph matrix measuring the similarity of samples for preserving the local structure among samples. Furthermore, we propose a new optimization algorithm to the resulting objective function, which iteratively updates the graph matrix and the intrinsic space so that collaboratively improving each of them. Experimental analysis on 12 benchmark datasets showed that the proposed method outperformed the state-of-the-art feature selection methods in terms of classification performance.

Previous spectral feature selection methods generate the similarity graph via ignoring the negative effect of noise and redundancy of the original feature space, and ignoring the association between graph matrix learning and feature selection, so that easily producing suboptimal results. To address these issues, this paper joints graph learning and feature selection in a framework to obtain optimal selected performance. More specifically, we use the least square loss function and an l 2,1-norm regularization to remove the effect of noisy and redundancy features, and use the resulting local correlations among the features to dynamically learn a graph matrix from a low-dimensional space of original data. Experimental results on real data sets show that our method outperforms the state-of-the-art feature selection methods for classification tasks.

Dimension reduction methods always catch many attentions, because it could effectively solve the curse of dimensionality problem. In this paper, we propose an unsupervised feature selection method which could efficiently select a subset of informative features from unlabeled data. We integrate the low-rank constraint, hypergraph theory, and the self-representation property of features in a unified framework to conduct unsupervised feature selection. Specifically, we represent each feature by other features to conduct unsupervised feature selection via the feature-level self-representation property. We then embed a low-rank constraint to consider the relations among features. Moreover, a hypergarph regularizer is utilized to consider both the high-order relations and the local structure of the data. This enables the proposed model to take into account both the global structure of the data (via the low-rank constraint) and the local structure of the data (via the hypgergraph regularizer). We use an 2, p-norm regularizer to satisfy the constraints. Therefore, the proposed model is more robust to the previous models due to achieving better feature selection model. Experimental results on benchmark datasets showed that the proposed method effectively selected the most informative features by removing the adverse effect of irrelevant/redundant features, compared to the state-of-the-art methods.

In order to select informative features from a high-dimensional multi-view dataset, we have proposed a feature selection method that simultaneously embedding the low-rank constraint, sparse representation, global and local structure learning into a unified framework. Firstly, we utilize the conventional regression function (i.e. the least square loss function) to form a novel regression framework by introducing a low-rank constraint and a relaxation term. And then we employ an l21-norm regularization term to filter out the redundant and irrelative features. Furthermore, we utilize a hypergraph based regularization term rather than the simple graph to construct a Laplacian matrix that will be used in enhancing the inherent association of data. Besides, we proposed a novel optimization algorithm to solve the objective function. Finally, we feed the reduced data got by the proposed feature selection method into Support Vector Machines (SVM) in term of classification accuracy. The experimental results showed that the proposed method achieved the best classification performance, compared with the state-of-the-art feature selection methods on real multi-view dataset.

Insulin resistance and diabetes can develop spontaneously with obesity and aging in rhesus monkeys, highly similar to the natural history of obesity, insulin resistance, and progression to type 2 diabetes in humans. The current studies in obese rhesus were undertaken to assess hepatic and adipose contributions to systemic insulin resistance—currently, a gap in our knowledge—and to benchmark the responses to pioglitazone (PIO). A two-step hyperinsulinemic-euglycemic clamp, with tracer-based glucose flux estimates, was used to measure insulin resistance, and in an intervention study was repeated following 6 wk of PIO treatment (3 mg/kg). Compared with lean healthy rhesus, obese rhesus has a 60% reduction of glucose utilization during a high insulin infusion and markedly impaired suppression of lipolysis, which was evident at both low and high insulin infusion. However, obese dysmetabolic rhesus manifests only mild hepatic insulin resistance. Six-week PIO treatment significantly improved skeletal muscle and adipose insulin resistance (by ~50%). These studies strengthen the concept that insulin resistance in obese rhesus closely resembles human insulin resistance and indicate the value of obese rhesus for appraising new insulin-sensitizing therapeutics.

Pancreatic cancer is one of the most aggressive cancers with poor prognosis and a low 5-year survival rate. The family of P21-activated kinases (PAKs) appears to modulate many signaling pathways that contribute to pancreatic carcinogenesis. In this work, we demonstrated that PAK1 is a critical regulator in pancreatic cancer cell growth. PAK1-targeted inhibition is therefore a new potential therapeutic strategy for pancreatic cancer. Our small molecule screening identified a relatively specific PAK1-targeted inhibitor, CP734. Pharmacological and biochemical studies indicated that CP734 targets residue V342 of PAK1 to inhibit its ATPase activity. Further in vitro and in vivo studies elucidated that CP734 suppresses pancreatic tumor growth through depleting PAK1 kinase activity and its downstream signaling pathways. Little toxicity of CP734 was observed in murine models. Combined with gemcitabine or 5-fluorouracil, CP734 also showed synergistic effects on the anti-proliferation of pancreatic cancer cells. All these favorable results indicated that CP734 is a new potential therapeutic candidate for pancreatic cancer., Graphical abstract CP734 was identified as a relatively selective inhibitor of PAK1 by structure-based virtual screening. It significantly suppressed pancreatic cancer cell growth both in vitro and in xenograft models. This study will provide molecular insights into the development of novel PAK1 kinase inhibitors, which may benefit anti-cancer drug candidate development.Image 1

Metabolic dysregulation and mitochondrial dysfunction are important features of acute and chronic tissue injury across species, and human genetics and preclinical data suggest that the master metabolic regulator 5'-adenosine monophosphate-activated protein kinase (AMPK) may be an effective therapeutic target for chronic kidney disease (CKD). We have recently disclosed a pan-AMPK activator, MK-8722, that was shown to have beneficial effects in preclinical models. In this study we investigated the effects of MK-8722 in a progressive rat model of diabetic nephropathy to determine whether activation of AMPK would be of therapeutic benefit. We found that MK-8722 administration in a therapeutic paradigm is profoundly renoprotective, as demonstrated by a reduction in proteinuria (63% decrease in MK-8722 10 mg/kg per day compared with vehicle group) and a significant improvement in glomerular filtration rate (779 and 430 μl/min per gram kidney weight in MK-8722 10 mg/kg per day and vehicle group, respectively), as well as improvements in kidney fibrosis. We provide evidence that the therapeutic effects of MK-8722 may be mediated by modulation of renal mitochondrial quality control as well by attenuating fibrotic and lipotoxic mechanisms in kidney cells. MK-8722 (10 mg/kg per day compared with vehicle group) achieved modest blood pressure reduction (10 mmHg lower for mean blood pressure) and significant metabolic improvements (decreased plasma glucose, triglyceride, and body weight) that could contribute to renoprotection. These data further validate the concept that targeting metabolic dysregulation in CKD could be a potential therapeutic approach. SIGNIFICANCE STATEMENT: We demonstrate in the present study that the pharmacological activation of AMPK using a small-molecule agent provided renoprotection and improved systemic and cellular metabolism. We further indicate that modulation of renal mitochondrial quality control probably contributed to renoprotection and was distinct from the effects of enalapril. Our findings suggest that improving renal mitochondrial biogenesis and function and attenuating fibrosis and lipotoxicity by targeting key metabolic nodes could be a potential therapeutic approach in management of CKD that could complement the current standard of care.

The generative adversarial networks (GANs) have demonstrated the ability to synthesize realistic images. However, there are few researches applying GANs into the field of food image synthesis. In this paper, we propose an extension to GANs for generating more realistic food dish images with rich detail, which adds a food condition that contains taste and other information. That makes the model generate images with rich details. To improve the quality of the generated image, the taste information condition is added to each stage of the generator and discriminator. First, the model learns embedding conditions of food information, including ingredients, cooking methods, tastes and cuisines. Secondly, the training model grows progressively, and the model learns details increasingly during the training process, which allows the model to generate images with rich details. To demonstrate the effectiveness of our proposed model, we collect a dataset called Food-121, which includes the names of the food, ingredients, cooking methods, tastes, and cuisines. The results of experiment show that our model can produce complex details of food dish image and obtain high inception score on the Food-121 dataset compared with other models.

Precise agricultural predictions of climate change effects on crop water productivity are essential to ensure food security and alleviate water scarcity. In this regard, the present study provides an overview of the future impacts of climate change on the irrigation of agricultural products such as rice, millet, maize, cassava, sorghum, and sugar cane. These crops are some of the most-consumed foodstuffs in countries of the Niger River basin. This study is realized throughout 2020 to 2080, and three Global Climate Models (GCMs) (CSIRO, MIROC5, and ECHAM. MPI-ESM-LR) have been used. The GCMs data have been provided by the IPCC5 database. The irrigation water requirement for each crop was calculated using Smith’s CROPWAT approach. The Penman–Monteith equation recommended by the FAO was used to calculate the potential evapotranspiration. The inter-annual results of the IWR, according to the set of models selected, illustrate that the largest quantities of water used for irrigation are generally observed between January and March, and the lowest quantities are the most often seen between July and September. The majority of models also illustrate a peak in the IWR between March and April. Sorghum and millet are the crops consuming the least amount of water for irrigation, followed by cassava, then rice and corn, and finally sugar cane. The most significant IWRs, which have been predicted, will be between 16.3 mm/day (MIROC5 model, RCP 4.5) and 45.9 mm/day (CSIRO model, RCP 4.5), particularly in Mali, Niger, Algeria, and rarely in Burkina-Faso (CSIRO model, RCP4.5 and 8.5). The lowest IWRs predicted by the models will be from 1.29 mm/day (MIROC5 model, RCP 4.5) to 33.4 mm/day (CSIRO model, RCP 4.5), they will be observed according to the models in Guinea, southern Mali, Ivory Coast, center and southern Nigeria, and Cameroon. However, models predict sugarcane to be the plant with the highest IWR, between 0.25 mm/day (Benin in 2020–2040) and 25.66 mm/day (Chad in 2060–2080). According to the models’ predictions, millet is the crop with the most IWR, between 0.20 mm/day (Benin from 2020 to 2060) and 19.37 mm/day (Chad in 2060–2080). With the results of this study, the countries belonging to the Niger River basin can put in place robust policies in the water resources and agriculture sectors, thus ensuring food security and high-quality production of staple crops, and avoiding water scarcity while facing the negative impacts of climate change.

Social media has become indispensable to people’s lives, where they can share their views and emotion with images and texts. Analyzing social images for sentiment prediction can help understand human social behavior and provide better recommendation results. Most current researches on image sentiment analysis have achieved quite good progress, which ignores the semantic correlation between an image and its corresponding descriptive sentences (caption). To capture the complementary multimodal information for joint sentiment classification, in this paper, we propose a novel cross-modal Semantic Content Correlation (SCC) method based on deep matching and hierarchical networks, which bridges the correlation between images and captions. Specifically, pre-trained convolutional neural networks (CNNs) are leveraged to encode the visual sub-regions contents, and a GloVe is employed to embed the textual semantic. Relying on visual contents and textual semantic, a joint attention network is proposed to learn the content correlation of the image and its caption, which is then exported as an image–text pair. To exploit the dependence of visual contents on textual semantic in caption effectively, the caption is processed by a Class-Aware Sentence Representation (CASR) network with a class dictionary, and a fully connected layer concatenates the outputs of CASR into a class-aware vector. Finally, the class-aware distributed vector is fed into an Inner-class Dependency Long Short-Term Memory network (IDLSTM) with the image–text pair as a query to further capture the cross-modal non-linear correlations for sentiment prediction. The performance of extensive experiments conducted on three datasets verifies the effectiveness of the model SCC.

The class distribution of imbalanced data sets is skewed in practical application. As traditional clustering methods mainly are designed for improving the overall learning performance, the majority class usually tends to be clustered and the minority class which is more valuable maybe ignored. Moreover, existing clustering methods can be limited for the performance of imbalanced and high-dimensional domains. In this paper, we present one-step spectral rotation clustering for imbalanced high-dimensional data (OSRCIH) by integrating self-paced learning and spectral rotation clustering in a unified learning framework, where sample selection and dimensionality reduction are simultaneously considered with mutual and iterative update. Specifically, the imbalance problem is considered by selecting the same number of training samples from each intrinsic group of the training data, where the sample-weight vector is obtained by self-paced learning. Moreover, dimensionality reduction is conducted by combining subspace learning and feature selection. Experimental analysis on synthetic datasets and real datasets showed that OSRCIH could recognize and enhance the weight of important samples and features so as to avoid the clustering method in favor of the majority class and to improve effectively the clustering performance.

Groundwater-dependent ecosystems (GDEs) in the lower reaches of arid inland river basins of Northwest China have degraded because of a decrease in water supply caused by excessive socio-economic development in the middle reaches of the river. The degraded GDEs have resulted in serious eco-environmental problems, such as drying of terminal lakes, river channel shifts, declines in riparian forest, desert activation, intensification of desert dust storms, and salinization of secondary soils. These problems have further hindered local and regional efforts at sustainable development. An ecosystem conservation and restoration project was initiated in 2000 in order to deal with the eco-environmental problems. The project has been underway for 15 years, but problems still exist. Using the Heihe River and Tarim River basins as examples, this paper explains the ecological–environmental problems and their causes, describes with the status of ecosystem restoration and conservation, analyzes the problems in the ongoing ecosystem conservation and restoration, and suggests strategies for achieving sustainable ecosystem restoration in the lower reaches of the arid inland river basins of Northwest China.

With the increasing of Web services on Internet, services composition requests considering not only the static information compatibility, such as interface grammar and semantic, but also the structural behavior compatibility, i.e., control flow and communication protocol. However, due to the uncertainty of Internet, application failures occurred in service-based software will cause kinds of economic losses. This stochastic phenomenon has been impacted on the structural behavior which makes services composition should take the probabilistic compatibility into account. The probabilistic compatibility needs to compute the degree of compatibility for the effective performance evaluation, which has been an important issue during services collaboration. In this paper, the probabilistic interface automaton is proposed to formalize service behaviors. Then, the synchronized product model is used to the qualitative checking for verifying whether they can interact with each other or not. Third, the probabilistic compatibilities of composite service and component service are discussed for the quantitative computing. Our method provides a reference to generate the correctness and reliable services composition.

This paper proposes a graph feature selection method for dementia diagnosis, by adding the information inherent in the observations into a sparse multi-task learning framework. Specifically, this paper first defines two relations (i.e., the feature-feature relation and the sample-sample relation, respectively) based on the prior knowledge in the data. The feature-feature selection enforces the similarity relationship between features to be preserved in the coefficient matrix while the sample-sample relation is designed to preserve the relation between samples invariant in the predicted space. Then we embed these two kinds of relations into a multi-task learning framework (i.e., a least square loss function plus an l2-norm regularization term) to conduct feature selection. Furthermore, we feed the reduced data into Support Vector Machine (SVM) for conducting the identification of Alzheimer's Disease (AD). Finally, the experimental results on a subset of the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset showed the effectiveness of the proposed method in terms of classification accuracy, by comparing with the state-of-the-art methods, including k Nearest Neighbor (kNN), ridge regression, SVM, and so on. Refereed/Peer-reviewed

The development of multimodal media compensates for the lack of information expression in a single modality and thus gradually becomes the main carrier of sentiment. In this situation, automatic assessment for sentiment information in multimodal contents is of increasing importance for many applications. To achieve this, we propose a joint sentiment part topic regression model (JSP) based on latent Dirichlet allocation (LDA), with a sentiment part, which effectively utilizes the complementary information between the modalities and strengthens the relationship between the sentiment layer and multimodal content. Specifically, a linear regression module is developed to share implicit variables between image&ndash, text pairs, so that one modality can predict the other. Moreover, a sentiment label layer is added to model the relationship between sentiment distribution parameters and multimodal contents. Experimental results on several datasets verify the feasibility of our proposed approach for multimodal sentiment analysis.

Previous spectral clustering methods sequentially conduct three steps, i.e., similarity matrix learning from original data, spectral representation learning, and K-means clustering on spectral representation, respectively, to difficultly output robust clustering result even though each of three steps achieves individual optimization. The reason is that each goal of former two steps is not focused on achieving optimal clustering result. Moreover, original data usually contains noise to affect the clustering result, as well as has high-dimensional representation to easily result in the curse of dimensionality. In this paper, we propose a deep spectral clustering method which embeds four parts (i.e., similarity matrix learning, spectral representation learning, optimized K-means clustering, and transformation matrix learning) in a unified framework with the following advantages: 1) similarity matrix is obtained from the low-dimensional feature space of original data where the influence of both noise and high-dimensional data are considered; 2) optimized K-means clustering rotates original result of K-means clustering to search optimized clustering hyperplane which partitions data points into clusters; and 3) each of four parts is iteratively updated so that the clustering result is obtained based on the feedback of other three parts. As a result, our proposed framework develops a two-task deep clustering model with linear activation functions to output effective clustering result. Experimental results on real data sets show the effectiveness of our method in terms of four clustering evaluation metrics, compared to state-of-the-art clustering methods.

Unsupervised Spectral Feature Selection (USFS) methods may output interpretable and discriminative results by integrating a Laplacian regularizer into the feature selection framework to keep the similarity of the training samples. To do this, USFS methods usually construct the Laplacian matrix using either a general graph or a Hyper-graph on the original data. Since the construction of the Laplacian matrix is separated from the process of feature selection and the original data usually contain noise, the performance of feature selection will be influenced. Based on previous observation that a hyper-graph measuring the relationship among more than two samples is able to characterize more complex (i.e. high-order) relationship of the samples than a general graph measuring the relationship between two samples, in this paper, we propose a novel feature selection method by dynamically learning a hyper-graph based Laplacian matrix to measure the relationship among the samples for selecting the features. Experimental results on real datasets showed that our proposed method outperforms the state-of-the-art feature selection methods in terms of both clustering and segmentation performance.

The prevalence that people share their opinions on the products and services in their daily lives on the Internet has generated a large quantity of comment data, which contain great business value. As for comment sentences, they often contain several comment aspects and the sentiment on these aspects are different, which makes it meaningless to give an overall sentiment polarity of the sentence. In this paper, we introduce Attention-based Aspect-level Recurrent Convolutional Neural Network (AARCNN) to analyze the remarks at aspect-level. The model integrates attention mechanism and target information analysis, which enables the model to concentrate on the important parts of the sentence and to make full use of the target information. The model uses bidirectional LSTM (Bi-LSTM) to build the memory of the sentence, and then CNN is applied to extracting attention from memory to get the attentive sentence representation. The model uses aspect embedding to analyze the target information of the representation and finally the model outputs the sentiment polarity through a softmax layer. The model was tested on multi-language datasets, and demonstrated that it has better performance than conventional deep learning methods.

There are several features in the content of microblog, including brief, flexible, too colloquial, too much noise caused by informality terms (abbreviations, cyber words), etc. These features are new challenges for named entity recognition. In this paper, a modified method based on cascade conditional random field (CRF) is presented to address the challenge. First of all, we preprocessed the captured microblog text to basic corpus, then we manually built common name, toponym and organization database. A sliding window was set to extract the world feature, sentence feature and context feature of the text, then we established atomic templates and fabricated templates to train cascaded CRF model. Finally, an active learning algorithm based on minimum confidence was adopted to enhance the training effect of the model. This method has been verified in People's Daily labeled corpus and microblog corpus. Experimental data show that our method has achieved outstanding experiment effect compared with the traditional CRF method, the precision, recall and F-value are all improved.

Feature selection is an indispensable preprocessing procedure for high-dimensional data analysis,but previous feature selection methods usually ignore sample diversity (i.e., every sample has individual contribution for the model construction) andhave limited ability to deal with incomplete datasets where a part of training samples have unobserved data. To address these issues, in this paper, we firstly propose a robust feature selectionframework to relieve the influence of outliers, andthen introduce an indicator matrix to avoid unobserved data to take participation in numerical computation of feature selection so that both our proposed feature selection framework and exiting feature selection frameworks are available to conductfeature selection on incomplete data sets. We further propose a new optimization algorithm to optimize the resulting objective function as well asprove our algorithm to converge fast. Experimental results on both real and artificial incompletedata sets demonstrated that our proposed methodoutperformed the feature selection methods undercomparison in terms of clustering performance.

Fibrosis, or the accumulation of extracellular matrix, causes loss of organ function and is a common feature of many chronic diseases. To interrogate the core molecular pathways of fibrosis, we cross-examined human primary cells from various tissues treated with TGFβ. Transcriptome analyses revealed that in addition to the known TGFβ signature, top-regulated cluster of genes are involved in fatty acid metabolism. To further evaluate this observation in vivo, we characterized a renal fibrosis model through unilateral ureteral obstruction (UUO) in mice. TGFβ signaling was greatly augmented in UUO kidneys and preventive treatment with an anti-TGFβ antibody resulted in significant reduction of fibrosis. Transcriptome analysis also identified fatty acid metabolism as one of the top dysregulated pathways in UUO kidneys, which was further supported by a substantial accumulation of acylcarnitines. Additional rodent models of liver fibrosis revealed a similar metabolic signature upon fibrosis induction. Lastly, a compound library phenotypic screen for suppressers of fibrosis identified AMPK and PPAR activators. As a proof of concept, we demonstrated that pharmacological treatment of telmisartan, an angiotensin receptor and PPARγ dual-agent, significantly reduced fibrosis in UUO kidneys, suggesting that metabolic defect is integral to TGFβ signaling and fibrosis. Altogether, our work has revealed a common node of metabolic signature underlying fibrosis that could represent a promising therapeutic target for multiple fibrotic diseases.

This paper addresses the problem of multi-class image classification by proposing a novel multi-view multi-sparsity kernel reconstruction (MMKR for short) model. Given images (including test images and training images) representing with multiple visual features, the MMKR first maps them into a high-dimensional space, e.g., a reproducing kernel Hilbert space (RKHS), where test images are then linearly reconstructed by some representative training images, rather than all of them. Furthermore a classification rule is proposed to classify test images. Experimental results on real datasets show the effectiveness of the proposed MMKR while comparing to state-of-the-art algorithms.

This paper proposed a novel feature selection method that includes a self-representation loss function, a graph regularization term and an $${l_{2,1}}$$l2,1-norm regularization term. Different from traditional least square loss function which focuses on achieving the minimal regression error between the class labels and their corresponding predictions, the proposed self-representation loss function pushes to represent each feature with a linear combination of its relevant features, aim at effectively selecting representative features and ensuring the robustness to outliers. The graph regularization terms include two kinds of inherent information, i.e., the relationship between samples (the sample---sample relation for short) and the relationship between features (the feature---feature relation for short). The feature---feature relation reflects the similarity between two features and preserves the relation into the coefficient matrix, while the sample---sample relation reflects the similarity between two samples and preserves the relation into the coefficient matrix. The $${l_{2,1}}$$l2,1-norm regularization term is used to conduct feature selection, aim at selecting the features, which satisfies the characteristics mentioned above. Furthermore, we put forward a new optimization method to solve our objective function. Finally, we feed reduced data into support vector machine (SVM) to conduct classification on real datasets. The experimental results showed that the proposed method has a better performance comparing with state-of-the-art methods, such as k nearest neighbor, ridge regression, SVM and so on.

This paper proposes an automated method for crystal structure refinement in powder diffraction to solve some outstanding issues including a great number of manual interventions, time-consuming and low efficiency caused by manual refinement of refinement software. The automated method is based on crystal parameters sorting space, black-box model and Fullprof software. The automated method can get the best refinement result in all kinds of order of crystal parameters refinement by “optimal R wp (the weighted profile R factor) search algorithm”. Meanwhile, an automated auxiliary system has been developed by the Python language to implement the application of the automated method to experimental data. The experimental results show that the automated method is feasible, which can save time and reduce the error rate caused by a lot of repeated manual refinement experiments, and improve the refinement efficiency.

Although variable refrigerant flow systems have become attractive due to good energy performance in part-load conditions, the shortcoming of the variable refrigerant flow system in outdoor air ventilation has not been well solved. A variable refrigerant flow and variable air volume combined air-conditioning system was proposed to solve this problem. The variable air volume part of the combined system consists of an outdoor air processing unit and air supply and distribution devices. The combined system was found having potential to obtain high-level energy efficiency without compromising thermal comfort by an intelligent control system. Therefore, a so-called global coordinated optimization strategy based on a local coordinated optimization strategy is proposed in this study. The local coordinated optimization strategy decreases energy consumption of the combined system through optimizing load allocation between the variable refrigerant flow unit and the outdoor air processing unit. The global coordinated...

In recent years, the destruction of the ecological environment in Huaibei Plain has limited the development of the economy. Doing research on eco-environment quality evaluation methods may be helpful to the recovery of the eco-environment in Huaibei Plain and the construction of ecological civilization. A new assessment system was introduced in this paper on the basis of a traditional eco-environmental evaluation method; the NPP index was used to replace biological abundance index and vegetation coverage index. This new method was used to evaluate the eco-environment quality of Huaibei Plain. Results indicate that: (a) the eco-environment of Huaibei Plain has been getting worse since 1990, but has improved since 2000; (b) the water-network density index is the key factor which affects the eco-environment of Huaibei Plain; (c) If human activities, pollution control, land degradation and urban area development are not taken into consideration, the eco-environment of Huaibei Plain in dry years will be serious.

The variable refrigerant flow (VRF) and variable air volume (VAV) combined air conditioning system can solve the problem of the VRF system in outdoor air ventilation while taking advantage of its high part load energy efficiency. Energy performance of the combined air conditioning system can also be optimized by joint control of both the VRF and the VAV parts. A model-based online optimal control strategy for the combined air conditioning system is presented. Simplified adaptive models of major components of the combined air conditioning system are firstly developed for predicting system performances. And a cost function in terms of energy consumption and thermal comfort is constructed. Genetic algorithm is used to search for the optimal control sets. The optimal control strategy is tested and evaluated through two case studies based on the simulation platform. Results show that the optimal strategy can effectively reduce energy consumption of the combined air conditioning system while maintaining acceptable thermal comfort.

This paper has proposed a novel spectral feature selection approach by embedding two modified subspace learning methods into a sparse feature selection framework. Specifically, we use an adaptive graph matrix learning and a low-rank constraint to preserve the local and global structure of data simultaneously. Sparse learning and low-rank constraint are used for relieving the impact of noise. Furthermore, we have coupled the graph matrix learning and low-dimensional feature space learning into an unified framework, aiming at achieving the global optimization of feature selection. By analysing the results of both proposed method and comparison methods on four realword and benchmark datasets, the proposed method achieves competitive results in term of classification performance.

Current unsupervised feature selection (UFS) methods learn the similarity matrix by using a simple graph which is learnt from the original data as well as is independent from the process of feature selection, and thus unable to efficiently remove the redundant/irrelevant features. To address these issues, we propose a new UFS method to jointly learn the similarity matrix and conduct both subspace learning (via learning a dynamic hypergraph) and feature selection (via a sparsity constraint). As a result, we reduce the feature dimensions using different methods (i.e., subspace learning and feature selection) from different feature spaces, and thus makes our method select the informative features effectively and robustly. We tested our method using benchmark datasets to conduct the clustering tasks using the selected features, and the experimental results show that our proposed method outperforms all the comparison methods.

Multiple integrins have been implicated in modulating renal function. Modulation of integrin function can lead to pathophysiological processes associated with diabetic nephropathy such as alterations in the glomerular filtration barrier and kidney fibrosis. The complexity of these pathophysiological changes implies that multiple integrin subtypes might need to be targeted to ameliorate the progression of renal disease. To address this hypothesis, we investigated the effects of MK‐0429, a compound that was originally developed as an αvβ3 inhibitor for the treatment of osteoporosis, on renal function and fibrosis. We demonstrated that MK‐0429 is an equipotent pan‐inhibitor of multiple av integrins. MK‐0429 dose‐dependently inhibited podocyte motility and also suppressed TGF‐β‐induced fibrosis marker gene expression in kidney fibroblasts. Moreover, in the obese ZSF1 rat model of diabetic nephropathy, chronic treatment with MK‐0429 resulted in significant reduction in proteinuria, kidney fibrosis, and collagen accumulation. In summary, our results suggest that inhibition of multiple integrin subtypes might lead to meaningful impact on proteinuria and renal fibrosis in diabetic nephropathy.

Insulin has a narrow therapeutic index, reflected in a small margin between a dose that achieves good glycemic control and one that causes hypoglycemia. Once injected, the clearance of exogenous insulin is invariant regardless of blood glucose, aggravating the potential to cause hypoglycemia. We sought to create a “smart” insulin, one that can alter insulin clearance and hence insulin action in response to blood glucose, mitigating risk for hypoglycemia. The approach added saccharide units to insulin to create insulin analogs with affinity for both the insulin receptor (IR) and mannose receptor C-type 1 (MR), which functions to clear endogenous mannosylated proteins, a principle used to endow insulin analogs with glucose responsivity. Iteration of these efforts culminated in the discovery of MK-2640, and its in vitro and in vivo preclinical properties are detailed in this report. In glucose clamp experiments conducted in healthy dogs, as plasma glucose was lowered stepwise from 280 mg/dL to 80 mg/dL, progressively more MK-2640 was cleared via MR, reducing by ∼30% its availability for binding to the IR. In dose escalations studies in diabetic minipigs, a higher therapeutic index for MK-2640 (threefold) was observed versus regular insulin (1.3-fold).