Search

Multi-Instance Multi-Label learning (MIML) is a popular framework for supervised classification where an example is described by multiple instances and associated with multiple labels. Previous MIML approaches have focused on predicting labels for instances. The idea of tackling the problem is to identify its equivalence in the traditional supervised learning framework. Motivated by the recent advancement in deep learning, in this paper, we still consider the problem of predicting labels and attempt to model deep learning in MIML learning framework. The proposed approach enables us to train deep convolutional neural network with images from social networks where images are well labeled, even labeled with several labels or uncorrelated labels. Experiments on real-world datasets demonstrate the effectiveness of our proposed approach.

P. aegyptiaca is one of the most destructive root parasitic plants worldwide, causing serious damage to many crop species. Under natural conditions P. aegyptiaca seeds must be conditioned and then stimulated by host root exudates before germinating. However, preliminary experiments indicated that TIS108 (a triazole-type inhibitor of strigolactone) and fluridone (FL, an inhibitor of carotenoid-biosynthesis) both stimulated the germination of P. aegyptiaca seeds without a water preconditioning step (i.e. unconditioned seeds). The objective of this study was to use deep RNA sequencing to learn more about the mechanisms by which TIS108 and FL stimulate the germination of unconditioned P. aegyptiaca seeds. Deep RNA sequencing was performed to compare the mechanisms of germination in the following treatments: (i) unconditioned P. aegyptiaca seeds with no other treatment, (ii) unconditioned seeds treated with 100 mg/L TIS108, (iii) unconditioned seeds treated with 100 mg/L FL + 100 mg/L GA3, (iv) conditioned seeds treated with sterile water, and (v) conditioned seeds treated with 0.03 mg/L GR24. The de novo assembled transcriptome was used to analyze transcriptional dynamics during seed germination. The key gene categories involved in germination were also identified. The results showed that only 119 differentially expressed genes were identified in the conditioned treatment vs TIS108 treatment. This indicated that the vast majority of conditions for germination were met during the conditioning stage. Abscisic acid (ABA) and gibberellic acid (GA) played important roles during P. aegyptiaca germination. The common pathway of TIS108, FL+GA3, and GR24 in stimulating P. aegyptiaca germination was the simultaneous reduction in ABA concentrations and increase GA concentrations. These results could potentially aid the identification of more compounds that are capable of stimulating P. aegyptiaca germination. Some potential target sites of TIS108 were also identified in our transcriptome data. The results of this experiment suggest that TIS108 and FL+GA3 could be used to control P. aegyptiaca through suicidal germination.

A novel noncyclic crown-type diethoxycarbonyl glycoluril derivative receptor was synthesised and its crystal structure was obtained. The study of its binding behaviour towards neutral guests showed that it can bind hydroquinone, resorcinol and catechol.

Standard Gibbs energies of transfer of potassium tetraphenylborate from water to (water + t -butyl alcohol) (TBA) mixtures at T = 298.15 K have been studied. The mole fractions of TBA in the mixed solvent are 0.00, 0.01, 0.02, 0.03, 0.04, 0.045, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.20, 0.40, 0.60, 0.80, and 1.00, respectively. The primary medium effect of potassium tetraphenylborate in the mixed solvent is discussed.

H-mode-like phenomenon is observed with positive limiter biasing on the CT-6B Tokamak. Poloidal rotation of plasma in the boundary region is slowed and even changed its direction with positive bias, effect of positive bias on the toroidal rotation is also presented.

This paper discusses the relationship between cigarette smoking and the p53 protein and P21 protein expression by the immunohistochemical analysis in 93 cases with lung cancer in which squamous cell carcinoma accounted for 45 cases, adenocarcinoma 48 cases. The results showed that positive proportion of p53 protein expression was 74.20% (28 of 37 squamous cell carcinoma, 21 of 30 adenocarcinomas) in cigarette smoking group with lung cancers, and 38.46% (3 of 8 squamous cell carcinoma, 7 of 18 adenocarcinomas) in nonsmoking group with lung cancers. The difference was statistically significant. Odds ratio was 4.14 and confidence limits for OR was 1.42-12.52. A dose-related presents in the p53 protein expression for the smoking amount and smoking years. The positive proportion of P21 protein expression was 79.31% (21 of 28 squamous cell carcinoma, 25 of 30 adenocarcinomas) in cigarette smoking group with lung cancers, and 82.75% (10 of 11 squamous, 14 of 18 adenocarcinomas) in nonsmoking group with lung cancers, the difference was not statistically significant. But their positive proportion of P21 protein expression were very high in both groups. It was indicated that no relationship between cigarette smoking and the P21 protein expression. We suggest that the p53 gene could be a common target of tobacco-associated carcinogenesis in lung cancer.

Barium(II) tetraphenylborate, Ba(Bph4))2·4H2O was prepared, and its decomposition mechanism was studied by means of TG and DTA. The products of thermal decomposition were examined by means of gas chromatography and chemical methods. A kinetic analysis of the first stage of thermal decomposition was made on the basis of TG and DTG curves and kinetic parameters were obtained from an analysis of the TG and DTG curves using integral and differential methods. The most probable kinetic function was suggested by comparison of kinetic parameters. A mathematical expression was derived for the kinetic compensation effect.

Aiming at the problem of the traditional dimensionality reduction methods cannot recover the inherent structure, and Scale Invariant Feature Transform(SIFT)achieving low precision when reinstating images, an Image Retrieval Method Based on Manifold Learning with Scale-Invariant Feature is proposed. It aims to find low-dimensional compact representations of high-dimensional observation data and explores the inherent low and intrinsic dimension of data. The feature extraction method-SIFT and the adaptive ISOMAP method are combined and conducted experiments on the ORL face image dataset. This paper analyzes and discusses the problem of effects of the neighborhood parameter and the intrinsic dimension size on the face image recognition.

This study aimed to investigate the subtype distribution of gp41 gene of human immunodeficiency virus-1 (HIV-1) among men who have sex with men (MSM) in Zhengzhou.Thirty blood samples were collected from men who have sex with men infected by HIV. The complete gp41 gene was amplified by RT-PCR and nested-PCR and sequenced. All sequences were edited by Bioedit and subtyped with HIV sequence library US Los Alamos National Laboratory and online genotyping software provided by American National Center of Biotechnology Information. Phylogenetic analysis of gp41 gene was performed using the MEGA 3.1 software, and the genic dispersion rates among subtype of gp41 gene were analyzed.A total of eighteen gene sequences of HIV-1 gp41 gene were obtained from thirty men who have sex with men infected by HIV, which belonged to subtype CRF15-01B (50% (9/18)), CRF01-AE (22% (4/18)), CRF07-B (22% (4/18)) and B (6% (1/18)), respectively. The intersubtype HIV-1 strains aggregate with according reference strains. The genetic distance inter-subtype of subtype CRF15-01B, CRF01-AE and CRF07-B were 0.050 ± 0.007, 0.052 ± 0.009 and 0.082 ± 0.012, respectively.The prevalent subtypes of HIV-1 among among MSM in Zhengzhou was complicated and recombinant HIV-1 strains were the most prevalent strains.

To investigate the subtype distribution of human immunodeficiency virus-1(HIV-1) infection among men who have sex with men (MSM) in Zhengzhou, Henan Province, forty blood samples were collected from HIV-1 carriers, who acknowledged to have sex with men. The complete gag gene was amplified by RT-PCR and nested-PCR and sequenced. All sequences were edited by BioEdit and subtyped by genotyping software. Phylogenetic analysis of gag gene were then performed using the MEGA 3.1 software, the gene distances were calculated by Distance program. There were three different HIV-1 subtypes including B, CRF01-AE and CRF07-BC present among twenty four MSMs in Zhengzhou. Genotyping results showed that 33.33% (8/24) were B, 41.67% (10/24) were CRF01-AE and 25% (6/24) were CRF07-BC, and subtype CRF01-AE had become the most prevalent HIV-1 subtype in Zhengzhou, Henan province. In conclusion, recombinant HIV-1 strains are circulating in Henan province and the epidemiology is complicated.

Grafting eggplant (Solanum melongena L.) on tomato rootstock is an effective method to control Verticillium wilt (caused by Verticillium dahliae), a soil-borne disease. This study investigated the fruit quality and biological characteristics associated with disease resistance of eggplants grafted onto tomato rootstocks ‘Lydl’. The plant height, stem diameter, leaf area and root volume of grafted eggplant were higher than those of non-grafted eggplant. And the increase rates were up as the date prolonged. The root active-absorbing area, chlorophyll content and peroxidase (POX) and phenylalanine ammonialyase (PAL) activities of grafted eggplants were higher than those of non-grafted eggplants, while the comparative conductivity and the free proline content decreased by 42.42 and 44.08%. Obviously, grafting on tomato rootstock improved the growth and development of eggplant. The soluble sugar, vitamin C and protein content in fruit of grafted plants were not lower than those in non-grafted ones. Fruit qualities were not significantly affected by grafting. At the all three sampling stages, grafted plants showed higher amounts of bacteria and actinomyces, total microbial population, and the ratios of actinomyces to fungi (A/F), bacteria to fungi (B/F) and microbial biomass (MBC, MBN, MBP, MQ), but lower fungi population. The activities of urease and catalase in rhizosphere were enhanced by grafting. Grafting could modify the plant growth and microbial characteristics which may enhance the resistance to Verticillium wilt. Key words: Verticillium wilt, grafted eggplant, microbial property, microbial biomass.

The testicular yolk sac tumor (TYST) is the most common neoplasm originated from germ cells differentiated abnormally, a major part of pediatric malignant testicular tumors. The present study aimed at developing and validating the in vitro and vivo models of TYST and evaluating the sensitivity of TYST to treatments, by cloning human TYST cells and investigating the histology, ultra-structure, growth kinetics and expression of specific proteins of cloned cells. We found biological characteristics of cloned TYST cells were similar to the yolk sac tumor and differentiated from the columnar to glandular-like or goblet cells-like cells. Chromosomes for tumor identification in each passage met nature of the primary tumor. TYST cells were more sensitive to all-trans-retinoic acid which had significantly inhibitory effects on cell proliferation. Cisplatin induced apoptosis of TYST cells through the activation of p53 expression and down-regulation of Bcl- expression. Thus, we believe that cloned TYST cells and the animal model developed here are useful to understand the molecular mechanism of TYST cells and develop potential therapies for human TYST.

Prediction of precipitation is very significant for dispatching and control of water in river to manage river and flood. RBF neural network trained with particle swarm optimization is presented to predict precipitation in the study. RBF neural network has strong nonlinear prediction ability, which has considerable applications. In the study, RBF neural network trained with particle swarm optimization is presented to predict precipitation. In RBF neural network, forecasting accuracy is affected by its operational parameters. Particle swarm optimization technique is used as a training phase of RBF neural network to determine its optimal parameters. A certain river irrigated area precipitation data from 1970 to 2002 are adopted to evaluate the performance of the proposed method. The testing results show that the proposed model is superior to BP neural network.

Data mining involves nontrivial process of extracting knowledge or patterns from large databases. Many searching and optimization methods are used in data mining. In this paper we propose a Self-Adaptive Hybrid GA (SAHGA), where parameters of population size, crossover rate and mutation rate for each individual in each generation are adaptively fixed. Further, the crossover operator and mutation operator are decided dynamically. Finally, the tabu strategy is involved in the process of evolution. The three measures mentioned above help to maintain the diversity of the population and smooth over premature convergence. The effective performance of the algorithm is then shown using standard testbed functions and a set of classification datamining problems with UCI datasets based on Weka Platform.