含有五个反馈权值的二维细胞神经网络可实现Mosaic模式

研究一类含有五个反馈权值二维细胞神经网络系统,以系统的输出反馈权值及阈值为参数构成参数空间,引入几何方法,将此空间分解分块成有限个区域,当输出反馈权值及阈值在某一区域上时,给出系统具有M osaic解所有可能的输出函数M osaic模式的充要条件.     

稀土萃取分离过程软测量方法的研究

分析了稀土组分含量在线检测技术进展缓慢的原因,分析了稀土萃取分离过程机制,找出影响稀土组分含量检测的主要因素,确定辅助变量。利用从包钢稀土厂采集的数据通过newrb函数创建径向基网络来实现稀土组分含量的软测量。文中通过流程图详细阐述了建模方法,采用50组样本数据对所建立的模型进行计算机仿真验证,仿真结果证明了模型的可行性。     

Development of models for prediction of the antioxidant activity of derivatives of natural compounds

Antioxidants are important for maintaining the appropriate balance between oxidizing and reducing species in the body and thus preventing oxidative stress. Many natural compounds are being screened for their possible antioxidant activity. It was found that a mushroom pigment Norbadione A, which is a pulvinic acid derivative, shows an antioxidant activity; the same was found for other pulvinic acid derivatives and structurally related coumarines. Based on the results of in vitro studies performed on these compounds as a part of this study quantitative structure–activity relationship (QSAR) predictive models were constructed using multiple linear regression, counter-propagation artificial neural networks and support vector regression (SVR). The models have been developed in accordance with current QSAR guidelines, including the assessment of the models applicability domains. A new approach for the graphical evaluation of the applicability domain for SVR models is suggested. The developed models show sufficient predictive abilities for the screening of virtual libraries for new potential antioxidants.     

Selective Template Removal by Thermal Depolymerization to Obtain Mesostructured Molybdenum Oxycarbide

The carbon content of mesostructured organic‐inorganic hybrid material of a cylindrical block copolymer template of poly(2‐vinylpyridine)‐block‐poly(allyl methacrylate) (P2VP‐b‐PAMA) and ammonium paramolybdate (APM) could be reduced by thermal depolymerization. By calcination in vacuo at 320 °C the PAMA core can be completely removed while the remaining P2VP brush preserves the mesostructure. The P2VP‐APM composite can then be carburized in‐situ to MoO_xC_y in a second pyrolysis step without any additional carbon source but P2VP. The molybdenum oxycarbide nanotubes obtained, form hierarchically porous non‐woven structures, which were tested as catalyst in the decomposition of NH₃. They proved to be catalytically active at temperatures above 450 °C. The activation energy was estimated from an Arrhenius Plot to be 127 kJ · mol^–1.     

Network regularised Cox regression and multiplex network models to predict disease comorbidities and survival of cancer

In cancer genomics, gene expression levels provide important molecular signatures for all types of cancer, and this could be very useful for predicting the survival of cancer patients. However, the main challenge of gene expression data analysis is high dimensionality, and microarray is characterised by few number of samples with large number of genes. To overcome this problem, a variety of penalised Cox proportional hazard models have been proposed. We introduce a novel network regularised Cox proportional hazard model and a novel multiplex network model to measure the disease comorbidities and to predict survival of the cancer patient. Our methods are applied to analyse seven microarray cancer gene expression datasets: breast cancer, ovarian cancer, lung cancer, liver cancer, renal cancer and osteosarcoma. Firstly, we applied a principal component analysis to reduce the dimensionality of original gene expression data. Secondly, we applied a network regularised Cox regression model on the reduced gene expression datasets. By using normalised mutual information method and multiplex network model, we predict the comorbidities for the liver cancer based on the integration of diverse set of omics and clinical data, and we find the diseasome associations (disease–gene association) among different cancers based on the identified common significant genes. Finally, we evaluated the precision of the approach with respect to the accuracy of survival prediction using ROC curves. We report that colon cancer, liver cancer and renal cancer share the CXCL5 gene, and breast cancer, ovarian cancer and renal cancer share the CCND2 gene. Our methods are useful to predict survival of the patient and disease comorbidities more accurately and helpful for improvement of the care of patients with comorbidity. Software in Matlab and R is available on our GitHub page: https://github.com/ssnhcom/NetworkRegularisedCox.git.     

Synthesis,structure and characterization of two new metal–organic coordination polymers based on the ligand 5‐iodobenzene‐1,3‐dicarboxylate

Two new coordination polymers (CPs) formed from 5‐iodobenzene‐1,3‐dicarboxylic acid (H₂iip) in the presence of the flexible 1,4‐bis(1H‐imidazol‐1‐yl)butane (bimb) auxiliary ligand, namely poly[[μ₂‐1,4‐bis(1H‐imidazol‐1‐yl)butane‐κ²N³:N^3′](μ₃‐5‐iodobenzene‐1,3‐dicarboxylato‐κ⁴O¹,O^1′:O³:O^3′)cobalt(II)], [Co(C₈H₃IO₄)(C₁₀H₁₄N₄)]_n or [Co(iip)(bimb)]_n, (1), and poly[[[μ₂‐1,4‐bis(1H‐imidazol‐1‐yl)butane‐κ²N³:N^3′](μ₂‐5‐iodobenzene‐1,3‐dicarboxylato‐κ²O¹:O³)zinc(II)] trihydrate], {[Zn(C₈H₃IO₄)(C₁₀H₁₄N₄)]·3H₂O}_n or {[Zn(iip)(bimb)]·3H₂O}_n, (2), were synthesized and characterized by FT–IR spectroscopy, thermogravimetric analysis (TGA), solid‐state UV–Vis spectroscopy, single‐crystal X‐ray diffraction analysis and powder X‐ray diffraction analysis (PXRD). The iip²⁻ ligand in (1) adopts the (κ¹,κ¹‐μ₂)(κ¹, κ¹‐μ₁)‐μ₃ coordination mode, linking adjacent secondary building units into a ladder‐like chain. These chains are further connected by the flexible bimb ligand in a trans–trans–trans conformation. As a result, a twofold three‐dimensional interpenetrating α‐Po network is formed. Complex (2) exhibits a two‐dimensional (4,4) topological network architecture in which the iip²⁻ ligand shows the (κ¹)(κ¹)‐μ₂ coordination mode. The solid‐state UV–Vis spectra of (1) and (2) were investigated, together with the fluorescence properties of (2) in the solid state.     

One‐dimensional CuII coordination polymers containing C2h‐symmetric 1,1′:4′,1′′‐terphenyl‐3,3′‐dicarboxylate linkers

Two new one‐dimensional Cu^II coordination polymers (CPs) containing the C_2h‐symmetric terphenyl‐based dicarboxylate linker 1,1′:4′,1′′‐terphenyl‐3,3′‐dicarboxylate (3,3′‐TPDC), namely catena‐poly[[bis(dimethylamine‐κN)copper(II)]‐μ‐1,1′:4′,1′′‐terphenyl‐3,3′‐dicarboxylato‐κ⁴O,O′:O′′:O′′′] monohydrate], {[Cu(C₂₀H₁₂O₄)(C₂H₇N)₂]·H₂O}_n, (I), and catena‐poly[[aquabis(dimethylamine‐κN)copper(II)]‐μ‐1,1′:4′,1′′‐terphenyl‐3,3′‐dicarboxylato‐κ²O³:O^3′] monohydrate], {[Cu(C₂₀H₁₂O₄)(C₂H₇N)₂(H₂O)]·H₂O}_n, (II), were both obtained from two different methods of preparation: one reaction was performed in the presence of 1,4‐diazabicyclo[2.2.2]octane (DABCO) as a potential pillar ligand and the other was carried out in the absence of the DABCO pillar. Both reactions afforded crystals of different colours, i.e. violet plates for (I) and blue needles for (II), both of which were analysed by X‐ray crystallography. The 3,3′‐TPDC bridging ligands coordinate the Cu^II ions in asymmetric chelating modes in (I) and in monodenate binding modes in (II), forming one‐dimensional chains in each case. Both coordination polymers contain two coordinated dimethylamine ligands in mutually trans positions, and there is an additional aqua ligand in (II). The solvent water molecules are involved in hydrogen bonds between the one‐dimensional coordination polymer chains, forming a two‐dimensional network in (I) and a three‐dimensional network in (II).     

结合区域提取和改进卷积神经网络的水下小目标检测

针对水下小目标信息量有限而难以提取有效特征导致的检测性能不佳问题,提出了一种结合区域提取和融合Hu矩特征的改进卷积神经网络水下小目标检测方法。该方法包含区域提取和分类两个步骤。首先以马尔可夫随机场分割算法为基础进行区域提取,对潜在目标定位的同时降低伪目标对后续分类的干扰;然后提取潜在目标区域的Hu矩特征并融入卷积神经网络,形成一种形状特征表征能力更强的改进卷积神经网络用于分类。声呐实测数据处理结果表明,该方法可以有效提升对水下小目标的发现概率和正确报警率,与其他目标检测方法相比,该方法具有更好的检测性能和泛化性。     

基于DCGANs的半片光伏组件电致发光图像增强技术

针对半片光伏组件电致发光（electroluminescence,EL）缺陷自动识别过程中训练用样本不足导致模型过拟合的问题,采用深度卷积生成对抗网络（deep convolutional generative adversarial networks,DCGANs）生成可控制属性的半片光伏组件EL图像,再采用多尺度结构相似性（multiscale structural similarity,MS-SSIM）指标对生成的EL图像与拍摄的EL图像之间的相似程度进行了评估。评估结果得到,使用DCGANs生成的所有类型半片光伏组件的EL图像与拍摄的EL图像的MS-SSIM指标都大于0.55,大部分的MS-SSIM值在0.7附近。在分类模型的训练过程中,测试集准确率随着训练集中生成图像数量的增加而升高,当生成图像数量达到6 000张时,测试集准确率达到97.92%。实验结果表明,采用DCGANs能够生成高质量且可控制属性的半片光伏组件EL图像,较好地解决因缺少训练样本而导致的模型过拟合问题。