Optimal Randomness for Stochastic Configuration Network (SCN) with Heavy-Tailed Distributions

Stochastic Configuration Network (SCN) has a powerful capability for regression and classification analysis. Traditionally, it is quite challenging to correctly determine an appropriate architecture for a neural network so that the trained model can achieve excellent performance for both learning and generalization. Compared with the known randomized learning algorithms for single hidden layer feed-forward neural networks, such as Randomized Radial Basis Function (RBF) Networks and Random Vector Functional-link (RVFL), the SCN randomly assigns the input weights and biases of the hidden nodes in a supervisory mechanism. Since the parameters in the hidden layers are randomly generated in uniform distribution, hypothetically, there is optimal randomness. Heavy-tailed distribution has shown optimal randomness in an unknown environment for finding some targets. Therefore, in this research, the authors used heavy-tailed distributions to randomly initialize weights and biases to see if the new SCN models can achieve better performance than the original SCN. Heavy-tailed distributions, such as Lévy distribution, Cauchy distribution, and Weibull distribution, have been used. Since some mixed distributions show heavy-tailed properties, the mixed Gaussian and Laplace distributions were also studied in this research work. Experimental results showed improved performance for SCN with heavy-tailed distributions. For the regression model, SCN-Lévy, SCN-Mixture, SCN-Cauchy, and SCN-Weibull used less hidden nodes to achieve similar performance with SCN. For the classification model, SCN-Mixture, SCN-Lévy, and SCN-Cauchy have higher test accuracy of 91.5%, 91.7% and 92.4%, respectively. Both are higher than the test accuracy of the original SCN.     

Why Do Big Data and Machine Learning Entail the Fractional Dynamics?

Fractional-order calculus is about the differentiation and integration of non-integer orders. Fractional calculus (FC) is based on fractional-order thinking (FOT) and has been shown to help us to understand complex systems better, improve the processing of complex signals, enhance the control of complex systems, increase the performance of optimization, and even extend the enabling of the potential for creativity. In this article, the authors discuss the fractional dynamics, FOT and rich fractional stochastic models. First, the use of fractional dynamics in big data analytics for quantifying big data variability stemming from the generation of complex systems is justified. Second, we show why fractional dynamics is needed in machine learning and optimal randomness when asking: “is there a more optimal way to optimize?”. Third, an optimal randomness case study for a stochastic configuration network (SCN) machine-learning method with heavy-tailed distributions is discussed. Finally, views on big data and (physics-informed) machine learning with fractional dynamics for future research are presented with concluding remarks.     

Preparative isolation and purification of anti‐tumor agent ansamitocin P‐3 from fermentation broth of Actinosynnema pretiosum using high‐performance counter‐current chromatography

Ansamitocin P‐3 is a potent anti‐tumor maytansinoid found in Actinosynnema pretiosum. However, due to the complexity of the fermentation broth of Actinomycete, how to effectively separate ansamitocin P‐3 is still a challenge. In this study, both analytical and preparative high‐performance counter‐current chromatography were successfully used to separate and purify ansamitocin P‐3 from fermentation broth. A total of 28.8 mg ansamitocin P‐3 with purity of 98.4% was separated from 160 mg crude sample of fermentation broth in less than 80 min with the two‐phase solvent system of hexane–ethyl acetate–methanol–water (0.6:1:0.6:1, v/v/v/v). The purity and structural identification were determined by HPLC, ¹H NMR, ¹³C NMR and mass spectroscopy.     

基于光电准直虚拟扩展成像的姿态角测量系统的误差分析

建立了基于光电准直原理的姿态角测量系统,提出一种虚拟扩展成像面技术,应用于测量系统中,对图像传感器成像面进行复用,虚拟扩大系统的测量视场.介绍了系统的主要工作原理,在系统模型下,对可能影响系统测量精度的误差源进行了分析,建立其数学关系,对精度影响因素进行了定量仿真实验.按照仿真结果,对系统设计方案进行了分析验证.结果表...     

X型Gemini表面活性剂改性类水滑石的制备及其吸附对甲酚性能

传统表面活性剂-类水滑石(HTlc)纳米复合物可有效吸附水中主要有机污染物酚类污染物,如酚类污染物,以高性能Gemini双子表面活性剂代替传统表面活性剂有望提高其去除效率。本文采用剥离-重组法制备了X型Gemini表面活性剂(MXC_6)-HTlc纳米复合物,并采用粉末X射线衍射(XRD)、傅里叶变换红外光谱(FTIR)、比表面和元素分析等技术手段对样品进行了表征。结果表明,MXC_6成功重组于HTlc层间。分别从吸附动力学和热力学研究了MXC_6-HTlc纳米复合物对对甲酚的吸附行为:吸附动力学曲线符合准一级动力学方程;吸附等温线符合Linear方程;吸附量显著高于HTlc,且随p H值和温度的增加而降低。实验表明,MXC_6-HTlc纳米复合物是一种新型高效水中酚类污染物处理剂。     

Coumarinolignans Isolated from the Seeds of Brucea javanica

A new coumarinolignan, cleomiscosin E ( 1 ), together with the known compound cleomiscosin A ( 2 ), has been isolated from the seeds of Brucea javanica (L.) Merr . Their structures were assigned on the basis of spectral studies. These two compounds exhibited potent anti‐inflammatory activities by inhibiting the nitric oxide (NO) production in lipopolysaccharide (LPS)‐activated RAW264.7 macrophages.     

光学元件损伤在线检测图像处理技术

针对光学元件损伤图像中损伤区域的高精度检测问题,对内全反射照明下光学元件损伤图像的处理技术进行了研究。根据在线检测图像中损伤区域中心峰值的信号强度高于局部背景的信号强度这一特点,利用高斯滤波器生成待检测图像的局部信号强度比图像,实现了对损伤区域的低漏检率自动定位;根据CCD的成像原理,利用辐射标定的方法建立起损伤区域的尺寸与其在图像中总灰度的关系方程,实现了损伤区域的亚像素高精度尺寸测量。实验结果表明,与传统的光学元件损伤图像处理算法相比,本文提出的算法在保持低漏检率的同时大大提高了损伤区域的测量精度。     

Patterning catalyst via inkjet printing to grow single-walled carbon nanotubes

We demonstrated a method to pattern catalyst via inkjet printing to grow SWNTs, using metal salt solutions as the inks and an ordinary office-use printer. We printed water solutions of cobalt acetate on hydrophilic Si substrates and grew high quality SWNT films.     

Synthesis and UCST‐type phase behavior of polypeptide with alkyl side‐chains in alcohol or ethanol/water solvent mixtures

A series of thermoresponsive polypeptides bearing 1‐butyl, 1‐hexyl, or 1‐dodecyl side‐chains (i.e., 6a ‐ 6c ) were synthesized by copper‐mediated 1,3‐dipolar cycloaddition with high grafting efficiency (>95%) between side‐chain “clickable” polypeptide, namely poly(γ‐4‐(propargoxycarbonyl)benzyl‐_L‐glutamate) ( 5 ) and 1‐azidoalkanes. 5 with different degree of polymerization (DP = 48–86) were prepared from triethylamine initiated ring‐opening polymerization of γ‐4‐(propargoxycarbonyl)benzyl‐_L‐glutamic acid based N‐carboxyanhydride ( 4 ). ¹H NMR, FTIR, and GPC results revealed the successful preparation of the resulting polypeptides. 6a ‐ 6c showed reversible UCST‐type phase behaviors in methanol, ethanol, and ethanol/water solvent mixtures depending on the polymer main‐chain length, alkyl side‐chain length, weight percentage of ethanol (f_w) in the binary solvent, and so forth. FTIR analysis revealed the presence of the van der Waals interaction between the alkyl pendants of polypeptides and alkyl groups of alcoholic solvents. Variable‐temperature UV‐vis spectroscopy revealed that the UCST‐type phase transition temperature (T_pt) increased as polymer main‐chain length or concentration increased. In ethanol/water solvent mixtures, polypeptide with short alkyl pendant (i.e., 1‐butyl group) and short main‐chain length (DP = 41) showed the widest f_w range and T_pts in the range of 61.0–71.1 °C. © 2016 Wiley Periodicals, Inc. J. Polym. Sci., Part A: Polym. Chem. 2016 , 54, 3425–3435     

On the Upper Limits of Oxidation States in Chemistry

The concept of oxidation state ( OS ) is based on the concept of Lewis electron pairs, in which the bonding electrons are assigned to the more electronegative element. This approach is useful for keeping track of the electrons, predicting chemical trends, and guiding syntheses. Experimental and quantum‐chemical results reveal a limit near +8 for the highest OS in stable neutral chemical substances under ambient conditions. OS =+9 was observed for the isolated [IrO₄]⁺ cation in vacuum. The prediction of OS =+10 for isolated [PtO₄]²⁺ cations is confirmed computationally for low temperatures only, but hasn't yet been experimentally verified. For high OS species, oxidation of the ligands, for example, of O^?2 with formation of ^.O^?1 and O?O bonds, and partial reduction of the metal center may be favorable, possibly leading to non‐Lewis type structures.