基于m×2正则化交叉验证的神经网络超参数调优方法

超参数调优是神经网络建模的关键问题。针对传统的超参数调优方法存在的问题,该文提出了一种基于m×2正则化交叉验证的超参数调优方法。目的是给出一种适用于复杂模型、大数据集背景下的计算开销较小且稳健的超参数调优方法。该方法的思想是从完整的数据集上选取少部分数据进行调优,避免模型在数据集较大时非常耗时的超参数调优难题;在m×2交叉验证的基础上设置正则化条件均衡训练集与验证集之间的分布差异,从而减少分布不一致带来的性能波动;使用信噪比作为调优的优化目标,从而可以综合考虑模型性能评价指标的均值和方差;并采用正交设计选择相关性较低的超参数组合以提高调优效率。以命名实体任务为例进行实验,在CoNLL 2003数据集上的实验结果显示,提出的调优方法能够选到和网格搜索性能上没有显著差异的超参数组合,且调优时间可显著降低约66%。

Numerical estimation of choice of the regularization parameter for NMR T2 inversion

Nuclear Magnetic inversion is the basis of NMR Resonance （NMR） T2 logging interpretation. The regularization parameter selection of the penalty term directly influences the NMR T2 inversion result. We implemented both norm smoothing and curvature smoothing methods for NMR T2 inversion, and compared the inversion results with respect to the optimal regular- ization parameters （（Xopt） which were selected by the dis- crepancy principle （DP）, generalized cross-validation （GCV）, S-curve, L-curve, and the slope of L-curve methods, respectively. The numerical results indicate that the DP method can lead to an oscillating or oversmoothed solution which is caused by an inaccurately estimated noise level. The （Xopt selected by the L-curve method is occa- sionally small or large which causes an undersmoothed or oversmoothed T2 distribution. The inversion results from GCV, S-curve and the slope of L-curve methods show satisfying inversion results. The slope of the L-curve method with less computation is more suitable for NMR T2 inversion. The inverted T2 distribution from norm smoothing is better than that from curvature smoothing when the noise level is high.

A Class of Regular Simple ω2-semigroups-Ⅱ

In this paper, we study regular simple ω2-semigroups in which D｜Es = Wd by the generalized Bruck-Reilly extension and obtain its structure theorem.

Source reconstruction for bioluminescence tomography via L_(1/2)regularization

Bioluminescence tomography(BLT)is an important noninvasive optical molecular imaging modality in preclinical research.To improve the image quality,reconstruction algorithms have to deal with the inherent ill-posedness of BLT inverse problem.The sparse characteristic of bioluminescent sources in spatial distribution has been widely explored in BLT and many L1-regularized methods have been investigated due to the sparsity-inducing properties of L1 norm.In this paper,we present a reconstruction method based on L_(1/2) regularization to enhance sparsity of BLT solution and solve the nonconvex L_(1/2) norm problem by converting it to a series of weighted L1 homotopy minimization problems with iteratively updated weights.To assess the performance of the proposed reconstruction algorithm,simulations on a heterogeneous mouse model are designed to compare it with three representative sparse reconstruction algorithms,including the weighted interior-point,L1 homotopy,and the Stagewise Orthogonal Matching Pursuit algorithm.Simulation results show that the proposed method yield stable reconstruction results under different noise levels.Quantitative comparison results demonstrate that the proposed algorithm outperforms the competitor algorithms in location accuracy,multiple-source resolving and image quality.

Wavelet-based L_(1/2) regularization for CS-TomoSAR imaging of forested area

Tomographic synthetic aperture radar(TomoSAR)imaging exploits the antenna array measurements taken at different elevation aperture to recover the reflectivity function along the elevation direction.In these years,for the sparse elevation distribution,compressive sensing(CS)is a developed favorable technique for the high-resolution elevation reconstruction in TomoSAR by solving an L_(1) regularization problem.However,because the elevation distribution in the forested area is nonsparse,if we want to use CS in the recovery,some basis,such as wavelet,should be exploited in the sparse L_(1/2) representation of the elevation reflectivity function.This paper presents a novel wavelet-based L_(2) regularization CS-TomoSAR imaging method of the forested area.In the proposed method,we first construct a wavelet basis,which can sparsely represent the elevation reflectivity function of the forested area,and then reconstruct the elevation distribution by using the L_(1/2) regularization technique.Compared to the wavelet-based L_(1) regularization TomoSAR imaging,the proposed method can improve the elevation recovered quality efficiently.

Counting Rooted Nearly 2-regular Planar Maps

The number of rooted nearly 2-regular maps with the valency of root-vertex, the number of non-rooted vertices and the valency of root-face as three parameters is obtained. Furthermore, the explicit expressions of the special cases including loopless nearly 2-regular maps and simple nearly 2-regular maps in terms of the above three parameters are derived.

Oviposition Regularity of Hybrid Silkworm Stock Moth of Liangguang No. 2

To cure pebrine using dry - hot air treatment of silkworm eggs, and control egg age within 12 hours, oviposition regularity of hybrid silkworm stock moth of Liangguang No. 2--current production variety in Guangdong Province was investigated. The experiment showed that most hybrid stock female moths of Liangguang No. 2 at 25 -28℃ oviposited in the first 5 hours, and egg production declined sharply after 5 hours. 7 · Xiang × Fu · 9 achieved the oviposition peak within 1 hour after casting the moth, the oviposition amount within 5 horns accounted for 92.94% of the total oviposition amount of female moth, that within 9 hours accounted for 96.47%. Fu · 9 × 7 · Xiang achieved the oviposition peak within 2 -4 hours after casting the moth, the oviposition amount within 5 hours accounted for 85.21% of the total oviposition amount of female moth, that within 9 hours accounted for 92.78%. Oviposition regularity of the hybrid silkworm stock moth of Liangguang No. 2 meets the need of dry-hot air treatment of silkworm pebrine-control egg age within 12 hours.

基于L_(1/2)正则化的抛物线Radon变换多次波压制方法

在地震数据处理中,多次波的存在会对地震数据成像和地震资料解释带来影响,如何有效地压制多次波干扰是地震勘探中的重要问题。抛物线Radon变换因其高效的特点被广泛应用于多次波压制中,但在野外地震数据采集时,炮检距的有限性会导致变换域中的能量扩散,产生假象,使多次波压制达不到理想的效果。针对此问题,提出一种基于L_(1/2)正则化的稀疏反演高分辨抛物线Radon变换,并应用广义迭代收缩算法(generalized iterated shrinkage algorithm,GISA)进行求解。研究结果表明,L_(1/2)正则化有很强的稀疏约束能力,能提高解的稀疏度,改进信噪分离的效果。与最小二乘反演和基于L_(1)正则化的稀疏反演相比,基于L_(1/2)正则化的稀疏反演高分辨抛物线Radon变换能更有效地压制多次波,并确保了重构数据与原始数据的一致性。

Adsorption Regularity and Characteristics of sp^3-Hybridized Gas Molecules on Anatase TiO_2(101) Surface

We report the anatase titanium dioxide （101） surface adsorption of sp3-hybridized gas molecules, including NH3, 1-12 0 and CH4, using first-principles plane-wave ultrasoft pseudopotential based on the density functional theory. The results show that it is much easier for a surface with oxygen vacancies to adsorb gas molecules than it is for a surface without oxygen vacancies. The main factor affecting adsorption stability and energy is the polarizability of molecules, and adsorption is induced by surface oxygen vacancies of the negatively charged center. The analyses of state densities and charge population show that charge transfer occurs at the molecule surface upon adsorption and that the number of transferred charge reduces in the order of N, 0 and C. Moreover, the adsorption method is chemical adsorption, and adsorption stability decreases in the order of NH3, tt2 0 and CH4. Analyses of absorption and reflectance spectra reveal that after absorbed CH4 and H2 O, compared with the surface with oxygen vacancy, the optical properties of materials surface, including its absorption coefficients and reflectivity index, have slight changes, however, absorption coefficient and reflectivity would greatly increase after NH3 adsorption. These findings illustrate that anatase titanium dioxide （101） surface is extremely sensitive to NH3.

REGULARITY OF (0,1,…,r-2,r)INTERPOLATION ON THE UNIT CIRCLE

The object of this paper is to show regularity of(0,1,...,r-2,r) interpolation on the set obtained by projecting vertically the zeros of (1-x2)pn(x)(λ≥1/2)onto the unit circle,where Pn(x)stands for the nth ultraspherical polynomial.

基于有约束L_(1/2)范数稀疏正则化的声源识别方法

基于等效源法(equivalent source method,ESM)的近场声全息(near field acoustic holography,NAH)是一种有效的声源识别技术。然而,针对空间稀疏分布的声源识别问题,传统基于L_(2)范数以及基于L_(1)范数的ESM方法分别存在声源幅值被低估与算法稳定性差等问题。因此,提出了基于有约束L_(1/2)范数稀疏正则化的声源识别方法,该方法具有强稀疏性与强抗干扰的优势,可以解决传统方法的声源识别精度低的问题。通过数值模拟试验以及普通室内的实测实验,验证了方法的有效性。

L1/2 Regularization Based on Bayesian Empirical Likelihood

Bayesian empirical likelihood is a semiparametric method that combines parametric priors and nonparametric likelihoods, that is, replacing the parametric likelihood function in Bayes theorem with a nonparametric empirical likelihood function, which can be used without assuming the distribution of the data. It can effectively avoid the problems caused by the wrong setting of the model. In the variable selection based on Bayesian empirical likelihood, the penalty term is introduced into the model in the form of parameter prior. In this paper, we propose a novel variable selection method, L1/2 regularization based on Bayesian empirical likelihood. The L1/2 penalty is introduced into the model through a scale mixture of uniform representation of generalized Gaussian prior, and the posterior distribution is then sampled using MCMC method. Simulations demonstrate that the proposed method can have better predictive ability when the error violates the zero-mean normality assumption of the standard parameter model, and can perform variable selection.

The Isomorphism Theorem of Regular Bisimple ω2-semigroups

In this paper we study the isomorphisms of two regular bisimple ω2- semigroups and obtain a criterion for two such semigroups to be isomorphic.

上海市金山区社区在管2型糖尿病患者药物治疗情况探析

目的:收集糖尿病患者的处方用药信息,了解社区卫生服务中心在管2型糖尿病患者的药物使用情况,总结高频用药组合,为全科医生更好地指导糖尿病患者合理用药提供依据。方法:基于2021年上海市金山区社区糖尿病管理系统的随访数据库、医院诊疗系统药物处方数据库,获取2型糖尿病患者的药物治疗情况。结果:28218例2型糖尿病患者中,正在使用降糖药22534例,药物治疗率为79.9%,规律服药率为56.4%;开药频次及开具率最多的降糖药是双胍类(15605次,53.6%)。最常见的降糖药使用方案为单药处方用药(32.9%),其次为二联处方用药(31.5%);单药处方高频用药是双胍类(28.4%)、磺脲类促泌剂(25.2%)、胰岛素类(20.0%);二联处方用药方案高频组合是双胍类/磺脲类促泌剂(34.2%);三联处方用药方案为双胍类/磺脲类促泌剂/α-糖苷酶抑制剂(19.4%)。随着糖尿病患者病程的增加,联合处方用药的比例上升。结论:上海市金山区社区卫生服务中心在管2型糖尿病患者的降糖药治疗率较高,但规律服药率偏低,存在四联甚至五联及以上处方用药的不规范用药情况。新型降糖药物占比较低,药物治疗率受年龄、病程、慢性病患病数量等因素影响。

Design of Polynomial Fuzzy Neural Network Classifiers Based on Density Fuzzy C-Means and L2-Norm Regularization

In this paper, polynomial fuzzy neural network classifiers (PFNNCs) is proposed by means of density fuzzy c-means and L2-norm regularization. The overall design of PFNNCs was realized by means of fuzzy rules that come in form of three parts, namely premise part, consequence part and aggregation part. The premise part was developed by density fuzzy c-means that helps determine the apex parameters of membership functions, while the consequence part was realized by means of two types of polynomials including linear and quadratic. L2-norm regularization that can alleviate the overfitting problem was exploited to estimate the parameters of polynomials, which constructed the aggregation part. Experimental results of several data sets demonstrate that the proposed classifiers show higher classification accuracy in comparison with some other classifiers reported in the literature.

转录因子HNF1A、HNF4A和FOXA2调节肝细胞蛋白质N-糖基化

Hepatocyte nuclear factor 1 alpha(HNF1A),hepatocyte nuclear factor 4 alpha(HNF4A),and forkhead box protein A2(FOXA2)are key transcription factors that regulate a complex gene network in the liver,cre-ating a regulatory transcriptional loop.The Encode and ChIP-Atlas databases identify the recognition sites of these transcription factors in many glycosyltransferase genes.Our in silico analysis of HNF1A,HNF4A.and FOXA2 binding to the ten candidate glyco-genes studied in this work confirms a significant enrich-ment of these transcription factors specifically in the liver.Our previous studies identified HNF1A as a master regulator of fucosylation,glycan branching,and galactosylation of plasma glycoproteins.Here,we aimed to functionally validate the role of the three transcription factors on downstream glyco-gene transcriptional expression and the possible effect on glycan phenotype.We used the state-of-the-art clus-tered regularly interspaced short palindromic repeats/dead Cas9(CRISPR/dCas9)molecular tool for the downregulation of the HNF1A,HNF4A,and FOXA2 genes in HepG2 cells-a human liver cancer cell line.The results show that the downregulation of all three genes individually and in pairs affects the transcrip-tional activity of many glyco-genes,although downregulation of glyco-genes was not always followed by an unambiguous change in the corresponding glycan structures.The effect is better seen as an overall change in the total HepG2 N-glycome,primarily due to the extension of biantennary glycans.We propose an alternative way to evaluate the N-glycome composition via estimating the overall complexity of the glycome by quantifying the number of monomers in each glycan structure.We also propose a model showing feedback loops with the mutual activation of HNF1A-FOXA2 and HNF4A-FOXA2 affecting glyco-genes and protein glycosylation in HepG2 cells.

具有ω_(d,d′)-型D关系的正则单ω^(2)-半群

为得到正则单ω^(2)-半群构造的刻画,研究了具有ω_(d,d′)-型D关系的正则单ω^(2)-半群:从群的长度为(d,d′)的ω-链及其到群的同态出发,借助Bruck-Reilly扩张,获得了具有ω_(d,d′)-型D关系的正则单ω^(2)-半群;证明了任一个具有ω_(d,d′)-型D关系的正则单ω^(2)-半群都可以用Bruck-Reilly扩张构造出来,得到了这类半群的结构定理。

基于BP-L_(2)混合算法的线性混合效应模型在帕金森病语音信号中的应用

通过对帕金森病语音信号多重共线性纵向高维数据的预处理和分析,可为相关医疗机构提供重要信息。首先采用随机森林对数据进行预处理,得到降维数据集,通过相关性分析得知该数据存在多重共线问题;然后在BP神经网络中引入L_(2)正则化改变其目标函数,解决临床数据经常存在的数据多重共线性,以便更好地拟合线性混合效应模型;最后对比分析引入BP-L_(2)混合算法前后线性混合效应模型的AIC、BIC和-LogLik指标,证明引入该算法的优势。

Study on algorithms of low SNR inversion of T_2 spectrum in NMR

The method of regularization factor selection determines stability and accuracy of the regularization method. A formula of regularization factor was proposed by analyzing the relationship between the improved SVD and regularization method. The improved SVD algorithm and regularization method could adapt to low SNR. The regularization method is better than the improved SVD in the case that SNR is below 30 and the improved SVD is better than the regularization method when SNR is higher than 30. The regularization method with the regularization factor proposed in this paper can be better applied into low SNR （5〈SNR） NMR logging. The numerical simulations and real NMR data process results indicated that the improved SVD algorithm and regularization method could adapt to the low signal to noise ratio and reduce the amount of computation greatly. These algorithms can be applied in NMR logging.

规律性体育行为对老年2型糖尿病患者的治疗作用

目的:研究规律性体育行为对老年2型糖尿病患者的治疗作用。方法:采用12周运动干预试验法,通过血糖、血脂等指标的检测,用SPSS 17.0统计软件进行分析,组间比较选用配对样本T检验。结果:通过规律体育行为干预后,受试者空腹血糖(FBG)、胆固醇(TC)、低密度脂蛋白(LDL)均有不同程度的下降;而对照组血糖、血脂等指标则无显著变化。结论:规律性体育对老年2型糖尿病患者有显著的治疗作用,可作为老年2型糖尿病重要的临床辅助疗法。