基于快速神经网络架构搜索的鲁棒图像水印网络算法

为解决深度学习在图像水印算法中计算量大且模型冗余的问题，提高图像水印算法在抵抗噪声、旋转和剪裁等攻击时的鲁棒性，提出基于快速神经网络架构搜索（neural architecture search，NAS）的鲁棒图像水印网络算法。通过多项式分布学习快速神经网络架构搜索算法，在预设的搜索空间中搜索最优网络结构，进行图像水印的高效嵌入与鲁棒提取。首先，将子网络中线性连接的全卷积层设置为独立的神经单元结构，并参数化表示结构单元内节点的连接，预先设定结构单元内每个神经元操作的搜索空间；其次，在完成一个批次的数据集训练后，依据神经元操作中的被采样次数和平均损失函数值动态更新概率；最后，重新训练搜索完成的网络。水印网络模型的参数量较原始网络模型缩减了92%以上，大大缩短了模型训练时间。由于搜索得到的网络结构更为紧凑，本文算法具有较高的时间性能和较好的实验效果，在隐藏图像时，对空域信息的依赖比原始网络更少。对改进前后的2个网络进行了大量鲁棒性实验，对比发现，本文算法在CIFAR-10数据集上对抵抗椒盐噪声和旋转、移除像素行（列）等攻击优势显著；在ImageNet数据集上对抵抗椒盐高斯噪声、旋转、中值滤波、高斯滤波、JPEG压缩、裁剪等攻击优势显著，特别是对随机移除行（列）和椒盐噪声有较强的鲁棒性。     

Optimization of Transducer Location for Novel Non-Intrusive Methodologies of Diagnosis in Diesel Engines

The health monitoring has been studied to ensure integrity of design of engine structure by detection, quantification, and prediction of damages. Early detection of faults may allow the downtime of maintenance to be rescheduled, thus preventing sudden shutdown of machines. In cylinder pressure developed, vibrations and noise emissions data provide a rich source of information about condition of engines. Monitoring of vibrations and noise emissions are novel non-intrusive methodologies for which positioning of various transducers are important issue. The presented work shows applicability of these diagnosis methodologies adopted in case of diesel engines. The effects of changing various fuel injection parameters was analyzed. Scope of using non-intrusive technique has been analyzed by changing locations of microphone. Novelty of this worklies in exploring signal processing methods for various locations around the engine test set up. Various frequency ranges of contributing noise and vibration sources were identified. Time-Frequency analysis showed the onset of various cyclic. Based on the identification of various frequency bands, it is possible to device suitable filters in order to extract more information.     

The coupled-cluster formalism – a mathematical perspective

ABSTRACT

The Coupled-Cluster (CC) theory is one of the most successful high precision methods used to solve the stationary Schrödinger equation. In this article, we address the mathematical foundation of this theory with focus on the advances made in the past decade. Rather than solely relying on spectral gap assumptions (non-degeneracy of the ground state), we highlight the importance of coercivity assumptions – Gårding type inequalities – for the local uniqueness of the CC solution. Based on local strong monotonicity, different sufficient conditions for a local unique solution are suggested. One of the criteria assumes the relative smallness of the total cluster amplitudes (after possibly removing the single amplitudes) compared to the Gårding constants. In the extended CC theory the Lagrange multipliers are wave function parameters and, by means of the bivariational principle, we here derive a connection between the exact cluster amplitudes and the Lagrange multipliers. This relation might prove useful when determining the quality of a CC solution. Furthermore, the use of an Aubin–Nitsche duality type method in different CC approaches is discussed and contrasted with the bivariational principle.     

Nano-sized Cu(II) and Zn(II) complexes and their use as a precursor for synthesis of CuO and ZnO nanoparticles: A study on their sonochemical synthesis,characterization, and DNA-binding/cleavage,anticancer, and antimicrobial activities

Two new divalent copper (C1) and zinc (C2) chelates having the formulae [M(PIMC)₂] (where M = Cu(II), Zn(II) and PIMC = Ligand [(E)-3-(((3-hydroxypyridin-2-yl)imino)methyl)-4H-chromen-4-one] were obtained and characterized by several techniques. Structures and geometries of the synthesized complexes were judged based on the results of alternative analytical and spectral tools supporting the proposed formulae. IR spectral data confirmed the coordination of the ligands to the copper and zinc centers as monobasic tridentate in the enol form. Thermal analysis, UV-Vis spectra and magnetic moment confirmed the geometry around the copper center to be tetrahedral, square pyramidal and octahedral. Study of the binding ability of the synthesized compounds with Circulating tumor DNA (CT-DNA) bas been evaluated applying UV-Vis spectral titration and viscosity measurements. The copper and zinc oxides were achieved from the copper and zinc nano-particles structures Schiff base complexes as the raw material after calcination for 5 hr at 600°C. On the other hand, synthesized of C1 and C2 NPs were used as suitable precursors to the preparation of CuO and ZnO NPs. Finally, the synthesized of the two complexes exhibited enhanced activity against the tested bacterial (Staphylococcus aureus and Escherichia Coli) and fungal strains (Candida albicans and Aspergillus fumigatus) as compared to HPIMC. Among all these synthesized compounds, C1 exhibits good cleaving ability compared to other newly synthesized C2.     

Properties of bounded stochastic processes employed in biophysics

Abstract

Realistic stochastic modeling is increasingly requiring the use of bounded noises. In this work, properties and relationships of commonly employed bounded stochastic processes are investigated within a solid mathematical ground. Four families are object of investigation: the Sine-Wiener (SW), the Doering–Cai–Lin (DCL), the Tsallis–Stariolo–Borland (TSB), and the Kessler–Sørensen (KS) families. We address mathematical questions on existence and uniqueness of the processes defined through Stochastic Differential Equations, which often conceal non-obvious behavior, and we explore the behavior of the solutions near the boundaries of the state space. The expression of the time-dependent probability density of the Sine-Wiener noise is provided in closed form, and a close connection with the Doering–Cai–Lin noise is shown. Further relationships among the different families are explored, pathwise and in distribution. Finally, we illustrate an analogy between the Kessler–Sørensen family and Bessel processes, which allows to relate the respective local times at the boundaries.     

Design and development of a two‐dimensional system based on hydrophilic and reversed‐phase liquid chromatography with on‐line sample treatment for the simultaneous separation of excreted xenobiotics and endogenous metabolites in urine

In the present work we describe a two‐dimensional liquid chromatographic system (2D‐LC) with detection by mass spectrometry (MS) for the simultaneous separation of endogenous metabolites of clinical interest and excreted xenobiotics deriving from exposure to toxic compounds. The 2D‐LC system involves two orthogonal chromatographic modes, hydrophilic interaction liquid chromatography (HILIC) to separate polar endogenous metabolites and reversed‐phase (RP) chromatography to separate excreted xenobiotics of low and intermediate polarity. Additionally, the present proposal has the novelty of incorporating an on‐line sample treatment based on the use of restricted access materials (RAMs), which permits the direct injection of urine samples into the system. The work is focused on the instrumental coupling, studying all possible options and attempting to circumvent the problems of solvent incompatibility between the RAM device and the two chromatographic columns, HILIC and RP. The instrumental configuration developed, RAM‐HILIC‐RPLC‐MS/MS, allows the simultaneous assessment of urinary metabolites of clinical interest and excreted compounds derived from exposure to toxic agents with minimal sample manipulation. Thus, it may be of interest in areas such as occupational and environmental toxicology in order to explore the possible relationship between the two types of compounds. Copyright © 2015 John Wiley & Sons, Ltd.     

一阶最优性条件研究

本对由Botsko的关于多变量函数取极值的一阶导数检验条件定理^[1]进行了分析研究，给出了更实用而简捷的差别条件。最后，举出若干例子予以说明。     

ON THE CONVERGENCE OF MULTIPLICATIVE ITERATIVE ALGORITHMS WITH INEXACT LINE SEARCH

1.IntroductionIntillspaperweanalyzetheconvergenceonmultiplicativeiterativealgorithmsfortheIninimizationofadiffcrentiablefunctiondefinedonthepositiveorthantofR".ThealgorithmissllggestedbyEggermolltl'],andisrelatedtotheEM[2](Expextation--Maximization)algoritllnlforPositronemissiontonlography[']andimagereconstructi..14].Wecollsidertheproblenl"linf(x)s.t.x20.Themultiplicativeiterativealgorithmshavethel'orlniforj=l,2,',n,withAhdeterminedthroughalinesearch.Whilelusem[5]establishedanelegantconv…