Extension of TSVM to multi-class and hierarchical text classification problems With general losses

Transductive SVM (TSVM) is a well known semi-supervised large margin learning method for binary text classification. In this paper we extend this method to multi-class and hierarchical classification problems. We point out that the determination of labels of unlabeled examples with fixed classifier weights is a linear programming problem. We devise an efficient technique for solving it. The method is applicable to general loss functions. We demonstrate the value of the new method using large margin loss on a number of multi-class and hierarchical classification datasets. For maxent loss we show empirically that our method is better than expectation regularization/constraint and posterior regularization methods, and competitive with the version of entropy regularization method which uses label constraints.

COMPUTATIONAL MODELING OF A MULTI-LAYERED PIEZO-COMPOSITE BEAM MADE UP OF MFC

The Variational Asymptotic Method (VAM) is used for modeling a coupled non-linear electromechanical problem finding applications in aircrafts and Micro Aerial Vehicle (MAV) development. VAM coupled with geometrically exact kinematics forms a powerful tool for analyzing a complex nonlinear phenomena as shown previously by many in the literature 3 - 7] for various challenging problems like modeling of an initially twisted helicopter rotor blades, matrix crack propagation in a composite, modeling of hyper elastic plates and various multi-physics problems. The problem consists of design and analysis of a piezocomposite laminate applied with electrical voltage(s) which can induce direct and planar distributed shear stresses and strains in the structure. The deformations are large and conventional beam theories are inappropriate for the analysis. The behavior of an elastic body is completely understood by its energy. This energy must be integrated over the cross-sectional area to obtain the 1-D behavior as is typical in a beam analysis. VAM can be used efficiently to approximate 3-D strain energy as closely as possible. To perform this simplification, VAM makes use of thickness to width, width to length, width multiplied by initial twist and strain as small parameters embedded in the problem definition and provides a way to approach the exact solution asymptotically. In this work, above mentioned electromechanical problem is modeled using VAM which breaks down the 3-D elasticity problem into two parts, namely a 2-D non-linear cross-sectional analysis and a 1-D non-linear analysis, along the reference curve. The recovery relations obtained as a by-product in the cross-sectional analysis earlier are used to obtain 3-D stresses, displacements and velocity contours. The piezo-composite laminate which is chosen for an initial phase of computational modeling is made up of commercially available Macro Fiber Composites (MFCs) stacked together in an arbitrary lay-up and applied with electrical voltages for actuation. The expressions of sectional forces and moments as obtained from cross-sectional analysis in closed-form show the electro-mechanical coupling and relative contribution of electric field in individual layers of the piezo-composite laminate. The spatial and temporal constitutive law as obtained from the cross-sectional analysis are substituted into 1-D fully intrinsic, geometrically exact equilibrium equations of motion and 1-D intrinsic kinematical equations to solve for all 1-D generalized variables as function of time and an along the reference curve co-ordinate, x(1).

Onset of plane layer magnetoconvection at low Ekman number

We study the onset of magnetoconvection between two infinite horizontal planes subject to a vertical magnetic field aligned with background rotation. In order to gain insight into the convection taking place in the Earth's tangent cylinder, we target regimes of asymptotically strong rotation. The critical Rayleigh number Ra-c and critical wavenumber a(c) are computed numerically by solving the linear stability problem in a systematic way, with either stress-free or no-slip kinematic boundary conditions. A parametric study is conducted, varying the Ekman number E (ratio of viscous to Coriolis forces) and the Elsasser number. (ratio of the Lorentz force to the Coriolis force). E is varied from 10(-9) to 10(-2) and. from 10(-3) to 1. For a wide range of thermal and magnetic Prandtl numbers, our results verify and confirm previous experimental and theoretical results showing the existence of two distinct unstable modes at low values of E-one being controlled by the magnetic field, the other being controlled by viscosity (often called the viscous mode). It is shown that oscillatory onset does not occur in the range of parameters we are interested in. Asymptotic scalings for the onset of these modes are numerically confirmed and their domain of validity is precisely quantified. We show that with no-slip boundary conditions, the asymptotic behavior is reached for E < 10(-6) and establish a map in the (E, Lambda) plane. We distinguish regions where convection sets in either through the magnetic mode or through the viscous mode. Our analysis gives the regime in which the transition between magnetic and viscous modes may be observed. We also show that within the asymptotic regime, the role played by the kinematic boundary conditions is minimal. (C) 2015 AIP Publishing LLC.

Instability of solution of phase retrieval in direct diffraction phase-contrast imaging with partially coherent X-ray source

The theoretical model of direct diffraction phase-contrast imaging with partially coherent x-ray source is expressed by an operator of multiple integral. It is presented that the integral operator is linear. The problem of its phase retrieval is described by solving an operator equation of multiple integral. It is demonstrated that the solution of the phase retrieval is unstable. The numerical simulation is performed and the result validates that the solution of the phase retrieval is unstable.

Otimização da programação das limpezas em redes de trocadores de calor utilizando técnicas de programação matemática

Redes de trocadores de calor são bastante utilizadas na indústria química para promover a integração energética do processo, recuperando calor de correntes quentes para aquecer correntes frias. Estas redes estão sujeitas à deposição, o que causa um aumento na resistência à transferência de calor, prejudicando-a. Uma das principais formas de diminuir o prejuízo causado por este fenômeno é a realização periódica de limpezas nos trocadores de calor. O presente trabalho tem como objetivo desenvolver um novo método para encontrar a programação ótima das limpezas em uma rede de trocadores de calor. O método desenvolvido utiliza o conceito de horizonte deslizante associado a um problema de programação linear inteira mista (MILP). Este problema MILP é capaz de definir o conjunto ótimo de trocadores de calor a serem limpos em um determinado instante de tempo (primeiro instante do horizonte deslizante), levando em conta sua influência nos instantes futuros (restante do horizonte deslizante). O problema MILP utiliza restrições referentes aos balanços de energia, equações de trocadores de calor e número máximo de limpezas simultâneas, com o objetivo de minimizar o consumo de energia da planta. A programação ótima das limpezas é composta pela combinação dos resultados obtidos em cada um dos instantes de tempo.O desempenho desta abordagem foi analisado através de sua aplicação em diversos exemplos típicos apresentados na literatura, inclusive um exemplo de grande porte de uma refinaria brasileira. Os resultados mostraram que a abordagem aplicada foi capaz de prover ganhos semelhantes e, algumas vezes, superiores aos da literatura, indicando que o método desenvolvido é capaz de fornecer bons resultados com um baixo esforço computacional

Syntheses and reactions of heterometallic complexes of nickel and copper with tetrathiotungstates. The molecular structures of [Ph(4)P](2)[S2W(mu-S)(2)Ni(S-2)] and W(mu-S)(4)Cu-2(PPh(3))(4)center dot py

Reaction of [Ph(4)P]2WS4 With NiCl2 in methanol solution in the presence of NaOCH3 leads to the formation of [Ph(4)P](2) [S2W(mu-S)(2)Ni(S-2)] (I) A Similar reaction between (NH4)(2)WS4 and NiCl2 under O-2 atmosphere in the presence of Ph(4)PCl or (n)Bu(4)NCl affords [Ph(4)P](2)([(S-2)W(O)(mu-S)(2)]Ni-2] (IIa) and [(n)Bu(4)N](2)([(S-2)W(O)(mu-S)(2)]Ni-2} (IIb) Under argon the same reaction gives [Ph(4)P](2)[Ni(WS4)(2)] (IIIa) and [(n)Bu(4)N](2)[Ni(WS4)(2)] (IIIb). [Ph(4)P](2)[Ni(WOS3)(2)] (IV) and [Ph(4)P](2)[Ni(WO2S2)(2)] (V) can be prepared from the reaction of [Ph(4)P]2WOS3 and [Ph(4)P]2WO2S2 with NiCl2. Treatment of (NH4)(2)WS4 with CuCl in the presence of PPh(3) in boiling pyridine produces W(mu-S)(4)Cu-2(PPh(3))(3) (VI), which can further react with excess PPh(3) to give W(mu-S)(4)Cu-2(PPh(3))(4) . py (VII). Complex I crystallizes in the space group P2(1)/n with the cell parameters: a = 20.049(4), b = 17.010(4), c = 14.311(7) Angstrom; beta = 110.24(3)degrees and Z = 4; R = 0.058 for 4267 independent reflections. The structural study confirms that complex I contains two terminal sulfide ligands, two bridging sulfide ligands, a side-on disulfide ligand, and a planar central W(mu-S)(2)Ni four membered ring. Complex VII crystallizes in the space group C2/c with the cell parameters: a = 26.436(8), b = 20.542(6), c = 19.095(8) Angstrom; beta = 125.00(3)degrees and Z = 4; R = 0.080 for 3802 independent reflections. The structural study reveals a perfect linear arrangement of the three metal atoms Cu-W-Cu.

电阻率测井反演方法应用研究

Formation resistivity is one of the most important parameters to be evaluated in the evaluation of reservoir. In order to acquire the true value of virginal formation, various types of resistivity logging tools have been developed. However, with the increment of the proved reserves, the thickness of interest pay zone is becoming thinner and thinner, especially in the terrestrial deposit oilfield, so that electrical logging tools, limited by the contradictory requirements of resolution and investigation depth of this kinds of tools, can not provide the true value of the formation resistivity. Therefore, resitivity inversion techniques have been popular in the determination of true formation resistivity based on the improving logging data from new tools. In geophysical inverse problems, non-unique solution is inevitable due to the noisy data and deficient measurement information. I address this problem in my dissertation from three aspects, data acquisition, data processing/inversion and applications of the results/ uncertainty evaluation of the non-unique solution. Some other problems in the traditional inversion methods such as slowness speed of the convergence and the initial-correlation results. Firstly, I deal with the uncertainties in the data to be processed. The combination of micro-spherically focused log (MSFL) and dual laterolog(DLL) is the standard program to determine formation resistivity. During the inversion, the readings of MSFL are regarded as the resistivity of invasion zone of the formation after being corrected. However, the errors can be as large as 30 percent due to mud cake influence even if the rugose borehole effects on the readings of MSFL can be ignored. Furthermore, there still are argues about whether the two logs can be quantitatively used to determine formation resisitivities due to the different measurement principles. Thus, anew type of laterolog tool is designed theoretically. The new tool can provide three curves with different investigation depths and the nearly same resolution. The resolution is about 0.4meter. Secondly, because the popular iterative inversion method based on the least-square estimation can not solve problems more than two parameters simultaneously and the new laterolog logging tool is not applied to practice, my work is focused on two parameters inversion (radius of the invasion and the resistivty of virgin information ) of traditional dual laterolog logging data. An unequal weighted damp factors- revised method is developed to instead of the parameter-revised techniques used in the traditional inversion method. In this new method, the parameter is revised not only dependency on the damp its self but also dependency on the difference between the measurement data and the fitting data in different layers. At least 2 iterative numbers are reduced than the older method, the computation cost of inversion is reduced. The damp least-squares inversion method is the realization of Tikhonov's tradeoff theory on the smooth solution and stability of inversion process. This method is realized through linearity of non-linear inversion problem which must lead to the dependency of solution on the initial value of parameters. Thus, severe debates on efficiency of this kinds of methods are getting popular with the developments of non-linear processing methods. The artificial neural net method is proposed in this dissertation. The database of tool's response to formation parameters is built through the modeling of the laterolog tool and then is used to training the neural nets. A unit model is put forward to simplify the dada space and an additional physical limitation is applied to optimize the net after the cross-validation method is done. Results show that the neural net inversion method could replace the traditional inversion method in a single formation and can be used a method to determine the initial value of the traditional method. No matter what method is developed, the non-uniqueness and uncertainties of the solution could be inevitable. Thus, it is wise to evaluate the non-uniqueness and uncertainties of the solution in the application of inversion results. Bayes theorem provides a way to solve such problems. This method is illustrately discussed in a single formation and achieve plausible results. In the end, the traditional least squares inversion method is used to process raw logging data, the calculated oil saturation increased 20 percent than that not be proceed compared to core analysis.

Techniques for engineering quantum states

In this thesis I theoretically study quantum states of ultracold atoms. The majority of the Chapters focus on engineering specific quantum states of single atoms with high fidelity in experimentally realistic systems. In the sixth Chapter, I investigate the stability and dynamics of new multidimensional solitonic states that can be created in inhomogeneous atomic Bose-Einstein condensates. In Chapter three I present two papers in which I demonstrate how the coherent tunnelling by adiabatic passage (CTAP) process can be implemented in an experimentally realistic atom chip system, to coherently transfer the centre-of-mass of a single atom between two spatially distinct magnetic waveguides. In these works I also utilise GPU (Graphics Processing Unit) computing which offers a significant performance increase in the numerical simulation of the Schrödinger equation. In Chapter four I investigate the CTAP process for a linear arrangement of radio frequency traps where the centre-of-mass of both, single atoms and clouds of interacting atoms, can be coherently controlled. In Chapter five I present a theoretical study of adiabatic radio frequency potentials where I use Floquet theory to more accurately model situations where frequencies are close and/or field amplitudes are large. I also show how one can create highly versatile 2D adiabatic radio frequency potentials using multiple radio frequency fields with arbitrary field orientation and demonstrate their utility by simulating the creation of ring vortex solitons. In the sixth Chapter I discuss the stability and dynamics of a family of multidimensional solitonic states created in harmonically confined Bose-Einstein condensates. I demonstrate that these solitonic states have interesting dynamical instabilities, where a continuous collapse and revival of the initial state occurs. Through Bogoliubov analysis, I determine the modes responsible for the observed instabilities of each solitonic state and also extract information related to the time at which instability can be observed.

Three-dimensional dispersive metallic photonic crystals with a bandgap and a high cutoff frequency.

The goal of this work is to analyze three-dimensional dispersive metallic photonic crystals (PCs) and to find a structure that can provide a bandgap and a high cutoff frequency. The determination of the band structure of a PC with dispersive materials is an expensive nonlinear eigenvalue problem; in this work we propose a rational-polynomial method to convert such a nonlinear eigenvalue problem into a linear eigenvalue problem. The spectral element method is extended to rapidly calculate the band structure of three-dimensional PCs consisting of realistic dispersive materials modeled by Drude and Drude-Lorentz models. Exponential convergence is observed in the numerical experiments. Numerical results show that, at the low frequency limit, metallic materials are similar to a perfect electric conductor, where the simulation results tend to be the same as perfect electric conductor PCs. Band structures of the scaffold structure and semi-woodpile structure metallic PCs are investigated. It is found that band structures of semi-woodpile PCs have a very high cutoff frequency as well as a bandgap between the lowest two bands and the higher bands.

Solution structure of the N-terminal transactivation domain of ERM modified by SUMO-1.

ERM is a member of the PEA3 group of the Ets transcription factor family that plays important roles in development and tumorigenesis. The PEA3s share an N-terminal transactivation domain (TADn) whose activity is inhibited by small ubiquitin-like modifier (SUMO). However, the consequences of sumoylation and its underlying molecular mechanism remain unclear. The domain structure of ERM TADn alone or modified by SUMO-1 was analyzed using small-angle X-ray scattering (SAXS). Low resolution shapes determined ab initio from the scattering data indicated an elongated shape and an unstructured conformation of TADn in solution. Covalent attachment of SUMO-1 does not perturb the structure of TADn as indicated by the linear arrangement of the SUMO moiety with respect to TADn. Thus, ERM belongs to the growing family of proteins that contain intrinsically unstructured regions. The flexible nature of TADn may be instrumental for ERM recognition and binding to diverse molecular partners.

Representing temporal knowledge in legal discourse

This paper presents a formalism for representing temporal knowledge in legal discourse that allows an explicit expression of time and event occurrences. The fundamental time structure is characterized as a well‐ordered discrete set of primitive times, i.e. non‐decomposable intervals with positive duration or points with zero duration), from which decomposable intervals can be constructed. The formalism supports a full representation of both absolute and relative temporal knowledge, and a formal mechanism for checking the temporal consistency of a given set of legal statements is provided. The general consistency checking algorithm which addresses both absolute and relative temporal knowledge turns out to be a linear programming problem, while in the special case where only relative temporal relations are involved, it becomes a simple question of searching for cycles in the graphical representation of the corresponding legal text.

Simulation of 2-D metal cutting by means of a distributed algorithm

Temperature distributions involved in some metal-cutting or surface-milling processes may be obtained by solving a non-linear inverse problem. A two-level concept on parallelism is introduced to compute such temperature distribution. The primary level is based on a problem-partitioning concept driven by the nature and properties of the non-linear inverse problem. Such partitioning results to a coarse-grained parallel algorithm. A simplified 2-D metal-cutting process is used as an example to illustrate the concept. A secondary level exploitation of further parallel properties based on the concept of domain-data parallelism is explained and implemented using MPI. Some experiments were performed on a network of loosely coupled machines consist of SUN Sparc Classic workstations and a network of tightly coupled processors, namely the Origin 2000.

Scheduling parallel dedicated machines with the speeding-up resource

We consider a problem of scheduling jobs on m parallel machines. The machines are dedicated, i.e., for each job the processing machine is known in advance. We mainly concentrate on the model in which at any time there is one unit of an additional resource. Any job may be assigned the resource and this reduces its processing time. A job that is given the resource uses it at each time of its processing. No two jobs are allowed to use the resource simultaneously. The objective is to minimize the makespan. We prove that the two-machine problem is NP-hard in the ordinary sense, describe a pseudopolynomial dynamic programming algorithm and convert it into an FPTAS. For the problem with an arbitrary number of machines we present an algorithm with a worst-case ratio close to 3/2, and close to 3, if a job can be given several units of the resource. For the problem with a fixed number of machines we give a PTAS. Virtually all algorithms rely on a certain variant of the linear knapsack problem (maximization, minimization, multiple-choice, bicriteria). © 2008 Wiley Periodicals, Inc. Naval Research Logistics, 2008

Adaptive calibration of a three-microphone system for acoustic waveguide characterization under time-varying conditions

The pressure and velocity field in a one-dimensional acoustic waveguide can be sensed in a non-intrusive manner using spatially distributed microphones. Experimental characterization with sensor arrangements of this type has many applications in measurement and control. This paper presents a method for measuring the acoustic variables in a duct under fluctuating propagation conditions with specific focus on in-system calibration and tracking of the system parameters of a three-microphone measurement configuration. The tractability of the non-linear optimization problem that results from taking a parametric approach is investigated alongside the influence of extraneous measurement noise on the parameter estimates. The validity and accuracy of the method are experimentally assessed in terms of the ability of the calibrated system to separate the propagating waves under controlled conditions. The tracking performance is tested through measurements with a time-varying mean flow, including an experiment conducted under propagation conditions similar to those in a wind instrument during playing.

An optimal algorithm for low power multiplierless FIR filter design using chebychev criterion

In this paper, we propose a novel finite impulse response (FIR) filter design methodology that reduces the number of operations with a motivation to reduce power consumption and enhance performance. The novelty of our approach lies in the generation of filter coefficients such that they conform to a given low-power architecture, while meeting the given filter specifications. The proposed algorithm is formulated as a mixed integer linear programming problem that minimizes chebychev error and synthesizes coefficients which consist of pre-specified alphabets. The new modified coefficients can be used for low-power VLSI implementation of vector scaling operations such as FIR filtering using computation sharing multiplier (CSHM). Simulations in 0.25um technology show that CSHM FIR filter architecture can result in 55% power and 34% speed improvement compared to carry save multiplier (CSAM) based filters.