Showing papers by "Technical University of Dortmund published in 1999"

Combining Statistical Learning with a Knowledge-Based Approach - A Case Study in Intensive Care Monitoring

Abstract: The paper describes a case study in combining di erent methods for acquiring medical knowledge. Given a huge amount of noisy, high dimensional numerical time series data describing patients in intensive care, the support vector machine is used to learn when and how to change the dose of which drug. Given medical knowledge about and expertise in clinical decision making, a rst-order logic knowledge base about e ects of therapeutical interventions has been built. As a preprocessing mechanism it uses another statistical method. The integration of numerical and knowledge-based procedures eases the task of validation in two ways. On one hand, the knowledge base is validated with respect to past patients' records. On the other hand, medical interventions that are recommended by learning results are justi ed by the knowledge base.

Pionic parton distributions revisited

Abstract: Using constituent quark model constraints we calculate the gluon and sea–quark content of pions solely in terms of their valence density (fixed by \(\pi N\) Drell–Yan data) and the known sea and gluon distributions of the nucleon, using the most recent updated valence–like input parton densities of the nucleon. The resulting small–x dynamical QCD predictions for \(g^{\pi}(x,Q^2)\) and \(\bar{q} ^{\pi}(x,Q^2)\) are unique and parameter free. Simple analytic parametrizations of the resulting parton distributions of the pion are presented in LO and NLO. These results and parametrizations will be important, among other things, for updated formulations of the parton distributions of real and virtual photons.

Stochastic Tunneling Approach for Global Minimization of Complex Potential Energy Landscapes

Abstract: We investigate a novel stochastic technique for the global optimization of complex potential energy surfaces (PES) that avoids the freezing problem of simulated annealing by allowing the dynamical process to tunnel energetically inaccessible regions of the PES by way of a dynamically adjusted nonlinear transformation of the original PES. We demonstrate the success of this approach, which is characterized by a single adjustable parameter, for three generic hard minimization problems.

Top-quark condensation.

Abstract: Top quark condensation, in particular the minimal framework where the neutral Higgs scalar is (predominantly) an effective ¯ tt condensate of the stan- dard model, is reviewed. Computational approaches are compared and sim- ilarities, differences and deficiencies pointed out. Extensions of the minimal framework, including scenarios with two composite Higgs doublets, additional neutrino condensates, andcondensation arising from four-fermion interac- tions with enlarged symmetries, are described. Possible renormalizable mod- els of underlying physics potentially responsible for the condensation, includ- ing topcolor assisted technicolor frameworks, are discussed. Phenomenological implications of top condensate models are outlined. Outstanding theoretical issues and problems for future investigation are pointed out. Progress in the field after this article was accepted has been briefly covered in a Note added at the end.

The distribution of points on the sphere and corresponding cubature formulae

Abstract: In applications, for instance in optics and astrophysics, there is a need for high-accuracy integration formulae for functions on the sphere. To construct better formulae than previously used, almost equidistantly spaced nodes on the sphere and weights belonging to these nodes are required. This problem is closely related to an optimal dispersion problem on the sphere and to the theories of spherical designs and multivariate Gauss quadrature formulae. We propose a two-stage algorithm to compute optimal point locations on the unit sphere and an appropriate algorithm to calculate the corresponding weights of the cubature formulae. Points as well as weights are computed to high accuracy. These algorithms can be extended to other integration problems. Numerical examples show that the constructed formulae yield impressively small integration errors of up to 10 -12 .

Limb Darkening of a K Giant in the Galactic Bulge: PLANET Photometry of MACHO 97-BLG-28

Abstract: We present the PLANET photometric data set10 for the binary-lens microlensing event MACHO 97- BLG-28, consisting of 696 I- and V -band measurements, and analyze it to determine the radial surface brightness pro—le of the Galactic bulge source star. The microlensed source, demonstrated to be a K giant by our independent spectroscopy, crossed an isolated cusp of the central caustic of the lensing binary, generating a sharp peak in the light curve that was well-resolved by dense (3¨30 minute) and continuous monitoring from PLANET sites in Chile, South Africa, and Australia. This is the —rst time that such a cusp crossing has been observed. Analysis of the PLANET data set has produced a measure- ment of the square-root limb-darkening coefficients of the source star in the I and V bands; the resulting stellar pro—les are in excellent agreement with those predicted by stellar atmospheric models for K giants. The limb-darkening coefficients presented here are the —rst derived from microlensing. They are also among the —rst found for normal giants by any technique and the —rst for any star as distant as the Galactic bulge. Modeling of our light curve for MACHO 97-BLG-28 indicates that the lensing binary has a mass ratio q \ 0.23 and an (instantaneous) separation in units of the angular Einstein ring radius of d \ 0.69. For a lens in the Galactic bulge, this corresponds to a typical stellar binary with a projected separation between 1 and 2 AU. If the lens lies closer (i.e., in the Galactic disk), the separation is smaller, and one or both of the lens objects is in the brown dwarf regime. Assuming that the source is a bulge K2 giant at 8 kpc, the relative lens-source proper motion is k \ 19.4 ^ 2.6 km s~1 kpc~1, consistent with a disk or bulge lens. If the nonlensed blended light is due to a single star, it is likely to be a young white dwarf in the bulge, consistent with the blended light coming from the lens itself. Subject headings: binaries: visualgravitational lensingstars: fundamental parameters ¨ stars: late-type

Excitons, biexcitons, and phonons in ultrathin CdSe/ZnSe quantum structures

Abstract: The optical properties of CdSe nanostructures grown by migration-enhanced epitaxy of CdSe on ZnSe are studied by time-, energy-, and temperature-dependent photoluminescence and excitation spectroscopy, as well as by polarization-dependent four-wave mixing and two-photon absorption experiments. The nanostructures consist of a coherently strained Zn1−xCdxSe/ZnSe quantum well with embedded islands of higher Cd content with sizes of a few nanometer due to strain-induced CdSe accumulation. The local increase in CdSe concentration results in a strong localization of the excitonic wave function, in an increase in radiative lifetime, and a decrease of the dephasing rate. Local LO-phonon modes caused by the strong modulation of the Cd concentration profile are found in phonon-assisted relaxation processes. Confined biexcitons with large binding energies between 20 and 24 meV are observed, indicating the important role of biexcitons even at room temperature.

An Automata-Theoretic Approach to Interprocedural Data-Flow Analysis

Abstract: We show that recent progress in extending the automata-theoretic approach to model-checking beyond the class of finite-state processes finds a natural application in the area of interprocedural data-flow analysis.

Nash moser methods for the solution of quasilinear schrödinger equations

Abstract: Using new Nash Moser techniques for Frechet spaces by M. Poppenberg the local existence, uniqueness and continuous dependence of smooth solutions of a special quasilinear evolutionary Schrodinger equation is proved; as basic function space H∞ (IRn), the intersection of all Sobolev spaces Hk(IRn), is used. The method consists in finding an appropriate linearization of the given nonlinear Schrodinger equation, and proving that this linear Schrodinger equation admits a strongly continuous evolution operator which provides the necessary a priori estimates for any derivative; this is shown by a transformation procedure using a time dependent metric which overcomes the difficulty arising from nondissipativity of the linearized Schrodinger equation

Effects of (N7)-Coordinated Nickel(II), Copper(II), or Platinum(II) on the Acid–Base Properties of Guanine Derivatives and Other Related Purines[≠]

Abstract: Guanine is an important nucleobase residue in nucleic acids; for example, the anticancer drug cisplatin, cis-[(NH3)2PtCl2], preferably binds to N7 of this site. In this study the effect of (N7)-coordinated Ni2+ (which is expected to correspond to that of Zn2+), Cu2+, and cis-a2Pt2+ or trans-a2Pt2+ (where a = NH3 or CH3NH2) on the acid–base properties of guanine derivatives (see diagram) is quantified. Several related purines are included for comparison. Micro-acidity-constant evaluations are presented and more than 60 acidity constants are listed.

Radiatively generated parton distributions of real and virtual photons

Abstract: The parton content of real ${(P}^{2}=0)$ and virtual ${(P}^{2}\ensuremath{ e}0)$ transverse photons $\ensuremath{\gamma}{(P}^{2})$ is expressed in terms of perturbative pointlike and nonperturbative hadronic (VMD) components, employing recently updated parton distributions of pions and protons. The resulting parameter-free and perturbatively stable LO and NLO parton densities ${f}^{\ensuremath{\gamma}{(P}^{2})}{(x,Q}^{2})$ are smooth in ${P}^{2}$ and apply to all ${P}^{2}g~0$ whenever $\ensuremath{\gamma}{(P}^{2})$ is probed at scales ${Q}^{2}\ensuremath{\gg}{P}^{2}$ where transverse photons also dominate physically relevant cross sections. Predictions are given for the structure function ${F}_{2}^{\ensuremath{\gamma}{(P}^{2})}{(x,Q}^{2})$ and ${f}^{\ensuremath{\gamma}{(P}^{2})}{(x,Q}^{2}),$ and are compared with all relevant data for real photons as well as with recent data for virtual photons as extracted from DIS $\mathrm{ep}$ dijet events. Simple analytic parametrizations of our predicted parton distributions are presented for the real photon in LO and NLO, and for the virtual photon in LO which, within sufficient accuracy, may be also used in NLO-QCD.

Bistable micromechanical fiber-optic switches on silicon with thermal actuators

Abstract: A new bistable moving-fiber switch for applications in optical communication systems is developed by employing advanced bulk silicon micromachining. The bistable switch utilizes actuators based on thermal expansion and the bimetal effect. Switches and switch arrays are successfully fabricated and tested. Due to the low number of process steps and the high yield and reliability, the process technology is suitable for low-cost single- and multi-mode fiber optic switches. A typical fiber-to-fiber insertion loss of 2±0.5 dB (single-mode fibers at 1300 nm wavelength, including MT-connector losses of typically 0.5–1.5 dB) is achieved in an array of 12 switches. The cross talk is below the detection limit (

How asymmetry helps load balancing

Abstract: This paper deals with balls and bins processes related to randomized load balancing, dynamic resource allocation and hashing. Suppose n balls have to be assigned to n bins, where each ball has to be placed without knowledge about the distribution of previously placed balls. The goal is to achieve an allocation that is as even as possible so that no bin gets much more balls than the average. A well known and good solution for this problem is to choose d possible locations for each ball at random, to look into each of these bins, and to place the ball into the least full among these bins. This class of algorithms has been investigated intensively in the past but almost all previous analyses assume that the d locations for each ball are chosen uniform and independently at random from the set of all bins. We investigate whether a non-uniform and possibly dependent choice of the d locations for a ball can improve the load balancing. Three types of selections are distinguished: 1) uniform and independent 2) non-uniform and independent 3) non-uniform and dependent. Our first result shows that choosing the locations in a non-uniform way (type 2) results in a better load balancing than choosing the locations uniformly (type 1). Surprising, this smooth load balancing is obtained by an algorithm called "Always-Go-Left" which creates an asymmetric assignment of the balls to the bins. Our second result is a lower bound on the smallest-possible maximum load that can be achieved by any allocation algorithm of type 1, 2, or 3.

Formation of ice XII at different conditions

Abstract: The structural versatility of solid water has been reinforced by the identification of a metastable, crystalline phase called ice XII (ref. 1). It was obtained by cooling liquid water to 260 K at a pressure of 0.55 GPa. We have found ice XII in a completely different region of water's phase diagram and show that it can be formed under different conditions from those previously reported1.

METAFrame in Practice: Design of Intelligent Network Services

Abstract: In this paper we present MetaFrame, an environment for formal methods-based, application-specific software design. Characteristic for MetaFrame are the following features: library-based development, meaning software construction by combination of components on a coarse granular level, incremental formalization, through successive enrichment of a special-purpose development environment, and library-based consistency checking, allowing continuous verification of application- and purpose-specific properties by means of model checking. These features and their impact for application developers and end users will be illustrated along an industrial application, the design of intelligent network (IN) services.

Quantum cluster equilibrium theory of liquids: molecular clusters and thermodynamics of liquid ethanol

Abstract: Quantum cluster equilibrium (QCE) theory is presented for liquid ethanol. The cluster equilibria that dictate phase composition are determined by the rigorous techniques of quantum statistical thermodynamics in the canonical ensemble, based on the ab initio partition function. The characteristic features of the supramolecular clusters which comprise the QCE model are discussed in terms of binding energies, geometries, cooperativity and charge transfer. The validity of the resulting QCE model is demonstrated by comparison with experimental ther-modynamic data: Clausius—Clapeyron pressure/temperature dependence and the specific heat. At room temperature, neat, liquid ethanol consists of approximately equal parts of monomer, cyclic tetramer and cyclic pentamer clusters.

Simulation Study on the Diffusive Motion in Deeply Supercooled Water

Abstract: The diffusion process in supercooled liquid water has been studied by a series of molecular dynamics simulations using systems of 216 ST2-model molecules in the temperature range between 255 and 360 K. In contrast to SPC/E water (Phys. Rev. E 1996, 54, 6331), the ST2 liquid does not show the appearance of a kinetic glass transition with structural arrest. Instead, we observe a significant change of the diffusion mechanism in the deeply supercooled region. The high-temperature microstep diffusion mechanism transforms continuously into a jump-diffusive behavior at low temperatures. By analyzing the intermediate incoherent structure factor of the center of mass motion we can characterize the hopping process by an average residence time τ0. The hopping process is found to control reorientational and translational motions, leading to a second region of apparent Arrhenius behavior with a high activation energy of about 115 kJ mol-1. The change in the dynamic behavior occurs in parallel to the structural and the...