Sandia National Laboratories California

An approach to the construction of classifiers from imbalanced datasets is described. A dataset is imbalanced if the classification categories are not approximately equally represented. Often real-world data sets are predominately composed of ``normal'' examples with only a small percentage of ``abnormal'' or ``interesting'' examples. It is also the case that the cost of misclassifying an abnormal (interesting) example as a normal example is often much higher than the cost of the reverse error. Under-sampling of the majority (normal) class has been proposed as a good means of increasing the sensitivity of a classifier to the minority class. This paper shows that a combination of our method of over-sampling the minority (abnormal) class and under-sampling the majority (normal) class can achieve better classifier performance (in ROC space) than only under-sampling the majority class. This paper also shows that a combination of our method of over-sampling the minority class and under-sampling the majority class can achieve better classifier performance (in ROC space) than varying the loss ratios in Ripper or class priors in Naive Bayes. Our method of over-sampling the minority class involves creating synthetic minority class examples. Experiments are performed using C4.5, Ripper and a Naive Bayes classifier. The method is evaluated using the area under the Receiver Operating Characteristic curve (AUC) and the ROC convex hull strategy.

This survey provides an overview of higher-order tensor decompositions, their applications, and available software. A tensor is a multidimensional or N-way array. Decompositions of higher-order tensors (i.e., N-way arrays with $N \geq 3$) have applications in psycho-metrics, chemometrics, signal processing, numerical linear algebra, computer vision, numerical analysis, data mining, neuroscience, graph analysis, and elsewhere. Two particular tensor decompositions can be considered to be higher-order extensions of the matrix singular value decomposition: CANDECOMP/PARAFAC (CP) decomposes a tensor as a sum of rank-one tensors, and the Tucker decomposition is a higher-order form of principal component analysis. There are many other tensor decompositions, including INDSCAL, PARAFAC2, CANDELINC, DEDICOM, and PARATUCK2 as well as nonnegative variants of all of the above. The N-way Toolbox, Tensor Toolbox, and Multilinear Engine are examples of software packages for working with tensors.

Standard sets of supplementary diffuse s and p functions, multiple polarization functions (double and triple sets of d functions), and higher angular momentum polarization functions (f functions) are defined for use with the 6-31G and 6-311G basis sets. Preliminary applications of the modified basis sets to the calculation of the bond energy and hydrogenation energy of N2 illustrate that these functions can be very important in the accurate computation of reaction energies.

The 6-31G* and 6-31G** basis sets previously introduced for first-row atoms have been extended through the second-row of the periodic table. Equilibrium geometries for one-heavy-atom hydrides calculated for the two-basis sets and using Hartree–Fock wave functions are in good agreement both with each other and with the experimental data. HF/6-31G* structures, obtained for two-heavy-atom hydrides and for a variety of hypervalent second-row molecules, are also in excellent accord with experimental equilibrium geometries. No large deviations between calculated and experimental single bond lengths have been noted, in contrast to previous work on analogous first-row compounds, where limiting Hartree–Fock distances were in error by up to a tenth of an angstrom. Equilibrium geometries calculated at the HF/6-31G level are consistently in better agreement with the experimental data than are those previously obtained using the simple split-valance 3-21G basis set for both normal- and hypervalent compounds. Normal-mode vibrational frequencies derived from 6-31G* level calculations are consistently larger than the corresponding experimental values, typically by 10%–15%; they are of much more uniform quality than those obtained from the 3-21G basis set. Hydrogenation energies calculated for normal- and hypervalent compounds are in moderate accord with experimental data, although in some instances large errors appear. Calculated energies relating to the stabilities of single and multiple bonds are in much better accord with the experimental energy differences.

We develop the embedded-atom method [Phys. Rev. Lett. 50, 1285 (1983)], based on density-functional theory, as a new means of calculating ground-state properties of realistic metal systems. We derive an expression for the total energy of a metal using the embedding energy from which we obtain several ground-state properties, such as the lattice constant, elastic constants, sublimation energy, and vacancy-formation energy. We obtain the embedding energy and accompanying pair potentials semiempirically for Ni and Pd, and use these to treat several problems: surface energy and relaxation of the (100), (110), and (111) faces; properties of H in bulk metal (H migration, binding of H to vacancies, and lattice expansion in the hydride phase); binding site and adsorption energy of hydrogen on (100), (110), and (111) surfaces; and lastly, fracture of Ni and the effects of hydrogen on the fracture. We emphasize problems with hydrogen and with surfaces because none of these can be treated with pair potentials. The agreement with experiment, the applicability to practical problems, and the simplicity of the technique make it an effective tool for atomistic studies of defects in metals.

ADVERTISEMENT RETURN TO ISSUEPREVReviewNEXTMetal–Organic Framework Materials as Chemical SensorsLauren E. Kreno†, Kirsty Leong‡, Omar K. Farha†, Mark Allendorf*‡, Richard P. Van Duyne†, and Joseph T. Hupp*†View Author Information† Department of Chemistry, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, United States‡ Sandia National Laboratories, Mail Stop 9291, Livermore, California 94551-0969, United States*Phone: (925) 294-2895 (M.A.); (847) 491-3504 (J.T.H.). E-mail: [email protected] (M.A.); [email protected] (J.T.H.).Cite this: Chem. Rev. 2012, 112, 2, 1105–1125Publication Date (Web):November 9, 2011Publication History Received18 August 2011Published online9 November 2011Published inissue 8 February 2012https://pubs.acs.org/doi/10.1021/cr200324thttps://doi.org/10.1021/cr200324treview-articleACS PublicationsCopyright © 2011 American Chemical SocietyRequest reuse permissionsArticle Views71971Altmetric-Citations6186LEARN ABOUT THESE METRICSArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated. Share Add toView InAdd Full Text with ReferenceAdd Description ExportRISCitationCitation and abstractCitation and referencesMore Options Share onFacebookTwitterWechatLinked InRedditEmail Other access optionsGet e-Alertsclose SUBJECTS:Luminescence,Metal organic frameworks,Molecules,Sensors,Thin films Get e-Alerts

Metal-organic frameworks (MOFs) display a wide range of luminescent behaviors resulting from the multifaceted nature of their structure. In this critical review we discuss the origins of MOF luminosity, which include the linker, the coordinated metal ions, antenna effects, excimer and exciplex formation, and guest molecules. The literature describing these effects is comprehensively surveyed, including a categorization of each report according to the type of luminescence observed. Finally, we discuss potential applications of luminescent MOFs. This review will be of interest to researchers and synthetic chemists attempting to design luminescent MOFs, and those engaged in the extension of MOFs to applications such as chemical, biological, and radiation detection, medical imaging, and electro-optical devices (141 references).

A consistent set of embedding functions and pair interactions for use with the embedded-atom method [M.S. Daw and M. I. Baskes, Phys. Rev. B 29, 6443 (1984)] have been determined empirically to describe the fcc metals Cu, Ag, Au, Ni, Pd, and Pt as well as alloys containing these metals. The functions are determined empirically by fitting to the sublimation energy, equilibrium lattice constant, elastic constants, and vacancy-formation energies of the pure metals and the heats of solution of the binary alloys. The validity of the functions is tested by computing a wide range of properties: the formation volume and migration energy of vacancies, the formation energy, formation volume, and migration energy of divacancies and self-interstitials, the surface energy and geometries of the low-index surfaces of the pure metals, and the segregation energy of substitutional impurities to (100) surfaces.

UNLABELLED: The recovery of genomes from metagenomic datasets is a critical step to defining the functional roles of the underlying uncultivated populations. We previously developed MaxBin, an automated binning approach for high-throughput recovery of microbial genomes from metagenomes. Here we present an expanded binning algorithm, MaxBin 2.0, which recovers genomes from co-assembly of a collection of metagenomic datasets. Tests on simulated datasets revealed that MaxBin 2.0 is highly accurate in recovering individual genomes, and the application of MaxBin 2.0 to several metagenomes from environmental samples demonstrated that it could achieve two complementary goals: recovering more bacterial genomes compared to binning a single sample as well as comparing the microbial community composition between different sampling environments. AVAILABILITY AND IMPLEMENTATION: MaxBin 2.0 is freely available at http://sourceforge.net/projects/maxbin/ under BSD license.

A new, semiempirical model of metals and impurities (embedded atom method) makes possible a static treatment of the brittle fracture of a transition metal in the presence of hydrogen. Results indicate that hydrogen can reduce the fracture stress in nickel.

We model indentation of a metal surface by combining an atomistic metal with a hard-sphere indenter. This work provides atomistic imaging of dislocation nucleation during displacement controlled indentation on a passivated surface. Dislocations and defects are located and imaged by local deviations from centrosymmetry. For a Au(111) surface, nucleation of partial dislocation loops occurs below the surface inside the indenter contact area. We compare and contrast these observations with empirical criteria for dislocation nucleation and corresponding continuum elasticity solutions.

In a comprehensive study, the modified embedded-atom method is extended to a variety of cubic materials and impurities. In this extension, all functions are analytic and computationally simple. The basic equations of the method are developed and applied to 26 elements: ten fcc, ten bcc, three diamond cubic, and three gaseous materials. The materials modeled include metals, semiconductors, and diatomic gases, all of which exhibit different types of bonding. Properties of these materials, including equation of state, elastic moduli, structural energies and lattice constants, simple defects, and surfaces, are calculated. The formalism for applying the method to combinations of these elements is developed and applied to the calculation of dilute heats of solution. In all cases, comparison is made to experiment or higher-level calculations when possible.

Simple and efficient methods of organizing materials are key to the realization of a nanotech world. These authors report on recent developments in simple evaporation-induced self-assembly processes, which enable the rapid production of patterned porous or nanocomposite materials. The Figure shows a calcined particle exhibiting vesicular mesophase, which was generated by aerosol self-assembly of the tri-block copolymer P123.

Diatoms are unicellular algae with plastids acquired by secondary endosymbiosis. They are responsible for approximately 20% of global carbon fixation. We report the 34 million-base pair draft nuclear genome of the marine diatom Thalassiosira pseudonana and its 129 thousand-base pair plastid and 44 thousand-base pair mitochondrial genomes. Sequence and optical restriction mapping revealed 24 diploid nuclear chromosomes. We identified novel genes for silicic acid transport and formation of silica-based cell walls, high-affinity iron uptake, biosynthetic enzymes for several types of polyunsaturated fatty acids, use of a range of nitrogenous compounds, and a complete urea cycle, all attributes that allow diatoms to prosper in aquatic environments.

A package of 38 low level subprograms for many of the basic operations of numerical linear algebra m presented. The package is intended to be used with Fortran. The operations m the package include dot product, elementary vector operation, Givens transformation, vector copy and swap, vector norm, vector scaling, and the determination of the index of the vector component of largest magnitude.

There exists significant demand for improved Reynolds-averaged Navier–Stokes (RANS) turbulence models that are informed by and can represent a richer set of turbulence physics. This paper presents a method of using deep neural networks to learn a model for the Reynolds stress anisotropy tensor from high-fidelity simulation data. A novel neural network architecture is proposed which uses a multiplicative layer with an invariant tensor basis to embed Galilean invariance into the predicted anisotropy tensor. It is demonstrated that this neural network architecture provides improved prediction accuracy compared with a generic neural network architecture that does not embed this invariance property. The Reynolds stress anisotropy predictions of this invariant neural network are propagated through to the velocity field for two test cases. For both test cases, significant improvement versus baseline RANS linear eddy viscosity and nonlinear eddy viscosity models is demonstrated.

IslandViewer (http://www.pathogenomics.sfu.ca/islandviewer/) is a widely-used webserver for the prediction and interactive visualization of genomic islands (GIs, regions of probable horizontal origin) in bacterial and archaeal genomes. GIs disproportionately encode factors that enhance the adaptability and competitiveness of the microbe within a niche, including virulence factors and other medically or environmentally important adaptations. We report here the release of IslandViewer 4, with novel features to accommodate the needs of larger-scale microbial genomics analysis, while expanding GI predictions and improving its flexible visualization interface. A user management web interface as well as an HTTP API for batch analyses are now provided with a secured authentication to facilitate the submission of larger numbers of genomes and the retrieval of results. In addition, IslandViewer's integrated GI predictions from multiple methods have been improved and expanded by integrating the precise Islander method for pre-computed genomes, as well as an updated IslandPath-DIMOB for both pre-computed and user-supplied custom genome analysis. Finally, pre-computed predictions including virulence factors and antimicrobial resistance are now available for 6193 complete bacterial and archaeal strains publicly available in RefSeq. IslandViewer 4 provides key enhancements to facilitate the analysis of GIs and better understand their role in the evolution of successful environmental microbes and pathogens.

It is postulated that steps on crystal surfaces capture atoms diffusing on the surface with certain probabilities and, in addition, that the capture probability depends on the direction from which adsorbed atoms approach the step. A general solution for the time-dependent step distribution is obtained in terms of these probabilities and an arbitrary initial distribution of an infinite sequence of parallel steps. It is shown that coalescence of steps or stabilization of step spacings can occur as a consequence of assuming that capture probabilities are directionally dependent. Some of the implications of the theoretical model are related to the growth of real crystal surfaces.

We simulate a method to teach a laser pulse sequences to excite specified molecular states. We use a learning procedure to direct the production of pulses based on ``fitness'' information provided by a laboratory measurement device. Over a series of pulses the algorithm learns an optimal sequence. The experimental apparatus, which consists of a laser, a sample of molecules and a measurement device, acts as an analog computer that solves Schr\"odinger's equation n/Iexactly, in real time. We simulate an apparatus that learns to excite specified rotational states in a diatomic molecule.

We report the creation of a nanoscale electrochemical device inside a transmission electron microscope--consisting of a single tin dioxide (SnO(2)) nanowire anode, an ionic liquid electrolyte, and a bulk lithium cobalt dioxide (LiCoO(2)) cathode--and the in situ observation of the lithiation of the SnO(2) nanowire during electrochemical charging. Upon charging, a reaction front propagated progressively along the nanowire, causing the nanowire to swell, elongate, and spiral. The reaction front is a "Medusa zone" containing a high density of mobile dislocations, which are continuously nucleated and absorbed at the moving front. This dislocation cloud indicates large in-plane misfit stresses and is a structural precursor to electrochemically driven solid-state amorphization. Because lithiation-induced volume expansion, plasticity, and pulverization of electrode materials are the major mechanical effects that plague the performance and lifetime of high-capacity anodes in lithium-ion batteries, our observations provide important mechanistic insight for the design of advanced batteries.