University of Washington Applied Physics Laboratory

Extracting governing equations from data is a central challenge in many diverse areas of science and engineering. Data are abundant whereas models often remain elusive, as in climate science, neuroscience, ecology, finance, and epidemiology, to name only a few examples. In this work, we combine sparsity-promoting techniques and machine learning with nonlinear dynamical systems to discover governing equations from noisy measurement data. The only assumption about the structure of the model is that there are only a few important terms that govern the dynamics, so that the equations are sparse in the space of possible functions; this assumption holds for many physical systems in an appropriate basis. In particular, we use sparse regression to determine the fewest terms in the dynamic governing equations required to accurately represent the data. This results in parsimonious models that balance accuracy with model complexity to avoid overfitting. We demonstrate the algorithm on a wide range of problems, from simple canonical systems, including linear and nonlinear oscillators and the chaotic Lorenz system, to the fluid vortex shedding behind an obstacle. The fluid example illustrates the ability of this method to discover the underlying dynamics of a system that took experts in the community nearly 30 years to resolve. We also show that this method generalizes to parameterized systems and systems that are time-varying or have external forcing.

Abstract The effect of disclosure level on the cost of equity capital is a matter of considerable interest and importance to the financial reporting community. However, the association between disclosure level and cost of equity capital is not well established and has been difficult to quantify. In this paper I examine the association between disclosure level and the cost of equity capital by regressing firm-specific estimates of cost of equity capital on market beta, firm size and a self-constructed measure of disclosure level. My measure of disclosure level is based on the amount of voluntary disclosure provided in the 1990 annual reports of a sample of 122 manufacturing firms. For firms that attract a low analyst following, the results indicate that greater disclosure is associated with a lower cost of equity capital. The magnitude of the effect is such that a one-unit difference in the disclosure measure is associated with a difference of approximately twenty-eight basis points in the cost of equity capital, after controlling for market beta and firm size. For firms with a high analyst following, however, I find no evidence of an association between my measure of disclosure level and cost of equity capital perhaps because the disclosure measure is limited to the annual report and accordingly may not provide a powerful proxy for overall disclosure level when analysts play a significant role in the communication process.

This review of 122 research reports (184 independent samples, 14,900 subjects) found average r = .274 for prediction of behavioral, judgment, and physiological measures by Implicit Association Test (IAT) measures. Parallel explicit (i.e., self-report) measures, available in 156 of these samples (13,068 subjects), also predicted effectively (average r = .361), but with much greater variability of effect size. Predictive validity of self-report was impaired for socially sensitive topics, for which impression management may distort self-report responses. For 32 samples with criterion measures involving Black-White interracial behavior, predictive validity of IAT measures significantly exceeded that of self-report measures. Both IAT and self-report measures displayed incremental validity, with each measure predicting criterion variance beyond that predicted by the other. The more highly IAT and self-report measures were intercorrelated, the greater was the predictive validity of each.

Finite mixture models are being used increasingly to model a wide variety of random phenomena for clustering, classification and density estimation. mclust is a powerful and popular package which allows modelling of data as a Gaussian finite mixture with different covariance structures and different numbers of mixture components, for a variety of purposes of analysis. Recently, version 5 of the package has been made available on CRAN. This updated version adds new covariance structures, dimension reduction capabilities for visualisation, model selection criteria, initialisation strategies for the EM algorithm, and bootstrap-based inference, making it a full-featured R package for data analysis via finite mixture modelling.

Voltage-gated Ca(2+) channels mediate Ca(2+) entry into cells in response to membrane depolarization. Electrophysiological studies reveal different Ca(2+) currents designated L-, N-, P-, Q-, R-, and T-type. The high-voltage-activated Ca(2+) channels that have been characterized biochemically are complexes of a pore-forming alpha1 subunit of approximately 190-250 kDa; a transmembrane, disulfide-linked complex of alpha2 and delta subunits; an intracellular beta subunit; and in some cases a transmembrane gamma subunit. Ten alpha1 subunits, four alpha2delta complexes, four beta subunits, and two gamma subunits are known. The Cav1 family of alpha1 subunits conduct L-type Ca(2+) currents, which initiate muscle contraction, endocrine secretion, and gene transcription, and are regulated primarily by second messenger-activated protein phosphorylation pathways. The Cav2 family of alpha1 subunits conduct N-type, P/Q-type, and R-type Ca(2+) currents, which initiate rapid synaptic transmission and are regulated primarily by direct interaction with G proteins and SNARE proteins and secondarily by protein phosphorylation. The Cav3 family of alpha1 subunits conduct T-type Ca(2+) currents, which are activated and inactivated more rapidly and at more negative membrane potentials than other Ca(2+) current types. The distinct structures and patterns of regulation of these three families of Ca(2+) channels provide a flexible array of Ca(2+) entry pathways in response to changes in membrane potential and a range of possibilities for regulation of Ca(2+) entry by second messenger pathways and interacting proteins.

We demonstrate an integrated approach to build, test, and refine a model of a cellular pathway, in which perturbations to critical pathway components are analyzed using DNA microarrays, quantitative proteomics, and databases of known physical interactions. Using this approach, we identify 997 messenger RNAs responding to 20 systematic perturbations of the yeast galactose-utilization pathway, provide evidence that approximately 15 of 289 detected proteins are regulated posttranscriptionally, and identify explicit physical interactions governing the cellular response to each perturbation. We refine the model through further iterations of perturbation and global measurements, suggesting hypotheses about the regulation of galactose utilization and physical interactions between this and a variety of other metabolic pathways.

Abstract Ensembles used for probabilistic weather forecasting often exhibit a spread-error correlation, but they tend to be underdispersive. This paper proposes a statistical method for postprocessing ensembles based on Bayesian model averaging (BMA), which is a standard method for combining predictive distributions from different sources. The BMA predictive probability density function (PDF) of any quantity of interest is a weighted average of PDFs centered on the individual bias-corrected forecasts, where the weights are equal to posterior probabilities of the models generating the forecasts and reflect the models' relative contributions to predictive skill over the training period. The BMA weights can be used to assess the usefulness of ensemble members, and this can be used as a basis for selecting ensemble members; this can be useful given the cost of running large ensembles. The BMA PDF can be represented as an unweighted ensemble of any desired size, by simulating from the BMA predictive distribution. The BMA predictive variance can be decomposed into two components, one corresponding to the between-forecast variability, and the second to the within-forecast variability. Predictive PDFs or intervals based solely on the ensemble spread incorporate the first component but not the second. Thus BMA provides a theoretical explanation of the tendency of ensembles to exhibit a spread-error correlation but yet be underdispersive. The method was applied to 48-h forecasts of surface temperature in the Pacific Northwest in January–June 2000 using the University of Washington fifth-generation Pennsylvania State University–NCAR Mesoscale Model (MM5) ensemble. The predictive PDFs were much better calibrated than the raw ensemble, and the BMA forecasts were sharp in that 90% BMA prediction intervals were 66% shorter on average than those produced by sample climatology. As a by-product, BMA yields a deterministic point forecast, and this had root-mean-square errors 7% lower than the best of the ensemble members and 8% lower than the ensemble mean. Similar results were obtained for forecasts of sea level pressure. Simulation experiments show that BMA performs reasonably well when the underlying ensemble is calibrated, or even overdispersed.

CONTEXT: Maternal depressive symptoms during pregnancy have been reported in some, but not all, studies to be associated with an increased risk of preterm birth (PTB), low birth weight (LBW), and intrauterine growth restriction (IUGR). OBJECTIVE: To estimate the risk of PTB, LBW, and IUGR associated with antenatal depression. DATA SOURCES AND STUDY SELECTION: We searched for English-language and non-English-language articles via the MEDLINE, PsycINFO, CINAHL, Social Work Abstracts, Social Services Abstracts, and Dissertation Abstracts International databases (January 1980 through December 2009). We aimed to include prospective studies reporting data on antenatal depression and at least 1 adverse birth outcome: PTB (<37 weeks' gestation), LBW (<2500 g), or IUGR (<10th percentile for gestational age). Of 862 reviewed studies, 29 US-published and non-US-published studies met the selection criteria. DATA EXTRACTION: Information was extracted on study characteristics, antenatal depression measurement, and other biopsychosocial risk factors and was reviewed twice to minimize error. DATA SYNTHESIS: Pooled relative risks (RRs) for the effect of antenatal depression on each birth outcome were calculated using random-effects methods. In studies of PTB, LBW, and IUGR that used a categorical depression measure, pooled effect sizes were significantly larger (pooled RR [95% confidence interval] = 1.39 [1.19-1.61], 1.49 [1.25-1.77], and 1.45 [1.05-2.02], respectively) compared with studies that used a continuous depression measure (1.03 [1.00-1.06], 1.04 [0.99-1.09], and 1.02 [1.00-1.04], respectively). The estimates of risk for categorically defined antenatal depression and PTB and LBW remained significant when the trim-and-fill procedure was used to correct for publication bias. The risk of LBW associated with antenatal depression was significantly larger in developing countries (RR = 2.05; 95% confidence interval, 1.43-2.93) compared with the United States (RR = 1.10; 95% confidence interval, 1.01-1.21) or European social democracies (RR = 1.16; 95% confidence interval, 0.92-1.47). Categorically defined antenatal depression tended to be associated with an increased risk of PTB among women of lower socioeconomic status in the United States. CONCLUSIONS: Women with depression during pregnancy are at increased risk for PTB and LBW, although the magnitude of the effect varies as a function of depression measurement, country location, and US socioeconomic status. An important implication of these findings is that antenatal depression should be identified through universal screening and treated.

Van der Waals bound heterostructures constructed with two-dimensional materials, such as graphene, boron nitride and transition metal dichalcogenides, have sparked wide interest in device physics and technologies at the two-dimensional limit. One highly coveted heterostructure is that of differing monolayer transition metal dichalcogenides with type-II band alignment, with bound electrons and holes localized in individual monolayers, that is, interlayer excitons. Here, we report the observation of interlayer excitons in monolayer MoSe2–WSe2 heterostructures by photoluminescence and photoluminescence excitation spectroscopy. We find that their energy and luminescence intensity are highly tunable by an applied vertical gate voltage. Moreover, we measure an interlayer exciton lifetime of ~1.8 ns, an order of magnitude longer than intralayer excitons in monolayers. Our work demonstrates optical pumping of interlayer electric polarization, which may provoke further exploration of interlayer exciton condensation, as well as new applications in two-dimensional lasers, light-emitting diodes and photovoltaic devices. Monolayer transition metal dichalcogenide heterostructures with type II band alignment have generated wide interest in device physics at the two-dimensional limit. Here, Rivera et al. observe interlayer excitons in vertically stacked MoSe2–WSe2 heterostructures and demonstrate tunability of the energy and luminescence.

Warming and Melting Mass loss from the ice sheets of Greenland and Antarctica account for a large fraction of global sea-level rise. Part of this loss is because of the effects of warmer air temperatures, and another because of the rising ocean temperatures to which they are being exposed. Joughin et al. (p. 1172 ) review how ocean-ice interactions are impacting ice sheets and discuss the possible ways that exposure of floating ice shelves and grounded ice margins are subject to the influences of warming ocean currents. Estimates of the mass balance of the ice sheets of Greenland and Antarctica have differed greatly—in some cases, not even agreeing about whether there is a net loss or a net gain—making it more difficult to project accurately future sea-level change. Shepherd et al. (p. 1183 ) combined data sets produced by satellite altimetry, interferometry, and gravimetry to construct a more robust ice-sheet mass balance for the period between 1992 and 2011. All major regions of the two ice sheets appear to be losing mass, except for East Antarctica. All told, mass loss from the polar ice sheets is contributing about 0.6 millimeters per year (roughly 20% of the total) to the current rate of global sea-level rise.

Cervical cancer remains one of the leading causes of cancer-related deaths worldwide. Here we report the extensive molecular characterization of 228 primary cervical cancers, one of the largest comprehensive genomic studies of cervical cancer to date. We observed notable APOBEC mutagenesis patterns and identified SHKBP1, ERBB3, CASP8, HLA-A and TGFBR2 as novel significantly mutated genes in cervical cancer. We also discovered amplifications in immune targets CD274 (also known as PD-L1) and PDCD1LG2 (also known as PD-L2), and the BCAR4 long non-coding RNA, which has been associated with response to lapatinib. Integration of human papilloma virus (HPV) was observed in all HPV18-related samples and 76% of HPV16-related samples, and was associated with structural aberrations and increased target-gene expression. We identified a unique set of endometrial-like cervical cancers, comprised predominantly of HPV-negative tumours with relatively high frequencies of KRAS, ARID1A and PTEN mutations. Integrative clustering of 178 samples identified keratin-low squamous, keratin-high squamous and adenocarcinoma-rich subgroups. These molecular analyses reveal new potential therapeutic targets for cervical cancers. This paper describes molecular subtypes of cervical cancers, including squamous cell carcinoma and adenocarcinoma clusters defined by HPV status and molecular features, and distinct molecular pathways that are activated in cervical carcinomas caused by different somatic alterations and HPV types. Cervical cancer is one of the main causes of cancer-related deaths worldwide, and 95% of cases result from human papilloma virus (HPV) infection. The Cancer Genome Atlas Research Network now reports the genomic and molecular characterization of 228 primary cervical cancers. The authors identify significantly mutated genes and pathways that differ by cervical cancer subtype, and find that keratin-low squamous, keratin-high squamous and adenocarcinoma-rich clusters are marked by different HPV types and molecular features.

We propose a sparse regression method capable of discovering the governing partial differential equation(s) of a given system by time series measurements in the spatial domain. The regression framework relies on sparsity-promoting techniques to select the nonlinear and partial derivative terms of the governing equations that most accurately represent the data, bypassing a combinatorially large search through all possible candidate models. The method balances model complexity and regression accuracy by selecting a parsimonious model via Pareto analysis. Time series measurements can be made in an Eulerian framework, where the sensors are fixed spatially, or in a Lagrangian framework, where the sensors move with the dynamics. The method is computationally efficient, robust, and demonstrated to work on a variety of canonical problems spanning a number of scientific domains including Navier-Stokes, the quantum harmonic oscillator, and the diffusion equation. Moreover, the method is capable of disambiguating between potentially nonunique dynamical terms by using multiple time series taken with different initial data. Thus, for a traveling wave, the method can distinguish between a linear wave equation and the Korteweg-de Vries equation, for instance. The method provides a promising new technique for discovering governing equations and physical laws in parameterized spatiotemporal systems, where first-principles derivations are intractable.

We present the third Fermi Large Area Telescope (LAT) source catalog (3FGL) of sources in the 100 MeV-300 GeV range. Based on the first 4 yr of science data from the Fermi Gamma-ray Space Telescope mission, it is the deepest yet in this energy range. Relative to the Second Fermi LAT catalog, the 3FGL catalog incorporates twice as much data, as well as a number of analysis improvements, including improved calibrations at the event reconstruction level, an updated model for Galactic diffuse -ray emission, a refined procedure for source detection, and improved methods for associating LAT sources with potential counterparts at other wavelengths. The 3FGL catalog includes 3033 sources above 4 significance, with source location regions, spectral properties, and monthly light curves for each. Of these, 78 are flagged as potentially being due to imperfections in the model for Galactic diffuse emission. Twenty-five sources are modeled explicitly as spatially extended, and overall 238 sources are considered as identified based on angular extent or correlated variability (periodic or otherwise) observed at other wavelengths. For 1010 sources we have not found plausible counterparts at other wavelengths. More than 1100 of the identified or associated sources are active galaxies of the blazar class; several other classes of non-blazar active galaxies are also represented in the 3FGL. Pulsars represent the largest Galactic source class. From source counts of Galactic sources we estimate that the contribution of unresolved sources to the Galactic diffuse emission is 3% at 1 GeV.

Medication adherence usually refers to whether patients take their medications as prescribed (eg, twice daily), as well as whether they continue to take a prescribed medication. Medication nonadherence is a growing concern to clinicians, healthcare systems, and other stakeholders (eg, payers) because of mounting evidence that it is prevalent and associated with adverse outcomes and higher costs of care. To date, measurement of patient medication adherence and use of interventions to improve adherence are rare in routine clinical practice. The goals of the present report are to address (1) different methods of measuring adherence, (2) the prevalence of medication nonadherence, (3) the association between nonadherence and outcomes, (4) the reasons for nonadherence, and finally, (5) interventions to improve medication adherence.

Abstract Low‐cost, environment‐friendly aqueous Zn batteries have great potential for large‐scale energy storage, but the intercalation of zinc ions in the cathode materials is challenging and complex. Herein, the critical role of structural H 2 O on Zn 2+ intercalation into bilayer V 2 O 5 ·nH 2 O is demonstrated. The results suggest that the H 2 O‐solvated Zn 2+ possesses largely reduced effective charge and thus reduced electrostatic interactions with the V 2 O 5 framework, effectively promoting its diffusion. Benefited from the “lubricating” effect, the aqueous Zn battery shows a specific energy of ≈144 Wh kg −1 at 0.3 A g −1 . Meanwhile, it can maintain an energy density of 90 Wh kg −1 at a high power density of 6.4 kW kg −1 (based on the cathode and 200% Zn anode), making it a promising candidate for high‐performance, low‐cost, safe, and environment‐friendly energy‐storage devices.

The United Nations (UN) recently released population projections based on data until 2012 and a Bayesian probabilistic methodology. Analysis of these data reveals that, contrary to previous literature, the world population is unlikely to stop growing this century. There is an 80% probability that world population, now 7.2 billion people, will increase to between 9.6 billion and 12.3 billion in 2100. This uncertainty is much smaller than the range from the traditional UN high and low variants. Much of the increase is expected to happen in Africa, in part due to higher fertility rates and a recent slowdown in the pace of fertility decline. Also, the ratio of working-age people to older people is likely to decline substantially in all countries, even those that currently have young populations.

BACKGROUND: Traumatic events are common globally; however, comprehensive population-based cross-national data on the epidemiology of posttraumatic stress disorder (PTSD), the paradigmatic trauma-related mental disorder, are lacking. METHODS: Data were analyzed from 26 population surveys in the World Health Organization World Mental Health Surveys. A total of 71 083 respondents ages 18+ participated. The Composite International Diagnostic Interview assessed exposure to traumatic events as well as 30-day, 12-month, and lifetime PTSD. Respondents were also assessed for treatment in the 12 months preceding the survey. Age of onset distributions were examined by country income level. Associations of PTSD were examined with country income, world region, and respondent demographics. RESULTS: The cross-national lifetime prevalence of PTSD was 3.9% in the total sample and 5.6% among the trauma exposed. Half of respondents with PTSD reported persistent symptoms. Treatment seeking in high-income countries (53.5%) was roughly double that in low-lower middle income (22.8%) and upper-middle income (28.7%) countries. Social disadvantage, including younger age, female sex, being unmarried, being less educated, having lower household income, and being unemployed, was associated with increased risk of lifetime PTSD among the trauma exposed. CONCLUSIONS: PTSD is prevalent cross-nationally, with half of all global cases being persistent. Only half of those with severe PTSD report receiving any treatment and only a minority receive specialty mental health care. Striking disparities in PTSD treatment exist by country income level. Increasing access to effective treatment, especially in low- and middle-income countries, remains critical for reducing the population burden of PTSD.

Identifying coordinate transformations that make strongly nonlinear dynamics approximately linear has the potential to enable nonlinear prediction, estimation, and control using linear theory. The Koopman operator is a leading data-driven embedding, and its eigenfunctions provide intrinsic coordinates that globally linearize the dynamics. However, identifying and representing these eigenfunctions has proven challenging. This work leverages deep learning to discover representations of Koopman eigenfunctions from data. Our network is parsimonious and interpretable by construction, embedding the dynamics on a low-dimensional manifold. We identify nonlinear coordinates on which the dynamics are globally linear using a modified auto-encoder. We also generalize Koopman representations to include a ubiquitous class of systems with continuous spectra. Our framework parametrizes the continuous frequency using an auxiliary network, enabling a compact and efficient embedding, while connecting our models to decades of asymptotics. Thus, we benefit from the power of deep learning, while retaining the physical interpretability of Koopman embeddings.

A Viking Lander 1 image was modeled as mixtures of reflectance spectra of palagonite dust, gray andesitelike rock, and a coarse rocklike soil. The rocks are covered to varying degrees by dust but otherwise appear unweathered. Rocklike soil occurs as lag deposits in deflation zones around stones and on top of a drift and as a layer in a trench dug by the lander. This soil probably is derived from the rocks by wind abrasion and/or spallation. Dust is the major component of the soil and covers most of the surface. The dust is unrelated spectrally to the rock but is equivalent to the global‐scale dust observed telescopically. A new method was developed to model a multispectral image as mixtures of end‐member spectra and to compare image spectra directly with laboratory reference spectra. The method for the first time uses shade and secondary illumination effects as spectral end‐members; thus the effects of topography and illumination on all scales can be isolated or removed. The image was calibrated absolutely from the laboratory spectra, in close agreement with direct calibrations. The method has broad applications to interpreting multispectral images, including satellite images.

Streams, rivers, lakes, and other inland waters are important agents in the coupling of biogeochemical cycles between continents, atmosphere, and oceans. The depiction of these roles in global‐scale assessments of carbon (C) and other bioactive elements remains limited, yet recent findings suggest that C discharged to the oceans is only a fraction of that entering rivers from terrestrial ecosystems via soil respiration, leaching, chemical weathering, and physical erosion. Most of this C influx is returned to the atmosphere from inland waters as carbon dioxide (CO 2 ) or buried in sedimentary deposits within impoundments, lakes, floodplains, and other wetlands. Carbon and mineral cycles are coupled by both erosion–deposition processes and chemical weathering, with the latter producing dissolved inorganic C and carbonate buffering capacity that strongly modulate downstream pH, biological production of calcium‐carbonate shells, and CO 2 outgassing in rivers, estuaries, and coastal zones. Human activities substantially affect all of these processes.