Marginal maximum likelihood (MML) estimation is the preferred approach to fitting item response theory models in psychometrics due to the MML estimators consistency, normality, and efficiency as the sample size tends to infinity. However, state-of-th
e-art MML estimation procedures such as the Metropolis-Hastings Robbins-Monro (MH-RM) algorithm as well as approximate MML estimation procedures such as variational inference (VI) are computationally time-consuming when the sample size and the number of latent factors are very large. In this work, we investigate a deep learning-based VI algorithm for exploratory item factor analysis (IFA) that is computationally fast even in large data sets with many latent factors. The proposed approach applies a deep artificial neural network model called an importance-weighted autoencoder (IWAE) for exploratory IFA. The IWAE approximates the MML estimator using an importance sampling technique wherein increasing the number of importance-weighted (IW) samples drawn during fitting improves the approximation, typically at the cost of decreased computational efficiency. We provide a real data application that recovers results aligning with psychological theory across random starts. Via simulation studies, we show that the IWAE yields more accurate estimates as either the sample size or the number of IW samples increases (although factor correlation and intercepts estimates exhibit some bias) and obtains similar results to MH-RM in less time. Our simulations also suggest that the proposed approach performs similarly to and is potentially faster than constrained joint maximum likelihood estimation, a fast procedure that is consistent when the sample size and the number of items simultaneously tend to infinity.
Assess the joint capabilities of emerging telescopes for near-Earth objects (NEOs) survey and characterization, and what they will add to the current capabilities or replace. NASA telescopes in prime mission, in development, or under study, and reque
sted for this assessment, include: - The Transiting Exoplanet Survey Satellite (TESS) - The James Webb Space Telescope (JWST) - The Wide Field Infrared Survey Telescope (WFIRST) - The Near-Earth Object Camera (NEOCam). Also requested for this assessment is the Large Synoptic Survey Telescope (LSST), an 8.4-meter ground-based telescope in development by the National Science Foundation and Department of Energy (DOE), with the capability to discover and catalogue NEOs.
The Near-Earth Object Wide-Field Infrared Survey Explorer (NEOWISE) mission continues to detect, track, and characterize minor planets. We present diameters and albedos calculated from observations taken during the second year since the spacecraft wa
s reactivated in late 2013. These include 207 near-Earth asteroids and 8,885 other asteroids. $84%$ of the near-Earth asteroids did not have previously measured diameters and albedos by the NEOWISE mission. Comparison of sizes and albedos calculated from NEOWISE measurements with those measured by occultations, spacecraft, and radar-derived shapes shows accuracy consistent with previous NEOWISE publications. Diameters and albedos fall within $ pm sim20%$ and $pmsim40%$, 1-sigma, respectively, of those measured by these alternate techniques. NEOWISE continues to preferentially discover near-Earth objects which are large ($>100$ m), and have low albedos.
We have investigated the unidirectional spin wave heat conveyer effect in sub-micron thick yttrium iron garnet (YIG) films using lock-in thermography (LIT). Although the effect is small in thin layers this technique allows us to observe asymmetric he
at transport by magnons which leads to asymmetric temperature profiles differing by several mK on both sides of the exciting antenna, respectively. Comparison of Damon-Eshbach and backward volume modes shows that the unidirectional heat flow is indeed due to non-reciprocal spin-waves. Because of the finite linewidth, small asymmetries can still be observed when only the uniform mode of ferromagnetic resonance is excited. The latter is of extreme importance for example when measuring the inverse spin-Hall effect because the temperature differences can result in thermovoltages at the contacts. Because of the non-reciprocity these thermovoltages reverse their sign with a reversal of the magnetic field which is typically deemed the signature of the inverse spin-Hall voltage.
We have carried out simulations to predict the performance of a new space-based telescopic survey operating at thermal infrared wavelengths that seeks to discover and characterize a large fraction of the potentially hazardous near-Earth asteroid (NEA
) population. Two potential architectures for the survey were considered: one located at the Earth-Sun L1 Lagrange point, and one in a Venus-trailing orbit. A sample cadence was formulated and tested, allowing for the self-follow-up necessary for objects discovered in the daytime sky on Earth. Synthetic populations of NEAs with sizes >=140 m in effective spherical diameter were simulated using recent determinations of their physical and orbital properties. Estimates of the instrumental sensitivity, integration times, and slew speeds were included for both architectures assuming the properties of new large-format 10 um detector arrays capable of operating at ~35 K. Our simulation included the creation of a preliminary version of a moving object processing pipeline suitable for operating on the trial cadence. We tested this pipeline on a simulated sky populated with astrophysical sources such as stars and galaxies extrapolated from Spitzer and WISE data, the catalog of known minor planets (including Main Belt asteroids, comets, Jovian Trojans, etc.), and the synthetic NEA model. Trial orbits were computed for simulated position-time pairs extracted from the synthetic surveys to verify that the tested cadence would result in orbits suitable for recovering objects at a later time. Our results indicate that the Earth-Sun L1 and Venus-trailing surveys achieve similar levels of integral completeness for potentially hazardous asteroids larger than 140 m; placing the telescope in an interior orbit does not yield an improvement in discovery rates. This work serves as a necessary first step for the detailed planning of a next-generation NEA survey.
NASAs Wide-field Infrared Survey Explorer (WISE) spacecraft has been brought out of hibernation and has resumed surveying the sky at 3.4 and 4.6 um. The scientific objectives of the NEOWISE reactivation mission are to detect, track, and characterize
near-Earth asteroids and comets. The search for minor planets resumed on December 23, 2013, and the first new near-Earth object (NEO) was discovered six days later. As an infrared survey, NEOWISE detects asteroids based on their thermal emission and is equally sensitive to high and low albedo objects; consequently, NEOWISE-discovered NEOs tend to be large and dark. Over the course of its three-year mission, NEOWISE will determine radiometrically-derived diameters and albedos for approximately 2000 NEOs and tens of thousands of Main Belt asteroids. The 32 months of hibernation have had no significant effect on the missions performance. Image quality, sensitivity, photometric and astrometric accuracy, completeness, and the rate of minor planet detections are all essentially unchanged from the prime missions post-cryogenic phase.
We present a novel scheme for an unbiased and non-perturbative treatment of strongly correlated fermions. The proposed approach combines two of the most successful many-body methods, i.e., the dynamical mean field theory (DMFT) and the functional ren
ormalization group (fRG). Physically, this allows for a systematic inclusion of non-local correlations via the flow equations of the fRG, after the local correlations are taken into account non-perturbatively by the DMFT. To demonstrate the feasibility of the approach, we present numerical results for the two-dimensional Hubbard model at half-filling.
Magnetic molecules adsorbed on a superconductor give rise to a local competition of Cooper pair and Kondo singlet formation inducing subgap bound states. For Manganese-phthalocyanine molecules on a Pb(111) substrate, scanning tunneling spectroscopy r
esolves pairs of subgap bound states and two Kondo screening channels. We show in a combined approach of scaling and numerical renormalization group calculations that the intriguing relation between Kondo screening and superconducting pairing is solely determined by the hybridization strength with the substrate. We demonstrate that an effective one-channel Anderson impurity model with a sizable particle-hole asymmetry captures universal and non-universal observations in the system quantitatively. The model parameters and disentanglement of the two screening channels are elucidated by scaling arguments.
We present the preliminary analysis of 1023 known asteroids in the Hilda region of the Solar System observed by the NEOWISE component of the Wide-field Infrared Survey Explorer (WISE). The sizes of the Hildas observed range from $sim 3 - 200$km. We f
ind no size - albedo dependency as reported by other projects. The albedos of our sample are low, with a weighted mean value $p_V = 0.055pm0.018$, for all sizes sampled by the NEOWISE survey. We observed a significant fraction of the objects in the two known collisional families in the Hilda population. It is found that the Hilda collisional family is brighter, with weighted mean albedo of $p_V = 0.061pm0.011$, than the general population and dominated by D-type asteroids, while the Schubart collisional family is darker, with weighted mean albedo of ($p_V = 0.039pm0.013$). Using the reflected sunlight in the two shortest WISE bandpasses we are able to derive a method for taxonomic classification of $sim 10%$ of the Hildas detected in the NEOWISE survey. For the Hildas with diameter larger than 30km there are $67^{+7}_{-15}%$ D-type asteroids and $26^{+17}_{-5}%$ C-/P-type asteroids (with the majority of these being P-types).
We present the preliminary analysis of over 1739 known and 349 candidate Jovian Trojans observed by the NEOWISE component of the Wide-field Infrared Survey Explorer (WISE). With this survey the available diameters, albedos and beaming parameters for
the Jovian Trojans have been increased by more than an order of magnitude compared to previous surveys. We find that the Jovian Trojan population is very homogenous for sizes larger than $sim10$km (close to the detection limit of WISE for these objects). The observed sample consists almost exclusively of low albedo objects, having a mean albedo value of $0.07pm0.03$. The beaming parameter was also derived for a large fraction of the observed sample, and it is also very homogenous with an observed mean value of $0.88pm0.13$. Preliminary debiasing of the survey shows our observed sample is consistent with the leading cloud containing more objects than the trailing cloud. We estimate the fraction to be N(leading)/N(trailing) $sim 1.4 pm 0.2$, lower than the $1.6 pm 0.1$ value derived by others.
