arXiv subjects

E. Romero

Publications and source records attributed to E. Romero.

13 recordsLinked to original sources

arXiv2024

Mean field initialization of the Annealed Importance Sampling algorithm for an efficient evaluation of the Partition Function of Restricted Boltzmann Machines

Probabilistic models in physics often require from the evaluation of normalized Boltzmann factors, which in turn implies the computation of the partition function Z. Getting the exact value of Z, though, becomes a forbiddingly expensive task as the system size increases. This problem is also present in probabilistic learning models such as the Restricted Boltzmann Machine (RBM), where the situation is even worse as the exact learning rules implies the computation of Z at each iteration. A possible way to tackle this problem is to use the Annealed Importance Sampling (AIS) algorithm, which provides a tool to stochastically estimate the partition function of the system. So far, the standard application of the AIS algorithm starts from the uniform probability distribution and uses a large number of Monte Carlo steps to obtain reliable estimations of Z following an annealing process. In this work we show that both the quality of the estimation and the cost of the computation can be significantly improved by using a properly selected mean-field starting probability distribution. We perform a systematic analysis of AIS in both small- and large-sized problems, and compare the results to exact values in problems where these are known. As a result of our systematic analysis, we propose two successful strategies that work well in all the problems analyzed. We conclude that these are good starting points to estimate the partition function with AIS with a relatively low computational cost.

physics.comp-ph↗

arXiv2022

Cascading of Nanomechanical Resonator Logic

Nanomechanical systems have been proposed as an alternative computing platform for high radiation environments, where semiconductor electronics traditionally fail, as well as to allow improved gate densities and energy consumption. While there have been numerous demonstrations of individual nanomechanical logic gates leveraging the Duffing nonlinearity, the development of useful nanomechanical logic circuits depends strongly on the ability to cascade multiple logic gates. Here we show theoretically that cascading nanomechanical logic gates, where the output of one gate is fed into the input of another, is a complex problem due to the transient dynamics of the collective system. These transient behaviours can lead to undesired bit flips, which precludes cascading altogether. We then show that this issue can be circumvented by carefully initialising the system prior to computation. We illustrate these salient features through the modelled dynamics of two cascaded nanomechanical NAND gates.

physics.app-ph↗

arXiv2015

The PROSPECT Physics Program

The Precision Reactor Oscillation and Spectrum Experiment, PROSPECT, is designed to make a precise measurement of the antineutrino spectrum from a highly-enriched uranium reactor and probe eV-scale sterile neutrinos by searching for neutrino oscillations over meter-long distances. PROSPECT is conceived as a 2-phase experiment utilizing segmented $^6$Li-doped liquid scintillator detectors for both efficient detection of reactor antineutrinos through the inverse beta decay reaction and excellent background discrimination. PROSPECT Phase I consists of a movable 3-ton antineutrino detector at distances of 7 - 12 m from the reactor core. It will probe the best-fit point of the $ν_e$ disappearance experiments at 4$σ$ in 1 year and the favored region of the sterile neutrino parameter space at $>$3$σ$ in 3 years. With a second antineutrino detector at 15 - 19 m from the reactor, Phase II of PROSPECT can probe the entire allowed parameter space below 10 eV$^{2}$ at 5$σ$ in 3 additional years. The measurement of the reactor antineutrino spectrum and the search for short-baseline oscillations with PROSPECT will test the origin of the spectral deviations observed in recent $θ_{13}$ experiments, search for sterile neutrinos, and conclusively address the hypothesis of sterile neutrinos as an explanation of the reactor anomaly.

physics.ins-det↗

arXiv2015

METing SUSY on the Z peak

Recently the ATLAS experiment announced a 3 $σ$ excess at the Z-peak consisting of 29 pairs of leptons together with two or more jets, $E_T^{\rm miss}> 225$ GeV and $H_T \geq 600$ GeV, to be compared with $10.6 \pm 3.2$ expected lepton pairs in the Standard Model. No excess outside the Z-peak was observed. By trying to explain this signal with SUSY we find that only relatively light gluinos, $m_{\tilde g} \lesssim 1.2$ TeV, together with a heavy neutralino NLSP of $m_{\tilde χ} \gtrsim 400$ GeV decaying predominantly to Z-boson plus a light gravitino, such that nearly every gluino produces at least one Z-boson in its decay chain, could reproduce the excess. We construct an explicit general gauge mediation model able to reproduce the observed signal overcoming all the experimental limits. Needless to say, more sophisticated models could also reproduce the signal, however, any model would have to exhibit the following features, light gluinos, or heavy particles with a strong production cross-section, producing at least one Z-boson in its decay chain. The implications of our findings for the Run II at LHC with the scaling on the Z peak, as well as for the direct search of gluinos and other SUSY particles, are pointed out.

hep-ph↗

arXiv2015

Background Radiation Measurements at High Power Research Reactors

Research reactors host a wide range of activities that make use of the intense neutron fluxes generated at these facilities. Recent interest in performing measurements with relatively low event rates, e.g. reactor antineutrino detection, at these facilities necessitates a detailed understanding of background radiation fields. Both reactor-correlated and naturally occurring background sources are potentially important, even at levels well below those of importance for typical activities. Here we describe a comprehensive series of background assessments at three high-power research reactors, including $γ$-ray, neutron, and muon measurements. For each facility we describe the characteristics and identify the sources of the background fields encountered. The general understanding gained of background production mechanisms and their relationship to facility features will prove valuable for the planning of any sensitive measurement conducted therein.

physics.ins-det↗

arXiv2015

Light Collection and Pulse-Shape Discrimination in Elongated Scintillator Cells for the PROSPECT Reactor Antineutrino Experiment

A meter-long, 23-liter EJ-309 liquid scintillator detector has been constructed to study the light collection and pulse-shape discrimination performance of elongated scintillator cells for the PROSPECT reactor antineutrino experiment. The magnitude and uniformity of light collection and neutron/gamma discrimination power in the energy range of antineutrino inverse beta decay products have been studied using gamma and spontaneous fission calibration sources deployed along the cell long axis. We also study neutron-gamma discrimination and light collection abilities for differing PMT and reflector configurations. Key design features for optimizing MeV-scale response and background rejection capabilities are identified.

physics.ins-det↗

arXiv2015

A Neighbourhood-Based Stopping Criterion for Contrastive Divergence Learning

Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain generative models of data distributions. RBMs are often trained using the Contrastive Divergence learning algorithm (CD), an approximation to the gradient of the data log-likelihood. A simple reconstruction error is often used as a stopping criterion for CD, although several authors \cite{schulz-et-al-Convergence-Contrastive-Divergence-2010-NIPSw, fischer-igel-Divergence-Contrastive-Divergence-2010-ICANN} have raised doubts concerning the feasibility of this procedure. In many cases the evolution curve of the reconstruction error is monotonic while the log-likelihood is not, thus indicating that the former is not a good estimator of the optimal stopping point for learning. However, not many alternatives to the reconstruction error have been discussed in the literature. In this manuscript we investigate simple alternatives to the reconstruction error, based on the inclusion of information contained in neighboring states to the training set, as a stopping criterion for CD learning.

cs.NE↗

arXiv2015

PROSPECT - A Precision Reactor Oscillation and Spectrum Experiment at Short Baselines

Current models of antineutrino production in nuclear reactors predict detection rates and spectra at odds with the existing body of direct reactor antineutrino measurements. High-resolution antineutrino detectors operated close to compact research reactor cores can produce new precision measurements useful in testing explanations for these observed discrepancies involving underlying nuclear or new physics. Absolute measurement of the 235U-produced antineutrino spectrum can provide additional constraints for evaluating the accuracy of current and future reactor models, while relative measurements of spectral distortion between differing baselines can be used to search for oscillations arising from the existence of eV-scale sterile neutrinos. Such a measurement can be performed in the United States at several highly-enriched uranium fueled research reactors using near-surface segmented liquid scintillator detectors. We describe here the conceptual design and physics potential of the PROSPECT experiment, a U.S.-based, multi-phase experiment with reactor-detector baselines of 7-20 meters capable of addressing these and other physics and detector development goals. Current R&D status and future plans for PROSPECT detector deployment and data-taking at the High Flux Isotope Reactor at Oak Ridge National Laboratory will be discussed.

physics.ins-det↗

arXiv2011

Determination of the physical characteristics of the variable stars in the direction of the open cluster NGC 6811 through $uvbyβ$ photoelectric photometry

The study of open clusters and their short period variable stars is fundamental in stellar evolution. Because the cluster members are formed in almost the same physical conditions, they share similar stellar properties such age and chemical composition. The assumption of common age, metallicity and distance impose strong constraints when modeling an ensemble of short period pulsators belonging to open clusters (e.g. Fox Machado et al., 2006). Very recently, Luo et al. (2009) carried out a search for variable stars in the direction of NGC 6811 with CCD photometry in B and V bands. They detected a total of sixteen variable stars. Among these variables, twelve were catalogued as $δ$ {\it Scuti} stars, while no variability type was assigned to the remaining stars. In this paper we present $uvbyβ$ photoelectric photometry for the variable stars in the direction of NGC 6811.

astro-ph.SR↗

arXiv2011

$uvby-β$ photoelectric photometry of the open clusters NGC 6811 and NGC 6830

From $uvby-β$ photometry of the open clusters NGC 6811 (75 stars), and NGC 6830 (19 stars) we were able to determine membership of the stars to each cluster, and fix the age and reddening for each. Since several short period stars have recently been found, we have carried out a study of these variables.

astro-ph.SR↗

arXiv2010

Weak ferromagnetism in cobalt oxalate crystals

Microcrystals of diaquocobalt(II) oxalate have been synthesized by the coprecipitation reaction of aqueous solutions of Cobalt (II) bromide and oxalic acid. Chemical analysis and thermal experiments revealed that there is only one phase present. X-ray powder diffraction studies show that this compound is orthorhombic with space group Cccm. Molar susceptibility versus temperature measurements show the existence of an antiferromagnetic ordering, however, the hysteresis measured in magnetization measurements as a function of magnetic field reveals a weak ferromagnetic behavior.

cond-mat.mtrl-sci↗

arXiv2008

The metallicity extremes of the Sagittarius dSph using SALT spectroscopy of PNe

In this work we present the first spectroscopic results obtained with the Southern African Large Telescope (SALT) telescope during its perfomance-verification phase. We find that the Sagittarius dwarf spheroidal galaxy (Sgr) Sgr contains a youngest stellar population with [O/H] -0.2 and age t>1 Gyr, and an oldest population with [O/H]=-2.0. The values are based on spectra of two planetary nebulae (PNe), using empirical abundance determinations. We calculated abundances for O, N, Ne, Ar, S, Cl, Fe, C and He. We confirm the high abundances of PN StWr2-21 with 12+log(O/H) = 8.57+/-0.02 dex. The other PN studied, BoBn1, is an extraordinary object in that the neon abundance exceeds that of oxygen. The abundances of S, Ar and Cl in BoBn1 yield the original stellar metallicity, corresponding to 12+log(O/H) = 6.72+/-0.16 dex which is 1/110 of the solar value. The actual [O/H] is much higher: third dredge-up enriched the material by a factor of ~12 in oxygen, ~240 in nitrogen and ~70 in neon. Neon as well as nitrogen and oxygen content may have been produced in the intershell of low-mass AGB stars. Well defined broad WR lines are present in the spectrum of StWr2-21 and absent in the spectrum of BoBn1. This puts the fraction of [WR]-type central PNe stars to 67% for dSph galaxies.

astro-ph↗

arXiv2007

The polar ring galaxy AM1934-563 revisited

We report long-slit spectroscopic observations of the dust-lane polar-ring galaxy AM1934-563 obtained with the Southern African Large Telescope (SALT) during its performance-verification phase. The observations target the spectral region of the Ha, [NII] and [SII] emission-lines, but show also deep NaI stellar absorption lines that we interpret as produced by stars in the galaxy. We derive rotation curves along the major axis of the galaxy that extend out to about 8 kpc from the center for both the gaseous and the stellar components, using the emission and absorption lines. We derive similar rotation curves along the major axis of the polar ring and point out differences between these and the ones of the main galaxy. We identify a small diffuse object visible only in Ha emission and with a low velocity dispersion as a dwarf HII galaxy and argue that it is probably metal-poor. Its velocity indicates that it is a fourth member of the galaxy group in which AM1934-563 belongs. We discuss the observations in the context of the proposal that the object is the result of a major merger and point out some observational discrepancies from this explanation. We argue that an alternative scenario that could better fit the observations may be the slow accretion of cold intergalactic gas, focused by a dense filament of galaxies in which this object is embedded (abridged).

astro-ph↗