Cosmology from HSC Y1 weak lensing data with combined higher-order statistics and simulation-based inference
Not Available
Novaes, Camila P. (ORCID:0000000258453649)↗
SEARCH · Engineering Papers
Results for “likelihood function”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Analysis of Polarimetry Data with Angular Uncertainties
For a track based polarimeter, such as the Imaging X-ray Polarimetry Explorer (IXPE), the sensitivity to polarization depends on the modulation factor, which is a strong function of energy. In previous work, a likelihood method was developed that would account for this variation in order to estimate the minimum detectable polarization (MDP). That method essentially required that the position angles of individual events should be known precisely. In a separate work, however, it was shown that using a machine-learning method for measuring event tracks can generate track angle uncertainties, which can be used in the analysis. Here, the maximum likelihood method is used as a basis for revising the estimate of the MDP in a general way that can include uncertainties in event track position angles. The resultant MDP depends solely upon the distribution of track angle uncertainties present in the input data. Due to the physics of the IXPE detectors, it is possible to derive a simple relationship between these angular uncertainties and the energy-dependent modulation function as a step in the process.
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
Learning likelihood ratios with neural network classifiers
The likelihood ratio is a crucial quantity for statistical inference in science that enables hypothesis testing, construction of confidence intervals, reweighting of distributions, and more. Many modern scientific applications, however, make use of data- or simulation-driven models for which computing the likelihood ratio can be very difficult or even impossible. By applying the so-called “likelihood ratio trick,” approximations of the likelihood ratio may be computed using clever parametrizations of neural network-based classifiers. A number of different neural network setups can be defined to satisfy this procedure, each with varying performance in approximating the likelihood ratio when using finite training data. We present a series of empirical studies detailing the performance of several common loss functionals and parametrizations of the classifier output in approximating the likelihood ratio of two univariate and multivariate Gaussian distributions as well as simulated high-energy particle physics datasets.
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
Solving inverse problems in stochastic models using deep neural networks and adversarial training
Inverse problems associated with stochastic models constitute a significant portion of scientific and engineering applications. In such cases the unknown quantities are distributions. The applicability of traditional methods is limited because of their demanding assumptions or prohibitive computational consumption; for example, maximum likelihood methods require closed-form density functions, and Markov Chain Monte Carlo needs a large number of simulations. We propose a new method that estimates the unknown distribution by matching the statistical properties between observed and simulated random processes. We leverage the expressive power of neural networks to approximate the unknown distribution and use a discriminative neural network for computing the statistical discrepancies between the observed and simulated random processes. Here we demonstrated numerically that the proposed methods can estimate both the model parameters and learn complicated unknown distributions.
42 ENGINEERING↗
Latent map Gaussian processes for mixed variable metamodeling
Gaussian processes (GPs) are ubiquitously used in sciences and engineering as metamodels. Standard GPs, however, can only handle numerical or quantitative variables. Here we introduce latent map Gaussian processes (LMGPs) that inherit the attractive properties of GPs and are also applicable to mixed data which have both quantitative and qualitative inputs. The core idea behind LMGPs is to learn a continuous, low-dimensional latent space or manifold which encodes all qualitative inputs. To learn this manifold, we first assign a unique prior vector representation to each combination of qualitative inputs. We then use a low-rank linear map to project these priors on a manifold that characterizes the posterior representations. As the posteriors are quantitative, they can be directly used in any standard correlation function such as the Gaussian or Matern. Hence, the optimal map and the corresponding manifold, along with other hyperparameters of the correlation function, can be systematically learned via maximum likelihood estimation. Through a wide range of analytic and real-world examples, we demonstrate the advantages of LMGPs over state-of-the-art methods in terms of accuracy and versatility. In particular, we show that LMGPs can handle variable-length inputs, have an explainable neural network interpretation, and provide insights into how qualitative inputs affect the response or interact with each other. We also employ LMGPs in Bayesian optimization and illustrate that they can discover optimal compound compositions more efficiently than conventional methods that convert compositions to qualitative variables via manual featurization.
42 ENGINEERING↗
Dark Energy Survey Year 3 Results: clustering redshifts – calibration of the weak lensing source redshift distributions with redMaGiC and BOSS/eBOSS
ABSTRACT We present the calibration of the Dark Energy Survey Year 3 (DES Y3) weak lensing (WL) source galaxy redshift distributions n(z) from clustering measurements. In particular, we cross-correlate the WL source galaxies sample with redMaGiC galaxies (luminous red galaxies with secure photometric redshifts) and a spectroscopic sample from BOSS/eBOSS to estimate the redshift distribution of the DES sources sample. Two distinct methods for using the clustering statistics are described. The first uses the clustering information independently to estimate the mean redshift of the source galaxies within a redshift window, as done in the DES Y1 analysis. The second method establishes a likelihood of the clustering data as a function of n(z), which can be incorporated into schemes for generating samples of n(z) subject to combined clustering and photometric constraints. Both methods incorporate marginalization over various astrophysical systematics, including magnification and redshift-dependent galaxy-matter bias. We characterize the uncertainties of the methods in simulations; the first method recovers the mean z of tomographic bins to RMS (precision) of ∼0.014. Use of the second method is shown to vastly improve the accuracy of the shape of n(z) derived from photometric data. The two methods are then applied to the DES Y3 data.
79 ASTRONOMY AND ASTROPHYSICS↗
Chemical Analysis of the Ultrafaint Dwarf Galaxy Grus II. Signature of High-mass Stellar Nucleosynthesis
We present a detailed abundance analysis of the three brightest member stars at the top of the giant branch of the ultrafaint dwarf (UFD) galaxy Grus II. All stars exhibit a higher than expected [Mg/Ca] ratio compared to metal-poor stars in other UFD galaxies and in the Milky Way (MW) halo. Nucleosynthesis in high-mass ( 20 M ⊙) core-collapse supernovae has been shown to create this signature. The abundances of this small sample (three) stars suggests the chemical enrichment of Grus II could have occurred through substantial high-mass stellar evolution, and is consistent with the framework of a top-heavy initial mass function. However, with only three stars it cannot be ruled out that the abundance pattern is the result of a stochastic chemical enrichment at early times in the galaxy. The most metal-rich of the three stars also possesses a small enhancement in rapid neutron-capture (r-process) elements. The abundance pattern of the r-process elements in this star matches the scaled r-process pattern of the solar system and r-process enhanced stars in other dwarf galaxies and in the MW halo, hinting at a common origin for these elements across a range of environments. All current proposed astrophysical sites of r-process element production are associated with high-mass stars, thus the possible top-heavy initial mass function of Grus II would increase the likelihood of any of these events occurring. The time delay between the α and r-process element enrichment of the galaxy favors a neutron star merger as the origin of the r-process elements in Grus II.
79 ASTRONOMY AND ASTROPHYSICS↗
Efficient high-dimensional variational data assimilation with machine-learned reduced-order models
Abstract. Data assimilation (DA) in geophysical sciences remains the cornerstone of robust forecasts from numerical models. Indeed, DA plays a crucial role in the quality of numerical weather prediction and is a crucial building block that has allowed dramatic improvements in weather forecasting over the past few decades. DA is commonly framed in a variational setting, where one solves an optimization problem within a Bayesian formulation using raw model forecasts as a prior and observations as likelihood. This leads to a DA objective function that needs to be minimized, where the decision variables are the initial conditions specified to the model. In traditional DA, the forward model is numerically and computationally expensive. Here we replace the forward model with a low-dimensional, data-driven, and differentiable emulator. Consequently, gradients of our DA objective function with respect to the decision variables are obtained rapidly via automatic differentiation. We demonstrate our approach by performing an emulator-assisted DA forecast of geopotential height. Our results indicate that emulator-assisted DA is faster than traditional equation-based DA forecasts by 4 orders of magnitude, allowing computations to be performed on a workstation rather than a dedicated high-performance computer. In addition, we describe accuracy benefits of emulator-assisted DA when compared to simply using the emulator for forecasting (i.e., without DA). Our overall formulation is denoted AIEADA (Artificial Intelligence Emulator-Assisted Data Assimilation).
58 GEOSCIENCES↗
A possible mass distribution of primordial black holes implied by LIGO-Virgo
The LIGO-Virgo Collaboration has so far detected around 90 black holes, some of which have masses larger than what were expected from the collapse of stars. The mass distribution of LIGO-Virgo black holes appears to have a peak at ~ 30 M ⊙ and two tails on the ends. By assuming that they all have a primordial origin, we analyze the GWTC-1 (O1&O2) and GWTC-2 (O3a) datasets by performing maximum likelihood estimation on a broken power law mass function f ( m ), with the result f ∝ m 1.2 for m < 35 M ⊙ and f ∝ m -4 for m > 35 M ⊙ . This appears to behave better than the popular log-normal mass function. Surprisingly, such a simple and unique distribution can be realized in our previously proposed mechanism of PBH formation, where the black holes are formed by vacuum bubbles that nucleate during inflation via quantum tunneling. Moreover, this mass distribution can also provide an explanation to supermassive black holes formed at high redshifts.
Astronomy & Astrophysics↗
Neural simulation-based inference of the neutron star equation of state directly from telescope spectra
Neutron stars provide a unique opportunity to study strongly interacting matter under extreme density conditions. The intricacies of matter inside neutron stars and their equation of state are not directly visible, but determine bulk properties, such as mass and radius, which affect the star's thermal X-ray emissions. However, the telescope spectra of these emissions are also affected by the stellar distance, hydrogen column, and effective surface temperature, which are not always well-constrained. Uncertainties on these nuisance parameters must be accounted for when making a robust estimation of the equation of state. In this study, we develop a novel methodology that, for the first time, can infer the full posterior distribution of both the equation of state and nuisance parameters directly from telescope observations. This method relies on the use of neural likelihood estimation, in which normalizing flows use samples of simulated telescope data to learn the likelihood of the neutron star spectra as a function of these parameters, coupled with Hamiltonian Monte Carlo methods to efficiently sample from the corresponding posterior distribution. Our approach surpasses the accuracy of previous methods, improves the interpretability of the results by providing access to the full posterior distribution, and naturally scales to a growing number of neutron star observations expected in the coming years.
79 ASTRONOMY AND ASTROPHYSICS↗
2D-FFTLog: efficient computation of real-space covariance matrices for galaxy clustering and weak lensing
ABSTRACT Accurate covariance matrices for two-point functions are critical for inferring cosmological parameters in likelihood analyses of large-scale structure surveys. Among various approaches to obtaining the covariance, analytic computation is much faster and less noisy than estimation from data or simulations. However, the transform of covariances from Fourier space to real space involves integrals with two Bessel integrals, which are numerically slow and easily affected by numerical uncertainties. Inaccurate covariances may lead to significant errors in the inference of the cosmological parameters. In this paper, we introduce a 2D-FFTLog algorithm for efficient, accurate, and numerically stable computation of non-Gaussian real-space covariances for both 3D and projected statistics. The 2D-FFTLog algorithm is easily extended to perform real-space bin-averaging. We apply the algorithm to the covariances for galaxy clustering and weak lensing for a Dark Energy Survey Year 3-like and a Rubin Observatory’s Legacy Survey of Space and Time Year 1-like survey, and demonstrate that for both surveys, our algorithm can produce numerically stable angular bin-averaged covariances with the flat sky approximation, which are sufficiently accurate for inferring cosmological parameters. The code CosmoCov for computing the real-space covariances with or without the flat-sky approximation is released along with this paper.
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
Prospects for Characterizing the Haziest Sub-Neptune Exoplanets with High-resolution Spectroscopy
Observations to characterize planets larger than Earth but smaller than Neptune have led to largely inconclusive interpretations at low spectral resolution due to hazes or clouds that obscure molecular features in their spectra. However, here we show that high-resolution spectroscopy (R ∼ 25,000–100,000) enables one to probe the regions in these atmospheres above the clouds where the cores of the strongest spectral lines are formed. We present models of transmission spectra for a suite of GJ 1214b–like planets with thick photochemical hazes covering 1–5 μm at a range of resolutions relevant to current and future ground-based spectrographs. Furthermore, we compare the utility of the cross-correlation function that is typically used with a more formal likelihood-based approach, finding that only the likelihood-based method is sensitive to the presence of haze opacity. We calculate the signal-to-noise ratio (S/N) of these spectra, including telluric contamination, Required to robustly detect a host of molecules such as CO, CO{sub 2}, H{sub 2}O, and CH{sub 4} and photochemical products like HCN as a function of wavelength range and spectral resolution. Spectra in the M band require the lowest S/N{sub res} to detect multiple molecules simultaneously. CH{sub 4} is only observable for the coolest models (T {sub eff} = 412 K) and only in the L band. We quantitatively assess how these requirements compare to what is achievable with current and future instruments, demonstrating that characterization of small cool worlds with ground-based high-resolution spectroscopy is well within reach.
79 ASTRONOMY AND ASTROPHYSICS↗
Pervasive alterations to snow-dominated ecosystem functions under climate change
Climate change projections consistently demonstrate that warming temperatures and dwindling seasonal snowpack will elicit cascading effects on ecosystem function and water resource availability. Despite this consensus, little is known about potential changes in the variability of ecohydrological conditions, which is also required to inform climate change adaptation and mitigation strategies. Considering potential changes in ecohydrological variability is critical to evaluating the emergence of trends, assessing the likelihood of extreme events such as floods and droughts, and identifying when tipping points may be reached that fundamentally alter ecohydrological function. Using a single-model Large Ensemble with sophisticated terrestrial ecosystem representation, we characterize projected changes in the mean state and variability of ecohydrological processes in historically snow-dominated regions of the Northern Hemisphere. Widespread snowpack reductions, earlier snowmelt timing, longer growing seasons, drier soils, and increased fire risk are projected for this century under a high-emissions scenario. In addition to these changes in the mean state, increased variability in winter snowmelt will increase growing-season water deficits and increase the stochasticity of runoff. Thus, with warming, declining snowpack loses its dependable buffering capacity so that runoff quantity and timing more closely reflect the episodic characteristics of precipitation. This results in a declining predictability of annual runoff from maximum snow water equivalent, which has critical implications for ecosystem stress and water resource management. So our results suggest that there is a strong likelihood of pervasive alterations to ecohydrological function that may be expected with climate change.
54 ENVIRONMENTAL SCIENCES↗
Orbits and Masses of Binaries from Speckle Interferometry at SOAR
We present results from Speckle inteferometric observations of 15 visual binaries and one double-line spectroscopic binary, carried out with the HRCam Speckle camera of the SOAR 4.1 m telescope. These systems were observed as a part of an on-going survey to characterize the binary population in the solar vicinity, out to a distance of 250 pc. We obtained orbital elements and mass sums for our sample of visual binaries. The orbits were computed using a Markov Chain Monte Carlo algorithm that delivers maximum likelihood estimates of the parameters, as well as posterior probability density functions that allow us to evaluate their uncertainty. Their periods cover a range from 5 yr to more than 500 yr; and their spectral types go from early A to mid M, implying total system masses from slightly more than 4M{sub ⊙} down to 0.2M {sub ⊙}. They are located at distances between approximately 12 and 200 pc, mostly at low Galactic latitude. For the double-line spectroscopic binary YSC8, we present the first combined astrometric/radial-velocity orbit resulting from a self-consistent fit, leading to individual component masses of 0.897 ± 0.027 M {sub ⊙} and 0.857 ± 0.026 M {sub ⊙}; and an orbital parallax of 26.61 ± 0.29 mas, which compares very well with the Gaia DR2 trigonometric parallax (26.55 ± 0.27 mas). In combination with published photometry and trigonometric parallaxes, we place our objects on an H-R diagram and discuss their evolutionary status. We also present a thorough analysis of the precision and consistency of the photometry available for them.
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
System risk quantification and decision making support using functional modeling and dynamic Bayesian network
Risk-informed decision-making requires a probabilistic assessment of the likelihood of success of control action, given the system status. This paper presents a systematic state transition modeling approach integrating dynamic probabilistic risk assessment with a decision-making process using a dynamic Bayesian network (DBN) coupled with functional modeling. A functional model designed with multilevel flow modeling (MFM) technique was used to build a system state structure inferred by energy, mass, and information flow so that one can verify the developed model with respect to system functionality. The MFM model represents the causal relationship among the nodes, which captures the structure of process parameters and control units. Each node may have multiple possible states, and the DBN structured by the MFM model represents the time-domain transitions among the defined states. Furthermore, the MFM-DBN integrated state transition modeling is a white-box approach that allows one to draw the system's risk profile by updating the system states and supports the decisions probabilistically with physical inference. An example of a simple heating system has been used to illustrate this process, including decision-making support based on quantitative risk profile. For demonstrating its applicability to a complex system operational decision making, a case study of station blackout accident scenario leading to the seal loss of coolant accident in a nuclear power plant is presented. The proposed approach effectively provided the risk profile along time for each option so that the operators can make the best decision, which minimizes the plant risk.
42 ENGINEERING↗
Massive νs through the CNN lens: interpreting the field-level neutrino mass information in weak lensing
Modern cosmological surveys probe the Universe deep into the nonlinear regime, where massive neutrinos suppress cosmic structure. Traditional cosmological analyses, which use the 2-point correlation function to extract information, are no longer optimal in the nonlinear regime, and there is thus much interest in extracting beyond-2-point information to improve constraints on neutrino mass. Quantifying and interpreting the beyond-2-point information is thus a pressing task. We study the field-level information in weak lensing convergence maps using convolution neural networks. We find that the network performance increases as higher source redshifts and smaller scales are considered — investigating up to a source redshift of 2.5 and ℓ max ≃ 10 4 — verifying that massive neutrinos leave a distinct effect on weak lensing. However, the performance of the network significantly drops after scaling out the 2-point information from the maps, implying that most of the field-level information can be found in the 2-point correlation function alone. We quantify these findings in terms of the likelihood ratio and also use Integrated Gradient saliency maps to interpret which parts of the map the network is learning the most from. We find that, in the absence of noise, the network extracts a similar amount of information from the most overdense and underdense regions. However, upon adding noise, the information in underdense regions is distorted as noise disproportionately washes out void-like structures.
Golshan, Malika [University of California, Berkele↗
A composite likelihood approach for inference under photometric redshift uncertainty
ABSTRACT Obtaining accurately calibrated redshift distributions of photometric samples is one of the great challenges in photometric surveys like LSST, Euclid, HSC, KiDS, and DES. We present an inference methodology that combines the redshift information from the galaxy photometry with constraints from two-point functions, utilizing cross-correlations with spatially overlapping spectroscopic samples, and illustrate the approach on CosmoDC2 simulations. Our likelihood framework is designed to integrate directly into a typical large-scale structure and weak lensing analysis based on two-point functions. We discuss efficient and accurate inference techniques that allow us to scale the method to the large samples of galaxies to be expected in LSST. We consider statistical challenges like the parametrization of redshift systematics, discuss and evaluate techniques to regularize the sample redshift distributions, and investigate techniques that can help to detect and calibrate sources of systematic error using posterior predictive checks. We evaluate and forecast photometric redshift performance using data from the CosmoDC2 simulations, within which we mimic a DESI-like spectroscopic calibration sample for cross-correlations. Using a combination of spatial cross-correlations and photometry, we show that we can provide calibration of the mean of the sample redshift distribution to an accuracy of at least 0.002(1 + z), consistent with the LSST-Y1 science requirements for weak lensing and large-scale structure probes.
(cosmology:) large-scale structure of Universe↗
Modeling the Functional Forms of Grid Disturbances
This report introduces a functional form that may be used to quantitatively predict the impacts of new grid tools and changing system qualities on the likelihoods, durations, and depths of various grid disturbances. Each disturbance scenario is modeled to have three component stages—avoidance, reactance and recovery, which together parametrically estimate one disturbance’s impacts. The modeled scenario is then placed and replicated within an analysis period to represent the likelihood or frequency of the scenario and its consequent impacts. Whereas analysts have struggled to define and apply metrics for grid resilience, the functional form introduced by this report shares units of measurement with accepted grid-status measures (e.g., numbers of customers currently experiencing a service outage). Furthermore, the integrated and averaged functional form over an analysis period provides a meaningful normalized performance metric (e.g., customer outage minutes per year) that is ultimately independent of the duration of the period of. The approach may be applied similarly regardless of the severity or frequency of the disturbances that are being analyzed. Because metrics can be chosen to be identical in both the hypothetical future and the actual historical past, the historical past eventually becomes the test of the future predictions, at least in a statistical sense. The authors originally developed this approach to facilitate analysis of the effects of transactive energy (TE) systems effects on electric power grid resilience. TE systems invite energy suppliers and consumers to actively collaborate toward the discovery of, and their responses to, the locational value of energy. The findings from this process are often embodied as energy prices, the dynamics of which indicate the locational value of energy and can further represent important grid service needs. While some academic papers claim to quantify the value of a specific TE system design toward grid resilience, the answer, in general, has been elusive. Not only do multiple and conflicting definitions of resilience and reliability exist, but countless TE systems are being invented. We conclude the following: (1) The ideal analysis should harmonize rather than differentiate resilience and reliability. Therefore, this report uses the more general term disturbance whenever the overloaded terms resilience and reliability can be avoided. (2) The effectiveness of TE systems must be mapped to underlying qualities of a TE system, thereby avoiding presumptions that every TE design offers similar advantages. The authors seek to evaluate the parametric effects of TE system qualities (e.g., spatial granularity, granularity of time steps, length of future prediction horizon) on avoiding, reacting to, and recovering from grid disturbances. Furthermore, any advantages (or disadvantages) must be fairly compared with the many alternative tools, systems, and strategies that might offer comparable benefits.
24 POWER TRANSMISSION AND DISTRIBUTION↗