Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Statistical error”

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.

At least 631 records · Page 35
All sourcesNASA NTRSDOE OSTI
DOE OSTI2018

Cross-polarization systematics due to Mizuguchi-Dragone condition breaking by a continuously rotating half-wave plate at prime focus in the Huan Tran telescope

Polarization modulation using a continuously rotating half-wave plate (HWP) is a promising technique to reduce both low-frequency noise and instrumental systematics for Cosmic Microwave Background (CMB) polarization measurements targeting in ationary B-modes. Although a HWP is best placed sky-side of the telescope optics in order to minimize systematics, >0.5 meter aperture class telescopes must put the HWP elsewhere in the optics chain due to current fabrication limitations in the available HWP size. Polarbear is a ground-based CMB experiment installed on the 2.5m aperture off-axis Gregorian-Dragone type Huan Tran Telescope (HTT) designed to satisfy the Mizuguchi-Dragone condition. Polarbear-2 is a receiver that will be installed on a second HTT in 2018. Polarbear-2 is designed to have a larger field-of-view (FOV) and vastly increased sensitivity to the polarized CMB compared to Polarbear. From the third season of observations, Polarbear has installed a continuously rotating HWP at the spatially localized focus plane between the HTT primary and secondary re ectors which is an optimal location for minimizing the HWP diameter. The HWP's polarization angle re ection with respect to its birefringent axis will theoretically break the Mizuguchi-Dragone condition when placed between the two reflectors and increase cross-polarization systematics. In this study, we analyze how the Mizuguchi-Dragone condition is violated due to a HWP at this location. We then estimate the crosspolarization systematics of the HTT using physical optics simulations. We model an ideal HWP at various angles to estimate the effects of demodulation. We evaluate the increased cross-polarization as the Stokes Q-U mixing term using the Mueller Matrix formalism. It is calculated that this term creates a varying dipole beam pattern whose amplitude ranges from 1% at the center to 10% at the edge FOV pixels for Polarbear and potentially up to 20% for Polarbear-2. We also estimate the leakage of the E-mode into the B-mode angular power spectrum measurements due to this cross-polarization. We show that the cross-polarization systematic error leakage is sufficiently lower than the Polarbear-2 statistical uncertainty thanks to mitigations such as focal plane averaging and sky rotation. Currently for Polarbear-2 we are planning to place the HWP at Gregorian focus, but keeping the HWP at prime focus as a back-up solution in case that there are unforeseen telescope spatial and HWP material size constraints. Through this study we find that even though a HWP between the two reflectors will violate the Mizuguchi-Dragone condition, this HWP at prime focus will still have sufficiently low cross-polarization for Polarbear-2. The prime focus HWP is a potential configuration that can be applied to similar off-axis Gregorian-Dragone telescopes in order to minimize the required HWP diameter.

Takakura, Satoru↗
DOE OSTI2014

Harmonic analysis for the characterization and correction of geometric distortion in MRI

Purpose: Magnetic resonance imaging (MRI) is gaining widespread use in radiation therapy planning, patient setup verification, and real-time guidance of radiation delivery. Successful implementation of these technologies relies on the development of simple and efficient methods to characterize and monitor the geometric distortions arising due to system imperfections and gradient nonlinearities. To this end, the authors present the theory and validation of a novel harmonic approach to the quantification of system-related distortions in MRI. Methods: The theory of spatial encoding in MRI is applied to demonstrate that the 3D distortion vector field (DVF) is given by the solution of a second-order boundary value problem (BVP). This BVP is comprised of Laplace’s equation and a limited measurement of the distortion on the boundary of a specified region of interest (ROI). An analytical series expansion solving this BVP within a spherical ROI is obtained, and a statistical uncertainty analysis is performed to determine how random errors in the boundary measurements propagate to the ROI interior. This series expansion is then evaluated to obtain volumetric DVF mappings that are compared to reference data obtained on a 3 T full-body scanner. This validation is performed within two spheres of 20 cm diameter (one centered at the scanner origin and the other offset +3 cm along each of the transverse directions). Initially, a high-order mapping requiring measurements at 5810 boundary points is used. Then, after exploring the impact of the boundary sampling density and the effect of series truncation, a reduced-order mapping requiring measurements at 302 boundary points is evaluated. Results: The volumetric DVF mappings obtained from the harmonic analysis are in good agreement with the reference data. Following distortion correction using the high-order mapping, the authors estimate a reduction in the mean distortion magnitude from 0.86 to 0.42 mm and from 0.93 to 0.39 mm within the central and offset ROIs, respectively. In addition, the fraction of points with a distortion magnitude greater than 1 mm is reduced from 35.6% to 2.8% and from 40.4% to 1.5%, respectively. Similarly, following correction using the reduced-order mapping, the mean distortion magnitude reduces to 0.45–0.42 mm within the central and offset ROIs, and the fraction of points with a distortion magnitude greater than 1 mm is reduced to 2.8% and 1.5%, respectively. Conclusions: A novel harmonic approach to the characterization of system-related distortions in MRI is presented. This method permits a complete and accurate mapping of the DVF within a specified ROI using a limited measurement of the distortion on the ROI boundary. This technique eliminates the requirement to exhaustively sample the DVF at a dense 3D array of points, thereby permitting the design of simple, inexpensive phantoms that may incorporate additional modules for auxiliary quality assurance objectives.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2025

CV4Quantum

CV4Quantum is a statistical technique for reducing the sampling overhead in probabilistic error cancellation, which is an error mitigation technique used in quantum computing. CV4Quantum is based on the control variates method, which is a Monte Carlo variance reduction technique. This repository contains the code and data associated with a demonstration of CV4Quantum using simulation experiments.

Shyamsundar, Prasanth [Fermi National Accelerator ↗
DOE OSTI2014

Data Assimilation - Advances and Applications

This presentation provides an overview of data assimilation (model calibration) for complex computer experiments. Calibration refers to the process of probabilistically constraining uncertain physics/engineering model inputs to be consistent with observed experimental data. An initial probability distribution for these parameters is updated using the experimental information. Utilization of surrogate models and empirical adjustment for model form error in code calibration form the basis for the statistical methodology considered. The role of probabilistic code calibration in supporting code validation is discussed. Incorporation of model form uncertainty in rigorous uncertainty quantification (UQ) analyses is also addressed. Design criteria used within a batch sequential design algorithm are introduced for efficiently achieving predictive maturity and improved code calibration. Predictive maturity refers to obtaining stable predictive inference with calibrated computer codes. These approaches allow for augmentation of initial experiment designs for collecting new physical data. A standard framework for data assimilation is presented and techniques for updating the posterior distribution of the state variables based on particle filtering and the ensemble Kalman filter are introduced.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2021

Scalable, In-situ Data Clustering Data Analysis for Extreme Scale Scientific Computing (Final Report)

The objective of this project is to address challenges in the design and development of scalable in-situ data clustering and analytics algorithms and software. Our goal is to develop parallel software consisting of a set of spatio-temporal data clustering and anomaly detection functions, both of which are very important for large-scale analysis and have wide applicability for in-situ runs as well as post-processing analysis. Our design principles for in-situ analysis consider the following: (1) identify parts of the computation can be done close to the data within the nodes, while it is still in memory; (2) extract analysis components can (and should) be performed in remote staging and analysis nodes; (3) develop error-bound approximation methods for applications tolerable for small errors; (4) identify the type of derived distributions and statistics, for spatio-temporal data, that can be kept locally in order to both accelerate computations and meet energy constraints in subsequent iterations and phases; (5) use a self-describing data format so that data can be consistent and understood among local storage (memory and SSDs) and at staging and analysis nodes, thereby providing portability and flexibility; (6) develop service-oriented functions that can schedule in-situ and post-hoc analysis tasks based on the dynamic requirements of applications. Our development focus is to produce the parallel data analysis software/library that will be scalable, reusable, extensible, and generic for applications in different disciplines. The software will be able to run in-situ with the simulations as well as post-hoc analysis. This approach will satisfy many synergistic requirements for data intensive applications executed on data coming from instruments and experiments. In particular, the proposed multilevel approach is directly applicable to perform design tradeoffs for running part of the algorithms near the instruments and the rest on remote (analysis) systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Implementing machine learning methods on QICK hardware for qubit readout & control

Quantum readout and control is a fundamental aspect of quantum computing that requires accurate measurement of qubit states. Errors emerge in all stages, from initialization to readout, and identifying errors in post-processing necessitates resource-intensive statistical analysis. In our work, we use a lightweight fully-connected neural network (NN) to classify states of a transmon system with no prior processing. Our NN accelerator yields higher fidelities (92%) than the classical matched filter method (84%). By exploiting the natural parallelism of NNs and their placement near the source of data on field-programmable gate arrays (FPGAs), we can achieve ultra-low latency on the Quantum Instrumentation Control Kit (QICK). Integrating machine learning methods on QICK opens several pathways for efficient real-time processing of quantum circuits.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2015

A collection of sub-daily pressure and temperature observations for the early instrumental period with a focus on the "year without a summer" 1816

The eruption of Mount Tambora (Indonesia) in April 1815 is the largest documented volcanic eruption in history. It is associated with a large global cooling during the following year, felt particularly in parts of Europe and North America, where the year 1816 became known as the "year without a summer". This paper describes an effort made to collect surface meteorological observations from the early instrumental period, with a focus on the years of and immediately following the eruption (1815–1817). Although the collection aimed in particular at pressure observations, correspondent temperature observations were also recovered. Some of the series had already been described in the literature, but a large part of the data, recently digitised from original weather diaries and contemporary magazines and newspapers, is presented here for the first time. The collection puts together more than 50 sub-daily series from land observatories in Europe and North America and from ships in the tropics. The pressure observations have been corrected for temperature and gravity and reduced to mean sea level. Moreover, an additional statistical correction was applied to take into account common error sources in mercury barometers. To assess the reliability of the corrected data set, the variance in the pressure observations is compared with modern climatologies, and single observations are used for synoptic analyses of three case studies in Europe. All raw observations will be made available to the scientific community in the International Surface Pressure Databank.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Qualitative and Quantitative Evaluation for Representative Human Reliability Analysis Methods

The Korea Institute of Nuclear Safety (KINS) is the regulatory expert organization established by the Korean government to strengthen the nation’s technical capabilities relating to nuclear safety regulation. KINS oversees the technical aspects of nuclear safety regulation, including safety reviews, inspections, education, and safety research—all conducted based on technical knowledge and accumulated regulatory experience. In 2023, KINS requested that Idaho National Laboratory (INL) validates representative human reliability analysis (HRA) methods used throughout the world, thus affording KINS with a basis for determining an HRA method adequate for its domestic regulatory purposes. The present paper mainly examines INL’s efforts in this regard. The resulting INL study covered four representative HRA methods widely used by nuclear utilities and regulatory institutes. These methods were qualitatively evaluated by applying specific evaluation criteria and determining how well each method reflected critical HRA issues. For this assessment, INL benchmarked the Halden International HRA Empirical Study. Using the Halden empirical data, along with information on human failure events (HFEs), the present study employed the selected HRA methods to estimate human error probabilities (HEPs) for the HFEs. It also performed statistical analyses to compare the HEPs predicted via the HRA methods against those from the Halden empirical data.

99 - GENERAL AND MISCELLANEOUS↗
DOE OSTI2019

Computational prediction of lattice thermal conductivity: A comparison of molecular dynamics and Boltzmann transport approaches

The predictive modeling of lattice thermal conductivity is of fundamental importance for the understanding and design of materials for a wide range of applications. Two major approaches, namely molecular dynamics (MD) simulations and calculations solving approximately the Boltzmann transport equation (BTE), have been developed to compute the lattice thermal conductivity. We present a detailed direct comparison of these two approaches, using as prototypical cases MgO and PbTe. The comparison, carried out using empirical potentials, takes into account the effects of fourth order phonon scattering, temperature-dependent phonon frequencies (phonon renormalization), and investigates the effects of quantum vs classical statistics. We clarify that equipartition, as opposed to Maxwell-Boltzmann, govern the statistics of phonons in MD simulations. We find that lattice thermal conductivity values from MD and BTE show an apparent, satisfactory agreement; however such an agreement is the result of error cancellations. As a result we also show that the primary effect of statistics on thermal conductivity is via the scattering rate dependence on phonon populations.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Optimizing Prediction Error for Time-dependent Solar Radiation Modeling

Numerical weather forecasting models and statistical methods have found wide use to help power companies estimate renewable output, but better methods are needed, particularly for extended forecasts. Machine learning approaches have been used here as well, but so far a major limitation is the ability to also predict the corresponding uncertainty in a forecast. Here we show that both can be done and demonstrate this using a long-term-short memory neural network where the difference between predicted and ground truth data are used to train a model for the corresponding forecast uncertainties.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

The orbit and transit prospects for β pictoris b constrained with one milliarcsecond astrometry

A principal scientific goal of the Gemini Planet Imager (GPI) is obtaining milliarcsecond astrometry to constrain exoplanet orbits. However, astrometry of directly imaged exoplanets is subject to biases, systematic errors, and speckle noise. Here, we describe an analytical procedure to forward model the signal of an exoplanet that accounts for both the observing strategy (angular and spectral differential imaging) and the data reduction method (Karhunen–Loève Image Projection algorithm). We use this forward model to measure the position of an exoplanet in a Bayesian framework employing Gaussian processes and Markov-chain Monte Carlo to account for correlated noise. In the case of GPI data on β Pic b, this technique, which we call Bayesian KLIP-FM Astrometry (BKA), outperforms previous techniques and yields 1σ errors at or below the one milliarcsecond level. We validate BKA by fitting a Keplerian orbit to 12 GPI observations along with previous astrometry from other instruments. The statistical properties of the residuals confirm that BKA is accurate and correctly estimates astrometric errors. Our constraints on the orbit of β Pic b firmly rule out the possibility of a transit of the planet at 10-σ significance. However, we confirm that the Hill sphere of β Pic b will transit, giving us a rare chance to probe the circumplanetary environment of a young, evolving exoplanet. As a result, we provide an ephemeris for photometric monitoring of the Hill sphere transit event, which will begin at the start of April in 2017 and finish at the end of January in 2018.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Evaluation of the Four-Dimensional Ensemble-Variational Hybrid Data Assimilation with Self-Consistent Regional Background Error Covariance for Improved Hurricane Intensity Forecasts

The feasibility of a hurricane initialization framework based on the Gridpoint Statistical Interpolation (GSI)-based four-dimensional ensemble-variational (GSI-4DEnVar) hybrid data assimilation system for the Hurricane Weather Research and Forecasting model (HWRF) model is evaluated in this study. The system considers the temporal evolution of error covariances via the use of four-dimensional ensemble perturbations that are provided by high-resolution, self-consistent HWRF ensemble forecasts. It is different from the configuration of the GSI-based three-dimensional ensemble-variational (GSI-3DEnVar) hybrid data assimilation system, similar to that used in the operational HWRF, which employs background error covariances provided by coarser-resolution global ensembles from the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS) ensemble Kalman filtering data assimilation system. In addition, our proposed initialization framework discards the empirical intensity correction in the vortex initialization package that is employed by the GSI-3DEnVar initialization framework in operational HWRF. Data assimilation and numerical simulation experiments for Hurricanes Joaquin (2015), Patricia (2015), and Matthew (2016) are conducted during their intensity changes. The impacts of two initialization frameworks on the HWRF analyses and forecasts are compared. It is found that GSI-4DEnVar leads to a reduction in track, minimum sea level pressure (MSLP), and maximum surface wind (MSW) forecast errors in all of the HWRF simulations, compared with the GSI-3DEnVar initialization framework. With assimilating high-resolution observations within the hurricane inner-core region, GSI-4DEnVar can produce the initial hurricane intensity reasonably well without the empirical vortex intensity correction. Further diagnoses with Hurricane Joaquin indicate that GSI-4DEnVar can significantly alleviate the imbalances in the initial conditions and enhance the performance of the data assimilation and subsequent hurricane intensity and precipitation forecasts

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2013

A Comparison of Forecast Error Generators for Modeling Wind and Load Uncertainty

This paper presents four algorithms to generate random forecast error time series, including a truncated-normal distribution model, a state-space based Markov model, a seasonal autoregressive moving average (ARMA) model, and a stochastic-optimization based model. The error time series are used to create real-time (RT), hour-ahead (HA), and day-ahead (DA) wind and load forecast time series that statistically match historically observed forecasting data sets, used for variable generation integration studies. A comparison is made using historical DA load forecast and actual load values to generate new sets of DA forecasts with similar stoical forecast error characteristics. This paper discusses and compares the capabilities of each algorithm to preserve the characteristics of the historical forecast data sets.

Lu, Ning↗
DOE OSTI2021

Reynolds stress tensor measurements using magnetic resonance velocimetry: expansion of the dynamic measurement range and analysis of systematic measurement errors

This study presents magnetic resonance velocimetry (MRV) Reynolds Stress measurements in a periodic hill channel with a hill Reynolds number of Re = 29,500. The velocity encoding scheme is based on the ICOSA6 method with six icosahedral encoding directions and multiple encoding values are measured to increase the dynamic range. The full Reynolds stress tensor is obtained from a voxel-wise three-dimensional Gaussian fit using the magnitude data of all acquisitions. The MRV results are compared to a wall-resolved large eddy simulation and laser Doppler velocimetry measurements conducted in the same channel. It is shown that the MRV Reynolds stress data have excellent precision and agree qualitatively with the reference data. However, there are apparent systematic deviations. One of the most prominent error contributions is the signal attenuation caused by higher orders of motion, which leads to an overestimation of the turbulence level. Another fundamental error is identified in the assumption that the turbulence is Gaussian distributed. With the presented reconstruction technique, the MRV data are fitted to a statistical model, and depending on the examined flow setup, the Gaussian model can lead to considerable errors. Possible ways of how to reduce all identified errors are presented. In summary, this technique enables Reynolds stress tensor measurements in complex internal flows with high dynamic range and excellent precision. However, several issues need to be resolved to make the turbulence quantification more accurate.

42 ENGINEERING↗
DOE OSTI2022

ARETE: Accurate Error Assessment via Machine Learning-Guided Dynamic-Timing Analysis

Nanometer circuits are increasingly prone to timing errors, escalating the need for fault injection frameworks to accurately evaluate their impact on applications. Here in this paper, we propose ARETE, a novel cross-layer, fault-injection framework that combines dynamic-binary instrumentation with machine learning-guided dynamic-timing analysis. ARETE enables accurate fault-injection into any application by estimating the location of the injecting errors via dynamic-timing analysis. To accelerate fault-injection, we develop a novel, data-aware, machine learning-based mechanism that dynamically pre-selects the error-prone instructions and limits the application of the costly dynamic-timing analysis only to them. To evaluate ARETE's accuracy, our fully automated toolflow is configured to support fault-injection based on detailed post-layout gate-level simulations as well as via existing workload-agnostic error models. Our results for various workloads, including an autonomous-driving library, show that the location and time of injected errors performed by ARETE, is 89.9% consistent with fault-injection based on full gate-level simulation. On average, ARETE executes 84.6x faster than gate-level simulation and at a cost of 3.4% loss in the program output quality estimation. When compared to the existing statistical fault-injection tools that are based on workload-agnostic error models, ARETE improves the accuracy of fault-injection rate and output quality estimation by 143.9% and 40.4% on average, respectively.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

KiDS-1000 cosmology: Combined second- and third-order shear statistics

Aims.In this work, we perform the first cosmological parameter analysis of the fourth release of Kilo Degree Survey (KiDS-1000) data with second- and third-order shear statistics. This paper builds on a series of studies aimed at describing the roadmap to third-order shear statistics. Methods.We derived and tested a combined model of the second-order shear statistic, namely, the COSEBIs and the third-order aperture mass statistics 〈ℳ ap 3 〉 in a tomographic set-up. We validated our pipeline withN-body mock simulations of the KiDS-1000 data release. To model the second- and third-order statistics, we used the latest version of HMCODE2020 for the power spectrum and BIHALOFITfor the bispectrum. Furthermore, we used an analytic description to model intrinsic alignments and hydro-dynamical simulations to model the effect of baryonic feedback processes. Lastly, we decreased the dimension of the data vector significantly by considering only equal smoothing radii for the 〈ℳ ap 3 〉 part of the data vector. This makes it possible to carry out a data analysis of the KiDS-1000 data release using a combined analysis of COSEBIs and third-order shear statistics. Results.We first validated the accuracy of our modelling by analysing a noise-free mock data vector, assuming the KiDS-1000 error budget, finding a shift in the maximum of the posterior distribution of the matter density parameter, ΔΩ m < 0.02 σ Ω m , and of the structure growth parameter, ΔS 8 < 0.05 σ S 8 . Lastly, we performed the first KiDS-1000 cosmological analysis using a combined analysis of second- and third-order shear statistics, where we constrained Ω m = 0.248 −0.055 +0.062 andS 8 = σ 8 √(Ω m /0.3 )= 0.772 ± 0.022. The geometric average on the errors of Ω m andS 8 of the combined statistics decreases, compared to the second-order statistic, by a factor of 2.2.

Astronomy & Astrophysics↗
DOE OSTI2013

Assessment of Interfraction Patient Setup for Head-and-Neck Cancer Intensity Modulated Radiation Therapy Using Multiple Computed Tomography-Based Image Guidance

Purpose: Various image guidance systems are commonly used in conjunction with intensity modulated radiation therapy (IMRT) in head-and-neck cancer irradiation. The purpose of this study was to assess interfraction patient setup variations for 3 computed tomography (CT)-based on-board image guided radiation therapy (IGRT) modalities. Methods and Materials: A total of 3302 CT scans for 117 patients, including 53 patients receiving megavoltage cone-beam CT (MVCBCT), 29 receiving kilovoltage cone-beam CT (KVCBCT), and 35 receiving megavoltage fan-beam CT (MVFBCT), were retrospectively analyzed. The daily variations in the mediolateral (ML), craniocaudal (CC), and anteroposterior (AP) dimensions were measured. The clinical target volume-to-planned target volume (CTV-to-PTV) margins were calculated using 2.5Σ + 0.7 σ, where Σ and σ were systematic and random positioning errors, respectively. Various patient characteristics for the MVCBCT group, including weight, weight loss, tumor location, and initial body mass index, were analyzed to determine their possible correlation with daily patient setup. Results: The average interfraction displacements (± standard deviation) in the ML, CC, and AP directions were 0.5 ± 1.5, −0.3 ± 2.0, and 0.3 ± 1.7 mm (KVCBCT); 0.2 ± 1.9, −0.2 ± 2.4, and 0.0 ± 1.7 mm (MVFBCT); and 0.0 ± 1.8, 0.5 ± 1.7, and 0.8 ± 3.0 mm (MVCBCT). The day-to-day random errors for KVCBCT, MVFBCT, and MVCBCT were 1.4-1.6, 1.7, and 2.0-2.1 mm. The interobserver variations were 0.8, 1.1, and 0.7 mm (MVCBCT); 0.5, 0.4, and 0.8 mm (MVFBCT); and 0.5, 0.4, and 0.6 mm (KVCBCT) in the ML, CC, and AP directions, respectively. The maximal calculated uniform CTV-to-PTV margins were 5.6, 6.9, and 8.9 mm for KVCBCT, MVFBCT, and MVCBCT, respectively. For the evaluated patient characteristics, the calculated margins for different patient parameters appeared to differ; analysis of variance (ANOVA) and/or t test analysis found no statistically significant setup difference in any direction. Conclusions: Daily random setup errors and CTV-to-PTV margins for treatment of head-and-neck cancer were affected by imaging quality. Our data indicated that larger margins were associated with MVFBCT and MVCBCT, compared with smaller margins for KVCBCT. IGRT modalities with better image quality are encouraged in clinical practice.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2021

Cosmic Inference: Constraining Parameters with Observations and a Highly Limited Number of Simulations

We look at cosmological probes that pose an inverse problem where the measurement result is obtained through observations, and the objective is to infer values of model parameters that characterize the underlying physical system-our universe, from these observations and theoretical forward-modeling. The only way to accurately forward-model physical behavior on small scales is via expensive numerical simulations, which are further “emulated” due to their high cost. Emulators are commonly built with a set of simulations covering the parameter space with Latin hypercube sampling and an interpolation procedure; the aim is to establish an approximately constant prediction error across the hypercube. In this paper, we provide a description of a novel statistical framework for obtaining accurate parameter constraints. The proposed framework uses multi-output Gaussian process emulators that are adaptively constructed using Bayesian optimization methods with the goal of maintaining a low emulation error in the region of the hypercube preferred by the observational data. In this paper, we compare several approaches for constructing multi-output emulators that enable us to take possible inter-output correlations into account while maintaining the efficiency needed for inference. Using a Lyα forest flux power spectrum, we demonstrate that our adaptive approach requires considerably fewer-by a factor of a few in the Lyα P(k) case considered here-simulations compared to the emulation based on Latin hypercube sampling, and that the method is more robust in reconstructing parameters and their Bayesian credible intervals.

79 ASTRONOMY AND ASTROPHYSICS↗