Search NASA⌕ Search

SEARCH · Search NASA

Results for “Covariance matrices”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

The XFaster Power Spectrum and Likelihood Estimator for the Analysis of Cosmic Microwave Background Maps

We present the XFaster analysis package, XFaster is a fast, iterative angular power spectrum estimator based on a diagonal approximation to the quadratic Fisher matrix estimator. XFaster uses Monte Carlo simulations to compute noise biases and filter transfer functions and is thus a hybrid of both Monte Carlo and quadratic estimator methods. In contrast to conventional pseudo-C ℓ based methods, the algorithm described here requires a minimal number of simulations, and does not require them to be precisely representative of the data to estimate accurate covariance matrices for the bandpowers. The formalism works with polarization-sensitive observations and also data sets with identical, partially overlapping, or independent survey regions. The method was first implemented for the analysis of BOOMERanG data (Netterfield et al. 2002; Jones et al. 2006), and also used as part of the Planck analysis (Rocha et al. 2011). Here, we describe the full, publicly available analysis package, written in Python, as developed for the analysis of data from the 2015 flight of the SPIDER instrument (SPIDER Collaboration 2021). The package includes extensions for self-consistently estimating null spectra and for estimating fits for Galactic foreground contributions. We show results from the extensive validation of XFaster using simulations, and its application to the SPIDER data set.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Testing the ΛCDM Cosmological Model with Forthcoming Measurements of the Cosmic Microwave Background with SPT-3G

We forecast constraints on cosmological parameters enabled by three surveys conducted with SPT-3G, the third-generation camera on the South Pole Telescope. The surveys cover separate regions of 1500, 2650, and 6000 deg$^{2}$ to different depths, in total observing 25% of the sky. These regions will be measured to white noise levels of roughly 2.5, 9, and 12μK-armin, respectively, in cosmic microwave background (CMB) temperature units at 150 GHz by the end of 2024. The survey also includes measurements at 95 and 220 GHz, which have noise levels a factor of ∼1.2 and 3.5 times higher than 150 GHz, respectively, with each band having a polarization noise level ∼2times higher than the temperature noise. We use a novel approach to obtain the covariance matrices for jointly and optimally estimated gravitational lensing potential band powers and unlensed CMB temperature and polarization band powers. We demonstrate the ability to test the ΛCDM model via the consistency of cosmological parameters constrained independently from SPT-3G and Planck data, and consider the improvement in constraints on ΛCDM extension parameters from a joint analysis of SPT-3G and Planck data. The ΛCDM cosmological parameters are typically constrained with uncertainties up to ∼2 times smaller with SPT-3G data, compared to Planck, with the two data sets measuring significantly different angular scales and polarization levels, providing additional tests of the standard cosmological model.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

TDCOSMO - XVII. New time delays in 22 lensed quasars from optical monitoring with the ESO-VST 2.6m and MPG 2.2m telescopes

We present new time delays, the main ingredient of time delay cosmography, for 22 lensed quasars resulting from high-cadence r-band monitoring on the 2.6 m ESO VLT Survey Telescope and Max-Planck-Gesellschaft 2.2 m telescope. Each lensed quasar was typically monitored for one to four seasons, often shared between the two telescopes to mitigate the interruptions forced by the COVID-19 pandemic. The sample of targets consists of 19 quadruply and 3 doubly imaged quasars, which received a total of 1918 hours of on-sky time split into 21 581 wide-field frames, each 320 seconds long. In a given field, the 5-σ depth of the combined exposures typically reaches the 27th magnitude, while that of single visits is 24.5 mag – similar to the expected depth of the upcoming Vera-Rubin LSST. The fluxes of the different lensed images of the targets were reliably de-blended, providing not only light curves with photometric precision down to the photon noise limit, but also high-resolution models of the targets whose features and astrometry were systematically confirmed in Hubble Space Telescope imaging. This was made possible thanks to a new photometric pipeline, lightcurver, and the forward modelling method STARRED. Finally, the time delays between pairs of curves and their uncertainties were estimated, taking into account the degeneracy due to microlensing, and for the first time the full covariance matrices of the delay pairs are provided. Of note, this survey, with 13 square degrees, has applications beyond that of time delays, such as the study of the structure function of the multiple high-redshift quasars present in the footprint at a new high in terms of both depth and frequency. The reduced images will be available through the European Southern Observatory Science Portal.Key words: methods: data analysis / surveys / distance scale

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2022

An influence of manufacturing tolerances on Pin-Cell k-infinity of MOX fuel using data from the FUBILA experiment program

An influence of manufacturing tolerances (MTs) was evaluated on the pin-cell k-infinity of MOX fuel through the random sampling of CASMO5 calculations. The data from the FUBILA experiment program was used as the manufacturing parameters. For the uncertainties of element/isotope mass fractions, their covariance matrices were calculated by the generalized least square method. The total k-infinity uncertainty was 120-250 pcm (percent mille). From the breakdown of k-infinity uncertainty per materials, the MTs of fuel pellet had a dominant influence. The individual influences were also evaluated for element/isotope mass fractions, an inner/outer diameter, and a density. Those of element/isotope mass fractions were less than a few dozen pcm. Since the perturbations of fuel pellet diameter and density caused large variations to the total amount of heavy metals and effect of spatial self-shielding, they had a negative correlation with the k-infinity perturbation. On the AG3 (Al-Mg alloy) over-cladding, the inner/outer diameter perturbations had both negative and positive correlations. The negative one was due to decrease in amount of light water and the positive one was due to increase in amount of AG3. Since the effect of neutron slowing-down by light water had a dominant influence, the influence of the former negative correlation was larger than that of the latter positive correlation. It is, therefore, concluded that the consideration of the MTs that had a large influence on the effect of neutron slowing-down is important for precise quantification of the pin-cell k-infinity uncertainty of MOX fuel. (author)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

The NuMI Flux Prediction at ICARUS

The Deep Underground Neutrino Experiment (DUNE) is a next-generation long-baseline neutrino oscillation experiment seeking to probe fundamental symmetries within the structure of the Pontecorvo-Maki-Nakagawa-Sakata (PMNS) mixing matrix, and perform precision measurements its parameters including the neutrino mass ordering via the sign of $\Delta m^2_{31}$, and the charge-parity violating phase, $\delta_{CP}$. To make these measurements with high precision, DUNE will require external $\nu$-Ar scattering cross section data as a crucial input to the oscillation fit. Imaging Cosmic And Rare Underground Signals (ICARUS) is a 476 t liquid argon neutrino detector located at Fermi National Accelerator Laboratory (FNAL) where it is serving as the far detector for the Short-Baseline Neutrino (SBN) program along the Booster Neutrino Beam (BNB) axis. ICARUS additionally lies 795 m downstream and 100.1 mrad off-axis of the Neutrinos at the Main Injector (NuMI) neutrino beam. From this position, ICARUS is exposed to a large flux of NuMI (anti-)electron and (anti-)muon neutrinos, and poses a unique opportunity to provide high-statistics measurements of quasi-elastic and single pion-production cross sections for four neutrino flavors ($\nu_{\mu}$, $\nu_{e}$, $\bar{\nu}_{\mu}$, $\bar{\nu}_{e}$). This dissertation is centered around accurately characterizing the models and estimating their precision for use in making these measurements. This includes identifying major sources of uncertainty in the models such that they can be properly propagated to the cross section measurements. Specifically, this work focused on the model of the NuMI beamline and its impact on the neutrino fluxes, but also delved into the detector response model and its impact on reconstructed observables in the detector. Significant efforts were made to improve the characterization to enhance precision and thus reduce the level of propagated uncertainty. In particular, the NuMI flux was determined to be composed of 57% $\nu_{\mu}$, 38% $\bar{\nu}_{\mu}$, 3% $\nu_{e}$, and 2% $\bar{\nu}_{e}$ while the horns are operating in the positive-particle focusing configuration. The total uncertainty on the $\nu_{\mu} + \bar{\nu}_{\mu}$ ($\nu_{e} + \bar{\nu}_{e}$) flux while operating in the forward horn operating mode was determined to be 10.84% (9.04%). Compared to the on-axis flux, mesons that eventually decay to neutrinos more frequently reinteract within the NuMI structure, resulting in elevated uncertainty as these processes are not well-constrained by existing hadron interaction cross section measurements. Covariance matrices were calculated to propagate the flux uncertainty characterization to NuMI analyses.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

Physical Layer Security at a Point-to-Point MIMO System With 1-Bit DACs and ADCs

Low-resolution converters reduce the power consumption of wireless devices and help enable large energy-efficient arrays. In this letter, we evaluate the secrecy rate for the extreme case of 1-bit single-user multiple-input multiple-output systems following two signaling approaches. In the first method, the system uses discrete signaling. In the second, the system combines Gaussian signaling with artificial noise alongside power allocation. In the first case, we give a closed-form expression of the secrecy rate and propose an algorithm that provides near-optimal results for low signal-to-noise ratios (SNRs). Further, in the second scenario, depending on the channel state information, we suggest different precoding plans and derive a secrecy rate lower bound. At low SNR, power is allocated to the main signal. Only at high SNR is optimal power allocation effective. Moreover, knowing the eavesdropper channel greatly improves the system’s secrecy.

1-bit digital-to-analog (DAC) and analog-to-digita↗
DOE OSTI2024

A Bayesian Learning Approach to Wireless Outdoor Heatmap Construction using Deep Gaussian Process

We present a novel Bayesian learning approach to outdoor radio heatmap construction utilizing deep Gaussian process (GP). The proposed approach employs a two-layer hierarchy which consists of two cascaded Gaussian processes that are capable of modeling more complex input-output relations than standard single-layer Gaussian processes. Since deriving the exact model likelihood is challenging, a lower bound is optimized instead so that gradient descent-based methods can be performed to find out the optimal model parameters. Typically, inducing points are used in GPs to facilitate low-rank approximation of covariance (kernel) matrices for computation speedup. However, the inaccuracy induced by inducing points can accumulate when stacking multiple layers of GP which may hinder the performance of deep GP. Moreover, since inducing points need to be learned, having them at all layers of deep GP also incurs computational burden. To overcome the above challenges, in contrast to the canonical deep GP model, we use a modified architecture where a full standard GP resides in the first layer and inducing points are only introduced for the second layer. This modified architecture strikes a balance between model accuracy and training complexity. In the proposed model, the noise parameter of the first GP layer is also eliminated to improve the training efficiency as the noise parameter at the output of the second layer suffices to model the uncertainty in the output. The proposed approach is evaluated on real-world datasets, in the form of location-Received Signal Strength (RSS) pairs, collected from the Platform for Open Wireless Data-driven Experimental Research (POWDER) located at the campus of the University of Utah. Experiment results show that the proposed approach can achieve smaller prediction errors on various training and testing data configurations than DNN-based and GP-based methods.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2024

Experimental Covariance Determination for Critical Integral Experiments

Integral benchmarks for criticality safety and nuclear data validation require expensive uncertainty quantification studies. In general, uncertainty quantification techniques ignore correlations between experiments and shared parts. Experiments, such as the TEX (Thermal/Epithermal eXperiments) campaigns, consist of many shared parts, such as the ‘Jemima’ HEU fuel plates, which create a strong correlation in their uncertainties. While these correlations are known to exist, they are often not estimated due to the complexity of such calculations. This paper describes an intuitive method of determining the covariance for each of the experimental components, providing a correlation for each family of parts across the multiple cases examined within a benchmark. A proof-of-principle study using the TEX-HEU experimental campaign was performed and verified that the correlations can be calculated with information commonly found in the ICSBEP (International Criticality Safety Benchmark Evaluation Project) benchmarks. This study showed that the introduction of model and experimental covariances reduces the χ 2 per degree of freedom from 2.203 to 1.179, indicating that the omission causes overly pessimistic bias quantifications. This technique can be seamlessly integrated to current benchmark evaluations as well as reevaluations of legacy benchmarks.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2020

NGEE Arctic CO2, CH4 and Energy Eddy-Covariance (EC) Flux Tower Auxiliary Measurements, Utqiagvik (Barrow), Alaska, 2012 - 2019

Measurements began in September 2012 and are generally made throughout the snow-free growing season starting from April (pre-thawing) until November (freeze-up) when the instrumentation is taken down for maintenance and instrument calibration. CO2, CH4 and energy fluxes using the Eddy Covariance (EC) technique (Baldocchi 2003) and meteorological measurements are reported as 30-minute averages. The tower site is registered with AmeriFlux as US-NGB. See http://ameriflux.lbl.gov/sites/siteinfo/US-NGB for more information. Reported data include: (1) additional/extra meteorological and eddy covariance CO2, CH4 and energy flux data, (2) binned spectra for the three wind components, the sonic temperature and gas densities together with the binned cospectra for covariances of w (vertical wind component) and gas densities, (3) binned ogives (cumulative (co)spectra) for the three wind components, the sonic temperature and gas densities, and (4) tower site footprint matrices. Meteorological and eddy covariance CO2, CH4 and energy flux data (AmeriFlux BASE data product) are available from AmeriFlux https://doi.org/10.17190/AMF/1436326. The tower site is adjacent to the NGEE Arctic Tram system with several complementary time-series measurements https://doi.org/10.5440/1558782.This dataset supersedes the previously released dataset "Eddy-Covariance and auxiliary measurements, NGEE-Barrow, 2012-2013" DOI:10.5440/1124200 as this new product includes data for the longer time period processed using consistent methods for the entire record. The Next-Generation Ecosystem Experiments: Arctic (NGEE Arctic), was a research effort to reduce uncertainty in Earth System Models by developing a predictive understanding of carbon-rich Arctic ecosystems and feedbacks to climate. NGEE Arctic was supported by the Department of Energy's Office of Biological and Environmental Research. The NGEE Arctic project had two field research sites: 1) located within the Arctic polygonal tundra coastal region on the Barrow Environmental Observatory (BEO) near Utqiagvik (Barrow), Alaska and 2) multiple areas in the discontinuous permafrost region of the Seward Peninsula north of Nome, Alaska. Through observations, experiments, and synthesis with existing datasets, NGEE Arctic provided an enhanced knowledge base for multi-scale modeling and contributed to improved process representation at global pan-Arctic scales within the Department of Energy's Earth system Model (the Energy Exascale Earth System Model, or E3SM), and specifically within the E3SM Land Model component (ELM).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Sparse Cholesky factorization for solving nonlinear PDEs via Gaussian processes

In recent years, there has been widespread adoption of machine learning-based approaches to automate the solving of partial differential equations (PDEs). Among these approaches, Gaussian processes (GPs) and kernel methods have garnered considerable interest due to their flexibility, robust theoretical guarantees, and close ties to traditional methods. They can transform the solving of general nonlinear PDEs into solving quadratic optimization problems with nonlinear, PDE-induced constraints. However, the complexity bottleneck lies in computing with dense kernel matrices obtained from pointwise evaluations of the covariance kernel, and its partial derivatives, a result of the PDE constraint and for which fast algorithms are scarce. The primary goal of this paper is to provide a near-linear complexity algorithm for working with such kernel matrices. We present a sparse Cholesky factorization algorithm for these matrices based on the near-sparsity of the Cholesky factor under a novel ordering of pointwise and derivative measurements. The near-sparsity is rigorously justified by directly connecting the factor to GP regression and exponential decay of basis functions in numerical homogenization. We then employ the Vecchia approximation of GPs, which is optimal in the Kullback-Leibler divergence, to compute the approximate factor. This enables us to compute ϵ-approximate inverse Cholesky factors of the kernel matrices with complexity O(N log d (N/ϵ)) in space and O(N log 2d (N/ϵ)) in time. We integrate sparse Cholesky factorizations into optimization algorithms to obtain fast solvers of the nonlinear PDE. We numerically illustrate our algorithm’s near-linear space/time complexity for a broad class of nonlinear PDEs such as the nonlinear elliptic, Burgers, and Monge-Ampère equations. In summary, we provide a fast, scalable, and accurate method for solving general PDEs with GPs and kernel methods.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

ENDF/B-VIII.1 Validation: Covariance [Slides]

This presentation touches on ENDF/B-VIII.1 Validation: Covariance. Topics include inconsistencies in the ENDF files, derived matrices, lumped reactions and process ENDF data file with PUFF. Finally correcting data with cognac and redundant covariance that are not consistent are discussed.

Cognac↗
DOE OSTI2022

Graph link prediction in computer networks using Poisson matrix factorisation

Graph link prediction is an important task in cybersecurity: relationships between entities within a computer network, such as users interacting with computers or system libraries and the corresponding processes that use them, can provide key insights into adversary behaviour. Poisson matrix factorisation (PMF) is a popular model for link prediction in large networks, particularly useful for its scalability. In this article PMF is extended to include scenarios that are commonly encountered in cybersecurity applications. Specifically, an extension is proposed to explicitly handle binary adjacency matrices and include known categorical covariates associated with the graph nodes. A seasonal PMF model is also presented to handle seasonal networks. To allow the methods to scale to large graphs, variational methods are discussed for performing fast inference. The results show an improved performance over the standard PMF model and other statistical network models.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Non-asymptotic analysis of ensemble Kalman updates: effective dimension and localization

Many modern algorithms for inverse problems and data assimilation rely on ensemble Kalman updates to blend prior predictions with observed data. Ensemble Kalman methods often perform well with a small ensemble size, which is essential in applications where generating each particle is costly. This paper develops a non-asymptotic analysis of ensemble Kalman updates, which rigorously explains why a small ensemble size suffices if the prior covariance has moderate effective dimension due to fast spectrum decay or approximate sparsity. Here, we present our theory in a unified framework, comparing everal implementations of ensemble Kalman updates that use perturbed observations, square root filtering and localization. As part of our analysis, we develop new dimension-free covariance estimation bounds for approximately sparse matrices that may be of independent interest.

Mathematics↗
DOE OSTI2025

Analytical and EZmock covariance validation for the DESI 2024 results

The estimation of uncertainties in cosmological parameters is an important challenge in Large-Scale-Structure (LSS) analyses. For standard analyses such as Baryon Acoustic Oscillations (BAO) and Full-Shape two approaches are usually considered. First: analytical estimates of the covariance matrix use Gaussian approximations and (nonlinear) clustering measurements to estimate the matrix, which allows a relatively fast and computationally cheap way to generate matrices that adapt to an arbitrary clustering measurement. On the other hand, sample covariances are an empirical estimate of the matrix based on an ensemble of clustering measurements from fast and approximate simulations. While more computationally expensive due to the large amount of simulations and volume required, these allow us to take into account systematics that are impossible to model analytically. In this work we compare these two approaches in order to enable DESI's key analyses. We find that the configuration space analytical estimate performs satisfactorily in BAO analyses and its flexibility in terms of input clustering makes it the fiducial choice for DESI's 2024 BAO analysis. On the contrary, the analytical computation of the covariance matrix in Fourier space does not reproduce the expected measurements in terms of Full-Shape analyses, which motivates the use of a corrected mock covariance for DESI's 2024 Full Shape analysis.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

NGEE Arctic CO2, CH4 and Energy Eddy-Covariance (EC) Flux Tower Auxiliary Measurements, Council Road Mile Marker 71, Seward Peninsula, Alaska, 2017 - 2019

Measurements began in July 2017 and are made year-round (2017 seasonal only) on the tussock tundra with discontinuous permafrost at the Council Road Site on the Seward Peninsula. CO2, CH4 and energy fluxes using the Eddy Covariance (EC) technique (Baldocchi 2003) and meteorological measurements are reported as 30-minute averages. The tower site is registered with AmeriFlux as US-NGC. See http://ameriflux.lbl.gov/sites/siteinfo/US-NGC for more information. Reported data include: (1) additional/extra meteorological and eddy covariance CO2, CH4 and energy flux data, (2) binned spectra for the three wind components, the sonic temperature and gas densities together with the binned cospectra for covariances of w (vertical wind component) and gas densities, (3) binned ogives (cumulative (co)spectra) for the three wind components, the sonic temperature and gas densities, and (4) tower site footprint matrices. The majority of the files are *.csv with paired visualization footprint files as *.jpg. Meteorological and eddy covariance CO2, CH4 and energy flux data (AmeriFlux BASE data product) are available from AmeriFlux BASE US-NGC NGEE Arctic Council (Torn, 2020) https://doi.org/10.17190/AMF/1634883. The Next-Generation Ecosystem Experiments: Arctic (NGEE Arctic), was a research effort to reduce uncertainty in Earth System Models by developing a predictive understanding of carbon-rich Arctic ecosystems and feedbacks to climate. NGEE Arctic was supported by the Department of Energy's Office of Biological and Environmental Research. The NGEE Arctic project had two field research sites: 1) located within the Arctic polygonal tundra coastal region on the Barrow Environmental Observatory (BEO) near Utqiagvik (Barrow), Alaska and 2) multiple areas in the discontinuous permafrost region of the Seward Peninsula north of Nome, Alaska. Through observations, experiments, and synthesis with existing datasets, NGEE Arctic provided an enhanced knowledge base for multi-scale modeling and contributed to improved process representation at global pan-Arctic scales within the Department of Energy's Earth system Model (the Energy Exascale Earth System Model, or E3SM), and specifically within the E3SM Land Model component (ELM).

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

On the Error Covariance Correction Step of an ESKF Attitude Update

The attitude states of an error-state Kalman filter (ESKF) behave differently than most other states in the system due to their multiplicative (rather than additive) nature. One way in which they differ is an error covariance correction step after an ESKF error reset, which is not required for, for example, position and velocity states. This covariance correction step is not intuitive, and it has only been recently derived for coordinate transform matrices. The author of this memo, however, found the provided derivation in [1] confusing due to a lack of clarity surrounding the invoked reference frames, and clarity is required as there are at least 4 different ways to parameterize small-angle attitude errors in an ESKF. Furthermore, while reproducing the work, the author of this memo found a more straightforward derivation that provides additional insight into the correction step. This memo offers a derivation of the attitude error covariance correction step of an ESKF, which pays specific attention to the coordinate reference frames.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Covariance Testing at BNL [Slides]

This presentation discusses testing implemented in FUDGE & ADVANCE in regard to proper ENDF formatting: STAN, STANEF, CHECKR, FUDGE, NJOY, convert to BOXR format, plotting, endfcovreport.py which is an overview of ENDF library’s covariance contents, and covariance Matrix class implements that are the core of all GNDS covariance. Several checks are performed, including symmetric, real, positive definite, condition # of Eigenvalues, and too big uncertainties. The presentation states that it can fix too big and/or small uncertainties and that it can plot matrices. It states that further testing is recommended by CSEWG but has not yet been implemented and suggests methods such as documentation which needs a human to execute, completeness which is partially complete, and reasonableness which has not been started. The presentation concludes by elaborating on reasonableness and states that only cross section is documented, that similar limits on CP reactions cross sections are needed, that maximum limits are needed too which is tricky since ENDF assumes Normal PDFs, and that there may be other observables.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Fitting Matérn smoothness parameters using automatic differentiation

The Mat$\acute{e}$rn covariance function is ubiquitous in the application of Gaussian processes to spatial statistics and beyond. Perhaps the most important reason for this is that the smoothness parameter $\nu$ gives complete control over the mean-square differentiability of the process, which has significant implications for the behavior of estimated quantities such as interpolants and forecasts. Unfortunately, derivatives of the Mat$\acute{e}$rn covariance function with respect to $\nu$ require derivatives of the modified second-kind Bessel function $K$ $\nu$ with respect to $\nu$. While closed form expressions of these derivatives do exist, they are prohibitively difficult and expensive to compute. For this reason, many software packages require fixing $\nu$ as opposed to estimating it, and all existing software packages that attempt to offer the functionality of estimating $\nu$ use finite difference estimates for $\partial$ $\nu$ $K$ $\nu$ . In this work, we introduce a new implementation of $K$$\nu$ that has been designed to provide derivatives via automatic differentiation (AD), and whose resulting derivatives are significantly faster and more accurate than those computed using finite differences. Here, we provide comprehensive testing for both speed and accuracy and show that our AD solution can be used to build accurate Hessian matrices for second-order maximum likelihood estimation in settings where Hessians built with finite difference approximations completely fail.

97 MATHEMATICS AND COMPUTING↗