Search NASA⌕ Search

SEARCH · Search NASA

Results for “density estimation”

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 109 records · Page 6

Method for estimating the density of high-level nuclear waste glass

A database of over 1100 silicate glass compositions and densities was compiled and used to evaluate the efficacy of an algorithm for estimating the density of silicate glass compositions. We sought to develop a parsimonious algorithm based on the additivity of partial molar volumes of individual oxide components weighted by their mole fraction in a glass composition. Bound molar volumes were used for oxides in which the density of the oxide bound in a glass matrix was previously determined. The bound molar volumes were known for oxides covering 97.5 mole percent of the database compositional space. The measured glass densities were plotted against the estimated glass densities and a linear regression yielded an R 2 adj. = 0.95 and a slope and intercept of approximately one and zero, respectively. This regression suggests that glass densities estimated by the algorithm, within analysis uncertainty, are equal to the measured densities of the glasses. In addition to the development of the density estimation, we corroborated many of the referenced bound molar volume data used in the parameterization of the estimation algorithm via linear regression of the individual partial molar volumes versus the inverse measured densities (specific volumes) of the glasses in the database.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

SU-F-J-213: Feasibility Study of Using a Dual-Energy Cone Beam CT (DECBCT) in Proton Therapy Treatment Planning

Purpose: The aim of this study is to evaluate the feasibility of using a dual-energy CBCT (DECBCT) in proton therapy treatment planning to allow for accurate electron density estimation. Methods: For direct comparison, two scenarios were selected: a dual-energy fan-beam CT (high: 140 kVp, low: 80 kVp) and a DECBCT (high: 125 kVp, low: 80 kVp). A Gammex 467 tissue characterization phantom was used, including the rods of air, water, bone (B2–30% mineral), cortical bone (SB3), lung (LN-300), brain, liver and adipose. For the CBCT, Hounsfield Unit (HU) numbers were first obtained from the reconstructed images after a calibration was made based on water (=0) and air materials (=−1000). For each tissue surrogate, region-of-interest (ROI) analyses were made to derive high-energy and low-energy HU values (HUhigh and HUlow), which were subsequently used to estimate electron density based on the algorithm as previously described by Hunemohr N., et al. Parameters k1 and k2 are energy dependent and can be derived from calibration materials. Results: While for the dual-energy FBCT, the electron density is found be within +/−3% error relative to the values provided by the phantom vendor: −1.8% (water), 0.03% (lung), 1.1% (brain), −2.82% (adipose), −0.49% (liver) and −1.89% (cortical bones). While for the DECBCT, the estimation of electron density exhibits a relatively larger variation: −1.76% (water), −36.7% (lung), −1.92% (brain), −3.43% (adipose), 8.1% (liver) and 9.5% (cortical bones). Conclusion: For DECBCT, the accuracy of electron density estimation is inferior to that of a FBCT, especially for materials of either low-density (lung) or high density (cortical bone) compared to water. Such limitation arises from inaccurate HU number derivation in a CBCT. Advanced scatter-correction and HU calibration routines, as well as the deployment of photon counting CT detectors need be investigated to minimize the difference between FBCT and CBCT.

60 APPLIED LIFE SCIENCES↗

A COMPARATIVE STUDY OF DENSITY FIELD ESTIMATION FOR GALAXIES: NEW INSIGHTS INTO THE EVOLUTION OF GALAXIES WITH ENVIRONMENT IN COSMOS OUT TO z ∼ 3

It is well-known that a galaxy’s environment has a fundamental influence in shaping its properties. We study the environmental effects on galaxy evolution, with an emphasis on the environment defined as the local number density of galaxies. The density field is estimated with different estimators (weighted adaptive kernel smoothing, 10th and 5th nearest neighbors, Voronoi and Delaunay tessellation) for a K{sub s} < 24 sample of ∼190,000 galaxies in the COSMOS field at 0.1 < z < 3.1. The performance of each estimator is evaluated with extensive simulations. We show that overall there is a good agreement between the estimated density fields using different methods over ∼2 dex in overdensity values. However, our simulations show that adaptive kernel and Voronoi tessellation outperform other methods. Using the Voronoi tessellation method, we assign surface densities to a mass complete sample of quiescent and star-forming galaxies out to z ∼ 3. We show that at a fixed stellar mass, the median color of quiescent galaxies does not depend on their host environment out to z ∼ 3. We find that the number and stellar mass density of massive (>10{sup 11} M{sub ⊙}) star-forming galaxies have not significantly changed since z ∼ 3, regardless of their environment. However, for massive quiescent systems at lower redshifts (z ≲ 1.3), we find a significant evolution in the number and stellar mass densities in denser environments compared to lower density regions. Our results suggest that the relation between stellar mass and local density is more fundamental than the color–density relation and that environment plays a significant role in quenching star-formation activity in galaxies at z ≲ 1.

79 ASTRONOMY AND ASTROPHYSICS↗

Estimating the density of states of frustrated spin systems

Abstract Estimating the density of states (DOS) of systems with rugged free energy landscapes is a notoriously difficult task of the utmost importance in many areas of physics ranging from spin glasses to biopolymers. DOS estimation has also recently become an indispensable tool for the benchmarking of quantum annealers when these function as samplers. Some of the standard approaches suffer from a spurious convergence of the estimates to metastable minima, and these cases are particularly hard to detect. Here, we introduce a sampling technique based on population annealing enhanced with a multi-histogram analysis and report on its performance for spin glasses. We demonstrate its ability to overcome the pitfalls of other entropic samplers, resulting in some cases in large scaling advantages that can lead to the uncovering of new physics. The new technique avoids some inherent difficulties in established approaches and can be applied to a wide range of systems without relevant tailoring requirements. Benchmarking of the studied techniques is facilitated by the introduction of several schemes that allow us to achieve exact counts of the degeneracies of the tested instances.

Barash, Lev (ORCID:000000022298785X)↗

Detecting Anomalies in Time Series Using Kernel Density Approaches

This paper introduces a novel anomaly detection approach tailored for time series data with exclusive reliance on normal events during training. Our key innovation lies in the application of kernel-density estimation (KDE) to scrutinize reconstruction errors, providing an empirically derived probability distribution for normal events post-reconstruction. This non-parametric density estimation technique offers a nuanced understanding of anomaly detection, differentiating it from prevalent threshold-based mechanisms in existing methodologies. In post-training, events are encoded, decoded, and evaluated against the estimated density, providing a comprehensive notion of normality. In addition, we propose a data augmentation strategy involving variational autoencoder-generated events and a smoothing step for enhanced model robustness. The significance of our autoencoder-based approach is evident in its capacity to learn normal representation without prior anomaly knowledge. Through the KDE step on reconstruction errors, our method addresses the versatility of anomalies, departing from assumptions tied to larger reconstruction errors for anomalous events. Our proposed likelihood measure then distinguishes normal from anomalous events, providing a concise yet comprehensive anomaly detection solution. The extensive experimental results support the feasibility of our proposed method, yielding significantly improved classification performance by nearly 10% on the UCR benchmark data.

Frehner, Robin↗

Directional Statistics of Preferential Orientations of Two Shapes in Their Aggregate and Its Application to Nanoparticle Aggregation

Nanoscientists have long conjectured that adjacent nanoparticles aggregate with one another in certain preferential directions during chemical syntheses of nanoparticles, which is so called the oriented attachment. For the study of the oriented attachment phenomenon, the microscopy and nanoscience community have used dynamic electron microscopy for direct observations of nanoparticle aggregation and have been so far relying on the manual and qualitative analysis of the observations. In this paper, we present a statistical approach for studying the oriented attachment phenomenon with multiple aggregation examples observed by dynamic microscope data. We abstract a nanoparticle aggregation by an event of two primary geometric objects merging into a secondary geometric object. We define a metric space for geometrical bodies, and the endowed metric in the space allows us to define the likelihood function of the orientation angles of two primary objects within the corresponding secondary object. We maximize the likelihood to estimate the orientation angles. We fit a mixture of bivariate von Mises distributions to the estimated orientation angles for the density estimation. By analyzing the density estimate, we are able to demonstrate that nanoparticles of certain geometries have indeed preferential orientations in their aggregates.

Mixture of bivariate von Mises distributions, Bi-a↗

Foundations of Feature Selection and Classification for Non-Gaussian Distributed Targets

This report describes the author’s research toward creating algorithms for selecting useful statistical features in pattern/target classification problems in which the features are non-Gaussian distributed. In engineering practice, it is common to either (1) Not perform any feature selection procedure, or (2) Use a feature selection algorithm that assumes the features are Gaussian distributed. This process can be far from optimal if the data are actually non-Gaussian, as they frequently are in practice. This research has the goal of mitigating that problem by creating a useful algorithm that can be used in practice. The approach is to focus on the performance indices used in common feature selection algorithms, such as Branch and Bound algorithms and the Sequential Forward Selection Algorithm. Ordinarily, the performance index used to measure the class separation in feature space involves assuming the data are Gaussian and deriving tractable performance indices that use closed form equations that represent the probability density functions of the class data. The advantage of this approach is that it produces feature selection algorithms that have low computational complexity and do not require knowledge of the actual data densities. The disadvantage is that these algorithms may not perform reasonably when the data are non-Gaussian, as they commonly are in practice. This research examines the use of information-theoretic class separability measures that can deal with the non-Gaussian case. In particular, this work shows that the Hellinger Distance (a divergence measure), has very desirable mathematical properties and can be useful for feature selection. The work in this paper employs a multivariate kernel density estimator along with the Hellinger Distance and an optimal subset selection algorithm to do feature selection. This report (1) Defines the classification and feature selection problems, (2) Describes the continuous variable Hellinger distance in one dimension and discusses its mathematical properties in comparison with other distance measures, (3) Develops the discrete variable Hellinger distance expression for one and two dimensions, and shows that the discrete Hellinger distance for multiple dimensions is a straightforward extension, and (4) Presents experimental density estimation results using a new MATLAB implementation of a multivariate kernel density estimator. The draft of this report served as a starting point for Ensign Matthew Wilder’s MSEE thesis work at the Naval Postgraduate School (NPS) in 2010 and 2011. Ensign Wilder completed his thesis and graduated from NPS in June 2011 (M. J. Wilder, “Automatic Target Recognition: Statistical Feature Selection of Non-Gaussian Distributed Target Classes,” MSEE Thesis, Naval Postgraduate School, Thesis Advisor: Grace A. Clark, Second Reader: Monique P. Fargues, June 2011). In the thesis, Ensign Wilder created implementations of the Branch and Bound and Sequential Feature Selection algorithms, extended the multivariate kernel density estimator code, and evaluated the performance of Hellinger Distance-based feature selection algorithms using both simulated data and data from a real-world benchmark data set. The new algorithm was shown to be very effective for both Gaussian and non-Gaussian data sets. The advantages gained for non-Gaussian data come at the cost of increased computational complexity. Using a laptop computer for practical problems, we generally use initial feature vectors containing no more than about ten or twelve features. Proposed future research lies in creating faster kernel density estimator algorithms that would allow for the use of higher dimensional data.

42 ENGINEERING↗

An analysis of lead-free (Bi{sub 0.5}Na{sub 0.5}){sub 0.915}-(Bi{sub 0.5}K{sub 0.5}){sub 0.05}Ba{sub 0.02}Sr{sub 0.015}TiO{sub 3} ceramic for efficient refrigeration and thermal energy harvesting

This article demonstrates the colossal energy harvesting capability of a lead-free (Bi{sub 0.5}Na{sub 0.5}){sub 0.915}-(Bi{sub 0.5}K{sub 0.5}){sub 0.05}Ba{sub 0.02}Sr{sub 0.015}TiO{sub 3} ceramic using the Olsen cycle. The maximum harvestable energy density estimated for this system is found to be 1523 J/L (1523 kJ/m{sup 3}) where the results are presented for extreme ambient conditions of 20–160 °C and electric fields of 0.1–4 MV/m. This estimated energy density is 1.7 times higher than the maximum reported to date for the lanthanum-doped lead zirconate titanate (thin film) system. Moreover, this study introduces a generalized and effective solid state refrigeration cycle in contrast to the ferroelectric Ericson refrigeration cycle. The cycle is based on a temperature induced polarization change on application of an unipolar electric field to ferroelectric ceramics.

36 MATERIALS SCIENCE↗

ESTIMATES OF DENSITIES AND FILLING FACTORS FROM A COOLING TIME ANALYSIS OF SOLAR MICROFLARES OBSERVED WITH RHESSI

We use more than 4500 microflares from the RHESSI microflare data set to estimate electron densities and volumetric filling factors of microflare loops using a cooling time analysis. We show that if the filling factor is assumed to be unity, the calculated conductive cooling times are much shorter than the observed flare decay times, which in turn are much shorter than the calculated radiative cooling times. This is likely unphysical, but the contradiction can be resolved by assuming that the radiative and conductive cooling times are comparable, which is valid when the flare loop temperature is a maximum and when external heating can be ignored. We find that resultant radiative and conductive cooling times are comparable to observed decay times, which has been used as an assumption in some previous studies. The inferred electron densities have a mean value of 10{sup 11.6} cm{sup -3} and filling factors have a mean of 10{sup -3.7}. The filling factors are lower and densities are higher than previous estimates for large flares, but are similar to those found for two microflares by Moore et al.

79 ASTRONOMY AND ASTROPHYSICS↗

High-dimensional and permutation invariant anomaly detection

Methods for anomaly detection of new physics processes are often limited to low-dimensional spaces due to the difficulty of learning high-dimensional probability densities. Particularly at the constituent level, incorporating desirable properties such as permutation invariance and variable-length inputs becomes difficult within popular density estimation methods. In this work, we introduce a permutation-invariant density estimator for particle physics data based on diffusion models, specifically designed to handle variable-length inputs. We demonstrate the efficacy of our methodology by utilizing the learned density as a permutation-invariant anomaly detection score, effectively identifying jets with low likelihood under the background-only hypothesis. To validate our density estimation method, we investigate the ratio of learned densities and compare to those obtained by a supervised classification algorithm.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Estimates of Ta strength at ultrahigh pressures and strain rates using thin-film graded-density impactors

In this paper, we present results from an experimental technique used to estimate the strength of Ta at extreme pressures (150 GPa) and strain rates (10 7 s -1 ). A graded-density impactor (GDI) was fabricated using sputter deposition to produce an approximately 40-μm-thick film containing alternating layers of Al and Cu. The thicknesses of the respective layers are adjusted to give an effective density gradient through the film. The GDIs were launched with a 2-stage light gas gun, and shock-ramp-release velocity profiles were measured over timescales of ~10 ns. Results are presented for the direct impact of the film onto LiF windows, which allows for a dynamic characterization of the GDI, as well as from impact onto thin (~40μm) sputtered Ta samples backed by a LiF window. The measurements were coupled with mesoscale numerical simulations to infer the strength of Ta, and the results agree well with other high-pressure platforms, particularly when strain-rate, microstructural, and thermodynamic-path differences are considered.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Robust Estimation of Electron Density From Anatomic Magnetic Resonance Imaging of the Brain Using a Unifying Multi-Atlas Approach

Purpose: To develop a reliable method to estimate electron density based on anatomic magnetic resonance imaging (MRI) of the brain. Methods and Materials: We proposed a unifying multi-atlas approach for electron density estimation based on standard T1- and T2-weighted MRI. First, a composite atlas was constructed through a voxelwise matching process using multiple atlases, with the goal of mitigating effects of inherent anatomic variations between patients. Next we computed for each voxel 2 kinds of conditional probabilities: (1) electron density given its image intensity on T1- and T2-weighted MR images; and (2) electron density given its spatial location in a reference anatomy, obtained by deformable image registration. These were combined into a unifying posterior probability density function using the Bayesian formalism, which provided the optimal estimates for electron density. We evaluated the method on 10 patients using leave-one-patient-out cross-validation. Receiver operating characteristic analyses for detecting different tissue types were performed. Results: The proposed method significantly reduced the errors in electron density estimation, with a mean absolute Hounsfield unit error of 119, compared with 140 and 144 (P<.0001) using conventional T1-weighted intensity and geometry-based approaches, respectively. For detection of bony anatomy, the proposed method achieved an 89% area under the curve, 86% sensitivity, 88% specificity, and 90% accuracy, which improved upon intensity and geometry-based approaches (area under the curve: 79% and 80%, respectively). Conclusion: The proposed multi-atlas approach provides robust electron density estimation and bone detection based on anatomic MRI. If validated on a larger population, our work could enable the use of MRI as a primary modality for radiation treatment planning.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Dwarf galaxies with ionizing radiation feedback. II. Spatially resolved star formation relation

We investigate the spatially resolved star formation relation using a galactic disk formed in a comprehensive high-resolution (3.8 pc) simulation. Our new implementation of stellar feedback includes ionizing radiation as well as supernova explosions, and we handle ionizing radiation by solving the radiative transfer equation rather than by a subgrid model. Photoheating by stellar radiation stabilizes gas against Jeans fragmentation, reducing the star formation rate (SFR). Because we have self-consistently calculated the location of ionized gas, we are able to make simulated, spatially resolved observations of star formation tracers, such as Hα emission. We can also observe how stellar feedback manifests itself in the correlation between ionized and molecular gas. Applying our techniques to the disk in a galactic halo of 2.3 × 10{sup 11} M {sub ☉}, we find that the correlation between SFR density (estimated from mock Hα emission) and H{sub 2} density shows large scatter, especially at high resolutions of ≲75 pc that are comparable to the size of giant molecular clouds (GMCs). This is because an aperture of GMC size captures only particular stages of GMC evolution and because Hα traces hot gas around star-forming regions and is displaced from the H{sub 2} peaks themselves. By examining the evolving environment around star clusters, we speculate that the breakdown of the traditional star formation laws of the Kennicutt-Schmidt type at small scales is further aided by a combination of stars drifting from their birthplaces and molecular clouds being dispersed via stellar feedback.

79 ASTRONOMY AND ASTROPHYSICS↗

Gradient-based stochastic estimation of the density matrix

Fast estimation of the single-particle density matrix is key to many applications in quantum chemistry and condensed matter physics. The best numerical methods leverage the fact that the density matrix elements f(H) ij decay rapidly with distance rij between orbitals. This decay is usually exponential. However, for the special case of metals at zero temperature, algebraic decay of the density matrix appears and poses a significant numerical challenge. Here, we introduce a gradient-based probing method to estimate all local density matrix elements at a computational cost that scales linearly with system size. For zero-temperature metals, the stochastic error scales like S -(d+2)/2d , where d is the dimension and S is a prefactor to the computational cost. The convergence becomes exponential if the system is at finite temperature or is insulating.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Application of the Redlich-Kister expansion for estimating the density of molten fluoride psuedo-ternary salt systems of nuclear industry interest

The development of next-generation molten salt reactors relies on accurate knowledge of the thermophysical properties of the candidate coolant and fueled molten salts. These thermophysical properties include density, viscosity, thermal conductivity, and heat capacity. Because of difficulties in measuring thermophysical properties of molten salts, there are many gaps in the current state of thermophysical property knowledge of these salts, particularly those that contain actinides or beryllium. Therefore, leveraging modeling techniques to estimate unknown molten salt thermophysical properties and guide future experimental measurements has high value for the nuclear industry. Here, the densities of molten fluoride pseudo-ternary salt systems, which are of interest to the nuclear industry, were estimated using Redlich-Kister expansion and Muggianu interpolation techniques. The pseudo-ternary systems considered for estimation in this study were NaF-LiF-ZrF 4 , LiF-BeF 2 -ZrF 4 , LiF-BeF 2 -ThF 4 , NaF-LiF-BeF 2 , NaF-KF-BeF 2 , NaF-ZrF 4 -UF 4 , and NaF-BeF 2 -UF 4 . This Redlich-Kister estimation approach accounts for nonideal mixing behavior based on pseudo-binary subsystem interaction parameters determined from experimentally measured pseudo-binary system density data sets. The Redlich-Kister estimation was compared with the method of additive molar volumes, which assumes ideal mixing. Additionally, the Redlich-Kister approach was used to determine previously unknown binary and ternary interaction parameters based on experimentally measured density data sets for select pseudo-ternary salt systems. The results of this study show improvement in density estimation using the Redlich-Kister approach for all systems considered compared with estimation by additive molar volumes. Furthermore, this analysis allowed for the estimation of nonideal density behavior in experimentally unstudied ZrF 4 -UF 4 and BeF 2 -UF 4 , as well as the quantification of ternary interaction in NaF-LiF-ZrF 4 , NaF-BeF 2 -UF 4 , and NaF-ZrF 4 -UF 4 .

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Functional Data Analysis in NTCP Modeling: A New Method to Explore the Radiation Dose-Volume Effects

Purpose/Objective(s): To describe a novel method to explore radiation dose-volume effects. Functional data analysis is used to investigate the information contained in differential dose-volume histograms. The method is applied to the normal tissue complication probability modeling of rectal bleeding (RB) for patients irradiated in the prostatic bed by 3-dimensional conformal radiation therapy. Methods and Materials: Kernel density estimation was used to estimate the individual probability density functions from each of the 141 rectum differential dose-volume histograms. Functional principal component analysis was performed on the estimated probability density functions to explore the variation modes in the dose distribution. The functional principal components were then tested for association with RB using logistic regression adapted to functional covariates (FLR). For comparison, 3 other normal tissue complication probability models were considered: the Lyman-Kutcher-Burman model, logistic model based on standard dosimetric parameters (LM), and logistic model based on multivariate principal component analysis (PCA). Results: The incidence rate of grade ≥2 RB was 14%. V{sub 65Gy} was the most predictive factor for the LM (P=.058). The best fit for the Lyman-Kutcher-Burman model was obtained with n=0.12, m = 0.17, and TD50 = 72.6 Gy. In PCA and FLR, the components that describe the interdependence between the relative volumes exposed at intermediate and high doses were the most correlated to the complication. The FLR parameter function leads to a better understanding of the volume effect by including the treatment specificity in the delivered mechanistic information. For RB grade ≥2, patients with advanced age are significantly at risk (odds ratio, 1.123; 95% confidence interval, 1.03-1.22), and the fits of the LM, PCA, and functional principal component analysis models are significantly improved by including this clinical factor. Conclusion: Functional data analysis provides an attractive method for flexibly estimating the dose-volume effect for normal tissues in external radiation therapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

PICS: SIMULATIONS OF STRONG GRAVITATIONAL LENSING IN GALAXY CLUSTERS

Gravitational lensing has become one of the most powerful tools available for investigating the “dark side” of the universe. Cosmological strong gravitational lensing, in particular, probes the properties of the dense cores of dark matter halos over decades in mass and offers the opportunity to study the distant universe at flux levels and spatial resolutions otherwise unavailable. Studies of strongly lensed variable sources offer even further scientific opportunities. One of the challenges in realizing the potential of strong lensing is to understand the statistical context of both the individual systems that receive extensive follow-up study, as well as that of the larger samples of strong lenses that are now emerging from survey efforts. Motivated by these challenges, we have developed an image simulation pipeline, Pipeline for Images of Cosmological Strong lensing (PICS), to generate realistic strong gravitational lensing signals from group- and cluster-scale lenses. PICS uses a low-noise and unbiased density estimator based on (resampled) Delaunay Tessellations to calculate the density field; lensed images are produced by ray-tracing images of actual galaxies from deep Hubble Space Telescope observations. Other galaxies, similarly sampled, are added to fill in the light cone. The pipeline further adds cluster member galaxies and foreground stars into the lensed images. The entire image ensemble is then observed using a realistic point-spread function that includes appropriate detector artifacts for bright stars. Noise is further added, including such non-Gaussian elements as noise window-paning from mosaiced observations, residual bad pixels, and cosmic rays. The aim is to produce simulated images that appear identical—to the eye (expert or otherwise)—to real observations in various imaging surveys.

79 ASTRONOMY AND ASTROPHYSICS↗

Mixture density network estimation of continuous variable maximum likelihood using discrete training samples

Abstract Mixture density networks (MDNs) can be used to generate posterior density functions of model parameters $$\varvec{\theta }$$ θ given a set of observables $${\mathbf {x}}$$ x . In some applications, training data are available only for discrete values of a continuous parameter $$\varvec{\theta }$$ θ . In such situations, a number of performance-limiting issues arise which can result in biased estimates. We demonstrate the usage of MDNs for parameter estimation, discuss the origins of the biases, and propose a corrective method for each issue.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗