Search NASA⌕ Search

SEARCH · Search NASA

Results for “additive Gaussian models”

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 181 records · Page 10
All sourcesNASA NTRSDOE OSTI
NASA NTRS2013

Herschel Observations of the Debris Disc Around HIP 92043

Context: Typical debris discs are composed of particles ranging from several micron sized dust grains to kilometer-sized asteroidal bodies, and their infrared emission peaks at wavelengths 60-100 microns. Recent Herschel DUNES (Dust Around Nearby Stars) observations have identified several debris discs around nearby Sun-like stars (F, G and K spectral type) with significant excess emission only at 160 microns. Aims: We observed HIP 92043 (110 Her, HD 173667) at far-infrared and sub-millimetre wavelengths with Herschel PACS (Photodetector Array Camera and Spectrometer) and SPIRE (Spectral and Photometric Imaging Receiver).Identification of the presence of excess emission from HIP 92043 and the origin and physical properties of any excess was undertaken through analysis of its spectral energy distribution (SED) and the PACS images. Methods: The PACS and SPIRE images were produced using the HIPE (Herschel Interactive Processing Environment) photProject map maker routine. Fluxes were measured using aperture photometry. A stellar photosphere model was scaled to optical and near infrared photometry and subtracted from the far-infared and sub-mm fluxes to determine the presence of excess emission. Source radial profiles were fitted using a 2D Gaussian and compared to a PSF (Point Spread Function) model based on Herschel observations of alpha Boo to check for extended emission. Results: Clear excess emission from HIP 92043 was observed at 70 and 100 microns. Marginal excess was observed at 160 and 250 microns.Analysis of the images reveals that the source is extended at 160 microns. A fit to the source SED is inconsistent with a photosphere and single temperature black body. Conclusions: The excess emission from HIP 92043 is consistent with the presence of an unresolved circumstellar debris disc at 70and 100 microns, with low probability of background contamination. The extended 160 micron emission may be interpreted as an additional cold component to the debris disc or as the result of background contamination along the line of sight. The nature of the 160 micron excess cannot be determined absolutely from the available data, but we favour a debris disc interpretation, drawing parallels with previously identified cold disc sources in the DUNES sample.

debris discs↗
NASA NTRSDate not supplied

Non-Gaussian Ensemble Filtering and Adaptive Inflation for Soil Moisture Data Assimilation

The rank histogram filter (RHF) and the ensemble Kalman filter (EnKF) are assessed for soil moisture estimation using perfect model (identical twin) synthetic data assimilation experiments. The primary motivation is to gauge the impact on analysis quality attributable to the consideration of non-Gaussian forecast error distributions. Using the NASA Catchment land surface model, the two filters are compared at 18 globally distributed single-catchment locations for a 10-yr experiment period. It is shown that both filters yield adequate estimates of soil moisture, with the RHF having a small but significant performance advantage. Most notably, the RHF consistently increases the normalized information contribution (NIC) score of the mean absolute bias by 0.05 over that of the EnKF for surface, root-zone, and profile soil moisture. The RHF also increases the NIC score for the anomaly correlation of surface soil moisture by 0.02 over that of the EnKF (at a 5% significance level). Results additionally demonstrate that the performance of both filters is somewhat improved when the ensemble priors are adaptively inflated to offset the negative effects of systematic errors.

Rolf Reichle↗
NASA NTRS2005

Extracting Quantitative Data from Lunar Soil Spectra

Using the modified Gaussian model (MGM) developed by Sunshine et al. [1] we compared the spectral properties of the Lunar Soil Characterization Consortium (LSCC) suite of lunar soils [2,3] with their petrologic and chemical compositions to obtain quantitative data. Our initial work on Apollo 17 soils [4] suggested that useful compositional data could be elicited from high quality soil spectra. We are now able to expand upon those results with the full suite of LSCC soils that allows us to explore a much wider range of compositions and maturity states. The model is shown to be sensitive to pyroxene abundance and can evaluate the relative portion of high-Ca and low-Ca pyroxenes in the soils. In addition, the dataset has provided unexpected insights into the nature and causes of absorption bands in lunar soils. For example, it was found that two distinct absorption bands are required in the 1.2 m region of the spectrum. Neither of these bands can be attributed to plagioclase or agglutinates, but both appear to be largely due to pyroxene.

Noble, S. K.↗
DOE OSTI2024

Smooth Dispersion Is Physically Appropriate: Assessing and Amending the D4 Dispersion Model

The addition of dispersion corrections to density functionals is essential for accurate energy and geometry predictions. Among them, the D4 scheme is popular due to its low computational cost and high accuracy. However, due to its design, the D4 correction can occasionally lead to anomalies, such as unphysical curvature and bumps in the potential energy surface. We find these anomalies are common in the D4 model, although observable consequences are rarer than in the D3 model for reasons we explain. Nevertheless, we uncover instances of unphysical local minima and stationary points with the D4 scheme and propose two solutions that yield smoother dispersion energy as a function of nuclear position. One is trivial to implement, based on a smoother reparametrization of Gaussian weighting (D4S) to find the effective coordination number. The other replaces Gaussian weighting with soft linear interpolation (D4SL). These new approaches usually remove artificial extremum points, while maintaining accuracy.

Tkachenko, Nikolay V↗
DOE OSTI2025

S AP F LOWER : an automated tool for sap flow data preprocessing, gap-filling, and analysis using deep learning

Sap flow, a critical process in plant water use and ecosystem water cycles, is often measured using thermal dissipation probes (TDP) due to their ease of installation and continuous data collection. However, sap flow data frequently include noise, outliers, and gaps, creating challenges for analysis and requiring substantial manual processing. We developed S AP F LOWER , a tool that automates data preprocessing, model training, gap-filling, sapwood area scaling and modeling, and water use analysis. It integrates autocleaning, machine learning and deep learning models (e.g. random forest, Gaussian process regression, long short-term memory (LSTM), bidirectional LSTM (BiLSTM)), and efficient workflows to process sap flow data. S AP F LOWER can remove over 90% of noisy data while preserving legitimate variations and achieve high accuracy in gap-filling based on user-determined parameters. Random forest, LSTM, and BiLSTM models reduced root mean square error to 10% or less for long-term gaps. Model training and prediction can be performed efficiently within seconds. S AP F LOWER significantly enhances the efficiency and accessibility of TDP data analysis by automating complex tasks, enabling researchers without programming expertise to employ advanced techniques. Future improvements will focus on species-specific corrections for TDP and support for additional measurement methods. S AP F LOWER is openly available on GitHub (https://github.com/JiaxinWang123/SapFlower) and Zenodo (doi: 10.5281/zenodo.13665919).

ecosystem water balance↗
NASA NTRS1990

Incorporation of shuttle CCT parameters in computer simulation models

Computer simulations of shuttle missions have become increasingly important during recent years. The complexity of mission planning for satellite launch and repair operations which usually involve EVA has led to the need for accurate visibility and access studies. The PLAID modeling package used in the Man-Systems Division at Johnson currently has the necessary capabilities for such studies. In addition, the modeling package is used for spatial location and orientation of shuttle components for film overlay studies such as the current investigation of the hydrogen leaks found in the shuttle flight. However, there are a number of differences between the simulation studies and actual mission viewing. These include image blur caused by the finite resolution of the CCT monitors in the shuttle and signal noise from the video tubes of the cameras. During the course of this investigation the shuttle CCT camera and monitor parameters are incorporated into the existing PLAID framework. These parameters are specific for certain camera/lens combinations and the SNR characteristics of these combinations are included in the noise models. The monitor resolution is incorporated using a Gaussian spread function such as that found in the screen phosphors in the shuttle monitors. Another difference between the traditional PLAID generated images and actual mission viewing lies in the lack of shadows and reflections of light from surfaces. Ray tracing of the scene explicitly includes the lighting and material characteristics of surfaces. The results of some preliminary studies using ray tracing techniques for the image generation process combined with the camera and monitor effects are also reported.

Huntsberger, Terry↗
DOE OSTI2026

Predicting Li-Ion Battery Capacity Fade Using Early-Life Data and a Hybrid Data-Driven Gaussian Process-Bayesian Regression Approach

Accurately predicting Li-ion battery capacity trajectories using early-life data can dramatically improve battery-life understandings and be used to rapidly evaluate design/cost/performance trade-offs when developing new battery materials. Accurate early-life predictions enable researchers to quickly iterate over cell designs and material precursor properties without consistently cycling cells to failure. To this end, we present a toolbox that uses a combined Gaussian Process and Bayesian regression approach that capitalizes on signals other than just capacity (e.g., dQ/dV, voltage drops) to rapidly predict capacity-fade trajectories. The prediction tool uses Bayesian regression to fit functional forms, e.g., power law, sigmoids, etc., to predict capacity-fade dynamics. By fitting functional forms, the capacity fade can be interrogated at any point in the future, allowing for early cell-failure prediction. Additionally, Bayesian regression allows for accurate uncertainty estimates that account for cell-to-cell variability (aleatoric uncertainty) and the lack of observation data (epistemic uncertainty). By only using early cycle data to predict the capacity fade trajectory, uncertainty bounds at end-of-life can be extremely large. The large uncertainty bounds are further exacerbated because there is no systematic way to define the prior distribution of the functional forms' parameters. We improve our the predicted trajectory confidence interval of our predicted trajectory using two methods. First, we shows that a small amount of held-out cycling data is sufficientuse some train cells, that have been cycled to failure to derive information regarding the appropriate prior distributions for the functional forms' parameters of the functional form, effectively leading to data-driven priors.. We propose constructing the data-driven priors by first running a Bayesian regression starting with uninformed priors to generate intermediate cell-specific posterior parameter distributions. These posterior distributions are combined using a Ggaussian mixture model for each parameter to create the data-driven priors. These mixture models serve as the data-driven prior distributions for the parameters for. Second, we derive multiple features, e.g., C_dchg 0.5 DoD 0.5, log (|mean(dQ/dV_(w_3-w_0 ) (V)|), etc., from the train cellsheld-out cycling data, identify which the features are that best predicting capacity at early/mid-life cycles, and then create Ggaussian process regression models that are used for predicting capacity at early/mid-life cycles for the test cells (see blue dots with error bars in Fig 1b). Finally, these predicted data-points are used in addition to the actual early cycle data capacity fade to construct the Bayesian regression trajectory for the test cell s. Notably. We note that these two methods are complementary and can be combined with each other. We evaluate the performance of our proposed method on an testing open-source dataset from Iowa State University and Iowa Lakes Community College (ISU-ILCC). This dataset comprises of 251 nickel-manganese-cobalt/graphite Lithium-ion cells that are cycled under 63 different conditions. We compute the mean average percentage error (MAPE) and negative log predictive density (NLPD) to quantify the efficacy of our method. Our initial findings suggest that, when only few observations are available, for test cells, when using only Bayesian regression with uninformed priors, a power law functional provides the most accurate predictions. with very few data points. However, asHowever, a the number of data points increases, a twin sigmoidal function becomes more accurate as the number of observations further increases. We also find that using as little as 10% of the data set towards generating data-driven priors can lead to significant improvement in prediction accuracy when using early cycle data. Lastly, we found that augmenting early-cycle data with Gaussian process-predicted capacity data for Bayesian regression greatly improves the prediction accuracy. We will present a comprehensive comparison of our methods to other methods available in the literature and apply this method to additional battery datasets.

42 ENGINEERING↗
NASA NTRS1993

The roughness of the Martian surface: A scale dependent model

In the coming decade, several lander missions to Mars are planned (e.g., MESUR Pathfinder, MESUR). One of the dangers facing planners of these missions is the rough topography observed at both Viking Lander sites. Both landing sites are ubiquitously covered with meter-scale boulders. Objects of this size pose obvious threats to soft landers, especially at Mars where the distance from Earth causes prohibitive time lags between the transmission of commands and feedback from the spacecraft. An obvious solution is to scout for a 'smooth' site prior to the landing. However, the best resolutions realizable on current and future missions (i.e., Mars Observer) are on the order of several meters. Even at this scale, boulders of 1-2 meters in size are unresolvable. Additionally, the amount of time and spacecraft resources required to search even a small area of the planet are unrealistic given other mission objectives. An alternative is to determine the 'roughness' of the surface at a subpixel scale using bidirectional reflectance observations. Much larger areas of the planet can be searched, and much of the search can easily be automated. The morphology of the martian plains observed by the Viking Landers is physically simple. The surface is covered with a layer (approximately flat lying) of aeolian sediment from which numerous outcrops of bedrock and boulders protrude. This morphology, while simple, will be difficult to characterize from orbit using traditional bidirectional reflectance models for two reasons. First, modeling the surface as facets with Gaussian or exponential slope distributions is not realistic given the morphology described above. Second, the roughness parameter is an 'average' of the roughness at scales ranging from the wavelength of light being scattered to the pixel size of the observation. Thus, there is no definite scale of roughness that can be extracted from the Hapke roughness parameter. Using the concepts of geometric and boolean models developed by several workers, we have developed a model for the bidirectional reflectance of a surface morphology comparable to that observed at the Viking Lander sites.

Shepard, M. K.↗
NASA NTRSDate not supplied

Characterizing Spatiotemporal Uncertainty in Interpolated Meteorological Data

Interpolated meteorological data invariably contain errors. These errors have structure in time and space, particularly autocorrelation, which can cause the effects of errors to compound when model outputs are aggregated temporally or spatially. One way to account for this uncertainty is with a probabilistic model from which samples can be drawn that are coherent with respect to underlying spatial and temporal covariance structure. This work describes a probabilistic method for spatial interpolation of point-wise meteorological time series. Observational data from weather stations are generally sparse in space and dense in time (but sometimes missing). The method works by projecting time series onto orthogonal basis vectors and spatially interpolating each resulting component independently. Under suitable assumptions, and data transformations to better satisfy those assumptions, Gaussian process regression provides a complete description of the joint predictive distribution over a Gaussian random field. Spatiotemporally coherent realizations are generated as the sum of conditional (spatial) simulations of each orthogonal (temporal) component. Data-derived and generic orthogonal bases are considered. In addition to spatial interpolation, imputation of missing observational data is examined. The method is applied using near-surface air temperature over the Western United States and validated by comparing theoretical versus actual coverage of predictive distributions and analyzing the degree to which spatial and temporal covariance structure is reproduced. Computational considerations, relating to conditional simulation of random fields, are also addressed.

Conor T Doherty↗
NASA NTRS1989

Analysis of Nd3+:glass, solar-pumped, high-powr laser systems

The operating characteristics of Nd(3+):glass lasers energized by a solar concentrator were analyzed for the hosts YAG, silicate glass, and phosphate glass. The modeling is based on the slab zigzag laser geometry and assumes that chemical hardening methods for glass are successful in increasing glass hardness by a factor of 4. On this basis, it was found that a realistic 1-MW solar-pumped laser might be constructed from phosphate glass 4 sq m in area and 2 mm thick. If YAG were the host medium, a 1-MW solar-pumped laser need only be 0.5 sq m in area and 0.5 cm thick, which is already possible. In addition, Nd(3+) doped glass fibers were found to be excellent solar-pumped laser candidates. The small diameter of fibers eliminates thermal stress problems, and if their diameter is kept small (10 microns), they propagate a Gaussian single mode which can be expanded and transmitted long distances in space. Fiber lasers could then be used for communications in space or could be bundled and the individual beams summed or phase-matched for high-power operation.

Zapata, L. E.↗
NASA NTRS2005

Research Supporting Satellite Communications Technology

This report describes the second year of research effort under the grant Research Supporting Satellite Communications Technology. The research program consists of two major projects: Fault Tolerant Link Establishment and the design of an Auto-Configurable Receiver. The Fault Tolerant Link Establishment protocol is being developed to assist the designers of satellite clusters to manage the inter-satellite communications. During this second year, the basic protocol design was validated with an extensive testing program. After this testing was completed, a channel error model was added to the protocol to permit the effects of channel errors to be measured. This error generation was used to test the effects of channel errors on Heartbeat and Token message passing. The C-language source code for the protocol modules was delivered to Goddard Space Flight Center for integration with the GSFC testbed. The need for a receiver autoconfiguration capability arises when a satellite-to-ground transmission is interrupted due to an unexpected event, the satellite transponder may reset to an unknown state and begin transmitting in a new mode. During Year 2, we completed testing of these algorithms when noise-induced bit errors were introduced. We also developed and tested an algorithm for estimating the data rate, assuming an NRZ-formatted signal corrupted with additive white Gaussian noise, and we took initial steps in integrating both algorithms into the SDR test bed at GSFC.

Horan Stephen↗
NASA NTRS2012

GX 3+1: The Stability of Spectral Index as a Function of Mass Accretion Rate

We present an analysis of the spectral and timing properties observed in X-rays from neutron star (NS) binary GX 3+1 (4U 1744-26) during long-term transitions between the faint and bright phases superimposed on short-term transitions between lower banana (LB) and upper banana (UB) branches in terms of its color-color diagram, We analyze all observations of this source obtained with the Rossi X-ray Timing Explorer and BeppoSAX satellites, We find that the X-ray broadband energy spectra during these spectral transitions can be adequately reproduced by a composition of a low-temperature blackbody component, a Comptonized component (COMPTB), and Gaussian component We argue that the electron temperature kTe of the Compton cloud monotonically increases from 2.3 keY to 4.5 keY, when GX 3+1 makes a transition from UB to LB. We also detect an evolution of noise components (a very low frequency noise and a high-frequency noise) during these LB-UB transitions. Using a disk seed photon normalization of COMPTB, which is proportional to the mass accretion rate, we find that the photon power-law index Gamma is almost constant (Gamma = 2.00 +/- 0.02) when mass accretion rate changes by factor four. In addition, we find that the emergent spectrum is dominated by the strong Comptonized component We interpret this quasi-stability of the index Gamma and a particular form of the spectrum in the framework of a model in which the energy release in the transition layer located between the accretion disk and NS surface dominates that in the disk. Moreover, this index stability effect now established for GX 3+ I was previously found in the atoll source 4U 1728-34 and suggested for a number of other low-mass X-ray NS binaries. This intrinsic behavior of NSs, in particular for atoll sources, is fundamentally different from that seen in black hole binary sources where the index monotonically increases during spectral transition from the low state to the high state and then finally saturates at high values of mass accretion rate.

Seifana, Elena↗
DOE OSTI2025

Dissipative Phase Transition in the Two-Photon Dicke Model

We explore the dissipative phase transition of the two-photon Dicke model, a topic that has garnered significant attention recently. Our analysis reveals that while single-photon loss does not stabilize the intrinsic instability in the model, the inclusion of two-photon loss restores stability, leading to the emergence of superradiant states, which coexist with the normal vacuum states. Using a second-order cumulant expansion for the photons, we derive an analytical description of the system in the thermodynamic limit, which agrees well with the exact calculation results. Additionally, we present the Wigner function for the system, shedding light on the breaking of the 𝑍4 symmetry inherent in the model. These findings offer valuable insights into stabilization mechanisms in open quantum systems and pave the way for exploring complex nonlinear dynamics in two-photon Dicke models.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

Challenges of COVID-19 Case Forecasting in the US, 2020–2021

During the COVID-19 pandemic, forecasting COVID-19 trends to support planning and response was a priority for scientists and decision makers alike. In the United States, COVID-19 forecasting was coordinated by a large group of universities, companies, and government entities led by the Centers for Disease Control and Prevention and the US COVID-19 Forecast Hub ( https://covid19forecasthub.org ). We evaluated approximately 9.7 million forecasts of weekly state-level COVID-19 cases for predictions 1–4 weeks into the future submitted by 24 teams from August 2020 to December 2021. We assessed coverage of central prediction intervals and weighted interval scores (WIS), adjusting for missing forecasts relative to a baseline forecast, and used a Gaussian generalized estimating equation (GEE) model to evaluate differences in skill across epidemic phases that were defined by the effective reproduction number. Overall, we found high variation in skill across individual models, with ensemble-based forecasts outperforming other approaches. Forecast skill relative to the baseline was generally higher for larger jurisdictions (e.g., states compared to counties). Over time, forecasts generally performed worst in periods of rapid changes in reported cases (either in increasing or decreasing epidemic phases) with 95% prediction interval coverage dropping below 50% during the growth phases of the winter 2020, Delta, and Omicron waves. Ideally, case forecasts could serve as a leading indicator of changes in transmission dynamics. However, while most COVID-19 case forecasts outperformed a naïve baseline model, even the most accurate case forecasts were unreliable in key phases. Further research could improve forecasts of leading indicators, like COVID-19 cases, by leveraging additional real-time data, addressing performance across phases, improving the characterization of forecast confidence, and ensuring that forecasts were coherent across spatial scales. In the meantime, it is critical for forecast users to appreciate current limitations and use a broad set of indicators to inform pandemic-related decision making.

59 BASIC BIOLOGICAL SCIENCES↗
NASA NTRS1982

Robust Kalman filter design for active flutter suppression systems

Additional insight is provided into the use of the Doyle-Stein (1979, 1981) technique in aeroelastic control problems by examining the application of the method to a flutter control problem. The system to be controlled consists of a full-size wind tunnel model of a wing, plus an aileron, an actuator, and an accelerometer used to sense the motion of the wing. A full-state feedback controller was designed using linear optimal control theory, and a Kalman filter was used in the feedback loop for state estimation. The filter design procedure is explained along with that to improve closed-loop properties of the system. The locus of the poles of the filter is examined as a scalar design parameter is varied. The Doyle-Stein design procedure is shown to substantially improve the stability properties of an active flutter controller designed using the linear quadratic Gaussian control theory.

Garrard, W. L.↗
NASA NTRS1987

Theoretical study of the dissociation energy and the red and violet band systems of CN

The dissociation energy (D sub O) of CN is determined to be 7.65 + or - 0.06 eV. This corresponds to delta H sub f (CN) = 105.3 + or - 1.5 kcal/mole, in excellent agreement with Engleman and Rouse (1975), but considerably larger than the recent value deduced from shock-tube studies by Colket (1984). The result is obtained not only from extensive ab initio MRCI calculations using a very large Gaussian basis set, but also from extrapolation of the directly computed value by comparison of computed and experimental results fo NO, C2, and N2. As an additional calibration of the methods, the D sub O value for CN was computed from the corresponding value for CN(-) using the experimental electron affinity data. The lifetime of the nu prime = 0 level of the violet (B 2 sigma + yields X 2 sigma +) system was computed to be 62.4 ns, in good agreement with both experiment and previous calculations. Lifetimes for the red (A 2 pi yields X 2 sigma +) system decrease with increasing nu prime, which is consistent both with the recent experiment and calculations. While the computed lifetimes are significantly longer that those obtained from the experiment, they are shorter than those deduced from an analysis of the solar spectrum. However the D sub O and f (sub OO) are consistent with Lambert's model for the solar spectrum.

Bauschlicher, Charles W., Jr.↗
NASA NTRS2004

Intrinsic Line Shape Measurements of the XRS Instrument on Astro-E2

The XRS instrument on the Astro-E2 observatory contains a substantially improved microcalorimeter array over the Astro-E mission. In addition to roughly a factor of 2 improvement in the detector resolution at 6 keV, the detector response is shown to be almost perfectly gaussian. We have made measurements of the detector response of the flight instrument, using a double crystal monochrometer at 4 and 8 keV, a 55-Fe internal conversion source, and x-ray induced fluorescence from a number of targets including Ti, Cu, and GaAs. The detector response has been measured to be entirely gaussian to at least 2 orders of magnitude down from the peak of the line or line complex. This is in sharp contrast to the results from the XRS on Astro-E where many channels exhibited excess counts on the high energy side of the spectral lines. Here we present details of the line shape measurement as well as the detector response as measured during the XRS ground calibration including details of the line fits and line models.

Porter, F. Scott↗
DOE OSTI2024

Tackling the curse of dimensionality in fractional and tempered fractional PDEs with physics-informed neural networks

Fractional and tempered fractional partial differential equations (PDEs) are effective models of long-range interactions, anomalous diffusion, and non-local effects. Traditional numerical methods for these problems are mesh-based, thus struggling with the curse of dimensionality (CoD). Physics-informed neural networks (PINNs) offer a promising solution due to their universal approximation, generalization ability, and mesh-free training. In principle, Monte Carlo fractional PINN (MC-fPINN) estimates fractional derivatives using Monte Carlo methods and thus could lift CoD. However, this may cause significant variance and errors, hence affecting convergence; in addition, MC-fPINN is sensitive to hyperparameters. In general, numerical methods and specifically PINNs for tempered fractional PDEs are under-developed. Herein, we extend MC-fPINN to tempered fractional PDEs to address these issues, resulting in the Monte Carlo tempered fractional PINN (MC-tfPINN). To reduce possible high variance and errors from Monte Carlo sampling, we replace the one-dimensional (1D) Monte Carlo with 1D Gaussian quadrature, applicable to both MC-fPINN and MC-tfPINN. We validate our methods on various forward and inverse problems of fractional and tempered fractional PDEs, scaling up to 100,000 dimensions. Our improved MC-fPINN/MC-tfPINN using quadrature consistently outperforms the original versions in accuracy and convergence speed in very high dimensions.

42 ENGINEERING↗