Search NASA⌕ Search

SEARCH · Search NASA

Results for “Linearization error”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,117 records · Page 62

Regression of exchangeable relational arrays

Relational arrays represent measures of association between pairs of actors, often in varied contexts or over time. Trade flows between countries, financial transactions between individuals, contact frequencies between school children in classrooms and dynamic protein-protein interactions are all examples of relational arrays. Elements of a relational array are often modelled as a linear function of observable covariates. Uncertainty estimates for regression coefficient estimators, and ideally the coefficient estimators themselves, must account for dependence between elements of the array, e.g., relations involving the same actor. Existing estimators of standard errors that recognize such relational dependence rely on estimating extremely complex, heterogeneous structure across actors. This paper develops a new class of parsimonious coefficient and standard error estimators for regressions of relational arrays. Here we leverage an exchangeability assumption to derive standard error estimators that pool information across actors, and are substantially more accurate than existing estimators in a variety of settings. This exchangeability assumption is pervasive in network and array models in the statistics literature, but not previously considered when adjusting for dependence in a regression setting with relational data. We demonstrate improvements in inference theoretically, via a simulation study, and by analysis of a dataset involving international trade.

97 MATHEMATICS AND COMPUTING↗

Towards Automated Memory Model Generation Via Event Tracing

The importance of memory performance and capacity is a growing concern for high performance computing laboratories around the world. It has long been recognized that improvements in processor speed exceed the rate of improvement in dynamic random access memory speed and, as a result, memory access times can be the limiting factor in high performance scientific codes. The use of multi-core processors exacerbates this problem with the rapid growth in the number of cores not being matched by similar improvements in memory capacity, increasing the likelihood of memory contention. In this paper, we present WMTools , a lightweight memory tracing tool and analysis framework for parallel codes, which is able to identify peak memory usage and also analyse per-function memory use over time. An evaluation of WMTools , in terms of its effectiveness and also its overheads, is performed using nine established scientific applications/benchmark codes representing a variety of programming languages and scientific domains. We also show how WMTools can be used to automatically generate a parameterized memory model for one of these applications, a two-dimensional non-linear magnetohydrodynamics application, Lare2D . Through the memory model we are able to identify an unexpected growth term which becomes dominant at scale. With a refined model we are able to predict memory consumption with under 7% error.

97 MATHEMATICS AND COMPUTING↗

Orbital-free molecular dynamics simulations of a warm dense mixture: Examination of the excess-pressure matching rule

A form of the linear mixing rule involving the equality of excess pressures is tested with various mole fractions and various types of orbital-free molecular dynamics simulations. For all the cases considered, this mixing rule yields, within statistical error, the pressure of a mixture of helium and iron obtained by a direct simulation. In an attempt to interpret the robustness of the mixing rule, we show that it can be derived from thermodynamic stability if the system is regarded as a mixture of independent effective average atoms. The success of the mixing rule applied with equations of state including various degrees of approximation leads us to suggest its use in the thermodynamic domain where quantum molecular dynamics can be implemented.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

TU-F-BRE-01: A High Resolution Micro Fiber Scintillator Detector Optimized for SRS and SBRT in Vivo Real Time Treatment Verification

Purpose: We have built a high resolution real time scintillating fiber detector prototype to determine in real time the accuracy of stereotactic radiosurgery (SRS) and stereotactic body radiotherapy (SBRT) treatments when only a fraction of the planned dose was delivered. The motivation of this work is to enhance dose delivery accuracy and to achieve error free radiosurgery. Methods: A high density array of scintillating fibers and a high speed photo detectors array were integrated to implement a high resolution real time dosimeter that can sample with high resolution pulsed SRS and SBRT beams cross sections. The high efficiency of the developed system allows to read each linac pulse in real time and to compute the accumulated dose and dose errors when only a fraction of the beam was delivered. The fibers are highly packed in a substrate that is directly coupled to two 128 pixel arrays with a pitch matching the fiber spacing to achieve accurate spatial localization. The small cross section of the fiber array allows stacking multiple fiber arrays to measure independent angular profiles that are digitally processed in parallel for real time dosimetry. Results: We implemented a high density array detector prototype with a pitch of 0.5 mm, readout speed of 1.2 msec, and a response time of 0.5 usec. The fast reading speed has the capability to determining the dose in flattening free filter beams. The detector can be installed in transmission mode at the output port of a micro-MLC. Treatment deviations smaller than 3% are detected when less than 1/100 of the planned dose was delivered. Conclusions: We built a prototype of a high resolution fiber scintillator array detector for SRS and SBRT in vivo dosimetry. Results show that the developed detector has the potential to assure error free SRS and SBRT treatments.

60 APPLIED LIFE SCIENCES↗

SU-E-J-28: Gantry Speed Significantly Affects Image Quality and Imaging Dose for 4D Cone-Beam Computed Tomography On the Varian Edge Platform

Purpose: 4D-CBCT facilitates assessment of tumor motion at treatment position. We investigated the effect of gantry speed on 4D-CBCT image quality and dose using the Varian Edge On-Board Imager (OBI). Methods: A thoracic protocol was designed using a 125 kVp spectrum. Image quality parameters were obtained via 4D acquisition using a Catphan phantom with a gating system. A sinusoidal waveform was executed with a five second period and superior-inferior motion. 4D-CBCT scans were sorted into 4 and 10 phases. Image quality metrics included spatial resolution, contrast-to-noise ratio (CNR), uniformity index (UI), Hounsfield unit (HU) sensitivity, and RMS error (RMSE) of motion amplitude. Dosimetry was accomplished using Gafchromic XR-QA2 films within a CIRS Thorax phantom. This was placed on the gating phantom using the same motion waveform. Results: High contrast resolution decreased linearly from 5.93 to 4.18 lp/cm, 6.54 to 4.18 lp/cm, and 5.19 to 3.91 lp/cm for averaged, 4 phase, and 10 phase 4DCBCT volumes respectively as gantry speed increased from 1.0 to 6.0 degs/sec. CNRs decreased linearly from 4.80 to 1.82 as the gantry speed increased from 1.0 to 6.0 degs/sec, respectively. No significant variations in UIs, HU sensitivities, or RMSEs were observed with variable gantry speed. Ion chamber measurements compared to film yielded small percent differences in plastic water regions (0.1–9.6%), larger percent differences in lung equivalent regions (7.5–34.8%), and significantly larger percent differences in bone equivalent regions (119.1–137.3%). Ion chamber measurements decreased from 17.29 to 2.89 cGy with increasing gantry speed from 1.0 to 6.0 degs/sec. Conclusion: Maintaining technique factors while changing gantry speed changes the number of projections used for reconstruction. Increasing the number of projections by decreasing gantry speed decreases noise, however, dose is increased. The future of 4DCBCT’s clinical utility relies on further investigation of image optimization.

60 APPLIED LIFE SCIENCES↗

Moment of Fluid Interface Reconstruction with Filaments [Slides]

A moving system made up of multiple fluids (e.g. air and water) may be defined by an evolving interface with a changing topology. MOF uses a piecewise linear interface reconstruction to numerically model deforming boundaries. Given a volume fraction V and reference centroid x for a material in cell Ω, we seek to find an interface Γ that exactly captures V and minimizes error in x. This differs from Volume of Fluid methods.

97 MATHEMATICS AND COMPUTING↗

Quantitative Non-Destructive Evaluation of Fatigue Damage Based on Multi-Sensor Fusion

Based on sensor fusion and machine learning, this project developed a novel non-destructive evaluation (NDE) methodology, which consists of a remaining useful life (RUL) prediction framework and regression models for predicting residual stress and full width at half maximum (FWHM). A series of fatigue testing experiments were conducted using 5052-H32 aluminum alloy specimens. All specimens were measured using linear ultrasonic (LU) and nonlinear ultrasonic (NLU) testing methods non-destructively. Machine learning models were developed to use LU and NLU measurements to predict loading condition, fatigue level, residual stress, and FWHM. It was demonstrated that the developed methodology could distinguish new and fatigue specimens with an accuracy of 97.53%. Also, the prediction errors for residual stress and FWHM were as low as 4.73% and 1.62%, respectively. An interactive database was created to publicly share the data generated from the project. It is envisioned that the developed NDE technology will equip manufacturers with a responsive screening system for incoming used metallic components, and potentially lead to a significant increase in using used metallic components for remanufacturing.

42 ENGINEERING↗

A Fusion of Geothermal and InSAR Data with Machine Learning for Enhanced Deformation Forecasting at the Geysers

The Geysers geothermal field in California is experiencing land subsidence due to the seismic and geothermal activities taking place. This poses a risk not only to the underlying infrastructure but also to the groundwater level which would reduce the water availability for the local community. Because of this, it is crucial to monitor and assess the surface deformation occurring and adjust geothermal operations accordingly. In this study, we examine the correlation between the geothermal injection and production rates as well as the seismic activity in the area, and we show the high correlation between the injection rate and the number of earthquakes. This motivates the use of this data in a machine learning model that would predict future deformation maps. First, we build a model that uses interferometric synthetic aperture radar (InSAR) images that have been processed and turned into a deformation time series using LiCSBAS, an open-source InSAR time series package, and evaluate the performance against a linear baseline model. The model includes both convolutional neural network (CNN) layers as well as long short-term memory (LSTM) layers and is able to improve upon the baseline model based on a mean squared error metric. Then, after getting preprocessed, we incorporate the geothermal data by adding them as additional inputs to the model. This new model was able to outperform both the baseline and the previous version of the model that uses only InSAR data, motivating the use of machine learning models as well as geothermal data in assessing and predicting future deformation at The Geysers as part of hazard mitigation models which would then be used as fundamental tools for informed decision making when it comes to adjusting geothermal operations.

Yazbeck, Joe (ORCID:0000000302235260)↗

Gas fraction and depletion time of massive star-forming galaxies at z ∼ 3.2: no change in global star formation process out to z > 3

The observed evolution of the gas fraction and its associated depletion time in main-sequence (MS) galaxies provides insights on how star formation proceeds over cosmic time. We report ALMA detections of the rest-frame ∼300 μm continuum observed at 240 GHz for 45 massive (〈log(M{sub ⋆}(M{sub ⊙}))〉=10.7), normal star-forming (〈log(sSFR(yr{sup −1}))〉=−8.6), i.e., MS, galaxies at z≈3.2 in the COSMOS field. From an empirical calibration between cold neutral, i.e., molecular and atomic, gas mass M{sub gas} and monochromatic (rest-frame) infrared luminosity, the gas mass for this sample is derived. Combined with stellar mass M{sub ⋆} and star formation rate (SFR) estimates (from MagPhys fits) we obtain a median gas fraction of μ{sub gas}=M{sub gas}/M{sub ⋆}=1.65{sub −0.19}{sup +0.18} and a median gas depletion time t{sub depl.}(Gyr)=M{sub gas/SFR}=0.68{sub −0.08}{sup +0.07}; correction for the location on the MS will only slightly change the values. The reported uncertainties are the 1σ error on the median. Our results are fully consistent with the expected flattening of the redshift evolution from the 2-SFM (2 star formation mode) framework which empirically prescribes the evolution assuming a universal, log-linear relation between SFR and gas mass coupled to the redshift evolution of the specific star formation rate (sSFR) of MS galaxies. While t{sub depl.} shows only a mild dependence on location within the MS, a clear trend of increasing μ{sub gas} across the MS is observed (as known from previous studies). Further, we comment on trends within the MS and (in)consistencies with other studies.

79 ASTRONOMY AND ASTROPHYSICS↗

Solving phase appearance/disappearance two-phase flow problems with high resolution staggered grid and fully implicit schemes by the Jacobian-free Newton–Krylov Method

The phase appearance/disappearance issue presents serious numerical challenges in two-phase flow simulations. Many existing reactor safety analysis codes use different kinds of treatments for the phase appearance/disappearance problem. However, to our best knowledge, there are no fully satisfactory solutions. Additionally, the majority of the existing reactor system analysis codes were developed using low-order numerical schemes in both space and time. In many situations, it is desirable to use high-resolution spatial discretization and fully implicit time integration schemes to reduce numerical errors. In this work, we adapted a high-resolution spatial discretization scheme on staggered grid mesh and fully implicit time integration methods (such as BDF1 and BDF2) to solve the two-phase flow problems. The discretized nonlinear system was solved by the Jacobian-free Newton Krylov (JFNK) method, which does not require the derivation and implementation of analytical Jacobian matrix. These methods were tested with a few two-phase flow problems with phase appearance/disappearance phenomena considered, such as a linear advection problem, an oscillating manometer problem, and a sedimentation problem. The JFNK method demonstrated extremely robust and stable behaviors in solving the two-phase flow problems with phase appearance/disappearance. No special treatments such as water level tracking or void fraction limiting were used. High-resolution spatial discretization and second- order fully implicit method also demonstrated their capabilities in significantly reducing numerical errors.

97 MATHEMATICS AND COMPUTING↗

QuantumNAT: quantum noise-aware training with noise injection, quantization and normalization

Parameterized Quantum Circuits (PQC) are promising towards quantum advantage on near-term quantum hardware. However, due to the large quantum noises (errors), the performance of PQC models has a severe degradation on real quantum devices. Take Quantum Neural Network (QNN) as an example, the accuracy gap between noise-free simulation and noisy results on IBMQ-Yorktown for MNIST-4 classification is over 60%. Existing noise mitigation methods are general ones without leveraging unique characteristics of PQC; on the other hand, existing PQC work does not consider noise effect. To this end, we present QuantumNAT, a PQC-specific framework to perform noise-aware optimizations in both training and inference stages to improve robustness. We experimentally observe that the effect of quantum noise to PQC measurement outcome is a linear map from noise-free outcome with a scaling and a shift factor. Motivated by that, we propose post-measurement normalization to mitigate the feature distribution differences between noise-free and noisy scenarios. Furthermore, to improve the robustness against noise, we propose noise injection to the training process by inserting quantum error gates to PQC according to realistic noise models of quantum hardware. Lastly, post-measurement quantization is introduced to quantize the measurement outcomes to discrete values, achieving the denoising effect. Extensive experiments on 8 classification tasks using 6 quantum devices demonstrate that QuantumNAT improves accuracy by up to 43%, and achieves over 94% 2-class, 80% 4-class, and 34% 10-class classification accuracy measured on real quantum computers. The code for construction and noise-aware training of PQC is available in the TorchQuantum library.

Wang, Hanrui↗

Improvements to the three-color optical CCD-based pyrometer system

We put forward an improved three-color measurement scheme with a color CCD sensor using the combination of effective wavelengths and blackbody calibrations. The process of effective wavelengths realizes the separation of the emissivity function from the measurement integral equation. This improved method not only effectively avoids errors arising from the traditional process adopting the basic wavelengths, but also simplifies the complex integral calculations. For a blackbody or graybody, the measurement performance of the pyrometer was experimentally investigated from the point of the temperature sensitivity. The results show that narrowing the spectrum bandwidth with different bandpass spectral filters may improve the temperature sensitivities. The spectrum adjustment not only makes the linear emissivity more suitable in the narrower waveband, but also reduces the numerical uncertainty of the effective wavelengths. For nongray objects, the effects of the numerical uncertainty of the effective wavelengths to measurements were quantificationally evaluated at different measurement conditions (different temperatures, emissivities, and spectrums). The results indicate that the reference values of 584, 555, and 511nm or 607, 560, and 506nm of the effective wavelengths are reasonable for the present system with the modified spectral response using Filter 1 or Filter 2. The maximum calculation errors arising from the reference effective wavelengths are less than 1.6%. In this paper, the technical realization steps of the improved method are also described. The measurement experiments of a metal sample in hot environments were carried out to further verify the applicability of this improved CCD-based pyrometer. The results agreed well with the spectrometer-based results.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

NSLS-II: Nonlinear Model Calibration for Synchrotrons

This tech note is essentially a summary of a lecture we delivered to the Acc. Phys. Journal Club Apr, 2010. However, since the estimated accuracy of these methods has been naive and misleading in the field of particle accelerators, i.e., ignores the impact of noise, we will elaborate on this in some detail. A prerequisite for a calibration of the nonlinear Hamiltonian is that the quadratic part has been understood, i.e., that the linear optics for the real accelerator has been calibrated. For synchrotron light source operations, this problem has been solved by the interactive LOCO technique/tool (Linear Optics from Closed Orbits). Before that, in the context of hadron accelerators, it has been done by signal processing of turn-by-turn BPM data. We have outlined how to make a basic calibration of the nonlinear model for synchrotrons. In particular, we have shown how this was done for LEAR, CERN (antiprotons) in the mid-80s. Specifically, our accuracy for frequency estimation was {approx} 1 x 10{sup -5} for 1024 turns (to calibrate the linear optics) and {approx} 1 x 10{sup -4} for 256 turns for tune footprint and betatron spectrum. For a comparison, the estimated tune footprint for stable beam for NSLS-II is {approx}0.1. Since the transverse damping time is {approx}20 msec, i.e., {approx}4,000 turns. There is no fundamental difference for: antiprotons, protons, and electrons in this case. Because the estimated accuracy for these methods in the field of particle accelerators has been naive, i.e., ignoring the impact of noise, we have also derived explicit formula, from first principles, for a quantitative statement. For e.g. N = 256 and 5% noise we obtain {delta}{nu} {approx} 1 x 10{sup -5}. A comparison with the state-of-the-arts in e.g. telecomm and electrical engineering since the 60s is quite revealing. For example, Kalman filter (1960), crucial for the: Ranger, Mariner, and Apollo (including the Lunar Module) missions during the 60s. Or Claude Shannon et al since the 40s for that matter. Conclusion: what's elementary in the latter is considered 'advanced', if at all, in the former. It is little surprise then that published measurements typically contains neither error bars (for the random errors) nor estimates for the systematic in the former discipline. We have also showed how to estimate the state space by turn-by-turn data from two adjacent BPMs. And how to improve the resolution of the nonlinear resonance spectrum by Fourier analyzing the linear action variables instead of the betatron motion. In fact, the state estimator could be further improved by adding a Kalman filter. For transparency, we have also summarized on how these techniques provide a framework- and method for a TQM (Total Quality Management) approach for the main ring. Of course, to make the ($2.5M) turn-by-turn data acquisition system that is being implemented (for all the BPMs) useful, a means ({approx}10% contingency for the BPM system) to drive the beam is obviously required.

43 PARTICLE ACCELERATORS↗

Isospin breaking and chiral symmetry restoration

We analyze quark condensates and chiral (scalar) susceptibilities including isospin-breaking effects at finite temperature T. These include m{sub u{ne}}m{sub d} contributions as well as electromagnetic (e{ne}0) corrections, both treated in a consistent chiral Lagrangian framework to leading order in SU(2) and SU(3) chiral perturbation theory, so that our predictions are model-independent. The chiral restoration temperature extracted from <qq>=<uu+dd> is almost unaffected, while the isospin-breaking order parameter <uu-dd> grows with T for the three-flavor case SU(3). We derive a sum rule relating the condensate ratio <qq>(e{ne}0)/<qq>(e=0) with the scalar susceptibility difference {chi}(T)-{chi}(0), directly measurable on the lattice. This sum rule is useful also for estimating condensate errors in staggered lattice analysis. Keeping m{sub u{ne}}m{sub d} allows one to obtain the connected and disconnected contributions to the susceptibility, even in the isospin limit, whose temperature, mass, and isospin-breaking dependence we analyze in detail. The disconnected part grows linearly, diverging in the chiral (infrared) limit as T/M{sub {pi}}, while the connected part shows a quadratic behavior, infrared regular as T{sup 2}/M{sub {eta}}{sup 2}, and coming from {pi}{sup 0{eta}} mixing terms. This smooth connected behavior suggests that isospin-breaking correlations are weaker than critical chiral ones near the transition temperature. We explore some consequences in connection with lattice data and their scaling properties, for which our present analysis for physical masses, i.e. beyond the chiral limit, provides a useful model-independent description for low and moderate temperatures.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A comparative analysis of different van der Waals treatments for molecular adsorption on the basal plane of 2H-MoS 2

The binding energy of hydrogen sulfide, ammonia, ethane, ethylene, butadiene, benzene, toluene, pyridine, pyrrole, and thiophene on the basal plane of the semi-conducting 2H-molybdenum sulfide (MoS 2 ) was calculated with the following flavors of Density Functional Theory (DFT): GGA-PW91, PBE-D3, vdW-DF, optPBE, optB86b, optB88, vdW-TS, and BEEF-vdW. The GGA-PW91 binding energies are negligible (<0.07 eV in magnitude) in all cases. The predictions with vdW-DF and PBE-D3 are the closest (error <0.05 eV) to the isosteric heats of adsorption calculated from reported temperature programmed desorption data for thiophene and butadiene. For all dispersion flavors examined here, the magnitude of the dispersion contribution to the binding energy increases linearly with the number of heavy atoms in the adsorbate, with each atom contributing 0.05 eV (BEEF-vdW) – 0.09 eV (optB88-vdW). Further, this implies that the calculated adsorption constants of molecules larger than acridine (i.e., comprising > 14 non-heavy atoms) can vary by more than four orders of magnitude at industrial conditions depending on the chosen method of dispersion correction. Further, dispersion effects fall off rapidly (>0.03 eV/ non-hydrogen atom/Å) as the adsorbate-surface distance increases.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Reduced order modeling for flow and transport problems with Barlow Twins self-supervised learning

Abstract We propose a unified data-driven reduced order model (ROM) that bridges the performance gap between linear and nonlinear manifold approaches. Deep learning ROM (DL-ROM) using deep-convolutional autoencoders (DC–AE) has been shown to capture nonlinear solution manifolds but fails to perform adequately when linear subspace approaches such as proper orthogonal decomposition (POD) would be optimal. Besides, most DL-ROM models rely on convolutional layers, which might limit its application to only a structured mesh. The proposed framework in this study relies on the combination of an autoencoder (AE) and Barlow Twins (BT) self-supervised learning, where BT maximizes the information content of the embedding with the latent space through a joint embedding architecture. Through a series of benchmark problems of natural convection in porous media, BT–AE performs better than the previous DL-ROM framework by providing comparable results to POD-based approaches for problems where the solution lies within a linear subspace as well as DL-ROM autoencoder-based techniques where the solution lies on a nonlinear manifold; consequently, bridges the gap between linear and nonlinear reduced manifolds. We illustrate that a proficient construction of the latent space is key to achieving these results, enabling us to map these latent spaces using regression models. The proposed framework achieves a relative error of 2% on average and 12% in the worst-case scenario (i.e., the training data is small, but the parameter space is large.). We also show that our framework provides a speed-up of $$7 \times 10^{6}$$ 7 × 10 6 times, in the best case, and $$7 \times 10^{3}$$ 7 × 10 3 times on average compared to a finite element solver. Furthermore, this BT–AE framework can operate on unstructured meshes, which provides flexibility in its application to standard numerical solvers, on-site measurements, experimental data, or a combination of these sources.

97 MATHEMATICS AND COMPUTING↗

Dual polarized receiving steering antenna array for measurement of ultrawideband pulse polarization structure

To measure simultaneously two orthogonal components of the electromagnetic field of nano- and subnano-second duration, an antenna array has been developed. The antenna elements of the array are the crossed dipoles of dimension 5 × 5 cm. The arms of the dipoles are connected to the active four-pole devices to compensate the frequency response variations of a short dipole in the frequency band ranging from 0.4 to 4 GHz. The dipoles have superimposed phase centers allowing measuring the polarization structure of the field in different directions. The developed antenna array is the linear one containing four elements. The pattern maximum position is controlled by means of the switched ultrawideband true time delay lines. Discrete steering in seven directions in the range from −40° to +40° has been realized. The error at setting the pattern maximum position is less than 4°. The isolation of the polarization exceeds 29 dB in the direction orthogonal to the array axis and in the whole steering range it exceeds 23 dB. Measurement results of the polarization structure of radiated and scattered pulses with different polarization are presented as well.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

ON THE LINEAR TERM CORRECTION FOR NEEDLET/WAVELET NON-GAUSSIANITY ESTIMATORS

We derive the linear correction term for needlet and wavelet estimators of the bispectrum and the nonlinearity parameter f{sub NL} on cosmic microwave background radiation data. We show that on masked WMAP-like data with anisotropic noise, the error bars improve by 10%-20% and almost reach the optimal error bars obtained with the bispectrum estimator also known as 'KSW'. In the limit of full-sky and isotropic noise, this term vanishes. We apply needlet and wavelet estimators to the WMAP 7-year data and obtain our best estimate f{sub NL} = 37.5 {+-} 21.8 (68% CL).

79 ASTRONOMY AND ASTROPHYSICS↗