Search NASA⌕ Search

SEARCH · Search NASA

Results for “Functional data analysis”

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

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

At least 631 records · Page 35

A modal approach to the numerical calculation of primordial non-Gaussianities

We propose a new method to numerically calculate higher-order correlation functions of primordial fluctuations generated from any early-universe scenario. Our key-starting point is the realization that the tree-level In-In formalism is intrinsically separable. This enables us to use modal techniques to efficiently calculate and represent non-Gaussian shapes in a separable form well suited to data analysis. We prove the feasibility and the accuracy of our method by applying it to simple single-field inflationary models in which analytical results are available, and we perform non-trivial consistency checks like the verification of the single field consistency relation. We also point out that the iε prescription is automatically taken into account in our method, preventing the need for ad-hoc tricks to implement it numerically.

79 ASTRONOMY AND ASTROPHYSICS↗

Architecture for interoperable software in biology

Understanding biological complexity demands a combination of high-throughput data and interdisciplinary skills. One way to bring to bear the necessary combination of data types and expertise is by encapsulating domain knowledge in software and composing that software to create a customized data analysis environment. To this end, simple flexible strategies are needed for interconnecting heterogeneous software tools and enabling data exchange between them. Drawing on our own work and that of others, we present several strategies for interoperability and their consequences, in particular, a set of simple data structures—list, matrix, network, table and tuple—that have proven sufficient to achieve a high degree of interoperability. We provide a few guidelines for the development of future software that will function as part of an interoperable community of software tools for biological data analysis and visualization.

59 BASIC BIOLOGICAL SCIENCES↗

Comprehensive Population-Averaged Arterial Input Function for Dynamic Contrast–Enhanced vMagnetic Resonance Imaging of Head and Neck Cancer

Purpose: To generate a population-averaged arterial input function (PA-AIF) for quantitative analysis of dynamic contrast-enhanced MRI data in head and neck cancer patients. Methods and Materials: Twenty patients underwent dynamic contrast-enhanced MRI during concurrent chemoradiation therapy. Imaging consisted of 2 baseline scans 1 week apart (B1/B2) and 1 scan after 1 week of chemoradiation therapy (Wk1). Regions of interest (ROIs) in the right and left carotid arteries were drawn on coronal images. Plasma concentration curves of all ROIs were averaged and fit to a biexponential decay function to obtain the final PA-AIF (AvgAll). Right-sided and left-sided ROI plasma concentration curves were averaged separately to obtain side-specific AIFs (AvgRight/AvgLeft). Regions of interest were divided by time point to obtain time-point-specific AIFs (AvgB1/AvgB2/AvgWk1). The vascular transfer constant (K{sub trans}) and the fractional extravascular, extracellular space volume (V{sub e}) for primaries and nodes were calculated using the AvgAll AIF, the appropriate side-specific AIF, and the appropriate time-point-specific AIF. Median K{sub trans} and V{sub e} values derived from AvgAll were compared with those obtained from the side-specific and time-point-specific AIFs. The effect of using individual AIFs was also investigated. Results: The plasma parameters for AvgAll were a{sub 1,2} = 27.11/17.65 kg/L, m{sub 1,2} = 11.75/0.21 min{sup −1}. The coefficients of repeatability (CRs) for AvgAll versus AvgLeft were 0.04 min{sup −1} for K{sub trans} and 0.02 for V{sub e}. For AvgAll versus AvgRight, the CRs were 0.08 min{sup −1} for K{sub trans} and 0.02 for V{sub e}. When AvgAll was compared with AvgB1/AvgB2/AvgWk1, the CRs were slightly higher: 0.32/0.19/0.78 min{sup −1}, respectively, for K{sub trans}; and 0.07/0.08/0.09 for V{sub e}. Use of a PA-AIF was not significantly different from use of individual AIFs. Conclusion: A PA-AIF for head and neck cancer was generated that accounts for differences in right carotid artery versus left carotid artery, day-to-day fluctuations, and early treatment-induced changes. The small CRs obtained for K{sub trans} and V{sub e} indicate that side-specific AIFs are not necessary. However, a time-point-specific AIF may improve pharmacokinetic accuracy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Machine-learning-enabled on-the-fly analysis of RHEED patterns during thin film deposition by molecular beam epitaxy

Thin film deposition is a fundamental technology for the discovery, optimization, and manufacturing of functional materials. Deposition by molecular beam epitaxy (MBE) typically employs reflection high-energy electron diffraction (RHEED) as a real-time in situ probe of the growing film. However, the state-of-the-art for RHEED analysis during deposition requires human observation. Here, we present an approach using machine learning (ML) methods to monitor, analyze, and interpret RHEED images on-the-fly during thin film deposition. In the analysis workflow, RHEED pattern images are collected at one frame per second and featurized using a pretrained deep convolutional neural network. The feature vectors are then statistically analyzed to identify changepoints; these changepoints can be related to changes in the deposition mode from initial film nucleation to a transition regime, smooth film deposition, and in some cases, an additional transition to a rough, islanded deposition regime. The feature vectors are additionally analyzed via graph analysis and community classification. The graph is quantified as a stabilization plot, and we show that inflection points in the stabilization plot correspond to changes in the growth regime. The full RHEED analysis workflow is termed RHAAPsody and includes data transfer and output to a visual dashboard. We demonstrate the functionality of RHAAPsody by analyzing the precaptured RHEED images from epitaxial depositions of anatase TiO2 on SrTiO3(001) and show that the analysis workflow can be executed in less than 1 s. Our approach shows promise as one component of ML-enabled real-time feedback control of the MBE deposition process.

36 MATERIALS SCIENCE↗

On estimating the curvature attributes and strain invariants of deformed surface through radial basis functions

Numerical methods in approximation of derivatives on regular or irregular gridded and scattered data points are of significance in the study of numerous problems in engineering and geoscience. A mathematically accurate and stable method is assessed to compute the differential quantities, while being flexible and easy to implement. Strain-invariant parameters in deformation analysis and curvature attributes in geometrical analysis of approximated surfaces are independent quantities which could be derived numerically through radial basis functions (RBFs) for scattered data points. Approximation of a function or its derivatives through RBFs is highly correlated to its shape parameter which depends on the number and distribution of the particles on the support domain. A procedure to find the optimal shape parameter for RBFs in the influence domain of each nodal point together with the presentation of a manner of calculating the curvature attributes and strain parameters through Gaussian RBFs is proposed here. Results indicate a significant improvement in the approximation accuracy of strain parameters and curvature attributes, while this approximation is more stable when RBFs with optimal shape parameter are implemented rather than the traditional moving least squares.

97 MATHEMATICS AND COMPUTING↗

Cluster spin glass correlations and dynamics in Zn 0.5⁢ Mn 0.5⁢ Te

Here, we present a combined magnetometry, muon spin-relaxation (𝜇⁢SR), and neutron-scattering study of the insulating spin glass Zn 0.5 ⁢Mn 0.5 ⁢Te, for which magnetic Mn 2+ and nonmagnetic Zn 2+ ions are randomly distributed on a face-centered cubic lattice. The magnetometry and 𝜇⁢SR results confirm a spin freezing transition around 𝑇 𝑓 ≈ 23 K, with the spin-fluctuation rate decreasing gradually and somewhat inhomogeneously through the sample volume as the temperature decreases toward 𝑇 𝑓 . Characteristic spin-correlation times well above 𝑇 𝑓 are on the order of 10 −10 s, much slower than typically observed in canonical spin glasses but in line with expectations for a cluster spin glass. Using magnetic pair distribution function (mPDF) analysis and reverse Monte Carlo (RMC) modeling of the magnetic diffuse neutron-scattering data, we show that the spin-glass ground state consists of clusters of spins exhibiting short-range-ordered type-III antiferromagnetic correlations with a locally ordered moment of 3.1⁢(1)⁢𝜇 B between nearest-neighbor spins. The type-III correlations decay exponentially as a function of spin separation distance with a correlation length of approximately 5 Å. The diffuse magnetic scattering and corresponding mPDF show no significant changes across 𝑇 𝑓 , indicating that the dynamically fluctuating short-range spin correlations in the paramagnetic state retain the same basic type-III configuration that characterizes the spin-glass state; the only change apparent from the neutron-scattering data is a gradual reduction of the correlation length and locally ordered moment with increasing temperature. Taken together, these results paint a unique and detailed picture of the local magnetic structure and dynamics in Zn 0.5 ⁢Mn 0.5⁢ Te and provide strong evidence that this material is best described as a cluster spin glass. In addition, this work showcases a statistical method for extracting diffuse scattering signals from neutron powder diffraction data, which we developed to facilitate the mPDF and RMC analysis of the neutron data. This method has the potential to be broadly useful for neutron powder diffraction experiments on a variety of materials with short-range atomic or magnetic order.

magnetism↗

Lynx web services for annotations and systems analysis of multi-gene disorders

Lynx is a web-based integrated systems biology platform that supports annotation and analysis of experimental data and generation of weighted hypotheses on molecular mechanisms contributing to human phenotypes and disorders of interest. Lynx has integrated multiple classes of biomedical data (genomic, proteomic, pathways, phenotypic, toxicogenomic, contextual and others) from various public databases as well as manually curated data from our group and collaborators (LynxKB). Lynx provides tools for gene list enrichment analysis using multiple functional annotations and network-based gene prioritization. Lynx provides access to the integrated database and the analytical tools via REST based Web Services (http://lynx.ci.uchicago.edu/webservices.html). This comprises data retrieval services for specific functional annotations, services to search across the complete LynxKB (powered by Lucene), and services to access the analytical tools built within the Lynx platform.

59 BASIC BIOLOGICAL SCIENCES↗

Topological data analysis of task-based fMRI data from experiments on schizophrenia

We use methods from computational algebraic topology to study functional brain networks, in which nodes represent brain regions and weighted edges represent similarity of fMRI time series from each region. With these tools, which allow one to characterize topological invariants such as loops in high-dimensional data, we are able to gain understanding into low-dimensional structures in networks in a way that complements traditional approaches based on pairwise interactions. In the present paper, we analyze networks constructed from task-based fMRI data from schizophrenia patients, healthy controls, and healthy siblings of schizophrenia patients using persistent homology, which allows us to explore the persistence of topological structures such as loops at different scales in the networks. We use persistence landscapes, persistence images, and Betti curves to create output summaries from our persistent-homology calculations, and we study the persistence landscapes and images using k-means clustering and community detection. Based on our analysis of persistence landscapes, we find that the members of the sibling cohort have topological features (specifically, their 1-dimensional loops) that are distinct from the other two cohorts. From the persistence images, we are able to distinguish all three subject groups and to determine the brain regions in the loops (with four or more edges) that allow us to make these distinctions.

60 APPLIED LIFE SCIENCES↗

Z-dependence analysis of M x-ray production cross sections for heavy elements with 60≤Z≤90 by protons impact

Motivated by the large deviation between the experiment and the predictions of the most often used model of ionization process by a charged particle, namely ECPSSR model, a large database of experimental M-shell X-ray production cross-sections by protons energies varying from 0.1 to 4.0 MeV for elements with atomic number 60 ≤ Z ≤ 90, is collected from various sources published from 1980 till 2009 to deduce an empirical M x-ray production cross section. This latter is then deduced from the available experimental data as a function of the scaled velocity parameter by using the whole range of elements (collective analysis) or by introducing the dependence of these cross sections on the atomic number of the target, noted as “Z-dependence analysis” in addition to the collective one. The corresponding results and their deviation from the experimental data are presented for selected elements. Also, a comparison is made for selected elements between our results and other theoretical as well as experimental works.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Spatial analysis of cell patterning to aid genetic and phenotypic understanding of grass stomatal density: A case study in maize

Biological processes involve complex hierarchies where composite traits result from multiple component traits. However, holistically understanding of how sets of component traits interact to underpin genotype-to-phenotype relationships is generally lacking. Stomatal density (SD) is a tractable model system for exploring how high-throughput phenotyping (HTP) data could be exploited by a new spatial analysis approach to better understand a developmentally and functionally important trait. SD is a composite trait, resulting from various components related to cell identity and size, which are themselves governed by a series of spatio-developmental processes. Data from 192 recombinant inbred lines of maize [Zea mays (L.)] were analyzed by a new stomatal patterning phenotype (SPP) to (1) describe the average spatial probability distribution of the nearest neighboring stomata; (2) derive a core set of component traits related to cell size, cell packing, and positional probabilities; (3) build a structural equation model of component traits underlying SD; and (4) identify stomatal patterning quantitative trait loci (QTL). The core set of SPP-derived traits explained 74% of the variation in SD. Analyzing SPP component traits allowed some loci previously identified as generic SD QTL to be recognized as specific to lateral versus longitudinal elements of stomatal patterning. Therefore, this study highlights how novel insights can be gained by decomposing a composite trait (e.g., SD) into a set of component traits that were present in HTP data but not previously exploited.

59 BASIC BIOLOGICAL SCIENCES↗

Optical characteristics of pulsed laser deposited Ge-Sb-Te thin films studied by spectroscopic ellipsometry

Pulsed laser deposition technique was used for the fabrication of (GeTe){sub 1-x}(Sb{sub 2}Te{sub 3}){sub x} (x = 0, 0.33, 0.50, 0.66, and 1) amorphous thin films. Scanning electron microscopy with energy-dispersive x-ray analysis, x-ray diffraction, optical reflectivity, and sheet resistance temperature dependences as well as variable angle spectroscopic ellipsometry measurements were used to characterize as-deposited (amorphous) and annealed (rocksaltlike) layers. In order to extract optical functions of the films, the Cody-Lorentz model was applied for the analysis of ellipsometric data. Fitted sets of Cody-Lorentz model parameters are discussed in relation with chemical composition and the structure of the layers. The GeTe component content was found to be responsible for the huge optical functions and thickness changes upon amorphous-to-fcc phase transition.

36 MATERIALS SCIENCE↗

Deciphering Reaction Mechanisms of Molecular Proton Reduction Catalysts with Cyclic Voltammetry: Kinetic vs Thermodynamic Control

The kinetics and thermodynamics of elementary reaction steps involved in the catalytic reduction of protons to hydrogen define the reaction landscape for catalysis. The mechanisms can differ in the order of the elementary proton transfer, electron transfer, and bond-forming steps and can be further differentiated by the sites at which protons and electrons localize. Access to fully elucidated mechanistic, kinetic, and thermochemical details of molecular catalysts is crucial to facilitate the development of new catalysts that operate with optimal efficiency, selectivity, and durability. The mechanism by which a catalyst operates, as well as the kinetics and thermodynamics associated with the individual steps, can often be accessed through electroanalytical studies. Here, this Account details the application of cyclic voltammetry to interrogate reaction mechanisms and quantify the kinetics and thermodynamics of elementary reaction steps for a series of molecular catalysts that mediate electrochemical proton reduction. I distinguish the limiting scenarios wherein a catalyst operates under kinetic control vs thermodynamic control, with a focus on detecting how cyclic voltammetry features shift with proton source strength and concentration, as well as scan rate. For systems that operate under kinetic control, catalytic currents are observed at, or slightly positive toward, the formal potential for the redox process that triggers catalysis. Under thermodynamic control, catalytic responses shift as a function of the proton source pKa and effective pH of the solution. After drawing this distinction, we introduce the appropriate voltammetry experiments and accompanying analytical expressions for extracting key metrics from the data. To illustrate analytical strategies to quantify elementary reaction steps of catalysts operating under kinetic control, I describe our studies of proton reduction catalysts Co(dmgBF 2 ) 2 (CH 3 CN) 2 (dmgBF 2 = difluoroboryl-dimethylglyoxime) and [Ni(P 2 Ph N 2 Ph ) 2 ] 2+ (P 2 Ph N 2 Ph = 1,5-phenyl-3,7-phenyl-1,5-diaza-3,7-diphosphacyclooctane). Here, peak shift analysis, foot-of-the-wave analysis, and plateau current analysis are applied to data sets wherein voltammetric response are recorded as a function of catalyst concentration, proton source concentration, proton source strength, and scan rate to quantify rate constants for elementary proton transfer and bond-forming steps in a catalytic cycle. Further, the case study of [Ni(P 2 Ph N 2 Ph ) 2 ] 2+ illustrates how complementary spectroscopic methods can bolster the mechanistic assignment. Collectively, these two studies showcase how detailed mechanistic studies inform on rate-limiting elementary steps in catalysis and other key processes underpinning catalysis. Second, I present analytical strategies to interrogate catalysts operating under thermodynamic control, centered on the case study of [Ni II (P 2 Ph N 2 Bn ) 2 ] 2+ (P 2 Ph N 2 Bn = 1,5-dibenzyl-3,7-diphenyl-1,5-diaza-3,7-diphosphacyclooctane). Here, the application of nonaqueous Pourbaix theory to extract thermodynamic information is introduced, and the construction of a coupled Pourbaix diagram is detailed. This study identifies ligand-based protonation as the key process that places catalysis under thermodynamic control and influences the reaction mechanism. Together, the work detailed in this Account showcases the utility of electroanalytical methods to disentangle complex reaction mechanisms and extract key thermochemical and kinetic parameters for elementary steps of catalysis. Through detailed presentation of the key analytical expressions that underpin these analyses, this Account seeks to facilitate the adoption of cyclic voltammetry by the community to fully extract kinetic, thermochemical, and mechanistic information on electrochemical small-molecule activation.

catalysts↗

Comment on “Solubilities of ammonia in basic imidazolium ionic liquids” [Fluid Phase Equilib. 297 (2010) 34–39]

The enthalpies of solvation of ammonia in imidazolium-based ionic liquids (and other associated thermodynamic properties) are re-analyzed based on the data of [Fluid Phase Equilib. 297 (2010) 34–39], correcting numerical errors in the original analysis. In contrast to the original conclusions, the re-analyzed data demonstrates that the enthalpy of solvation is not a strong function of alkyl chain length. Here, the errors and uncertainties of the analysis procedure are also discussed, suggesting that the strong observed temperature dependence of enthalpy may in fact be an artifact of the systematic neglect of the non-ideality of the ammonia solute.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Data-driven analysis and prediction of stable phases for high-entropy alloy design

High-entropy alloys (HEAs) represent a promising class of materials with exceptional structural and functional properties. However, their design and optimization pose challenges due to the large composition-phase space coupled with the complex and diverse nature of the phase formation dynamics. In this study, a data-driven approach that utilizes machine learning (ML) techniques to predict HEA phases and their composition-dependent phases is proposed. By employing a comprehensive dataset comprising 5692 experimental records encompassing 50 elements and 11 phase categories, we compare the performance of various ML models. Our analysis identifies the most influential features for accurate phase prediction. Furthermore, the class imbalance is addressed by employing data augmentation methods, raising the number of records to 1500 in each category, and ensuring a balanced representation of phase categories. The results show that XGBoost and Random Forest consistently outperform the other models, achieving 86% accuracy in predicting all phases. Additionally, this work provides an extensive analysis of HEA phase formers, showing the contributions of elements and features to the presence of specific phases. We also examine the impact of including different phases on ML model accuracy and feature significance. Notably, the findings underscore the need for ML model selection based on specific applications and desired predictions, as feature importance varies across models and phases. This study significantly advances the understanding of HEA phase formation, enabling targeted alloy design and fostering progress in the field of materials science.

36 MATERIALS SCIENCE↗

Assembly-Driven Community Genomics of a Hypersaline Microbial Ecosystem

Microbial populations inhabiting a natural hypersaline lake ecosystem in Lake Tyrrell, Victoria, Australia, have been characterized using deep metagenomic sampling, iterative de novo assembly, and multidimensional phylogenetic binning. Composite genomes representing habitat-specific microbial populations were reconstructed for eleven different archaea and one bacterium, comprising between 0.6 and 14.1% of the planktonic community. Eight of the eleven archaeal genomes were from microbial species without previously cultured representatives. These new genomes provide habitat-specific reference sequences enabling detailed, lineage-specific compartmentalization of predicted functional capabilities and cellular properties associated with both dominant and less abundant community members, including organisms previously known only by their 16S rRNA sequences. Together, these data provide a comprehensive, culture-independent genomic blueprint for ecosystem-wide analysis of protein functions, population structure, and lifestyles of co-existing, co-evolving microbial groups within the same natural habitat. The ‘‘assembly-driven’’ community genomic approach demonstrated in this study advances our ability to push beyond single gene investigations, and promotes genome-scale reconstructions as a tangible goal in the quest to define the metabolic, ecological, and evolutionary dynamics that underpin environmental microbial diversity.

54 ENVIRONMENTAL SCIENCES↗

Three-pion contribution to hadronic vacuum polarization

We address the contribution of the 3π channel to hadronic vacuum polarization (HVP) using a dispersive representation of the e + e - → 3π amplitude. This channel gives the second-largest individual contribution to the total HVP integral in the anomalous magnetic moment of the muon (g - 2) μ , both to its absolute value and uncertainty. It is largely dominated by the narrow resonances ω and Φ, but not to the extent that the off-peak regions were negligible, so that at the level of accuracy relevant for (g - 2) μ an analysis of the available data as model independent as possible becomes critical. Here, we provide such an analysis based on a global fit function using analyticity and unitarity of the underlying γ* → 3π amplitude and its normalization from a chiral low-energy theorem, which, in particular, allows us to check the internal consistency of the various e + e - → 3π data sets. Overall, we obtain a$3π\atop{μ}$|≤1.8 GeV = 46.2(6)(6) × 10 -10 as our best estimate for the total 3π contribution consistent with all (low-energy) constraints from QCD. In combination with a recent dispersive analysis imposing the same constraints on the 2π channel below 1 GeV, this covers nearly 80% of the total HVP contribution, leading to a$HVP\atop{μ}$ = 692.3(3.3) × 10 -10 when the remainder is taken from the literature, and thus reaffirming the (g-2) μ anomaly at the level of at least 3.4σ. As side products, we find for the vacuum-polarization-subtracted masses M ω = 782.63(3)(1) MeV and M Φ = 1019.20(2)(1) MeV, confirming the tension to the ω mass as extracted from the 2π channel.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Try-A Global Database of Plant Traits

Plant traits the morphological, anatomical, physiological, biochemical and phenological characteristics of plants and their organs determine how primary producers respond to environmental factors, affect other trophic levels, influence ecosystem processes and services and provide a link from species richness to ecosystem functional diversity. Trait data thus represent the raw material for a wide range of research from evolutionary biology, community and functional ecology to biogeography. Here we present the global database initiative named TRY, which has united a wide range of the plant trait research community worldwide and gained an unprecedented buy-in of trait data: so far 93 trait databases have been contributed. The data repository currently contains almost three million trait entries for 69 000 out of the world s 300 000 plant species, with a focus on 52 groups of traits characterizing the vegetative and regeneration stages of the plant life cycle, including growth, dispersal, establishment and persistence. A first data analysis shows that most plant traits are approximately log-normally distributed, with widely differing ranges of variation across traits. Most trait variation is between species (interspecific), but significant intraspecific variation is also documented, up to 40% of the overall variation. Plant functional types (PFTs), as commonly used in vegetation models, capture a substantial fraction of the observed variation but for several traits most variation occurs within PFTs, up to 75% of the overall variation. In the context of vegetation models these traits would better be represented by state variables rather than fixed parameter values. The improved availability of plant trait data in the unified global database is expected to support a paradigm shift from species to trait-based ecology, offer new opportunities for synthetic plant trait research and enable a more realistic and empirically grounded representation of terrestrial vegetation in Earth system models.

60 APPLIED LIFE SCIENCES↗

Landscaper v1

Understanding the inner workings of machine learning models through their loss landscapes offers crucial insights into model properties, optimization dynamics, and generalizability. However, accessing these insights has traditionally required specialized mathematical expertise, limiting broader adoption. Landscaper is an open-source Python package designed to bridge this gap. Landscaper seamlessly integrates a suite of multi-dimensional loss landscape analyses with cutting-edge topological data analysis (TDA) methods. This powerful combination makes both fundamental loss landscape analysis and advanced TDA techniques accessible to the broader scientific ML community, without requiring deep pre-existing mathematical knowledge. Landscaper offers three key functionalities: * Construction: Builds detailed loss landscape representations through versatile low and high-dimensional sampling techniques. * Quantification: Applies advanced metrics, including a novel topological data analysis (TDA) based smoothness metric, enabling new perspectives on model behavior. * Visualization: Offers intuitive tools to visualize and interpret loss landscapes, providing actionable insights beyond traditional performance metrics.

Weber, Gunther [Lawrence Berkeley National Laborat↗