Results for “sample complexity analysis” — Page 4

NASA NTRS1983

Stillwater anorthosites - A lunar analog?

Samples of anorthosite from Stillwater complex units AN-I and AN-II have been studied by instrumental neutron activation analysis for FeO, CaO, Na2O, REE, and other trace elements, and the results are discussed. No systematic variations of chemical composition with stratigraphic position were observed. The major and trace compositions of both units appear to be identical, with plagioclase compositions virtually constant throughout both. Thus, there is almost no evidence for evolution of parent magma composition during the genesis of these units. The bulk of the anorthosites appear to consist of cumulus-adcumulus plagioclase and pyroxene. Very little trapped liquid component was found. The chemical elements are not normally or lognormally distributed among samples taken from either anorthosite unit. The irregular distribution of pyroxene oikocrysts indicates that interstitial melt migrated substantially prior to final consolidation of the rock.

Salpas, P. A.↗

NASA NTRS2010

Extraterrestrial Amino Acids in the Almahata Sitta Meteorite

Amino acid analysis of a meteorite fragment of asteroid 2008 TC3 called Almahata Sitta was carried out using reverse-phase liquid chromatography coupled with UV fluorescence detection and time-of-flight mass spectrometry (LC-FD/ToF-MS) as part of a sample analysis consortium. LC-FD/ToF-MS analyses of hot-water extracts from the meteorite revealed a complex distribution of two- to seven-carbon aliphatic amino acids and one- to three-carbon amines with abundances ranging from 0.5 to 149 parts-per-billion (ppb). The enantiomeric ratios of the amino acids alanine, R-amino-n-butyric acid (beta-ABA), 2-amino-2-methylbutanoic acid (isovaline), and 2-aminopentanoic acid (norvaline) in the meteorite were racemic (D/L approximately 1), indicating that these amino acids are indigenous to the meteorite and not terrestrial contaminants. Several other non-protein amino acids were also identified in the meteorite above background levels including alpha-aminoisobutyric acid (alpha-AIB), 4-amino-2- methylbutanoic acid, 4-amino-3-methylbutanoic acid, and 3-, 4-, and 5-aminopentanoic acid. The total abundances of isovaline and alpha-AIB in Almahata Sitta are 1000 times lower than the abundances of these amino acids found in the CM carbonaceous chondrite Murchison. The extremely low abundances and unusual distribution of five carbon amino acids in Almahata Sitta compared to Cl, CM, and CR carbonaceous chondrites may reflect extensive thermal alteration of amino acids on the parent asteroid by partial melting during formation or subsequent impact shock heating. It is also possible that amino acids were synthesized by catalytic reactions on the parent body after asteroid 2008 TC3 cooled to lower temperatures.

Glavin, Daniel P.↗

NASA NTRSDate not supplied

Discovery and Analysis of Rare High-Impact Failure Modes using Adversarial RL-Informed Sampling

Adaptive learning agents have tremendous potential to handle critical tasks currently performed by humans. Unfortunately, due to their complexity, it can be difficult to verify that these learning agents do not have critical failure modes. Standard verification and validation methods often do not apply directly to learning agents and Monte Carlo methods have difficulty covering even a small fraction of the state space, especially in multiagent systems or over long time horizons. To overcome this difficulty, we demonstrate an adaptive stress-testing method based on reinforcement learning of correlations that raise the probability of failure. This approach has three key properties: (1) it is able to find rare failure modes with far greater sample efficiency than Monte Carlo methods, (2) it can estimate the true probability of a failure mode despite the inherent bias in the learning method, and (3) it is capable of learning and resampling compact representations of multimodal failure spaces. These properties are important in practice as we need to find disparate failure modes while accounting for their actual relevance. This is a significant advantage over traditional adaptive stress testing methods that give abstract likelihoods of particular failure instances, but cannot estimate the probability of a broader failure mode. We test our algorithm on a simple problem from the aviation domain where an autonomous aircraft lands in gusty wind conditions. The results suggest that we can find failure modes with far fewer samples than the Monte Carlo approach and simultaneously estimate the probability of failure.

reinforcement learning↗

NASA NTRS1975

Computer program for analysis of vectorcardiograms (VECTAN II)

VECTAN II accepts as input digitized three-lead VCG data sampled at 320 samples/second/lead, analyzing one VCG complex in each 5-second interval for experiments of up to 25 minutes duration. Program calibrates these data, locates major waveforms, performs waveform analyses, and produces statistical summary of analyzed data.

Hoffler, G. W.↗

NASA NTRS2001

Determining Biosignatures by Complexity Analysis in Antarctic Cryptoendolithic Communities

One of the most difficult problems of life detection is that of identifying biosignatures across a wide range of scales using multiple co-registered probes. The technique should be of equal utility across a wide range of search spaces from remote sensors probing volumes of space or planetary surfaces, visual eye or camera searches across the surface of a rock in Antarctica, low resolution microscopic scanning of a rock or a space craft in situ, or high resolution electron microscope and computerized tomography scanning of geobiological samples. We describe here an approach to this problem which derives in large part from past work done in the area of astrophysics - namely the analysis of complexity in galactic signals by data compression methods. This approach is a radically new one for geobiology and astrobiology, and allows us to assess the complexity (and thus potential biogenicity) of an object being examined. This is done by considering the information within pixels of an image (regardless the sensor used to gather the information) as an energetic system capable of description in terms of classical thermodynamics. The image data space is searched by an algorithm that judges complexity via data compression (e.g., the more compressible it is, the less complex, and vice versa) and maximum entropy as originally outlined by Shannon. At present we are implementing methods to utilize images from multiple sensors gathering different kinds of information (e.g., visible gray-scale data, color analyses, UV fluorescence, chemical information, etc). We present here preliminary data from deep UV fluorescence and ESEM (Environmental Scanning Electron Microscope) images from a layered cryptoendolithic community of an Antarctic rock.

Storrie-Lombardi, M. C.↗

NASA NTRS1987

Computational analyses of crystal growth

Two important aspects of Hg/Cd/Te crystal growth processes are discussed. First, the thermal field and second, the fluid movement in the melt zone. The thermal analysis includes numerical calculation of axisymmetric heat conduction within the sample. It also includes a three-dimensional radiation model to calculate the radiative heat exchange between the furnace and the crystal as determined by the complex geometry of the furnace and the adiabatic shield. The thermal analysis also includes a crystal conductivity which is dependent on temperature and composition. To tackle the fluid flow aspect of the problem, an attempt was made to use a newly developed incompressible flow code based on the slight compressibility, and hence the finite sound speed, of all real fluids.

Dakhoul, Youssef M.↗

NASA NTRS2014

Coordinated Mineralogical and Isotopic Analysis of a Cosmic Symplectite Identified in a Stardust Terminal Particle

Comet Wild-2 samples returned by the Stardust spacecraft contain a chemically diverse mixture of material, underscoring the complex nature of comets. Studies of entire Stardust aerogel tracks afford the opportunity to examine the fine-grained particle fragments distributed along the length of the track as well as the terminal particles. Previous TEM characterization of a terminal particle (TP) in track #147 revealed a symplectically intergrown iron sulfide and oxide assemblage. Mineralogically similar assemblages, known as cosmic symplectites (COS, formerly termed "new-PCP"), have only been identified in the primitive carbonaceous chondrite Acfer 094. Meteoritic COS have isotopically heavy O compositions (delta (sup 17), O-18 = 180per mille) that point to interactions with early solar system primordial water. In this study we report mineralogical and O isotopic measurements of the Wild-2 COS assemblage. Experimental: Track #147 is a "bulbous"-type track (4600 microns long) containing 7 terminal particles. The TPs were removed from the track, embedded in epoxy, and ultramicrotomed. A JEOL 2500SE 200 keV field-emission scanning-transmission electron microscope was used to obtain quantitative elemental maps and detailed mineralogical characterization. Following TEM analysis, two thin sections of TP4 (12 microns) were analyzed for O isotopes by raster ion imaging with the JSC NanoSIMS 50L. All three O isotopes were measured simultaneously using electron multipliers. San Carlos olivine grains were used as isotopic standards. Results and Discussion: The COS in the Wild-2 track #147 TP4 sample consists of symplectically intergrown pentlandite and nanocrystalline maghemite which coexists with high-Ca pyroxene with Na and Cr (kosmochlor component). This kosmochlor component could have a nebular origin and be precursors to type II chondrules in ordinary chondrites. Yet pentlandite is not a stable phase in the nebula. The COS in Acfer 094 also consists of pentlandite, but contains magnetite [4] rather than the more oxidized maghemite observed in the Wild-2 COS. The Acfer 094 COS display heavy O isotopic compositions that are the result of sulfidization and oxidation of Fe, Ni-metal grains and sulfides by O-17- and O-18-rich water in the solar nebula or possibly on the parent body. The O isotopic composition of the Wild-2 COS, however, is indistinguishable from terrestrial, indicating it was not altered by the same primordial aqueous reservoir as the Acfer 094 COS. The alteration could have occurred on the parent body by isotopically equilibrated ice. The mineralogy and petrography of Wild-2 samples suggests an incomplete or nascent hydration process. In future work we will analyze S isotopes in the track #147 TP4 COS and search for additional COS in Stardust samples.

Nguyen, A. N.↗

NASA NTRS2010

The Proposed Mars Astrobiology Explorer - Cacher [MAX-C] Rover: First Step in a Potential Sample Return Campaign

Sample return from Mars has been advocated by numerous scientific advisory panels for over 30 years, most prominently beginning with the National Research Council s [1] strategy for the exploration of the inner solar system, and most recently by the Mars Exploration Program Analysis Group (MEPAG s) Next Decade Science Analysis Group [2]. Analysis of samples here on Earth would have enormous advantages over in situ analyses in producing the data quality needed to address many of the complex scientific questions the community has posed about Mars. Instead of a small, predetermined set of analytical techniques, state of the art preparative and instrumental resources of the entire scientific community could be applied to the samples. The analytical emphasis could shift as the meaning of each result becomes better appreciated. These arguments apply both to igneous rocks and to layered sedimentary materials, either of which could contain water and other volatile constituents. In 2009 MEPAG formed the Mid-Range Rover Science Analysis Group (MRR-SAG) to formulate a mission concept that would address two general objectives: (1) conduct high-priority in situ science and (2) make concrete steps towards the potential return of samples to Earth. This analysis resulted in a mission concept named the Mars Astrobiology Explorer-Cacher (MAX-C), which was envisioned for launch in the 2018 opportunity. After extensive discussion, this group concluded that by far the most definitive contribution to sample return by this mission would be to collect and cache, in an accessible location, a suite of compelling samples that could potentially be recovered and returned by a subsequent mission. This would have the effect of separating two of the essential functions of MSR, the acquisition of the sample collection and its delivery to martian orbit, into two missions.

Allen, Carlton C.↗

NASA NTRSDate not supplied

Discovery and Analysis of Rare High-Impact Failure Modes using Adversarial RL-Informed Sampling

Adaptive learning agents have tremendous potential to handle critical tasks currently performed by humans. Unfortunately, due to their complexity, it can be difficult to verify that these learning agents do not have critical failure modes. Standard verification and validation methods often do not apply directly to learning agents and Monte Carlo methods have difficulty covering even a small fraction of the state space, especially in multiagent systems or over long time horizons. To overcome this difficulty, we demonstrate an adaptive stress-testing method based on reinforcement learning of correlations that raise the probability of failure. This approach has three key properties: (1) it is able to find rare failure modes with far greater sample efficiency than Monte Carlo methods, (2) it can estimate the true probability of a failure mode despite the inherent bias in the learning method, and (3) it is capable of learning and resampling compact representations of multimodal failure spaces. These properties are important in practice as we need to find disparate failure modes while accounting for their actual relevance. This is a significant advantage over traditional adaptive stress testing methods that give abstract likelihoods of particular failure instances, but cannot estimate the probability of a broader failure mode. We test our algorithm on a simple problem from the aviation domain where an autonomous aircraft lands in gusty wind conditions. The results suggest that we can find failure modes with far fewer samples than the Monte Carlo approach and simultaneously estimate the probability of failure.

Validation↗

NASA NTRSDate not supplied

Discovery and Analysis of Rare High-Impact Failure Modes using Adversarial RL-Informed Sampling

Adaptive learning agents have tremendous potential to handle critical tasks currently performed by humans. Unfortunately, due to their complexity, it can be difficult to verify that these learning agents do not have critical failure modes. Standard verification and validation methods often do not apply directly to learning agents and Monte Carlo methods have difficulty covering even a small fraction of the state space, especially in multiagent systems or over long time horizons. To overcome this difficulty, we demonstrate an adaptive stress-testing method based on reinforcement learning of correlations that raise the probability of failure. This approach has three key properties: (1) it is able to find rare failure modes with far greater sample efficiency than Monte Carlo methods, (2) it can estimate the true probability of a failure mode despite the inherent bias in the learning method, and (3) it is capable of learning and resampling compact representations of multimodal failure spaces. These properties are important in practice as we need to find disparate failure modes while accounting for their actual relevance. This is a significant advantage over traditional adaptive stress testing methods that give abstract likelihoods of particular failure instances, but cannot estimate the probability of a broader failure mode. We test our algorithm on a simple problem from the aviation domain where an autonomous aircraft lands in gusty wind conditions. The results suggest that we can find failure modes with far fewer samples than the Monte Carlo approach and simultaneously estimate the probability of failure.

Validation↗

NASA NTRS2013

Community Supported Stardust Compendia

The Stardust cometary and interstellar collections present unprecedented challenges in sample preparation and analysis. The ensemble of approx.80 tracks and dozens of foil craters from the cometary collection for which we have analyses exhibits a bewildering complexity and diversity of materials. The interstellar collection is even more challenging, because of the extremely low fluence of interstellar dust, a relatively large background of secondary ejecta from impacts on the spacecraft, and the small size of interstellar dust, approximately three orders of magnitude smaller in mass than typical cometary particles. Unlike with the other returned sample collections, characterization of these samples beyond basic photo-documentation is not generally practical at JSC. Even among the other small-particle collections, currently the cosmic dust and Hayabusa samples, SEM/EDX can provide basic chemistry. This is not possible with Stardust particles without destructive and invasive sample preparation. Furthermore, SEM/EDX requires isolating small grains from adhering aerogel. A reliable technique to carry out this task does not exist. Complete characterization of particles requires coordinated analyses using synchrotron and electron-beam microprobes, which do not exist at any one lab. Thus, it was recognized since the Stardust Preliminary Examination in 2006 that characterization of the samples would rely on the worldwide community of Stardust Investigators. Here we announce the development of community-editable, wiki-style Stardust compendia that will support this effort. Our intention is that this will facilitate sample requests by providing basic characterization of tracks. We expect that this will also support comprehensive meta-analyses (global syntheses of analyses) of the collections.

Westphal, Andrew J.↗

NASA NTRS1988

Fine-scale traverses in cumulate rocks, Stillwater Complex: A lunar analogue study

The objective was to document finite-scale compositional variations in cumulate rocks from the Stillwater Complex in Montana and to interpret these data in the context of planetary magma fractionation processes such as those operative during the formation of the Earth's Moon. This research problem involved collecting samples in the Stillwater Complex and analyzing them by electron microprobe, X-ray fluorescence (XRF), and instrumental neutron activation analysis (INAA). The electron microprobe is used to determine the compositions of cumulus and intercumulus phases in the rocks, the XRF is used to determine the bulk-rock major element and trace element (Y, Sr, Rb, Zr, Ni, and Cr) abundances, and the INAA lab. is used to determine the trace element (Sc, Co, Cr, Ni, Ta, Hf, U, Th, and the REE) abundances of mineral separates and bulk rocks.

Elthon, Donald↗

NASA NTRS1983

Lunar sample analysis

The evolution of the lunar regolith under solar wind and micrometeorite bombardment is discussed as well as the size distribution of ultrafine iron in lunar soil. The most important characteristics of complex graphite, sulfide, arsenide, palladium, and platinum mineralization in a pegmatoid pyroxenite of the Stillwater Complex in Montana are examined. Oblique reflected light micrographs and backscattered electron SEM images of the graphite associations are included.

Housley, R. M.↗

NASA NTRSDate not supplied

Uncertainty Estimates for Fitting Zernike Polynomials to Discrete Data

Zernike polynomials are a widely used metric in modern optical analysis. They conveniently represent surfaces as a series of weighted terms corresponding to various optical aberrations. Ideally, each term is independent of others in the series, but Zernike polynomials lose this property when working with sets of discrete data. This gives rise to uncertainty in each polynomial’s actual contribution and affects metrology and simulation estimates of their relative weights. Several factors influencing these estimates are the number and arrangement of sample locations, the method for calculating the weights, and the total number of Zernike terms used in the calculation. Discussed is the uncertainty associated with linear regression using random sampling. Other topics reviewed are complex Zernike polynomials and vector spaces of functions.

Zernike Polynomials↗

NASA NTRSDate not supplied

Extraterrestrial Molecular Indicators of Life Investigation (EMILI)

Future missions to Enceladus, Europa, Mars, and beyond may seek the molecular signs of extraterrestrial life through chemical analysis of acquired samples. Particularly on ocean worlds such as Enceladus and Europa, samples may contain trace ocean-borne molecular biosignatures of extant life that may or may not share similarities to those of terrestrial life. In situ analyses must be prepared to detect and characterize a wide range of possible molecular species, structures, and patterns, typically with exquisite sensitivity and within a complex, poorly-characterized planetary environment. The Extraterrestrial Molecular Indicators of Life Investigation (EMILI) is designed to meet or exceed the requirements of such missions for organic molecular analysis through a powerful combination of dual chemical separation and both optical and mass spectrometry detection techniques, realized in an integrated, compact instrument package fully compatible with anticipated flight resources and conditions. The full EMILI instrument combines two sample analysis subsystems to provide wide-ranging and complementary detection of organic compounds and inorganic salts. The Gas Analysis Processing System (GAPS) uses a chemical derivatization protocol with gas chromatography (GC) separation prior to detection in an ion trap mass spectrometer (ITMS) to enable full characterization of lower-polarity, volatile and semi-volatile molecules such as fatty acids and hydrocarbons. The Organic Capillary Electrophoresis ANalysis System (OCEANS) uses a liquid-based extraction protocol with CE separation to enable precise analysis of more water-soluble/polar compounds. OCEANS features a laser-induced fluorescence detection mode to perform ultra-sensitive quantitative analysis of chiral amino acids. In EMILI, OCEANS is additionally coupled to the same ITMS through a novel electrospray ionization interface. The common ITMS allows EMILI to identify and cross-correlate molecular species and patterns, detected through either or both protocols, of molecular weights to over 1000 u, potentially even revealing complex biosignatures such as alien oligopeptides and informational polymers.

Europa↗

NASA NTRS2023

The Complex Exhumation History of Jezero Crater Floor Unit and Its Implication for Mars Sample Return

During the first year of NASA's Mars 2020 mission, Perseverance rover has investigated the dark crater floor unit of Jezero crater and four samples of this unit have been collected. The focus of this paper is to assess the potential of these samples to calibrate the crater-based Martian chronology. We first review the previous estimation of crater-based model age of this unit. Then, we investigate the impact crater density distribution across the floor unit. It reveals that the crater density is heterogeneous from areas which have been exposed to the bombardment during the last 3 Ga to areas very recently exposed to bombardment. It suggests a complex history of exposure to impact cratering. We also display evidence of several remnants of deposits on the top of the dark floor unit across Jezero below which the dark floor unit may have been buried. We propose the following scenario of burying/exhumation: the dark floor unit would have been initially buried below a unit that was a few tens of meters thick. This unit then gradually eroded away due to Aeolian processes from the northeast to the west, resulting in uneven exposure to impact bombardment over 3 Ga. A cratering model reproducing this scenario confirms the feasibility of this hypothesis. Due to the complexity of its exposure history, the Jezero dark crater floor unit will require additional detailed analysis to understand how the Mars 2020 mission samples of the crater floor can be used to inform the Martian cratering chronology.

Mars 2020↗

NASA NTRS1972

A comparison of the capabilities of three finite element programs

Three finite element programs are compared to assess their capabilities as an analysis tool in a structural design process. Because of the need for repetitive analyses as an integral part of a design loop, a candidate program must be capable of handling large problems, operate efficiently, and be readily adaptable for use in computer-aided design. The three programs considered in the study, ELAS,SNAP, and NASTRAN, range from a relatively small finite element program limited to static structural analysis (ELAS) to a large complex general analysis system (NASTRAN). Results are given for comparative speeds and computer resources required for each program in the analysis of sample fuselage problems representative of practical aircraft design.

Loendorf, D. D.↗

NASA NTRS1988

Environment Monitor

Viking landers touched down on Mars equipped with a variety of systems to conduct automated research, each carrying a compact but highly sophisticated instrument for analyzing Martian soil and atmosphere. Instrument called a Gas Chromatography/Mass Spectrometer (GC/MS) had to be small, lightweight, shock resistant, highly automated and extremely sensitive, yet require minimal electrical power. Viking Instruments Corporation commercialized this technology and targeted their primary market as environmental monitoring, especially toxic and hazardous waste site monitoring. Waste sites often contain chemicals in complex mixtures, and the conventional method of site characterization, taking samples on-site and sending them to a laboratory for analysis is time consuming and expensive. Other terrestrial applications are explosive detection in airports, drug detection, industrial air monitoring, medical metabolic monitoring and for military, chemical warfare agents.

Source record↗