SEARCH · Search NASA

Results for “Generalization Bounds”

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.

SourceResearch connectionFrom yearThrough year

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

At least 109 records · Page 6

All sources NASA NTRS DOE OSTI

NASA NTRS1989

Global models: Robot sensing, control, and sensory-motor skills

Robotics research has begun to address the modeling and implementation of a wide variety of unstructured tasks. Examples include automated navigation, platform servicing, custom fabrication and repair, deployment and recovery, and science exploration. Such tasks are poorly described at onset; the workspace layout is partially unfamiliar, and the task control sequence is only qualitatively characterized. The robot must model the workspace, plan detailed physical actions from qualitative goals, and adapt its instantaneous control regimes to unpredicted events. Developing robust representations and computational approaches for these sensing, planning, and control functions is a major challenge. The underlying domain constraints are very general, and seem to offer little guidance for well-bounded approximation of object shape and motion, manipulation postures and trajectories, and the like. This generalized modeling problem is discussed, with an emphasis on the role of sensing. It is also discussed that unstructured tasks often have, in fact, a high degree of underlying physical symmetry, and such implicit knowledge should be drawn on to model task performance strategies in a methodological fashion. A group-theoretic decomposition of the workspace organization, task goals, and their admissible interactions are proposed. This group-mechanical approach to task representation helps to clarify the functional interplay of perception and control, in essence, describing what perception is specifically for, versus how it is generically modeled. One also gains insight how perception might logically evolve in response to needs of more complex motor skills. It is discussed why, of the many solutions that are often mathematically admissible to a given sensory motor-coordination problem, one may be preferred over others.

Schenker, Paul S.↗

NASA NTRS1993

Convex set and linear mixing model

A major goal of optical remote sensing is to determine surface compositions of the earth and other planetary objects. For assessment of composition, single pixels in multi-spectral images usually record a mixture of the signals from various materials within the corresponding surface area. In this report, we introduce a closed and bounded convex set as a mathematical model for linear mixing. This model has a clear geometric implication because the closed and bounded convex set is a natural generalization of a triangle in n-space. The endmembers are extreme points of the convex set. Every point in the convex closure of the endmembers is a linear mixture of those endmembers, which is exactly how linear mixing is defined. With this model, some general criteria for selecting endmembers could be described. This model can lead to a better understanding of linear mixing models.

Xu, P.↗

NASA NTRS1998

NAESA Augmentation Pilot Project

This project was one project within the Native American Earth and Space Academy (NAESA). NAESA is a national initiative comprised of several organizations that support programs which focus on 1) enhancing the technological, scientific and pedagogical skills of K-14 teachers who instruct Native Americans, 2) enhancing the understanding and applications of science, technology, and engineering of college-bound Native Americans and teaching them general college "survival skills" (e.g., test taking, time management, study habits), 3) enhancing the scientific and pedagogical skills of the faculty of tribally-controllcd colleges and community colleges with large Native American enrollments, and 4) strengthening the critical relationships between students, their parents, tribal elders, and their communities. This Augmentation Pilot Project focused on the areas of community-school alliances and intemet technology use in teaching and learning and daily living addressing five major objectives.

Hoover, John J.↗

NASA NTRS2001

Complex Correlation Calculation of e(-) - H Total Cross Sections

Calculation of e(-) - H total and elastic partial wave cross sections is being carried out using the complex correlation variational T-matrix method. In this preliminary study, elastic partial wave phase shifts are calculated with the correlation functions which are confined to be real. In that case the method reduces to the conventional optical potential approach with 2 projection operators. The number of terms in the Hylleraas-type wave function for the S-1 phase shifts is 95 while for the S-3 it is 56, except for k = 0.8 where it is 84. Our results, which are rigorous lower bounds, are seen to be in general agreement with those of Schwartz, but they are of greater accuracy and outside of his error limits for k = 0.3 and 0.4 for S-1. The main aim of this approach is the application to higher energy scattering. By virtue of the complex correlation functions, the T-matrix is not unitary so that elastic and total scattering cross sections are independent of each other. Our results will be compared specifically with those of Bray and Stelbovics.

Bhatia, A. K.↗

NASA NTRS2003

Bayes Error Rate Estimation Using Classifier Ensembles

The Bayes error rate gives a statistical lower bound on the error achievable for a given classification problem and the associated choice of features. By reliably estimating th is rate, one can assess the usefulness of the feature set that is being used for classification. Moreover, by comparing the accuracy achieved by a given classifier with the Bayes rate, one can quantify how effective that classifier is. Classical approaches for estimating or finding bounds for the Bayes error, in general, yield rather weak results for small sample sizes; unless the problem has some simple characteristics, such as Gaussian class-conditional likelihoods. This article shows how the outputs of a classifier ensemble can be used to provide reliable and easily obtainable estimates of the Bayes error with negligible extra computation. Three methods of varying sophistication are described. First, we present a framework that estimates the Bayes error when multiple classifiers, each providing an estimate of the a posteriori class probabilities, a recombined through averaging. Second, we bolster this approach by adding an information theoretic measure of output correlation to the estimate. Finally, we discuss a more general method that just looks at the class labels indicated by ensem ble members and provides error estimates based on the disagreements among classifiers. The methods are illustrated for artificial data, a difficult four-class problem involving underwater acoustic data, and two problems from the Problem benchmarks. For data sets with known Bayes error, the combiner-based methods introduced in this article outperform existing methods. The estimates obtained by the proposed methods also seem quite reliable for the real-life data sets for which the true Bayes rates are unknown.

Tumer, Kagan↗

NASA NTRS2013

Tuning and Robustness Analysis for the Orion Absolute Navigation System

The Orion Multi-Purpose Crew Vehicle (MPCV) is currently under development as NASA's next-generation spacecraft for exploration missions beyond Low Earth Orbit. The MPCV is set to perform an orbital test ight, termed Exploration Flight Test 1 (EFT-1), some time in late 2014. The navigation system for the Orion spacecraft is being designed in a Multi-Organizational Design Environment (MODE) team including contractor and NASA personnel. The system uses an Extended Kalman Filter to process measurements and determine the state. The design of the navigation system has undergone several iterations and modi cations since its inception, and continues as a work-in-progress. This paper seeks to show the e orts made to-date in tuning the lter for the EFT-1 mission and instilling appropriate robustness into the system to meet the requirements of manned space ight. The results generally show Monte Carlo error performance bounded by the lter uncertainty for all phases of ight. Some future items of investigation are presented related to suspected anomalies in the trajectory truth reference le.

Holt, Greg N.↗

NASA NTRS2016

Binary Black Hole Mergers in the First Advanced LIGO Observing Run

The first observational run of the Advanced LIGO detectors, from September 12, 2015 to January 19, 2016, saw the first detections of gravitational waves from binary black hole mergers. In this paper we present full results from a search for binary black hole merger signals with total masses up to 100M solar mass and detailed implications from our observations of these systems. Our search, based on general-relativistic models of gravitational wave signals from binary black hole systems, unambiguously identified two signals, GW150914 and GW151226, with a significance of greater than 5 alpha over the observing period. It also identified a third possible signal, LVT151012, with substantially lower significance, which has a 87 probability of being of astrophysical origin. We provide detailed estimates of the parameters of the observed systems. Both GW150914 and GW151226 provide an unprecedented opportunity to study the two-body motion of a compact-object binary in the large velocity, highly nonlinear regime. We do not observe any deviations from general relativity, and place improved empirical bounds on several high-order post-Newtonian coefficients. From our observations we infer stellar-mass binary black hole merger rates lying in the range 9-240 Gpc-3 yr-1. These observations are beginning to inform astrophysical predictions of binary black hole formation rates, and indicate that future observing runs of the Advanced detector network will yield many more gravitational wave detections.

Abbott, B. P.↗

NASA NTRS1994

Empirical study of parallel LRU simulation algorithms

This paper reports on the performance of five parallel algorithms for simulating a fully associative cache operating under the LRU (Least-Recently-Used) replacement policy. Three of the algorithms are SIMD, and are implemented on the MasPar MP-2 architecture. Two other algorithms are parallelizations of an efficient serial algorithm on the Intel Paragon. One SIMD algorithm is quite simple, but its cost is linear in the cache size. The two other SIMD algorithm are more complex, but have costs that are independent on the cache size. Both the second and third SIMD algorithms compute all stack distances; the second SIMD algorithm is completely general, whereas the third SIMD algorithm presumes and takes advantage of bounds on the range of reference tags. Both MIMD algorithm implemented on the Paragon are general and compute all stack distances; they differ in one step that may affect their respective scalability. We assess the strengths and weaknesses of these algorithms as a function of problem size and characteristics, and compare their performance on traces derived from execution of three SPEC benchmark programs.

Carr, Eric↗

NASA NTRS2014

Towards Co-evolution of Membrane Transport and Metabolism

Protocellular boundaries were inextricably connected to the metabolism they encapsulated: to be inheritable, early metabolism must have led to an increased rate of growth and division of vesicles and, similarly, transport through vesicle boundaries must have supported the evolution of metabolism. Even though explaining how this coupling emerged and evolved in the absence of the complex machinery of modern cells is one of the key issues in studies on the origin of life, little is known about the biochemical and biophysical processes that might have been involved. This gap in our knowledge is a major impediment in efforts to construct scenarios for the origin of life and laboratory models of protocells. A combination of experimental and computational studies carried out by us and our collaborators is aimed at helping to close this gap. Properties of membranes might have contributed to the selection of RNA as an early biopolymer. A kinetic mechanism was proposed (Sacerdote & Szostak, 2005) in which ribose was supplied more quickly than other aldopentoses to primordial cells for preferential incorporation of ribonucleotides into nucleic acids. This proposal is based on a finding that ribose permeates membranes an order of magnitude faster than its diastereomers, arabinose and xylose. Our computer simulations, which yield permeation rates in excellent agreement with experiment, and kinetic modeling explain this phenomenon in terms of inter- and intramolecular interactions involving exocyclic hydroxyl groups attached to carbon atoms of the pyranose ring (Wei and Pohorille, 2009). They also constrain scenarios for the formation of the earliest nucleic acids (Wei and Pohorille, 2013). In one scenario, sugars permeate protocellular walls and subsequently are used to synthesize nucleic acids inside protocells. As long as this process proceeds at the rate faster than 6x10(exp -3)/s, ribose derivatives will be available for synthesis easier than their diastereomers. If nucleosides or their activated derivatives are synthesized outside protocells and subsequently transported across protocellular membranes the kinetic mechanism does not apply because all diastereomers, which have their sugars in the furanose rather than pyranose form, permeate the membrane at approximately the same rate. Properties of membranes might have been also coupled to metabolism involving peptides. Recently, Adamala and Szostak (2013) have shown that a dipeptide inside fatty-acid vesicles catalyzes the formation of another dipeptide that binds to vesicle walls and, by doing so, promotes their growth at the expense of other vesicles. This coupling of metabolism, permeability of vesicles and their growth is the first demonstration of evolutionary advantage imparted by small, membrane-bound peptides. Building on this work we have calculated the rate at which different blocked amino acids are delivered to a protocell for synthesis of dipeptides. We have further shown that the dipeptides are located at the water-membrane interface rather than in the center of the bilayer. On these basis it is anticipated that other dipeptides containing aromatic, but not necessarily hydrophobic amino acids (e.g. tyrosine) could have the same catalytic effects. Insight from these studies allows for estimating the rate of vesicle growth and the rates of dipeptide synthesis required to keep the system in balance. These results, in combination with our earlier studies, lead to a general scenario for evolution from membrane-bound dipeptides to ion channels in the origin of life.

Metabolism↗

NASA NTRS1987

Constructions for finite-state codes

A class of codes called finite-state (FS) codes is defined and investigated. These codes, which generalize both block and convolutional codes, are defined by their encoders, which are finite-state machines with parallel inputs and outputs. A family of upper bounds on the free distance of a given FS code is derived from known upper bounds on the minimum distance of block codes. A general construction for FS codes is then given, based on the idea of partitioning a given linear block into cosets of one of its subcodes, and it is shown that in many cases the FS codes constructed in this way have a d sub free which is as large as possible. These codes are found without the need for lengthy computer searches, and have potential applications for future deep-space coding systems. The issue of catastropic error propagation (CEP) for FS codes is also investigated.

Pollara, F.↗

NASA NTRS1979

A multiloop generalization of the circle stability criterion

A frequency-domain stability criterion is presented, generalizing the well-known circle stability criterion to multiloop feedback systems having bounded nonlinearity, parameter variations, and/or frequency-dependent ignorance of component dynamics. Unlike previous generalizations, the theory is not restricted to weakly-coupled, diagonally dominant or nearly normal systems. Potential applications include the analysis of feedback system integrity and multiloop feedback system stability margins.

Safonov, M. G.↗

NASA NTRS1992

Information models of software productivity - Limits on productivity growth

Research into generalized information-metric models of software process productivity establishes quantifiable behavior and theoretical bounds. The models establish a fundamental mathematical relationship between software productivity and the human capacity for information traffic, the software product yield (system size), information efficiency, and tool and process efficiencies. An upper bound is derived that quantifies average software productivity and the maximum rate at which it may grow. This bound reveals that ultimately, when tools, methodologies, and automated assistants have reached their maximum effective state, further improvement in productivity can only be achieved through increasing software reuse. The reuse advantage is shown not to increase faster than logarithmically in the number of reusable features available. The reuse bound is further shown to be somewhat dependent on the reuse policy: a general 'reuse everything' policy can lead to a somewhat slower productivity growth than a specialized reuse policy.

Tausworthe, Robert C.↗

NASA NTRS1973

Quadrature solution for the general relativistic motion of a satellite or a planet.

The general relativistic motion of a particle in a bounded orbit around a central mass M is derived with the aid of the Schwarzschild field of M. A quadrature gives the central angle as a quasi-periodic function of an effective true anomaly. The mean rate of advance of the pericenter is investigated. Classical and general relativistic orbiters with the same initial conditions are compared, taking into account the treatment of the general relativistic effect by means of a differential correction. General relativistic variations are discussed, giving attention to the variation of the eccentric anomaly and the variation of the central angle.

Vinti, J. P.↗

NASA NTRS1974

Upper bounds on sequential decoding performance parameters

This paper presents the best obtainable random coding and expurgated upper bounds on the probabilities of undetectable error, of t-order failure (advance to depth t into an incorrect subset), and of likelihood rise in the incorrect subset, applicable to sequential decoding when the metric bias G is arbitrary. Upper bounds on the Pareto exponent are also presented. The G-values optimizing each of the parameters of interest are determined, and are shown to lie in intervals that in general have nonzero widths. The G-optimal expurgated bound on undetectable error is shown to agree with that for maximum likelihood decoding of convolutional codes, and that on failure agrees with the block code expurgated bound. Included are curves evaluating the bounds for interesting choices of G and SNR for a binary-input quantized-output Gaussian additive noise channel.

Jelinek, F.↗

NASA NTRS1976

A new formulation for anisotropic radiative transfer problems. I - Solution with a variational technique

The equations of radiative transfer in anisotropically scattering media are reformulated as linear operator equations in a single independent variable. The resulting equations are suitable for solution by a variety of standard mathematical techniques. The operators appearing in the resulting equations are in general nonsymmetric; however, it is shown that every bounded linear operator equation can be embedded in a symmetric linear operator equation and a variational solution can be obtained in a straightforward way. For purposes of demonstration, a Rayleigh-Ritz variational method is applied to three problems involving simple phase functions. It is to be noted that the variational technique demonstrated is of general applicability and permits simple solutions for a wide range of otherwise difficult mathematical problems in physics.

Cheyney, H., III↗

NASA NTRS1972

A critical examination of the validity of simplified models for radiant heat transfer analysis.

Examination of the directional effects of the simplified models by comparing the experimental data with the predictions based on simple and more detailed models for the radiation characteristics of surfaces. Analytical results indicate that the constant property diffuse and specular models do not yield the upper and lower bounds on local radiant heat flux. In general, the constant property specular analysis yields higher values of irradiation than the constant property diffuse analysis. A diffuse surface in the enclosure appears to destroy the effect of specularity of the other surfaces. Semigray and gray analyses predict the irradiation reasonably well provided that the directional properties and the specularity of the surfaces are taken into account. The uniform and nonuniform radiosity diffuse models are in satisfactory agreement with each other.

Toor, J. S.↗

NASA NTRS1976

On optimal soft-decision demodulation

A necessary condition is derived for optimal J-ary coherent demodulation of M-ary (M greater than 2) signals. Optimality is defined as maximality of the symmetric cutoff rate of the resulting discrete memoryless channel. Using a counterexample, it is shown that the condition derived is generally not sufficient for optimality. This condition is employed as the basis for an iterative optimization method to find the optimal demodulator decision regions from an initial 'good guess'. In general, these regions are found to be bounded by hyperplanes in likelihood space; the corresponding regions in signal space are found to have hyperplane asymptotes for the important case of additive white Gaussian noise. Some examples are presented, showing that the regions in signal space bounded by these asymptotic hyperplanes define demodulator decision regions that are virtually optimal.

Lee, L.-N.↗

NASA NTRS1977

On the error probability of general trellis codes with applications to sequential decoding

An upper bound on the average probability of error for maximum-likelihood decoding of the ensemble of random L-branch binary trellis codes of rate R = 1/n with distinction between memory length and tail length is given. It is shown that the bound is independent of the length L of the information sequence if the memory length exceeds the tail length by a specified amount that depends on L. Sequential decoding simulations using the stack algorithm were conducted to test the dependence of the undetected error probability on tail length and memory length, and the results corroborated the theory.

Johannesson, R.↗