Search NASASearch

SEARCH · Search NASA

Results for “Factorization machine”

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 109 records · Page 6

Army-NASA aircrew/aircraft integration program: Phase 4 A(3)I Man-Machine Integration Design and Analysis System (MIDAS) software detailed design document

The Man-Machine Integration Design and Analysis System (MIDAS) is an integrated suite of software components that constitutes a prototype workstation to aid designers in applying human factors principles to the design of complex human-machine systems. MIDAS is intended to be used at the very early stages of conceptual design to provide an environment wherein designers can use computational representations of the crew station and operator, instead of hardware simulators and man-in-the-loop studies, to discover problems and ask 'what if' questions regarding the projected mission, equipment, and environment. This document is the Software Product Specification for MIDAS. Introductory descriptions of the processing requirements, hardware/software environment, structure, I/O, and control are given in the main body of the document for the overall MIDAS system, with detailed discussion of the individual modules included in Annexes A-J.

Banda, Carolyn

Machine Learning Approaches to Increasing Value of Spaceflight Omics Databases

The number of spaceflight bioscience mission opportunities is too small to allow all relevant biological and environmental parameters to be experimentally identified. Simulated spaceflight experiments in ground-based facilities (GBFs), such as clinostats, are each suitable only for particular investigations -- a rotating-wall vessel may be 'simulated microgravity' for cell differentiation (hours), but not DNA repair (seconds) -- and introduce confounding stimuli, such as motor vibration and fluid shear effects. This uncertainty over which biological mechanisms respond to a given form of simulated space radiation or gravity, as well as its side effects, limits our ability to baseline spaceflight data and validate mission science. Machine learning techniques autonomously identify relevant and interdependent factors in a data set given the set of desired metrics to be evaluated: to automatically identify related studies, compare data from related studies, or determine linkages between types of data in the same study. System-of-systems (SoS) machine learning models have the ability to deal with both sparse and heterogeneous data, such as that provided by the small and diverse number of space biosciences flight missions; however, they require appropriate user-defined metrics for any given data set. Although machine learning in bioinformatics is rapidly expanding, the need to combine spaceflight/GBF mission parameters with omics data is unique. This work characterizes the basic requirements for implementing the SoS approach through the System Map (SM) technique, a composite of a dynamic Bayesian network and Gaussian mixture model, in real-world repositories such as the GeneLab Data System and Life Sciences Data Archive. The three primary steps are metadata management for experimental description using open-source ontologies, defining similarity and consistency metrics, and generating testing and validation data sets. Such approaches to spaceflight and GBF omics data may soon enable unique insight into which measured phenomena correlate to biological mechanisms that are truly affected by spaceflight conditions; which are most likely to be confounded by other variables; and which are insufficiently characterized, significantly increasing existing and future science return from ISS and spaceflight missions.

Gentry, Diana

Advanced automated glass cockpit certification: Being wary of human factors

This paper presents some facets of the French experience with human factors in the process of certification of advanced automated cockpits. Three types of difficulties are described: first, the difficulties concerning the hotly debated concept of human error and its non-linear relationship to risk of accident; a typology of errors to be taken into account in the certification process is put forward to respond to this issue. Next, the difficulties connected to the basically gradual and evolving nature of pilot expertise on a given type of aircraft, which contrasts with the immediate and definitive style of certifying systems. The last difficulties to be considered are those related to the goals of certification itself on these new aircraft and the status of findings from human factor analyses (in particular, what should be done with disappointing results, how much can the changes induced by human factors investigation economically affect aircraft design, how many errors do we need to accumulate before we revise the system, what should be remedied when human factor problems are discovered at the certification stage: the machine? pilot training? the rules? or everything?). The growth of advanced-automated glass cockpits has forced the international aeronautical community to pay more attention to human factors during the design phase, the certification phase and pilot training. The recent creation of a human factor desk at the DGAC-SFACT (Official French services) is a direct consequence of this. The paper is divided into three parts. Part one debates human error and its relationship with system design and accident risk. Part two describes difficulties connected to the basically gradual and evolving nature of pilot expertise on a given type of aircraft, which contrasts with the immediate and definitive style of certifying systems. Part three focuses on concrete outcomes of human factors for certification purposes.

Amalberti, Rene

A Recursive Multi-step Machine Learning Approach for Airport Configuration Prediction

Airport configuration selection is a complex decision-making process that involves several operational and human factors. In this paper we propose a novel recursive multi-step machine learning (ML) approach to predict airport configuration. The multi-step approach guarantees stability of the predicted configuration by taking as input the configuration predicted at the previous time step. The features of the proposed model include weather data, future arrival and departure counts and current configuration. Due to the importance of arrival and departure counts in predicting the airport configuration, arrival counts are calculated using landing time predictions selected from physics-based landing time predictions available in FAA System Wide Information Management data feeds for each flight. The selection rules were developed and refined to select the most accurate time for different phases of flight. The proposed model predicts the airport configurations up to 6 hours ahead. In this paper we show the predictive performance of the proposed model for six major US airports, including Charlotte Douglas International Airport (CLT), Dallas/Fort Worth International Airport (DFW), John F. Kennedy International Airport (JFK), Newark Liberty International Airport (EWR), LaGuardia Airport (LGA) and Dallas Love Field Airport (DAL). We trained and evaluated models on 2019 and 2020 data in order to study the effect of the pandemic and how changes in traffic patterns affected the performance of the proposed model. Results are compared with a baseline assuming no airport configuration changes. In our results for DFW, we obtained a prediction accuracy of 89.3% for 3 hours ahead prediction, and 82.8% for 6 hours ahead when applied on 2019 data.

machine learning

A Recursive Multi-step Machine Learning Approach for Airport Configuration Prediction

Airport configuration selection is a complex decision-making process that involves several operational and human factors. In this paper we propose a novel recursive multi-step machine learning (ML) approach to predict airport configuration. The multi-step approach guarantees stability of the predicted configuration by taking as input the configuration predicted at the previous time step. The features of the proposed model include weather data, future arrival and departure counts and current configuration. Due to the importance of arrival and departure counts in predicting the airport configuration, arrival counts are calculated using landing time predictions selected from physics-based landing time predictions available in FAA System Wide Information Management data feeds for each flight. The selection rules were developed and refined to select the most accurate time for different phases of flight. The proposed model predicts the airport configurations up to 6 hours ahead. In this paper we show the predictive performance of the proposed model for six major US airports, including Charlotte Douglas International Airport (CLT), Dallas/Fort Worth International Airport (DFW), John F. Kennedy International Airport (JFK), Newark Liberty International Airport (EWR), LaGuardia Airport (LGA) and Dallas Love Field Airport (DAL). We trained and evaluated models on 2019 and 2020 data in order to study the effect of the pandemic and how changes in traffic patterns affected the performance of the proposed model. Results are compared with a baseline assuming no airport configuration changes. In our results for DFW, we obtained a prediction accuracy of 89.3% for 3 hours ahead prediction, and 82.8% for 6 hours ahead when applied on 2019 data.

machine learning

Human factors issues in the use of night vision devices

An account is given of the critical human factors that arise in field data on the differences between night vision displays and unaided day vision. Attention is given to the findings of empirical studies of performance on rotorcraft-flight-relevant perceptual tasks in which depth and distance perception are critical factors. Suggestions are made for man-machine-critical component design modifications in current night vision systems.

Kaiser, Mary K.

Bulk-Flow Analysis of Hybrid Thrust Bearings for Advanced Cryogenic Turbopumps

A bulk-flow analysis and computer program for prediction of the static load performance and dynamic force coefficients of angled injection, orifice-compensated hydrostatic/hydrodynamic thrust bearings have been completed. The product of the research is an efficient computational tool for the design of high-speed thrust bearings for cryogenic fluid turbopumps. The study addresses the needs of a growing technology that requires of reliable fluid film bearings to provide the maximum operating life with optimum controllable rotordynamic characteristics at the lowest cost. The motion of a cryogenic fluid on the thin film lands of a thrust bearing is governed by a set of bulk-flow mass and momentum conservation and energy transport equations. Mass flow conservation and a simple model for momentum transport within the hydrostatic bearing recesses are also accounted for. The bulk-flow model includes flow turbulence with fluid inertia advection, Coriolis and centrifugal acceleration effects on the bearing recesses and film lands. The cryogenic fluid properties are obtained from realistic thermophysical equations of state. Turbulent bulk-flow shear parameters are based on Hirs' model with Moody's friction factor equations allowing a simple simulation for machined bearing surface roughness. A perturbation analysis leads to zeroth-order nonlinear equations governing the fluid flow for the thrust bearing operating at a static equilibrium position, and first-order linear equations describing the perturbed fluid flow for small amplitude shaft motions in the axial direction. Numerical solution to the zeroth-order flow field equations renders the bearing flow rate, thrust load, drag torque and power dissipation. Solution to the first-order equations determines the axial stiffness, damping and inertia force coefficients. The computational method uses well established algorithms and generic subprograms available from prior developments. The Fortran9O computer program hydrothrust runs on a Windows 95/NT personal computer. The program, help files and examples are licensed by Texas A&M University Technology License Office. The study of the static and dynamic performance of two hydrostatic/hydrodynamic bearings demonstrates the importance of centrifugal and advection fluid inertia effects for operation at high rotational speeds. The first example considers a conceptual hydrostatic thrust bearing for an advanced liquid hydrogen turbopump operating at 170,000 rpm. The large axial stiffness and damping coefficients of the bearing should provide accurate control and axial positioning of the turbopump and also allow for unshrouded impellers, therefore increasing the overall pump efficiency. The second bearing uses a refrigerant R134a, and its application in oil-free air conditioning compressors is of great technological importance and commercial value. The computed predictions reveal that the LH2 bearing load capacity and flow rate increase with the recess pressure (i.e. increasing orifice diameters). The bearing axial stiffness has a maximum for a recess pressure rati of approx. 0.55. while the axial damping coefficient decreases as the recess pressure ratio increases. The computer results from three flow models are compared. These models are a) inertialess, b) fluid inertia at recess edges only, and c) full fluid inertia at both recess edges and film lands. The full inertia model shows the lowest flow rates, axial load capacity and stiffness coefficient but on the other hand renders the largest damping coefficients and inertia coefficients. The most important findings are related to the reduction of the outflow through the inner radius and the appearance of subambient pressures. The performance of the refrigerant hybrid thrust bearing is evaluated at two operating speeds and pressure drops. The computed results are presented in dimensionless form to evidence consistent trends in the bearing performance characteristics. As the applied axial load increases, the bearing film thickness and flow rate decrease while the recess pressure increases. The axial stiffness coefficient shows a maximum for a certain intermediate load while the damping coefficient steadily increases. The computed results evidence the paramount of centrifugal fluid inertia at low recess pressures (i.e. low loads), and where there is actually an inflow through the bearing inner diameter, accompanied by subambient pressures just downstream of the bearing recess edge. These results are solely due to centrifugal fluid inertia and advection transport effects. Recommendations include the extension of the computer program to handle flexure pivot tilting pad hybrid bearings and the ability to calculate moment coefficients for shaft angular misalignments.

SanAndres, Luis

Arithmetic Data Cube as a Data Intensive Benchmark

Data movement across computational grids and across memory hierarchy of individual grid machines is known to be a limiting factor for application involving large data sets. In this paper we introduce the Data Cube Operator on an Arithmetic Data Set which we call Arithmetic Data Cube (ADC). We propose to use the ADC to benchmark grid capabilities to handle large distributed data sets. The ADC stresses all levels of grid memory by producing 2d views of an Arithmetic Data Set of d-tuples described by a small number of parameters. We control data intensity of the ADC by controlling the sizes of the views through choice of the tuple parameters.

Frumkin, Michael A.

Optimization of Neutral Atom Imagers

The interactions between plasma structures and neutral atom populations in interplanetary space can be effectively studied with energetic neutral atom imagers. For neutral atoms with energies less than 1 keV, the most efficient detection method that preserves direction and energy information is conversion to negative ions on surfaces. We have examined a variety of surface materials and conversion geometries in order to identify the factors that determine conversion efficiency. For chemically and physically stable surfaces smoothness is of primary importance while properties such as work function have no obvious correlation to conversion efficiency. For the noble metals, tungsten, silicon, and graphite with comparable smoothness, conversion efficiency varies by a factor of two to three. We have also examined the way in which surface conversion efficiency varies with the angle of incidence of the neutral atom and have found that the highest efficiencies are obtained at angles of incidence greater then 80deg. The conversion efficiency of silicon, tungsten and graphite were examined most closely and the energy dependent variation of conversion efficiency measured over a range of incident angles. We have also developed methods for micromachining silicon in order to reduce the volume to surface area over that of a single flat surface and have been able to reduce volume to surface area ratios by up to a factor of 60. With smooth micro-machined surfaces of the optimum geometry, conversion efficiencies can be increased by an order of magnitude over instruments like LENA on the IMAGE spacecraft without increase the instruments mass or volume.

Shappirio, M.

Investigation of stress concentration factor - Stress intensity factor interaction for flaws in filleted rods

To predict the fatigue life of machine parts which contain flaws in regions of stress concentration, the engineer needs to treat a series of crack sizes and/or shapes quickly and economically. In this paper the stress intensity factor distributions for circular surface flaws in the roots of fillets in rods loaded by remote tension are presented and examined. The examination reveals insight into the interaction of cracks with other stress concentrating discontinuities, and based on this it is suggested that weight function methods used in conjunction with known crack solutions may provide needed, efficient stress intensity factor prediction methods for such three-dimensional geometries.

Springfield, C. W., Jr.

Concurrent Cholesky factorization of positive definite banded Hermitian matrices

First, the Cholesky factorization is extended to cover uniformly partitioned banded positive definite matrices of rank n which may be real symmetric or Hermitian. Then, two stratagems are given for the use of the algorithm in concurrent machines where the number of processing elements is less than required to factor the matrix in as few serial steps as possible, and where uniformly high efficiency is expected from all processing elements. Expressions are given for the efficiency factor e appearing in the speed-up expression q = eN, and these are specialized for the N node hypercube machine as a function of partition size s, the number N of processing elements of the hypercube machine, and the cost mu of interelement transmission relative to computation. It is shown that the efficiency factor e is inversely proportional to mu/s, and that e is almost independent of N when N is large and mu/s = 0. The task is completed in n/s serial steps with no limit on n. The half bandwidth b of the matrix is 2 Ns.

Utku, S.

Automatic Speech Recognition in Air Traffic Control: a Human Factors Perspective

The introduction of Automatic Speech Recognition (ASR) technology into the Air Traffic Control (ATC) system has the potential to improve overall safety and efficiency. However, because ASR technology is inherently a part of the man-machine interface between the user and the system, the human factors issues involved must be addressed. Here, some of the human factors problems are identified and related methods of investigation are presented. Research at M.I.T.'s Flight Transportation Laboratory is being conducted from a human factors perspective, focusing on intelligent parser design, presentation of feedback, error correction strategy design, and optimal choice of input modalities.

Karlsson, Joakim

Micro-structural features and material properties impact on adhesive metal joints via computational modeling and machine learning

The quality of structural bonding in practical applications depends on various factors arising from materials, pre-processing conditions, and manufacturing. Understanding how these factors influence bonding performance and determining their relative importance are of significant interest. Thus, this study evaluates the effects of microstructural features and material properties on the structural strength of adhesively-bonded metal joints at the submillimeter scale, utilizing a combination of Finite Element Modeling (FEM) and Machine Learning (ML) with Gradient Boosting Regression (GBR). The microstructural features include adhesive thickness, internal voids within the adhesive, adherend-adhesive interfacial voids, void size and volume fraction, and surface roughness. The material properties include the constitutive behavior of the adhesive, as well as the adherend-adhesive interfacial strength and fracture energy. The changes in structural strength and morphologies of the bonded metal structures with respect to different microstructural features and material properties were clarified by FEM. By further leveraging ML-GBR, the sequence of importance of these factors affecting bonding performance across various scenarios was summarized. This work provides valuable insights into the development of improved structural bonding for adhesive joints in industries such as automotive , aerospace, and beyond.

36 MATERIALS SCIENCE

Human factors in the presentation of computer-generated information - Aspects of design and application in automated flight traffic

The man-machine interface and its influence on the characteristics of computer displays in automated air traffic is discussed. The graphical presentation of spatial relationships and the problems it poses for air traffic control, and the solution of such problems are addressed. Psychological factors involved in the man-machine interface are stressed.

Roske-Hofstrand, Renate J.

Highly parallel sparse Cholesky factorization

Several fine grained parallel algorithms were developed and compared to compute the Cholesky factorization of a sparse matrix. The experimental implementations are on the Connection Machine, a distributed memory SIMD machine whose programming model conceptually supplies one processor per data element. In contrast to special purpose algorithms in which the matrix structure conforms to the connection structure of the machine, the focus is on matrices with arbitrary sparsity structure. The most promising algorithm is one whose inner loop performs several dense factorizations simultaneously on a 2-D grid of processors. Virtually any massively parallel dense factorization algorithm can be used as the key subroutine. The sparse code attains execution rates comparable to those of the dense subroutine. Although at present architectural limitations prevent the dense factorization from realizing its potential efficiency, it is concluded that a regular data parallel architecture can be used efficiently to solve arbitrarily structured sparse problems. A performance model is also presented and it is used to analyze the algorithms.

Gilbert, John R.

Explaining word embeddings with perfect fidelity: a case study in predicting research impact

The best-performing approaches for scholarly document quality prediction are based on embedding models. In addition to their performance when used in classifiers, embedding models can also provide predictions even for words that were not contained in the labelled training data for the classification model, which is important in the context of the ever-evolving research terminology. Although model-agnostic explanation methods, such as Local interpretable model-agnostic explanations, can be applied to explain machine learning classifiers trained on embedding models, these produce results with questionable correspondence to the model. We introduce a new feature importance method, Self-Model Entities Rated (SMER), for logistic regression-based classification models trained on word embeddings. We show that SMER has theoretically perfect fidelity with the explained model, as the average of logits of SMER scores for individual words (SMER explanation) exactly corresponds to the logit of the prediction of the explained model. Quantitative and qualitative evaluation is performed through five diverse experiments conducted on 50,000 research articles (papers) from the CORD-19 corpus. In conclusion, through an AOPC curve analysis, we experimentally demonstrate that SMER produces better explanations than LIME, SHAP and global tree surrogates.

Coarse-grained models

Using Filter Methods to Guide Convergence for ADMM, with Applications to Nonnegative Matrix Factorization Problems

Nonconvex, nonlinear optimization problems arise naturally in parameter fitting and machine learning. While augmented Lagrangian methods have demonstrated robust convergence for classes of these problems, their convergence for block updates has been relatively unexplored outside of the context of the alternating direction method of multipliers (ADMM). ADMM has seen extensive use in these applications, but may exhibit uncertain convergence behavior in many practical nonconvex settings, and struggles with general nonlinear constraints. In contrast, filter methods have proved effective in enforcing convergence for sequential quadratic programming methods and interior point methods with feasibility criteria. We develop an ADMM-filter method for highly nonlinear and nonconvex problems. Here, we show convergence under mild assumptions for several types of coordinate descent schemes, and demonstrate our algorithm on nonnegative matrix factorization and completion problems in imaging and chemical spectrum analysis.

Nonconvex optimization