Search NASA⌕ Search

SEARCH · Search NASA

Results for “knowledge representation”

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 847 records · Page 47

Identifying Meteorological Influences on Marine Low Cloud Mesoscale Morphology Using Satellite Classifications

Marine low cloud mesoscale morphology in the southeastern Pacific Ocean is analyzed using a large dataset of machine-learning generated classifications spanning three years. Meteorological variables and cloud properties are composited 10by mesoscale cloud type, showing distinct meteorological regimes of marine low cloud organization from the tropics to the midlatitudes. The presentation of mesoscale cellular convection, with respect to geographic distribution, boundary layer structure, and large-scale environmental conditions, agrees with prior knowledge. Two tropical and subtropical cumuliform boundary layer regimes, suppressed cumulus and clustered cumulus, are studied in detail. The patterns in precipitation, circulation, column water vapor, and cloudiness are consistent with the representation of marine shallow mesoscale convective 15 self-aggregation by large eddy simulations of the boundary layer. Although they occur under similar large-scale conditions, the suppressed and clustered low cloud types are found to be well-separated by variables associated with low-level mesoscale circulation, with surface wind divergence being the clearest discriminator between them, whether reanalysis or satellite observations are used. Clustered regimes are associated with surface convergence and suppressed regimes are associated with surface divergence.

Johannes Mohrmann↗

Ocean Carbon From Space: Current Status and Priorities for the Next Decade

The ocean plays a central role in modulating the Earth’s carbon cycle. Monitoring how the ocean carbon cycle is changing is fundamental to managing climate change. Satellite remote sensing is currently our best tool for viewing the ocean surface globally and systematically, at high spatial and temporal resolutions, and the past few decades have seen an exponential growth in studies utilising satellite data for ocean carbon research. Satellite-based observations must be combined with in-situ observations and models, to obtain a comprehensive view of ocean carbon pools and fluxes. To help prioritise future research in this area, a workshop was organised that assembled leading experts working on the topic, from around the world, including remote-sensing scientists, field scientists and modellers, with the goal to articulate a collective view of the current status of ocean carbon research, identify gaps in knowledge, and formulate a scientific roadmap for the next decade, with an emphasis on evaluating where satellite remote sensing may contribute. A total of 449 scientists and stakeholders participated (with balanced gender representation), from North and South America, Europe, Asia, Africa, and Oceania. Sessions targeted both inorganic and organic pools of carbon in the ocean, in both dissolved and particulate form, as well as major fluxes of carbon between reservoirs (e.g., primary production) and at interfaces (e.g., air-sea and land–ocean). Extreme events, blue carbon and carbon budgeting were also key topics discussed. Emerging priorities identified include: expanding the networks and quality of in-situ observations; improved satellite retrievals; improved uncertainty quantification; improved understanding of vertical distributions; integration with models; improved techniques to bridge spatial and temporal scales of the different data sources; and improved fundamental understanding of the ocean carbon cycle, and of the interactions among pools of carbon and light. We also report on priorities for the specific pools and fluxes studied, and highlight issues and concerns that arose during discussions, such as the need to consider the environmental impact of satellites or space activities; the role satellites can play in monitoring ocean carbon dioxide removal approaches; economic valuation of the satellite based information; to consider how satellites can contribute to monitoring cycles of other important climatically-relevant compounds and elements; to promote diversity and inclusivity in ocean carbon research; to bring together communities working on different aspects of planetary carbon; maximising use of international bodies; to follow an open science approach; to explore new and innovative ways to remotely monitor ocean carbon; and to harness quantum computing. Overall, this paper provides a comprehensive scientific roadmap for the next decade on how satellite remote sensing could help monitor the ocean carbon cycle, and its links to the other domains, such as terrestrial and atmosphere.

oceanography↗

Scalable Knowledge Graph Analytics at 136 Petaflop/s

We are motivated by newly proposed methods for data mining large-scale corpora of scholarly publications, such as the full biomedical literature, which may consist of tens of millions of papers spanning decades of research. In this setting, analysts seek to discover how concepts relate to one another. They construct graph representations from annotated text databases and then formulate the relationship-mining problem as one of computing all-pairs shortest paths (APSP), which becomes a significant bottleneck. In this context, we present a new high-performance algorithm and implementation of the Floyd-Warshall algorithm for distributed-memory parallel computers accelerated by GPUs, which we call DSNAPSHOT (Distributed Accelerated Semiring All-Pairs Shortest Path). For our largest experiments, we ran DSNAPSHOT on a connected input graph with millions of vertices using 4, 096nodes (24,576GPUs) of the Oak Ridge National Laboratory's Summit supercomputer system. We find DSNAPSHOT achieves a sustained performance of 136×1015 floating-point operations per second (136petaflop/s) at a parallel efficiency of 90% under weak scaling and, in absolute speed, 70% of the best possible performance given our computation (in the single-precision tropical semiring or “min-plus” algebra). Looking forward, we believe this novel capability will enable the mining of scholarly knowledge corpora when embedded and integrated into artificial intelligence-driven natural language processing workflows at scale.

Kannan, Ramakrishnan {ramki}↗

X-ray diffraction under grazing incidence conditions

Material properties frequently relate to structures at or near surfaces, particularly in thin films. As a result, it is essential to understand these structures at the molecular and atomistic levels. The most accurate and widely used techniques for characterizing crystallographic order are based on X-ray diffraction. When dealing with thin films or interfaces, standard approaches, such as single crystal or powder diffraction, are not suitable. However, X-ray diffraction under grazing incidence conditions can provide the required information. Here, in this Primer, grazing incidence X-ray diffraction (GIXD) is comprehensively introduced, starting from basic considerations on X-ray diffraction at crystals with reduced dimensionality and the optical properties of X-rays, followed by a more in-depth description of an experimental performance, including X-ray sources, goniometers and detectors. Experimental errors, limitations and reproducibility are discussed. Various applications, from highly ordered inorganic single crystal surfaces to weakly ordered polymer thin films, are presented to illustrate the potential of GIXD. Data visualizations, representations and evaluation strategies are summarized, based on the example of anthracene thin films. The Primer compiles information relevant to perform high-quality GIXD experiments, evaluate data and interpret results, to extend knowledge about X-ray diffraction from surfaces, interfaces and thin films.

36 MATERIALS SCIENCE↗

Energy Efficient Stochastic-Based Deep Spiking Neural Networks for Sparse Datasets

With large deep neural networks (DNNs) neces- sary to solve complex and data-intensive problems, energy efficiency is a key bottleneck for effectively deploying DL in the real world. Deep spiking NNs have gained much research at- tention recently due to the interest in building biological neural networks and the availability of neuromorphic platforms, which can be orders of magnitude more energy efficient compared to CPUs and GPUs. Although spiking NNs have proven to be an efficient technique for solving many machine learning and computer vision problems, to the best of our knowledge, this is the first attempt to adapt spiking NNs to sparse datasets. In this paper, we study the behaviour of spiking NNs in handling NLP datasets and the sparsity in their data representation. Then, we propose a novel framework for spiking NN using the concept of stochastic computing. Specifically, instead of generating spike trains with firing rates proportional to the intensity of each value in the feature set separately, the whole feature set is treated as a distribution function and a stochastic spiking train that follow this distribution is generated. This framework reduces the connectivity between NN layers from O(N) to O(logN). Also, it encodes input data differently and make suitable to handle sparse datasets. Finally, the framework achieves high energy efficiency since it uses Integrate and Fire neurons same as conventional spiking NNs. The results show that our proposed stochastic-based SNN achieves nearly the same accuracy as the original DNN on MNIST dataset, and it has better performance than state- of-the-art SNN. Besides that stochastic-based SNN is energy efficient, where the fully connected DNN, the conventional SNN, and the data normalized SNN consume 38.24, 1.83, and 1.85 -times more energy than the stochastic-based SNN, respectively. For sparse datasets, including IMDb and In-House clinical datasets, stochastic-based SNN achieves performance comparable to that of the conventional DNN. However, the conventional spiking NN has a significant decline in classification accuracy.

Alawad, Mohammed↗

User interface support

Space construction will require heavy investment in the development of a wide variety of user interfaces for the computer-based tools that will be involved at every stage of construction operations. Using today's technology, user interface development is very expensive for two reasons: (1) specialized and scarce programming skills are required to implement the necessary graphical representations and complex control regimes for high-quality interfaces; (2) iteration on prototypes is required to meet user and task requirements, since these are difficult to anticipate with current (and foreseeable) design knowledge. We are attacking this problem by building a user interface development tool based on extensions to the spreadsheet model of computation. The tool provides high-level support for graphical user interfaces and permits dynamic modification of interfaces, without requiring conventional programming concepts and skills.

Lewis, Clayton↗

When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling

Recent advances in differentiable modeling, a genre of physics-informed machine learning that trains neural networks (NNs) together with process-based equations, have shown promise in enhancing hydrological models' accuracy, interpretability, and knowledge-discovery potential. Current differentiable models are efficient for NN-based parameter regionalization, but the simple explicit numerical schemes paired with sequential calculations (operator splitting) can incur numerical errors whose impacts on models' representation power and learned parameters are not clear. Implicit schemes, however, cannot rely on automatic differentiation to calculate gradients due to potential issues of gradient vanishing and memory demand. Here we propose a “discretize-then-optimize” adjoint method to enable differentiable implicit numerical schemes for the first time for large-scale hydrological modeling. The adjoint model demonstrates comprehensively improved performance, with Kling–Gupta efficiency coefficients, peak-flow and low-flow metrics, and evapotranspiration that moderately surpass the already-competitive explicit model. Therefore, the previous sequential-calculation approach had a detrimental impact on the model's ability to represent hydrological dynamics. Furthermore, with a structural update that describes capillary rise, the adjoint model can better describe baseflow in arid regions and also produce low flows that outperform even pure machine learning methods such as long short-term memory networks. The adjoint model rectified some parameter distortions but did not alter spatial parameter distributions, demonstrating the robustness of regionalized parameterization. Despite higher computational expenses and modest improvements, the adjoint model's success removes the barrier for complex implicit schemes to enrich differentiable modeling in hydrology.

58 GEOSCIENCES↗

Computer-implemented method and apparatus for autonomous position determination using magnetic field data

A computer-implemented method and apparatus for determining position of a vehicle within 100 km autonomously from magnetic field measurements and attitude data without a priori knowledge of position. An inverted dipole solution of two possible position solutions for each measurement of magnetic field data are deterministically calculated by a program controlled processor solving the inverted first order spherical harmonic representation of the geomagnetic field for two unit position vectors 180 degrees apart and a vehicle distance from the center of the earth. Correction schemes such as a successive substitutions and a Newton-Raphson method are applied to each dipole. The two position solutions for each measurement are saved separately. Velocity vectors for the position solutions are calculated so that a total energy difference for each of the two resultant position paths is computed. The position path with the smaller absolute total energy difference is chosen as the true position path of the vehicle.

Ketchum, Eleanor A.↗

Narrowing Uncertainties in the Effects of Elevated CO2 on Crops

Plant responses to rising atmospheric carbon dioxide (CO2) concentrations, together with projected variations in temperature and precipitation will determine future agricultural production. Estimates of the impacts of climate change on agriculture provide essential information to design effective adaptation strategies, and develop sustainable food systems. Here, we review the current experimental evidence and crop models on the effects of elevated CO2 concentrations. Recent concerted efforts have narrowed the uncertainties in CO2-induced crop responses so that climate change impact simulations omitting CO2 can now be eliminated. To address remaining knowledge gaps and uncertainties in estimating the effects of elevated CO2 and climate change on crops, future research should expand experiments on more crop species under a wider range of growing conditions, improve the representation of responses to climate extremes in crop models, and simulate additional crop physiological processes related to nutritional quality.

carbon dioxide (CO2)↗

Cloud Influence on ERA5 and AMPS Surface Downwelling Longwave Radiation Biases in West Antarctica

The surface downwelling longwave radiation component (LW?) is crucial for the determination of the surface energy budget and has significant implications for the resilience of ice surfaces in the polar regions. Accurate model evaluation of this radiation component requires knowledge about the phase, vertical distribution, and associated temperature of water in the atmosphere, all of which control the LW? signal measured at the surface. In this study, we examine the LW? model errors found in the Antarctic Mesoscale Prediction System (AMPS) operational forecast model and the ERA5 reanalysis model relative to observations from the AWARE campaign at McMurdo Station and the West Antarctic Ice Sheet (WAIS) Divide. The errors are calculated separately for observed clear sky conditions, ice cloud occurrences, and liquid-bearing cloud layer (LBCL) occurrences. The analysis results show a tendency (in both models at each site) to underestimate the LW? during clear sky conditions, high error variability (standard deviations > 20 W/m2) during any type of cloud occurrence, and negative LW? biases when LBCLs are observed (bias magnitudes > 15 W/m2 in tenuous LBCL cases; > 43 W/m2 in optically thick/opaque LBCLs instances). We suggest that a generally dry and liquid water-deficient atmosphere responsible for the identified LW? biases in both models is the result of excessive ice formation and growth, which could stem from model initial and lateral boundary conditions, microphysics scheme, aerosol representation, and/or limited vertical resolution.

Silber, Israel↗

Dynamic edge warping - An experimental system for recovering disparity maps in weakly constrained systems

Dynamic edge warping (DEW), a technique for recovering reasonably accurate disparity maps from uncalibrated stereo image pairs, is presented. No precise knowledge of the epipolar camera geometry is assumed. The technique is embedded in a system including structural stereopsis on the front end and robust estimation in digital photogrammetry on the other for the purpose of self-calibrating stereo image pairs. Once the relative camera orientation is known, the epipolar geometry is computed and the system can use this information to refine its representation of the object space. Such a system will find application in the autonomous extraction of terrain maps from stereo aerial photographs, for which camera position and orientation are unknown a priori, and for online autonomous calibration maintenance for robotic vision applications, in which the cameras are subject to vibration and other physical disturbances after calibration. This work thus forms a component of an intelligent system that begins with a pair of images and, having only vague knowledge of the conditions under which they were acquired, produces an accurate, dense, relative depth map. The resulting disparity map can also be used directly in some high-level applications involving qualitative scene analysis, spatial reasoning, and perceptual organization of the object space. The system as a whole substitutes high-level information and constraints for precise geometric knowledge in driving and constraining the early correspondence process.

Boyer, K. L.↗

Climate-invariant machine learning

Projecting climate change is a generalization problem: We extrapolate the recent past using physical models across past, present, and future climates. Current climate models require representations of processes that occur at scales smaller than model grid size, which have been the main source of model projection uncertainty. Recent machine learning (ML) algorithms hold promise to improve such process representations but tend to extrapolate poorly to climate regimes that they were not trained on. To get the best of the physical and statistical worlds, we propose a framework, termed “climate-invariant” ML, incorporating knowledge of climate processes into ML algorithms, and show that it can maintain high offline accuracy across a wide range of climate conditions and configurations in three distinct atmospheric models. Our results suggest that explicitly incorporating physical knowledge into data-driven models of Earth system processes can improve their consistency, data efficiency, and generalizability across climate regimes.

54 ENVIRONMENTAL SCIENCES↗

Essential Function Analysis Capability

The Essential Function Analysis Capability (EFAC) is an extension of the All Hazards Knowledge Framework (AHA) and is a dynamic analytical framework that enables critical infrastructure knowledge discovery and decision support across the five mission areas – prevention, protection, mitigation, response and recovery. EFAC provides the ability to store and model infrastructure systems as a linked multigraphs providing an intuitive and natural representation. These infrastructure systems can then be tied to the organizational and essential function breakdown of an entity. This capability provides the foundation to rapidly evaluate and understand the potential consequences of manmade and natural disaster on these infrastructure systems and essential functions.

Hoover, MichaelJ.↗

All Hazards Analysis Core

The All Hazards Knowledge Framework (AHA) is a dynamic analytical framework that enables critical infrastructure knowledge discovery and decision support across the five mission areas – prevention, protection, mitigation, response and recovery. AHA provides the ability to store and model infrastructure systems as a linked multigraphs providing an intuitive and natural representation. This capability provides the foundation to rapidly evaluate and understand the potential consequences of manmade and natural disaster on these systems.

McGillivary, KentE.↗

High temperature environmental effects on metals

The current status of knowledge and ability to predict high-temperature environmental attack of metals is reviewed with particular reference to the gas turbine engine. Environmental attack is caused by high temperatures, combustion products, and impurities. A schematic representation of life-limiting factors of turbine components shows that environmental attack can lead to very early failures. Attention is given to high-temperature oxidation with prevailing modes of oxidation attack, and to hot corrosion and other impurity effects. Erosion attack results from the direct mechanical removal of component material by impact of hard substances like ash, sand, or dirt. Solutions to hot-corrosion problems can be found semiempirically by using improved alloys or ceramics, protective surface coatings, additives to the engine environment, and air/fuel cleanup to eliminate detrimental impurities.

Grisaffe, S. J.↗

The discrete correlation function - A new method for analyzing unevenly sampled variability data

A method for measuring correlation functions without interpolating in the temporal domain is proposed which provides an assumption-free representation of the correlation measured in the data and allows meaningful error estimates. Physical interpretation of the cross-correlation function of two series believed to be related by a convolution is shown to require knowledge of the input function's fluctuation power spectrum. Application of the method to two systems reveals no correlation for the optical data of Akn 120, but a strong correlation for the UV data of NGC 4151, placing bounds of between 1.2 and 20 light days on the size of the line-emitting region.

Edelson, R. A.↗

Navigating a Mobile Robot Across Terrain Using Fuzzy Logic

A strategy for autonomous navigation of a robotic vehicle across hazardous terrain involves the use of a measure of traversability of terrain within a fuzzy-logic conceptual framework. This navigation strategy requires no a priori information about the environment. Fuzzy logic was selected as a basic element of this strategy because it provides a formal methodology for representing and implementing a human driver s heuristic knowledge and operational experience. Within a fuzzy-logic framework, the attributes of human reasoning and decision- making can be formulated by simple IF (antecedent), THEN (consequent) rules coupled with easily understandable and natural linguistic representations. The linguistic values in the rule antecedents convey the imprecision associated with measurements taken by sensors onboard a mobile robot, while the linguistic values in the rule consequents represent the vagueness inherent in the reasoning processes to generate the control actions. The operational strategies of the human expert driver can be transferred, via fuzzy logic, to a robot-navigation strategy in the form of a set of simple conditional statements composed of linguistic variables. These linguistic variables are defined by fuzzy sets in accordance with user-defined membership functions. The main advantages of a fuzzy navigation strategy lie in the ability to extract heuristic rules from human experience and to obviate the need for an analytical model of the robot navigation process.

Seraji, Homayoun↗

Uncertainty reasoning in expert systems

Intelligent control is a very successful way to transform the expert's knowledge of the type 'if the velocity is big and the distance from the object is small, hit the brakes and decelerate as fast as possible' into an actual control. To apply this transformation, one must choose appropriate methods for reasoning with uncertainty, i.e., one must: (1) choose the representation for words like 'small', 'big'; (2) choose operations corresponding to 'and' and 'or'; (3) choose a method that transforms the resulting uncertain control recommendations into a precise control strategy. The wrong choice can drastically affect the quality of the resulting control, so the problem of choosing the right procedure is very important. From a mathematical viewpoint these choice problems correspond to non-linear optimization and are therefore extremely difficult. In this project, a new mathematical formalism (based on group theory) is developed that allows us to solve the problem of optimal choice and thus: (1) explain why the existing choices are really the best (in some situations); (2) explain a rather mysterious fact that fuzzy control (i.e., control based on the experts' knowledge) is often better than the control by these same experts; and (3) give choice recommendations for the cases when traditional choices do not work.

Kreinovich, Vladik↗