Search NASA⌕ Search

SEARCH · Search NASA

Results for “Domain knowledge”

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 1,207 records · Page 67

Towards Automated Reasoning Chains for Verification of LLM-Generated Scientific Code

With the rise of Large Language Model (LLM) generated code, including in domains like scientific computing, ensuring not only syntactical, but also mathematical correctness, has become a critical task. Traditional formal methods approaches often struggle with the ambiguity of floating-point code, and full symbolic execution is extremely costly and limited. We propose a chain-of-reasoning approach that iteratively lifts basic semantics from code into the SPIRAL system and then establishes numerical equivalency to the desired mathematical operation. Here, we leverage the ample mathematical knowledge already formalized in SPIRAL to enable the system to recognize not just different implementations of the same algorithm but fully separate approaches to solving the given problem. The chain establishes tight error bounds on the output of given code with respect to the true continuous solution it approximates, quantifying all sources of error. We demonstrate this approach by establishing the correctness of a pseudospectral solver for a simple 1-dimensional Poisson problem.

Oschatz, Quentin [Carnegie Mellon University,Pitts↗

Developing a SARS-CoV-2 main protease binding prediction random forest model for drug repurposing for COVID-19 treatment

The coronavirus disease 2019 (COVID-19) global pandemic resulted in millions of people becoming infected with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus and close to seven million deaths worldwide. It is essential to further explore and design effective COVID-19 treatment drugs that target the main protease of SARS-CoV-2, a major target for COVID-19 drugs. In this study, machine learning was applied for predicting the SARS-CoV-2 main protease binding of Food and Drug Administration (FDA)-approved drugs to assist in the identification of potential repurposing candidates for COVID-19 treatment. Ligands bound to the SARS-CoV-2 main protease in the Protein Data Bank and compounds experimentally tested in SARS-CoV-2 main protease binding assays in the literature were curated. These chemicals were divided into training (516 chemicals) and testing (360 chemicals) data sets. To identify SARS-CoV-2 main protease binders as potential candidates for repurposing to treat COVID-19, 1188 FDA-approved drugs from the Liver Toxicity Knowledge Base were obtained. A random forest algorithm was used for constructing predictive models based on molecular descriptors calculated using Mold2 software. Model performance was evaluated using 100 iterations of fivefold cross-validations which resulted in 78.8% balanced accuracy. The random forest model that was constructed from the whole training dataset was used to predict SARS-CoV-2 main protease binding on the testing set and the FDA-approved drugs. Model applicability domain and prediction confidence on drugs predicted as the main protease binders discovered 10 FDA-approved drugs as potential candidates for repurposing to treat COVID-19. Our results demonstrate that machine learning is an efficient method for drug repurposing and, thus, may accelerate drug development targeting SARS-CoV-2.

Research & Experimental Medicine↗

Relation Inference among Sensor Time Series in Smart Buildings with Metric Learning

Smart Building Technologies hold promise for better livability for residents and lower energy footprints. Yet, the rollout of these technologies, from demand response controls to fault detection and diagnosis, significantly lags behind and is impeded by the current practice of manual identification of sensing point relationships, e.g., how equipment is connected or which sensors are co-located in the same space. This manual process is still error-prone, albeit costly and laborious.We study relation inference among sensor time series. Our key insight is that, as equipment is connected or sensors co-locate in the same physical environment, they are affected by the same real-world events, e.g., a fan turning on or a person entering the room, thus exhibiting correlated changes in their time series data. To this end, we develop a deep metric learning solution that first converts the primitive sensor time series to the frequency domain, and then optimizes a representation of sensors that encodes their relations. Built upon the learned representation, our solution pinpoints the relationships among sensors via solving a combinatorial optimization problem. Extensive experiments on real-world buildings demonstrate the effectiveness of our solution.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

Using AI to build a hydrobiogeochemical soil model

Soil water content is a function of inputs from precipitation and outputs via evaporation, transpiration, lateral flow, and vertical percolation, and is sensitive to biogeochemical processes. As such, soils serve as an ideal integrator of atmospheric, hydrological, and biogeochemical processes affecting the water cycle. In addition, soil water retention capacity, infiltration rates, and hydraulic conductivity can buffer or exacerbate the effects of extreme precipitation events (e.g., flooding, runoff, subsurface transport, erosion, greenhouse gas emissions) and mitigate the impact of droughts and heat waves on land systems (e.g., fire, crop failure). However, integrating water cycle measurements spanning different land atmosphere compartments across scales is a fundamental barrier for numerical model predictability. A significant challenge is that each domain (soil, hydrology, biology, and atmosphere) typically collects different sets of data at different temporal and spatial frequencies/scales, and even different dimensionalities (2D vs 3D). To implement soil as an integrator of the water cycle in land models, we suggest that novel machine learning (ML) tools can be developed to effectively simulate complex landscapes across various domains and scales, extended to regions with sparse or no data. The ultimate goals are to improve predictive understanding of land-atmosphere interactions and to extend the predictability of current Earth System Models (ESMs) through better integration of hydrological and biogeochemical data. We envision a framework in which: (1) ML-aided data reconstructions enable the merger of data sources into a unified geospatial product; (2) automated detection techniques are used to improve the knowledge of complex soil processes and interactions; and (3) this knowledge is leveraged and incorporated into models through AI-based emulators to distinctly connect the land and atmospheric compartments of the water cycle in models.

54 ENVIRONMENTAL SCIENCES↗

Sandia National Laboratories Ecosystem for Open Science: Metadata Schema v0.2 Description.

The Ecosystem for Open Science (eOS) initiative was established in 2019. Its objective is improving openness and sharing of data and information across Defense Nuclear Nonproliferation (DNN) Research and Development (R&D) activities. To support this initiative, the eOS team at Sandia National Laboratories (SNL) developed metadata and data standards and proposed a machine-readable metadata schema. The nuclear explosion monitoring field was selected as a focus area due to its the wide range of pertinent phenomenologies.We developed the DCAT-eOS-AP metadata schema extending the Data Catalog Vocabulary version 2 (DCATv2) standard using an application profile (AP), to fit the needs of multi-disciplinary NA-22 projects. The DCAT-eOS-AP metadata schema describes data at different levels of granularity ranging from general descriptions to more domain-specific granular metadata. Its implementation and serialization is flexible with the ability to include new file or data types. Thus, it will scale with the ever-increasing data management needs of government research. Due to the multitude of phenomenologies represented in the DCAT-eOS-AP schema, we anticipate that it will be easily extensible to various projects across many DOE mission areas. This document describes data management challenges faced within the DNN R&D portfolio and provides insight on how metadata and data standards/guidelines combined with a comprehensive metadata schema can add value to programs throughout the Department of Energy (DOE). It reviews the importance of metadata standards, FAIR (Findability, Accessibility, Interoperability, and Reusability) data principles, and metadata schemas. Additionally, it summarizes input from subject matter experts (SME) at SNL and other National Laboratories that resulted in metadata and data standards/guidelines encompassing domains relevant to NA-22 projects. Finally, we discuss the DCAT-eOS-AP metadata development. Implementation recommendations and future development directions are included for those keen on adopting the DCAT-eOS-AP metadata schema.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

Program for Experimentation With Expert Systems

CERBERUS is forward-chaining, knowledge-based system program useful for experimentation with expert systems. Inference-engine mechanism performs deductions according to user-supplied rule set. Information stored in intermediate area, and user interrogated only when no applicable data found in storage. Each assertion posed by CERBERUS answered with certainty ranging from 0 to 100 percent. Rule processor stops investigating applicable rules when goal reaches certainty of 95 percent or higher. Capable of operating for wide variety of domains. Sample rule files included for animal identification, pixel classification in image processing, and rudimentary car repair for novice mechanic. User supplies set of end goals or actions. System complexity decided by user's rule file. CERBERUS written in FORTRAN 77.

Engle, S. W.↗

A knowledge-based system design/information tool for aircraft flight control systems

Research aircraft have become increasingly dependent on advanced control systems to accomplish program goals. These aircraft are integrating multiple disciplines to improve performance and satisfy research objectives. This integration is being accomplished through electronic control systems. Because of the number of systems involved and the variety of engineering disciplines, systems design methods and information management have become essential to program success. The primary objective of the system design/information tool for aircraft flight control system is to help transfer flight control system design knowledge to the flight test community. By providing all of the design information and covering multiple disciplines in a structured, graphical manner, flight control systems can more easily be understood by the test engineers. This will provide the engineers with the information needed to thoroughly ground test the system and thereby reduce the likelihood of serious design errors surfacing in flight. The secondary objective is to apply structured design techniques to all of the design domains. By using the techniques in the top level system design down through the detailed hardware and software designs, it is hoped that fewer design anomalies will result. The flight test experiences of three highly complex, integrated aircraft programs are reviewed: the X-29 forward-swept wing, the advanced fighter technology integration (AFTI) F-16, and the highly maneuverable aircraft technology (HiMAT) program. Significant operating anomalies and the design errors which cause them, are examined to help identify what functions a system design/information tool should provide to assist designers in avoiding errors.

Mackall, Dale A.↗

A knowledge-based system design/information tool for aircraft flight control systems

Research aircraft have become increasingly dependent on advanced electronic control systems to accomplish program goals. These aircraft are integrating multiple disciplines to improve performance and satisfy research objective. This integration is being accomplished through electronic control systems. Systems design methods and information management have become essential to program success. The primary objective of the system design/information tool for aircraft flight control is to help transfer flight control system design knowledge to the flight test community. By providing all of the design information and covering multiple disciplines in a structured, graphical manner, flight control systems can more easily be understood by the test engineers. This will provide the engineers with the information needed to thoroughly ground test the system and thereby reduce the likelihood of serious design errors surfacing in flight. The secondary object is to apply structured design techniques to all of the design domains. By using the techniques in the top level system design down through the detailed hardware and software designs, it is hoped that fewer design anomalies will result. The flight test experiences are reviewed of three highly complex, integrated aircraft programs: the X-29 forward swept wing; the advanced fighter technology integration (AFTI) F-16; and the highly maneuverable aircraft technology (HiMAT) program. Significant operating technologies, and the design errors which cause them, is examined to help identify what functions a system design/informatin tool should provide to assist designers in avoiding errors.

Mackall, Dale A.↗

ICADS: A cooperative decision making model with CLIPS experts

A cooperative decision making model is described which is comprised of six concurrently executing domain experts coordinated by a blackboard control expert. The focus application field is architectural design, and the domain experts represent consultants in the area of daylighting, noise control, structural support, cost estimating, space planning, and climate responsiveness. Both the domain experts and the blackboard were implemented as production systems, using an enhanced version of the basic CLIPS package. Acting in unison as an Expert Design Advisor, the domain and control experts react to the evolving design solution progressively developed by the user in a 2-D CAD drawing environment. A Geometry Interpreter maps each drawing action taken by the user to real world objects, such as spaces, walls, windows, and doors. These objects, endowed with geometric and nongeometric attributes, are stored as frames in a semantic network. Object descriptions are derived partly from the geometry of the drawing environment and partly from knowledge bases containing prototypical, generalized information about the building type and site conditions under consideration.

Pohl, Jens↗

In Situ Geophysical Exploration by Humans in Mars Analog Environments

We carried out three geophysical experiments in Mars analog environments in order to better understand the challenges future astronauts will face when conducting similar surveys on Mars or the Moon. The experiments included a passive seismometer deployment and a time-domain electromagnetic survey at the Flashline Mars Arctic Research Station (FMARS) on Devon Island, Canada and a seismic refraction survey in southeastern Utah at the Mars Desert Research Station (MDRS). FMARS is located on the rim of the 23 Ma Haughton Crater in a polar desert environment. MDRS is located in an area with sedimentary plateaus and canyons of Jurassic to Cretaceous age. Both facilities were built by The Mars Society to help develop key knowledge about human Mars exploration. Crews of six spend 2-4 weeks in the habitats and conduct eld research on simulated extravehicular activities (EVAs) wearing mock spacesuits. The work reported here was conducted in July 2009 at FMARS and February 2010 at MDRS.

Shiro, B. R.↗

A dynamic Bayesian optimized active recommender system for curiosity-driven partially Human-in-the-loop automated experiments

Optimization of experimental materials synthesis and characterization through active learning methods has been growing over the last decade, with examples ranging from measurements of diffraction on combinatorial alloys at synchrotrons, to searches through chemical space with automated synthesis robots for perovskites. In virtually all cases, the target property of interest for optimization is defined a priori with the ability to shift the trajectory of the optimization based on human-identified findings during the experiment is lacking. Thus, to highlight the best of both human operators and AI-driven experiments, here we present the development of a human–AI collaborated experimental workflow, via a Bayesian optimized active recommender system (BOARS), to shape targets on the fly with human real-time feedback. Here, the human guidance overpowers AI at early iteration when prior knowledge (uncertainty) is minimal (higher), while the AI overpowers the human during later iterations to accelerate the process with the human-assessed goal. We showcase examples of this framework applied to pre-acquired piezoresponse force spectroscopy of a ferroelectric thin film, and in real-time on an atomic force microscope, with human assessment to find symmetric hysteresis loops. It is found that such features appear more affected by subsurface defects than the local domain structure. This work shows the utility of human–AI approaches for curiosity driven exploration of systems across experimental domains.

36 MATERIALS SCIENCE↗

Terahertz Characteristics of Liquid D{sub 2}O in H{sub 2}O

D{sub 2}O used in nuclear power generation is very slightly different from H{sub 2}O. Deuterium (D) is the isotope of hydrogen (H) and has one more neutron than H. D{sub 2}O is heaver than H{sub 2}O by about 11% and its melting and boiling points are 3.81 deg. C and 101.42 deg. C, respectively. Moreover, D{sub 2}O is toxic. It is harmful for human body if it replaces more than approximately 25-50% of a human's body water. So, knowledge of the properties of D{sub 2}O included in H{sub 2}O is necessary. The confinement energy and hydrogen bonding energy of large molecules are in the terahertz frequency region (2{approx}10 THz, 3{approx}66 cm{sup -1}). Particularly, the polar molecules such as H{sub 2}O and D{sub 2}O react to this frequency region well because the time of hydrogen relaxation motion is subpicosecond. Terahertz time-domain spectroscopy (THz-TDS) using terahertz pulses has been used in the study of ultrafast dynamics of hydrogen bonds in biomolecules and carriers in semiconductors. We obtained the refractive index and power absorption of H{sub 2}O and D{sub 2}O at room temperature. The results show that the refractive index and power absorption of H{sub 2}O are larger than for D{sub 2}O.

74 ATOMIC AND MOLECULAR PHYSICS↗

Scaling Graph Community Detection on the Tilera Many-core Architecture

In an era when power constraints and data movement are proving to be significant barriers for the application of high-end computing, the Tilera many-core architecture offers a low-power platform exhibiting many important characteristics of future systems, including a large number of simple cores, a sophisticated network-on-chip, and fine-grained control over memory and caching policies. While this emerging architecture has been previously studied for structured compute-intensive kernels, benchmarking the platform for data-bound, irregular applications present significant challenges that have remained unexplored. Community detection is an advanced prototypical graph-theoretic operation with applications in numerous scientific domains including life sciences, cyber security, and power systems. In this work, we explore multiple design strategies toward developing a scalable tool for community detection on the Tilera platform. Using several memory layout and work scheduling techniques we demonstrate speedups of up to 46x on 36 cores of the Tilera TileGX36 platform over the best serial implementation, and also show results that have comparable quality and performance to mainstream x86 platforms. To the best of our knowledge this is the first work addressing graph algorithms on the Tilera platform. This study demonstrates that through careful design space exploration, low-power many-core platforms like Tilera can be effectively exploited for graph algorithms that that embody all the essential characteristics of an irregular application.

community detection↗

Generic Multi-Frequency Modelling of Converter-Connected Renewable Energy Generators Considering Frequency and Sequence Couplings

Frequency and sequence couplings can compromise the trustworthiness of multi-frequency models for converter-based systems. There have been effective attempts to address the couplings mainly by linearized averaged models. Only a few studies have been conducted on practical optimization of such models with enormous matrices and experimental results. This paper provides a generic theory for coupling patterns and proposes a multi-frequency modelling method to detect and address only the main couplings in the sequence domain for converter-connected renewable energy generators. The proposed generic model is based on empirical tests using small-signal perturbations and adopting Fourier transform on the switching converter response. The proposed theory and modelling methodology are verified using a 7MVA grid emulator for voltage perturbation tests on a 2MVA photo-voltaic converter. Accordingly, the couplings can exist in more generic forms, including multiples of perturbation and fundamental frequencies. To the best of our knowledge, the patterns with the multiples of the perturbation frequency have been overlooked in the literature. Furthermore, the mirror frequency concept is valid for all coupling patterns and is included in the proposed model. Besides, the proposed patterns and the environment noise levels have been practical criteria for selecting the main couplings.

frequency and sequence couplings↗

Developing Guidelines for Assessing Visual Analytics Environments

Visual analytic systems can be evaluated from a user perspective with quantitative metrics (i.e., time to complete the analysis or the accuracy of the solution found). However, qualitative measures are also useful in a user assessment. These include such measures as the utility of the interactive visualizations in the analysis process and the user's assessment of the efficiency of the analytic process. Quantitative measures can be found if data sets with embedded ground truth are used for analysis. Qualitative measures are more elusive. In this paper we report on an experiment with professional analysts who ranked five of submissions to the VAST 2009 Challenge and provided the rationale for their rankings. Their comments were used in conjunction with a meta-analysis of the 2009 VAST Challenge reviews to produce a set of guidelines for visual analytic systems. As visual analytic software is expected to eventually help in all aspects of analysis, we expect to see future systems provide more help with generating the final report. Hence, researchers also need to have an understanding of what makes a good analytic product. Therefore we asked the analysts to rank the situational assessments of four grand challenge entries and to provide comments on those assessments. We used these comments to produce guidelines for researchers to use in evaluating their analytic reports.

99 GENERAL AND MISCELLANEOUS↗

Post-Viking microbiology: new approaches, new data, new insights

In the 20 years since the Viking experiments, major advances have been made in the areas of microbial systematics, microbial metabolism, microbial survival capacity, and the definition of environments on earth, suggesting that life is more versatile and tenacious than was previously appreciated. Almost all niches on earth which have available energy, and which are compatible with the chemistry of carbon-carbon bonds, are known to be inhabited by bacteria. The oldest known bacteria on earth apparently evolved soon after the formation of the planet, and are heat loving, hydrogen and/or sulfur metabolizing forms. Among the two microbial domains (kingdoms) is a great deal of metabolic diversity, with members of these forms being able to grow on almost any known energy source, organic or inorganic, and to utilize an impressive array of electron acceptors for anaerobic respiration. Both hydrothermal environments and the deep subsurface environments have been shown to support large populations of bacteria, growing on energy supplied by geothermal energy, thus isolating these ecosystems from the rest of the global biogeochemical cycles. This knowledge, coupled with new insights into the history of the solar system, allow one to speculate on possible evolution and survival of life forms on Mars.

Review↗

Conservation of Fold and Topology of Functional Elements in Thiamin Pyrophosphate Enzymes

Thiamin pyrophosphate (TPP)-dependent enzymes are a highly divergent family of proteins binding both TPP and metal ions. They perform decarboxylation-hydroxyaldehydes. Prior -ketoacids and of a common - (O=)C-C(OH)- fragment of to knowledge of three-dimensional structures of these enzmes, the GDGY25-30NN sequence was used to identify these enzymes. Subsequently, a number of structural studies on those enzymes revealed multi-subunit organization and the features of the two duplicate cofactor binding sites. Analyzing the structures of 44 structurally known enzymes, we found that the common structure of these enzymes is reduced to 180-220 amino acid long fragments of two PP and two PYR domains that form the [PP:PYR]2 binding center of two cofactor molecules. The structures of PP and PYR are arranged in a similar fold-sheet with triplets of helices on both sides.Dconsisting of a six-stranded Residues surrounding the cofactors are not strictly conserved, but they provide the same interatomic contacts required for the catalytic functions that these enzymes perform while maintaining interactive structural integrity. These structural and functional amino acids are topological counterparts located in the same positions of the conserved fold of sets of PP and PYR domains. Additional parallels include short fragments of sequences that link these amino acids to the fold and function. This report on the structural commonalities amongst TPP dependent enzymes is thought to contribute new approaches to annotation that may assist in advancing the functional proteomics of TPP dependent enzymes, and trace their complexity within evolutionary context.

Dominiak, P.↗

A Lagrangian, stochastic modeling framework for multi-phase flow in porous media

Many of the complex physical processes relevant for compositional multi-phase flow in porous media are well understood at the pore-scale level. In order to study CO{sub 2} storage in sub-surface formations, however, it is not feasible to perform simulations at these small scales directly and effective models for multi-phase flow description at Darcy scale are needed. Unfortunately, in many cases it is not clear how the micro-scale knowledge can rigorously be translated into consistent macroscopic equations. Here, we present a new methodology, which provides a link between Lagrangian statistics of phase particle evolution and Darcy scale dynamics. Unlike in finite-volume methods, the evolution of Lagrangian particles representing small fluid phase volumes is modeled. Each particle has a state vector consisting of its position, velocity, fluid phase information and possibly other properties like phase composition. While the particles are transported through the computational domain according to their individual velocities, the properties are modeled via stochastic processes honoring specified Lagrangian statistics. Note that the conditional expectations of the particle velocities are different for different fluid phases. The goal of this paper is to present the general framework for this alternative modeling approach. Various one and two-dimensional numerical experiments demonstrate that with appropriate stochastic rules the particle solutions are consistent with a standard two-phase Darcy flow formulation. In the end, we demonstrate how to model non-equilibrium phenomena within the stochastic particle framework, which will be the main focus of the future work.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗