Search NASA⌕ Search

SEARCH · Search NASA

Results for “Large language model”

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.

809 records · Page 45

Open Building Control

Best practice control sequences are often not implemented correctly, or are not implemented at all, in large commercial buildings. This typically leads to 10-30 percent energy waste, along with reduced occupant productivity and unnecessary equipment wear. The current process of designing and implementing such control sequences is a manual process that starts with designers who often don’t have adequate training, then requires controls programmers to interpret and program a verbose written sequence. This process has been shown to fail to deliver high performance control sequences at scale. The Open Building Control project digitizes the current control delivery process. The project is developing tools for system designers to select control sequences, assess their energy performance and load flexibility potential using whole building simulation, specify the sequence for implementation using machine-to-machine translation by a control provider and formally testing the as-installed sequences by a commissioning agent. The project developed tools for each stage of this delivery process. The key innovation of the project is the development of the Control Description Language, a language that allows such a digitized control delivery process with end-to-end verification. Libraries of control sequences have been implemented using the Control Description Language, and their performance has been demonstrated using whole building energy simulation. An automated translation of such sequences to a commercial control product line has been conducted using a prototype translator. Tools for formal verification of as-installed control sequences relative to their specification have been developed and demonstrated. The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) started the process of forming a committee to make this language an ASHRAE/ANSI Standard. This new standard will complement existing and emerging ASHRAE standards for building communication and semantic modeling by providing a standard for expressing the control logic - the actual brain of the building. We expect this language and the process it enables to be an important contribution to the deployment of high performance building control sequences at scale because it allows taming the complexity of the control delivery process, which is continually increasing due to the need for higher performance and increased load flexibility to meet goals for net zero energy and increased renewable integration.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

On the Language of Reliability: A System Engineer Perspective

In its classical definition, risk is defined by three elements: what can go wrong, what are its consequences, and how likely is it to occur? While this definition makes sense in a regulatory-based framework where for the current fleet of operating light water reactors (LWRs), the risks associated with nuclear power plants typically are characterized in terms of core damage and large early release frequency (LERF), this approach does not provide a useful snapshot of the health of the plant from a broader perspective. This is due to the very narrow context in which the term “risk” typically is defined as nuclear safety aspects that have the potential to impact public health. In this paper, we take the viewpoint of nuclear safety that is reflective of the current fleet of operating LWRs for which core damage frequency and LERF are appropriate metrics. For other advanced reactor designs, other more applicable technology neutral metrics of reactor safety metrics would be specified. A possible alternate path would start by redefining the word risk with a broader meaning that better reflects the needs of a system health and asset management decision-making process. Rather than asking how likely an event could occur (in probabilistic terms), we can ask how far this event is from occurring. Our approach starts by defining and quantifying component and system health in terms of a “distance” between its actual and limiting conditions, i.e., determination of the margin that exists between the current state/condition and the state where the component/system is no longer capable of achieving its intended function. A margin is a measure that is more reflective of the current state or performance of a component, and therefore more closely tied to decisions that are made on an ongoing basis. Further, we will show how, given the data available from plant equipment reliability and monitoring (e.g., pump vibration data) and prognostic (e.g., component remaining useful life estimation) data, a margin can be described and determined for all types of maintenance approaches (e.g., corrective or predictive maintenance). We show how classical reliability models (e.g., fault trees) can be used to quantify the system margin provided component margin values. In the approach described in this paper, the propagation of margin values through classical reliability models are not performed using classical probabilistic calculations applied to sets (as performed in a typical plant probabilistic risk assessment). Instead, we show how it is possible to propagate margin values through Boolean logic gates (i.e., AND and OR operators) through distance-based operations.

97 MATHEMATICS AND COMPUTING↗

Autonomous Command Operations of the WIRE Spacecraft

This paper presents operational innovations which will be introduced on NASA's Wide Field Infrared Explorer (WIRE) mission. These innovations include an end-to-end design architecture for an autonomous commanding capability for the uplink of command loads during unattended station contacts. The WIRE mission is the fifth and final mission of NASA's Goddard Space Flight Center Small Explorer (SMEX) series to be launched in March of 1999. Its primary mission is the targeting of deep space fields using an ultra-cooled infrared telescope. Due to its mission design WIRE command loads are large (approximately 40 Kbytes per 24 hours) and must be performed daily. To reduce the cost of mission operations support that would be required in order to uplink command loads, the WIRE Flight Operations Team has implemented all autonomous command loading capability. This capability allows completely unattended operations over a typical two-day weekend period. The key factors driving design and implementation of this capability were: 1) integration with already existing ground system autonomous capabilities and systems, 2) the desire to evolve autonomous operations capabilities based upon previous SMEX operations experience - specifically the TRACE mission, 3) integration with ground station operations - both autonomous and man-tended, 4) low cost and quick implementation, and 5) end-to-end system robustness. A trade-off study was performed to examine these factors in light of the low-cost, higher-risk SMEX mission philosophy. The study concluded that a STOL (Spacecraft Test and Operations Language) based script, highly integrated with other scripts used to perform autonomous operations, was best suited given the budget and goals of the mission. Each of these factors is discussed in addition to use of the TRACE mission as a testbed for autonomous commanding prior to implementation on WIRE. The capabilities implemented on the WIRE mission are an example of a low-cost, robust, and efficient method for autonomous command loading when implemented with other autonomous features of the ground system. They call be used as a design and implementation template by other missions interested in evolving toward autonomous and lower cost operations. Additionally, the WIRE spacecraft will be used as an operational testbed upon completion of its nominal mission later in 1999. One idea being studied is advanced on-board modeling. Advanced on-board modeling techniques will be used to more efficiently display the spacecraft state. This health and safety information could be used by engineers on the ground or could be used by tile spacecraft for its own assessments. Additionally, this same state information could also be input into the event-driven scheduling system, as the scheduling system will need to assess the spacecraft state before undertaking a new activity. Advanced modeling techniques are being evaluated for a number of NASA missions including The Next Generation Space Telescope (NGST), which is scheduled to launch in 2007.

Walyus, Keith↗

Evaluating proxies for the drivers of natural gas productivity using machine-learning models

We report the extensive development of unconventional reservoirs using horizontal drilling and multistage hydraulic fracturing has generated large volumes of reservoir characterization and production data. The analysis of this abundant data using statistical methods and advanced machine-learning (ML) techniques can provide data-driven insights into well performance. Most predictive modeling studies have focused on the impact that different well completion and stimulation strategies have on well production but have not fully exploited the available in situ rock property data to determine its role in reservoir productivity. We have used machine-learning techniques to rank rock mechanical properties, microseismic attributes, and stimulation parameters in the order of their significance for predicting natural gas production from an unconventional reservoir. The data for this study came from a hydraulically fractured well in the Marcellus Shale in Monongalia County, West Virginia. The data classes included measurements aggregated by well completion stage that included (1) gas production, (2) well-log-derived measurements including bulk density, elastic moduli, shear impedance, compressional impedance, brittleness, and gamma measurements, (3) microseismic attributes, (4) long-period long-duration (LPLD) event counts, (5) fracture counts, and (6) stimulation parameters that included the fluid injection volume and average pumping pressure. To identify observable proxies for the drivers of gas production, we evaluated five commonly used ML approaches including multivariate adaptive regression spline, Gaussian mixture model, random forest, gradient boosting, and neural network. We selected five variables including LPLD event count, seismogenic b-value, hydraulic diffusivity, cumulative moment, and fluid volume as the features most likely to impact gas productivity at the stage level in the study area. The data-driven selection of these parameters for their importance in determining gas production can help reservoir engineers design more effective hydraulic-fracture treatments in the Marcellus Shale and other similar unconventional reservoirs. Plain language summary: We use machine-learning methods and data-driven selection of reservoir parameters to rank and better understand their importance in determining gas production, which can help reservoir engineers design more effective hydraulic-fracture treatments in the Marcellus Shale and other similar unconventional reservoirs.

58 GEOSCIENCES↗

Performance and Portability of a Linear Solver Across Emerging Architectures

A linear solver algorithm used by a large-scale unstructured-grid computational fluid dynamics application is examined for a broad range of familiar and emerging architectures. Efficient implementation of a linear solver is challenging on recent CPUs offering vector architectures. Vector loads and stores are essential to effectively utilize available memory bandwidth on CPUs, and maintaining performance across different CPUs can be difficult in the face of varying vector lengths offered by each. A similar challenge occurs on GPU architectures, where it is essential to have coalesced memory accesses to utilize memory bandwidth effectively. In this work, we demonstrate that restructuring a computation, and possibly data layout, with regard to architecture is essential to achieve optimal performance by establishing a performance benchmark for each target architecture in a low level language such as vector intrinsics or CUDA. In doing so, we demonstrate how a linear solver kernel can be mapped to Intel® Xeon™ and Xeon Phi™, Marvell® ThunderX2®, NEC® SX-Aurora™ TSUBASA Vector Engine, and NVIDIA® and AMD® GPUs. We further demonstrate that the required code restructuring can be achieved in higher level programming environments such as OpenACC, OCCA, and Intel® OneAPI™/SYCL, and that each generally results in optimal performance on the target architecture. Relative performance metrics for all implementations are shown, and subjective ratings for ease of implementation and optimization are suggested.

Programming models↗

AI Meets Exascale Computing: Advancing Cancer Research With Large-Scale High Performance Computing

The application of data science in cancer research has been boosted by major advances in three primary areas: (1) Data: diversity, amount, and availability of biomedical data; (2) Advances in Artificial Intelligence (AI) and Machine Learning (ML) algorithms that enable learning from complex, large-scale data; and (3) Advances in computer architectures allowing unprecedented acceleration of simulation and machine learning algorithms. These advances help build in silico ML models that can provide transformative insights from data including: molecular dynamics simulations, next-generation sequencing, omics, imaging, and unstructured clinical text documents. Unique challenges persist, however, in building ML models related to cancer, including: (1) access, sharing, labeling, and integration of multimodal and multi-institutional data across different cancer types; (2) developing AI models for cancer research capable of scaling on next generation high performance computers; and (3) assessing robustness and reliability in the AI models. In this paper, we review the National Cancer Institute (NCI) -Department of Energy (DOE) collaboration, Joint Design of Advanced Computing Solutions for Cancer (JDACS4C), a multi-institution collaborative effort focused on advancing computing and data technologies to accelerate cancer research on three levels: molecular, cellular, and population. This collaboration integrates various types of generated data, pre-exascale compute resources, and advances in ML models to increase understanding of basic cancer biology, identify promising new treatment options, predict outcomes, and eventually prescribe specialized treatments for patients with cancer.

60 APPLIED LIFE SCIENCES↗

AI Model Benchmarking for Nonproliferation Applications: Steel Thread Benchmarking Task Force Technical Report (Rev. 2)

Steel Thread is a NA-22 venture that seeks to build trustworthy, reliable AI models that can be used in a wide variety of nonproliferation tasks. A key aspect of building these models is developing appropriate benchmarks and evaluation methods, which will enable the venture to identify and adapt models to provide the most value in the nonproliferation domain. Benchmarks must be relevant to key tasks in this domain, such as question answering, information retrieval, document summarization and classification, consensus analysis, and image and data analysis. This report 1) provides an overview of benchmark design, evaluation, and challenges; 2) reviews a variety of open benchmarks, with a focus on language models and tasks; and 3) identifies benchmarks that are most relevant to Steel Thread. This report is intended to serve as a basis for further efforts to classify and evaluate benchmarks and their correlation with success on nonproliferation-specific tasks. The Steel Thread venture has defined benchmarks to be a particular combination of a dataset (or datasets) and a metric (or metrics) conceptualized as representing one or more specific tasks or sets of abilities for a specific modality. It is adopted by a research community as a shared framework for comparing methods.1 It includes 1) Data: Labeled (a designated subset not used for training, which could be all the data), 2) Metric: A way to quantify performance, 3) Task/Ability: What the benchmark is testing, 4) Protocol: A structured and repeatable evaluation process, 5) Baseline/Reference Model: For comparison; could be statistical, rule-based, SME-derived, or another model, and 6) Maintenance Plan: to update with new information over time; important for long-term utility. For further clarity, the definition includes what a benchmark, in this context, is not. It is not a corpus of training data, specific to a model (it is intended to apply to a range of models), a universal evaluation of performance, a guarantee that the ‘top’ model on the leaderboard will be the best fit for every specific use case, an all-encompassing proof of a model’s universal quality, nor is it a one-size-fits-all measure of success. It does not cover every real-world constraint (like operational, ethical, or cost considerations), a systems integration test, or a unit test. This definition was inspired by and resulted from discussions within the Steel Thread Benchmarking Task Force. This group was formed to define what we would mean as a benchmark within Steel Thread but persisted as the need to develop a thorough understanding of the large and expanding existing benchmarking space. This technical report is a result of the group’s divide and conquer approach to exploring this space. The release of benchmarks might not be progressing as quickly as model development, but it is moving very fast, as many benchmarks quickly become saturated, when state-of-the-art models score so close to the benchmark’s ceiling that their results are virtually indistinguishable. At that point, the test no longer differentiates between new systems, so researchers usually stop reporting scores as the benchmark no longer informs about improvements from the next generation of models. In the OpenAI announcement of GPT-5, they reported results on six flagship public benchmarks (AIME 2025, SWE-bench Verified, Aider Polyglot, MMMU, HealthBench Hard, GPQA) but the full system-card covers roughly thirty-five separate evaluations, comprising hundreds of test task items in total. There have been some efforts to summarize benchmarks in specific fields, like for text-to-image generation, but these surveys have had a narrow methodology scope. Therefore, a comprehensive survey of all benchmarks or even all benchmarks that could be relevant to Steel Thread is outside of the scope of this report. We chose some specific benchmarks to investigate in detail.

97 MATHEMATICS AND COMPUTING↗

Translating MAPGEN to ASPEN for MER

This software translates MAPGEN (Europa and APGEN) domains to ASPEN, and the resulting domain can be used to perform planning for the Mars Exploration Rover (MER). In other words, this is a conversion of two distinct planning languages (both declarative and procedural) to a third (declarative) planning language in order to solve the problem of faithful translation from mixed-domain representations into the ASPEN Modeling Language. The MAPGEN planning system is an example of a hybrid procedural/declarative system where the advantages of each are leveraged to produce an effective planner/scheduler for MER tactical planning. The adaptation of the planning system (ASPEN) was investigated, and, with some translation, much of the procedural knowledge encoding is amenable to declarative knowledge encoding. The approach was to compose translators from the core languages used for adapting MAGPEN, which consists of Europa and APGEN. Europa is a constraint- based planner/scheduler where domains are encoded using a declarative model. APGEN is also constraint-based, in that it tracks constraints on resources and states and other variables. Domains are encoded in both constraints and code snippets that execute according to a forward sweep through the plan. Europa and APGEN communicate to each other using proxy activities in APGEN that represent constraints and/or tokens in Europa. The composition of a translator from Europa to ASPEN was fairly straightforward, as ASPEN is also a declarative planning system, and the specific uses of Europa for the MER domain matched ASPEN s native encoding fairly closely. On the other hand, translating from APGEN to ASPEN was considerably more involved. On the surface, the types of activities and resources one encodes in APGEN appear to match oneto- one to the activities, state variables, and resources in ASPEN. But, when looking into the definitions of how resources are profiled and activities are expanded, one sees code snippets that access various information available during planning for the moment in time being planned to decide at the time what the appropriate profile or expansion is. APGEN is actually a forward (in time) sweeping discrete event simulator, where the model is composed of code snippets that are artfully interleaved by the engine to produce a plan/schedule. To solve this problem, representative code is simulated as a declarative series of task expansions. Predominantly, three types of procedural models were translated: loops, if statements, and code blocks. Loops and if statements were handled using controlled task expansion, and code blocks were handled using constraint networks that maintained the generation of results based on what the order of execution would be for a procedural representation. One advantage with respect to performance for MAPGEN is the use of APGEN s GUI. This GUI is written in C++ and Motif, and performs very well for large plans.

Rabideau, Gregg R.↗

Experience Report for WOPR

One of the purposes of the SQA effort at LLNL is to attempt to determine the 'goodness' of the research codes used for various scientific applications. Typically these are two and three dimensional multi-physics simulation and modeling codes. These legacy research codes are used for applciations such as atmospheric dispersion modeling and analysis and prediction of the performance of engineered systems. These codes are continually subjected to automated regression test suites consisting of verified and validated expected results. Code is managed in repositories. Experience level of developers is high in the knowledge domain, platforms, and languages used. Code size of the multi-physics code used in this study was 578,242 lines excluding comment and blank lines or 5538.7 function points. Languages were 70% C++, 20% C, and 10% Fortran. The code has 130 users and a development team of 14 and an embedded SQE. The code has achieved 100% prime feature test coverage, 73.6% functional test coverage, and 71.5% statement test coverage. The average cyclomatic complexity of the code was 6.25. The codes have evolved over 10 years. Research codes are challenging because there is a desire to balance agility with discipline as well as compliance with DOE standards. Agility is important to allow experimentation with new algorithms and addition of the latest physics features. Discipline is important to increase the quality of the codes. Automation of processes and defect prevention/detection are deployed throughout the software development process. Since resarch codes are a small segment of the software industry, not much information exists in terms of reliability studies on these types of codes. This paper describes attempts to determine the goodness of these research codes. Goodness defined as both correctness of the codes and their fault densities. Correctness is determined by user interviews, peer review; feature based automated testing, and coverage measurement. This paper focuses on the fault density aspect of goodness and reliability of the codes in particular. The approach taken was to use multiple fault density prediction methods and compare results to actual experimentation and other industry studies on fault density. As a result of the predictions and experiments our confidence in the prediction methods was increased and our confidence in the goodness of the code from a fault density perspective was given more context. A large unintended benefit of these experiments was to find defects hidden for years in the codes when using the Monte Carlo reliability testing results to develop heuristic based bug driven tests.

99 GENERAL AND MISCELLANEOUS↗

Machine-Learning-Based Mapping and Modeling of Solar Energy with Ultra-High Spatiotemporal Granularity

Despite the rapid growth of solar energy, we still lack a dynamic, high-fidelity database that tracks the spatiotemporal variations of solar PVs and their associated infrastructures across different places at a spatially resolved scale. The absence of such data presents a barrier to various applications such as solar PV growth projection, solar energy integration, solar incentive design, and climate risk assessment. In this project, we aim to bridge this gap by developing AI-based algorithms to extract granular information about solar PV installations and their associated infrastructures (i.e., distribution grids) from widely available unstructured data like remote sensing images and street views. As a result, we have built the Solar Energy Atlas, a fine-grained, large-scale geospatial overlay of distributed solar PVs and distribution grids. On top of it, we have advanced the understanding of solar adoption and distribution grid vulnerability to climate-induced extremes. Our major contributions can be summarized as follow: (1) By developing new AI algorithms, we have built the most comprehensive solar PV spatiotemporal database covering the entire US. This is the first time we obtained the exact GPS locations, size, subtype, and installation year information for rooftop solar PVs across the US. This database can be used for solar PV growth projection, solar energy integration, solar energy policy analysis and design, and spatially-resolved climate risk assessment. (2) Leveraging this database, we have uncovered the socioeconomic driving factors that are correlated with earlier onset of solar adoption and higher saturated adoption levels. We have identified the heterogeneity in the effects of different types of financial incentives on solar adoption and provided implications for tailoring incentive design based on local income levels to promote equitable solar adoption. (3) We have developed a distribution grid GIS mapping algorithm which can obtain granular geospatial and topology information about distribution grids using multi-modal open data, reducing the dependency on hard-to-obtain smart meter data of conventional approaches. It shows effectiveness in both the U.S. and Sub-Saharan Africa. Using this algorithm, we have uncovered the non-uniform vulnerability of distribution grids to wildfires in California in the aspects of undergrounding protection and Distributed Energy Resources (DER) preparedness. This has provided important implications for improving the affordability and equity of grid adaptation approaches. (3) We have made our produced database publicly available and provided user-friendly interface to enable various stakeholders and the general public to interact with the data. We have also integrated the produced data into the Data Commons platform to enable the public to access the data and correlate it with other location-specific characteristics simply using natural language as queries. The impact of our project is three-fold: (1) New algorithms for mapping solar PVs and distribution grids across space and time, which are open source to facilitate researchers and industry; (2) New databases of solar PVs and distribution grids that have been made publicly available for engineering, social, and policy applications; (3) New understandings and actionable insights on the potential approaches to promoting solar adoption and reducing energy infrastructure vulnerabilities. In this report, we start by discussing the project background and motivation (section 5), followed by the overview of project objectives (section 6). Results and discussion for each task are presented in section 7. Significant accomplishments are summarized in section 8. This report will be concluded by discussing the paths forwards (section 9), products (section 10), and team roles (section 11).

14 SOLAR ENERGY↗

The Accelerator Neutrino Neutron InteractionExperiment: Design, Construction and Preparationfor First Physics

This thesis covers work performed for the Accelerator Neutrino Neutron Interaction Experiment (ANNIE) between 2015 and 2020. ANNIE is a beamline neutrino experiment located at the Fermi National Accelerator Laboratory in Chicago. Its primary physics goals are the measurement of neutrino interaction cross-sections on oxygen, and a measurement of the multiplicity of final state neutrons from these interactions, both as functions of final state kinematics. In order to achieve these goals ANNIE will be pioneering the use of Large Area Picosecond Photodetectors and the use of gadolinium doping to enhance neutron visibility in a water Cherenkov detector. Over the time period of this thesis the ANNIE detector was constructed, an initial 'Phase I' background run was completed, the detector was decommissioned and then substantially upgraded in preparation for the main 'Phase II' physics run. At the time of writing ANNIE is being re-commissioned, with Phase II data taking anticipated to begin in 2021. During this time I spent 21/2 years on-site contributing to almost all aspects of hardware work, while performing software development for the simulation and analysis frameworks. As my first software task I developed Monte Carlo simulations of the detector response, based on the WCSim framework. While thismore » framework provided a starting point for the tank simulation significant modifications were required to implement the complete ANNIE detector system. This included a rewrite of the geometry, along with modifications to the photosensor definitions, hit collections, digitization, triggering, input and output interfaces, etc. To take advantage of improved modelling of neutrons in Geant4.10 the physics lists were also reimplemented in the new version format. As part of this work I determined performance metrics based on the output files, calculating kinematic acceptance, detector efficiency and expected event rates. My contributions on reconstruction focused primarily on the MRD, for which I implemented two track reconstruction algorithms, specialised for high and low energy muons. The high energy algorithm achieves >90% efficiency for tracks with at least 8 hits, and is robust against dead paddles and unrelated hits. The low energy algorithm performs a search in combination with tank information to provide sensitivity down to a single MRD hit. Early data handling work for the Phase I background run included writing the code for handling TDC hits and performing studies on neutron detection without the requirement of multi-hit coincidence. As the collaboration moved to a new framework for the Phase II analysis I performed a substantial amount of work to develop the codebase, implementing tools to perform tasks such as file IO, data parsing, waveform processing, event building, timestamp alignment, time clustering, hit residual calculation, event visualisation etc. I helped develop the python API for integrating multi-language tools, and ported the framework into a Singularity container to provide a self-contained environment that…« less

43 PARTICLE ACCELERATORS↗

Navigating Team Dynamics: Automated Detection of Micro-Behaviors Between Team Members Through Longitudinal Interaction Data

The success in future long term space exploration missions will depend on the cooperation, coordination, and mutual understanding among the crew members. Micro-behaviors are momentary, subtle linguistic and paralinguistic indicators of thinking and feeling toward another member of the team (Cortina et al., 2001; Smith & Griffiths, 2022) that can significantly impact team dynamics and influence the overall team performance (Paromita & Chaspari, 2024). Due to their interactive nature, micro-behaviors have a sender (i.e., the team member expressing the micro-behavior) and a target (the team member impacted by the micro-behavior). Detection of these behaviors can assist in avoiding possible conflict among crew members and promoting the overall team success. Our prior research focused on an initial proof of concept of machine learning (ML) models and natural language processing (NLP) techniques that were used for automatically detect micro-behaviors among crew members of the US National Aeronautics and Space Administration’s (NASA) Human Exploration Research Analog (HERA) Campaigns 4 and 5 missions (Paromita et al., 2023). Results underscored the importance of incorporating contextual information in the ML models in the form of sentiment analysis, type of task, and dyadic interaction among team members. Here, we expand the scope of our prior work in two ways. First, we assess ML/NLP methods on new behavioral annotations coded using an adapted version of Smith & Griffins (2022) theoretical framework in terms of Violation (i.e., presence of valenced behavior, uplifting/positive or discouraging/negative), Intensity (i.e., force of behavior in terms of how uplifting or discouraging is the behavior), and Intent (i.e., motive of the behavior in terms of whether it was deliberate or unintentional). Second, we expand the design of the ML model to preserve information about the role of each team member within the occurrence of the micro-behavior (in contrast to the previous model that only considered the sender and the target without determining the team member role). This allows to consider all team members' contributions in the conversation and model long-term dependencies in the dialogue. Our experiments for this study are conducted on data from 5 teams of the NASA HERA C4 (NASA grant NNX16AQ48G (PI: Bell)). Conversations were extracted from the 1.5 hour Team Interaction Battery (TIB) task that occurred 5 times in-mission per crew. This resulted in a total of 13,058 conversational turns (i.e., 17.8% uplifting, 3.3% discouraging, 75.76% neutral, 3.14% nulls). Our findings with the revised behavioral coding and ML/NLP models indicate a 43.66% macro F1-score (i.e., 38.29% precision (P), 50.8% recall (R)) for a dialog state-tracking model that includes information from the sender only, and a 40.9% F1-score (i.e., 38.7% P, 43.36% R) for the same model that includes information from both the sender and the target of the micro-behavior. These are significantly higher compared to simple random forest models that classify behaviors strictly based on speech content and do not consider iterative team dynamics, achieving a 36.07% F1-score (i.e., 39.04% R, 33.53% P). Our findings demonstrate potential ways to leverage large conversational datasets to better capture complex team dynamics. We will discuss future directions including proposed models that can incorporate additional mission days and tasks beyond the TIB for objectively quantifying team behavior at high temporal resolution in space exploration missions.

Projna Paromita↗

On the nature and evolution of the neural bases of human language

The traditional theory equating the brain bases of language with Broca's and Wernicke's neocortical areas is wrong. Neural circuits linking activity in anatomically segregated populations of neurons in subcortical structures and the neocortex throughout the human brain regulate complex behaviors such as walking, talking, and comprehending the meaning of sentences. When we hear or read a word, neural structures involved in the perception or real-world associations of the word are activated as well as posterior cortical regions adjacent to Wernicke's area. Many areas of the neocortex and subcortical structures support the cortical-striatal-cortical circuits that confer complex syntactic ability, speech production, and a large vocabulary. However, many of these structures also form part of the neural circuits regulating other aspects of behavior. For example, the basal ganglia, which regulate motor control, are also crucial elements in the circuits that confer human linguistic ability and abstract reasoning. The cerebellum, traditionally associated with motor control, is active in motor learning. The basal ganglia are also key elements in reward-based learning. Data from studies of Broca's aphasia, Parkinson's disease, hypoxia, focal brain damage, and a genetically transmitted brain anomaly (the putative "language gene," family KE), and from comparative studies of the brains and behavior of other species, demonstrate that the basal ganglia sequence the discrete elements that constitute a complete motor act, syntactic process, or thought process. Imaging studies of intact human subjects and electrophysiologic and tracer studies of the brains and behavior of other species confirm these findings. As Dobzansky put it, "Nothing in biology makes sense except in the light of evolution" (cited in Mayr, 1982). That applies with as much force to the human brain and the neural bases of language as it does to the human foot or jaw. The converse follows: the mark of evolution on the brains of human beings and other species provides insight into the evolution of the brain bases of human language. The neural substrate that regulated motor control in the common ancestor of apes and humans most likely was modified to enhance cognitive and linguistic ability. Speech communication played a central role in this process. However, the process that ultimately resulted in the human brain may have started when our earliest hominid ancestors began to walk.

Review, Tutorial↗

Award ER25844: Minimizing System Noise Effects for Extreme-Scale Scientific Simulation Through Function Delegation

In software running on distributed computing clusters, time spent on communication between nodes in the cluster can be a significant portion of the overall computation time; background operating system tasks and other computational noise on the nodes of the system can have a significant impact on the amount of time this communication takes, especially on large systems. The research completed in this period has improved understanding of when such noise will have a significant impact. Specifically, it was demonstrated that not just noise on the nodes, but also noise on the network between nodes can have a significant impact on computation time. It was also demonstrated that noise patterns matter more than noise intensity: very regular noise can cause less disruption than lighter (on average) but less regular noise. It was also demonstrated that the effect of noise is more prominent as the speed of the network between nodes is increased. Furthermore, a tracing tool, Netgauge, was improved via our work, and a system simulator, LogGOPSim, was developed; they can be used by application developers to improve performance of their program and by system designers to mitigate the effects of noise by adjusting the noise characteristics of the operating system. Both have been made freely available as open source programs. In the course of developing these tools, we demonstrated weaknesses in existing methodologies for modeling communication, and we introduced a more detailed model, LogGOPS, for simulating systems. Not only were the deleterious effects of noise explored but we have also offered solutions. Our studies of simulations of system noise have led to specific recommendations on tuning systems to mitigate noise. We have also improved existing approaches to mitigating noise. Non-blocking collective communication avoids the effects of noise by letting communication continue simultaneously with computation (thus being non-blocking), so that the delays in communication introduced by noise have a smaller impact on overall computation time. Potentially, noise can be reduced much further by offloading communication tasks to a separate processing element than the operating system is using. We have improved our library LibNBC, which provides an implementation of non-blocking collectives, via this work. During this research, our proposal to include non-blocking collectives (which used LibNBC as a reference implementation) in the upcoming MPI-3 standard was accepted. As MPI is a ubiquitous and important standard for communication in parallel computing, this demonstrates a certain acceptance of the practicality and desirability of non-blocking collectives. Now that non-blocking collectives are a part of the standard we can expect to see optimized platform-specific implementations of non-blocking collectives. Also as part of this work we have also developed a language GOAL (Global Operation Assembly Language) that can be used as a starting point for defining languages to express optimized communication algorithms.

97 MATHEMATICS AND COMPUTING↗

ARCH: Large-scale knowledge graph via aggregated narrative codified health records analysis

Objective: Electronic health record (EHR) systems contain a wealth of clinical data stored as both codified data and free-text narrative notes (NLP). The complexity of EHR presents challenges in feature representation, information extraction, and uncertainty quantification. Here, to address these challenges, we proposed an efficient Aggregated naRrative Codified Health (ARCH) records analysis to generate a large-scale knowledge graph (KG) for a comprehensive set of EHR codified and narrative features. Methods: Using data from 12.5 million Veterans Affairs patients, ARCH first derives embedding vectors and generates similarities along with associated p-values to measure the strength of relatedness between clinical features with statistical certainty quantification. Next, ARCH performs a sparse embedding regression to remove indirect linkage between features to build a sparse KG. Finally, ARCH was validated on various clinical tasks, including detecting known relationships between entity pairs, predicting drug side effects, disease phenotyping, as well as sub-typing Alzheimer’s disease patients. Results: ARCH produces high-quality clinical embeddings and KG for over 60,000 codified and narrative EHR concepts. The KG and embeddings are visualized in the R-shiny powered web-API.3 ARCH achieved high accuracy in detecting EHR concept relationships, with AUCs of 0.926 (codified) and 0.861 (NLP) for similar EHR concepts, and 0.810 (codified) and 0.843 (NLP) for related pairs. It detected drug side effects with a 0.723 AUC, which improved to 0.826 after fine-tuning. Using both codified and NLP features, the detection power increased significantly. Compared to other methods, ARCH has superior accuracy and enhances weakly supervised phenotyping algorithms’ performance. Notably, it successfully categorized Alzheimer’s patients into two subgroups with varying mortality rates. Conclusion: The proposed ARCH algorithm generates large-scale high-quality semantic representations and knowledge graph for both codified and NLP EHR features, useful for a wide range of predictive modeling tasks.

Electronic health records↗

Distributions of the Mean for Lognormally Distributed Data - 19427

In environmental sciences, a distribution of the mean is often needed for right-skewed data. For example, risk assessments are often based on distributions of the mean to obtain upper confidence limits (UCLs) for analyte concentrations, and, in probabilistic performance assessment (PPA) models, parameters are given distributions rather than deterministic values, and they are usually given distributions of their mean rather than the population distributions to account for the large spatial and temporal domains to which they are applied. How this distribution of the mean is developed depends on the characteristics of the parameter's full distribution and how the parameter relates to its use. In this presentation, I will discuss the implications of a right skew in distribution development. Parameters may vary over orders of magnitude and/or may be right-skewed. The material property of saturated hydraulic conductivity, Ks, is a prime example of such a parameter. Right-skewed parameters are often assumed to follow lognormal distributions. If X follows a lognormal distribution, then log(X) follows a normal distribution. For some right-skewed parameters, practitioners go back and forth between the two forms of the data depending on their needs. For example, it may be preferred to use the untransformed version of X in a model or equation, but it is easier to visualize the parameter values on a logarithmic scale. However, distribution development for the average value of the parameter will differ depending on whether X or log(X) is modeled. This difference is due to Jensen's Inequality, a theorem which states that the expected value of a concave function of a variable is less than or equal to the same function of the expected value of the variable; a logarithm is a concave function. In other words, the mean of log(X) is always less than or equal to the log of the mean of X and therefore the mean estimated from a distribution based on log(X) is less than or equal to the mean estimated from a distribution based on X. If a parameter is naturally considered on a log scale by subject matter experts, then distributions developed for the mean of log(X) may be more consistent with the use of the parameters in the literature than distributions developed for X. The statistical decisions of distribution development will be discussed within the context of right-skewed, lognormal data. These issues include method-of-moments versus maximum likelihood estimation of the distribution parameters and the use of bootstrapping to aid in the visualization of a distribution of the mean. The details surrounding Jensen's Inequality will be explained and visualized within the context of distribution development of a right-skewed parameter such as saturated hydraulic conductivity. (authors)

54 ENVIRONMENTAL SCIENCES↗

Model Predictive Control of HVAC Systems: Implementation and Testing at the University of California, Merced

A Model Predictive Control algorithm was developed for the UC Merced campus chilled water plant. Model predictive control (MPC) is an advanced control technology that has proven successful in the chemical process industry and other industries. The main goal of the research was to demonstrate the practical and commercial viability of MPC for optimization of building energy systems. The control algorithms were developed and implemented in MATLAB, allowing for rapid development, performance, and robustness assessment. The UC Merced chilled water plant includes three water-cooled chillers and a two million gallon chilled water storage tank. The tank is charged during the night to minimize on-peak electricity consumption and take advantage of the lower ambient wet bulb temperature. The control algorithms determined the optimal chilled water plant operation including chilled water supply (CHWS) temperature set-point, condenser water supply (CWS) temperature set-point and the charging start and stop times to minimize a cost function that includes energy consumption and peak electrical demand over a 3-day prediction horizon. A detailed model of the chilled water plant and simplified models of the buildings served by the plant were developed using the equation-based modeling language Modelica. Steady state models of the chillers, cooling towers and pumps were developed, based on manufacturers performance data, and calibrated using measured data collected and archived by the control system. A detailed dynamic model of the chilled water storage tank was also developed and calibrated. Simple, semi-empirical models were developed to predict the temperature and flow rate of the chilled water returning to the plant from the buildings. These models were then combined and simplified for use in a model predictive control algorithm that determines the optimal chiller start and stop times and set-points for the condenser water temperature and the chilled water supply temperature. The report describes the development and testing of the algorithm and evaluates the resulting performance, concluding with a discussion of next steps in further research. The experimental results show a small improvement in COP over the baseline policy but it is difficult to draw any strong conclusions about the energy savings potential for MPC with this system only four days of suitable experimental data were obtained once correct operation of the MPC system had been achieved. These data show an improvement in COP of 3.1% {+-} 2.2% relative to a baseline established immediately prior to the period when the MPC was run in its final form. This baseline includes control policy improvements that the plant operators learned by observing the earlier implementations of MPC, including increasing the temperature of the water supplied to the chiller condensers from the cooling towers. The process of data collection and model development, necessary for any MPC project, resulted in the team uncovering various problems with the chilled water system. Although it is difficult to quantify the energy savings resulting from these problems being remedied, they were likely on the same order as the energy savings from the MPC itself. Although the types of problems uncovered and the level of energy savings may differ significantly from other projects, some of the benefits of detecting and diagnosing problems are expected from the use of MPC for any chilled water plant. The degree of chiller loading was found to be a key factor for efficiency. It is more efficient to operate the chillers at or near full load. In order to maximize the chiller load, one would maximize the temperature difference across chillers and the chilled water flow rate through the chillers. Thus, the CHWS set-point and the chilled water flow-rate can be used to limit the chiller loading to prevent chiller surging. Since the flow rate has an upper bound and the CHWS set point has a lower bound, the chiller loading is constrained and often determined by the chilled water return temperature (CHWR). The CHWR temperature is primarily comprised of warm water from the top of the TES tank. The CHWR temperature falls substantially as the thermocline approaches the top of the tank, which reduces the chiller loading. As a result, it has been determined that overcharging the TES tank can be detrimental to the chilled water plant efficiency. The resulting MPC policy differs from the current practice of fully charging the TES tank. A heuristic rule could possible avoid this problem without using predictive control. Similarly, the COP improvements from the change in CWS set-point were largely captured by a static set-point change by the operators. Further research is required to determine how much of the MPC savings could be garnered through simplified rules (based on the MPC study), with and without prediction.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗