SEARCH · Search NASA

Results for “multiple instance learning”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

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

At least 37 records · Page 2

All sources NASA NTRS DOE OSTI

NASA NTRS2005

Lessons Learned From Developing Three Generations of Remote Sensing Science Data Processing Systems

The Biospheric Information Systems Branch at NASA s Goddard Space Flight Center has developed three generations of Science Investigator-led Processing Systems for use with various remote sensing instruments. The first system is used for data from the MODIS instruments flown on NASA s Earth Observing Systems @OS) Terra and Aqua Spacecraft launched in 1999 and 2002 respectively. The second generation is for the Ozone Measuring Instrument flying on the EOS Aura spacecraft launched in 2004. We are now developing a third generation of the system for evaluation science data processing for the Ozone Mapping and Profiler Suite (OMPS) to be flown by the NPOESS Preparatory Project (NPP) in 2006. The initial system was based on large scale proprietary hardware, operating and database systems. The current OMI system and the OMPS system being developed are based on commodity hardware, the LINUX Operating System and on PostgreSQL, an Open Source RDBMS. The new system distributes its data archive across multiple server hosts and processes jobs on multiple processor boxes. We have created several instances of this system, including one for operational processing, one for testing and reprocessing and one for applications development and scientific analysis. Prior to receiving the first data from OMI we applied the system to reprocessing information from the Solar Backscatter Ultraviolet (SBUV) and Total Ozone Mapping Spectrometer (TOMS) instruments flown from 1978 until now. The system was able to process 25 years (108,000 orbits) of data and produce 800,000 files (400 GiB) of level 2 and level 3 products in less than a week. We will describe the lessons we have learned and tradeoffs between system design, hardware, operating systems, operational staffing, user support and operational procedures. During each generational phase, the system has become more generic and reusable. While the system is not currently shrink wrapped we believe it is to the point where it could be readily adopted, with substantial cost savings, for other similar tasks.

Tilmes, Curt↗

NASA NTRS1997

Using APEX to Model Anticipated Human Error: Analysis of a GPS Navigational Aid

The interface development process can be dramatically improved by predicting design facilitated human error at an early stage in the design process. The approach we advocate is to SIMULATE the behavior of a human agent carrying out tasks with a well-specified user interface, ANALYZE the simulation for instances of human error, and then REFINE the interface or protocol to minimize predicted error. This approach, incorporated into the APEX modeling architecture, differs from past approaches to human simulation in Its emphasis on error rather than e.g. learning rate or speed of response. The APEX model consists of two major components: (1) a powerful action selection component capable of simulating behavior in complex, multiple-task environments; and (2) a resource architecture which constrains cognitive, perceptual, and motor capabilities to within empirically demonstrated limits. The model mimics human errors arising from interactions between limited human resources and elements of the computer interface whose design falls to anticipate those limits. We analyze the design of a hand-held Global Positioning System (GPS) device used for radical and navigational decisions in small yacht recalls. The analysis demonstrates how human system modeling can be an effective design aid, helping to accelerate the process of refining a product (or procedure).

VanSelst, Mark↗

DOE OSTI2024

Progressive transfer learning for advancing machine learning-based reduced-order modeling

Abstract To maximize knowledge transfer and improve the data requirement for data-driven machine learning (ML) modeling, a progressive transfer learning for reduced-order modeling (p-ROM) framework is proposed. A key concept of p-ROM is to selectively transfer knowledge from previously trained ML models and effectively develop a new ML model(s) for unseen tasks by optimizing information gates in hidden layers. The p-ROM framework is designed to work with any type of data-driven ROMs. For demonstration purposes, we evaluate the p-ROM with specific Barlow Twins ROMs (p-BT-ROMs) to highlight how progress learning can apply to multiple topological and physical problems with an emphasis on a small training set regime. The proposed p-BT-ROM framework has been tested using multiple examples, including transport, flow, and solid mechanics, to illustrate the importance of progressive knowledge transfer and its impact on model accuracy with reduced training samples. In both similar and different topologies, p-BT-ROM achieves improved model accuracy with much less training data. For instance, p-BT-ROM with four-parent (i.e., pre-trained models) outperforms the no-parent counterpart trained on data nine times larger. The p-ROM framework is poised to significantly enhance the capabilities of ML-based ROM approaches for scientific and engineering applications by mitigating data scarcity through progressively transferring knowledge.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Semi-automatic image annotation using 3D LiDAR projections and depth camera data

Efficient image annotation is necessary to utilize deep learning object recognition neural networks in nuclear safeguards, such as for the detection and localization of target objects like nuclear material containers (NMCs). This capability can help automate the inventory accounting of different types of NMCs within nuclear storage facilities. The conventional manual annotation process is labor-intensive and time-consuming, hindering the rapid deployment of deep learning models for NMC identifications. This paper introduces a novel semi-automatic method for annotating 2D images of nuclear material containers (NMCs) by combining 3D light detection and ranging (LiDAR) data with color and depth camera images collected from a handheld scan system. The annotation pipeline involves an operator manually marking new target objects on a LiDAR-generated map, and projecting these 3D locations to images, thereby automatically creating annotations from the projections. The semi-automatic approach significantly reduces manual efforts and the expertise in image annotation that is required to perform the task, allowing deep learning models to be trained on-site within a few hours. The paper compares the performance of models trained on datasets annotated through various methods, including semi-automatic, manual, and commercial annotation services. The evaluation demonstrates that the semi-automatic annotation method achieves comparable or superior results, with a mean average precision (mAP) above 0.9, showcasing its efficiency in training object recognition models. Additionally, the paper explores the application of the proposed method to instance segmentation, achieving promising results in detecting multiple types of NMCs in various formations.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

DOE OSTI2025

Investigation of the Performance and Explainability Tradeoffs for Machine-Learning Models for Predictive Maintenance of Circulating Water Systems in Nuclear Power Plants

Predictive maintenance (PdM) has shown great potential for achieving substantial cost savings and enhancing the economic competitiveness of nuclear power plants (NPPs) in today's energy market. Among the different modeling approaches that exist, machine learning (ML) tools in particular have a demonstrated ability to handle high dimensional and multivariate data and to extract hidden relationships within data in industrial environments. While ML methods show great potential, their lack of explainability---especially for black-box models---is a major hurdle to their adoption. Moreover, considering the supposed trade-off between explainability and performance challenges, careful consideration must be made as to which of these quality aspects takes precedence in light of multiple modeling options, resource availability, and domain characteristics. The present work evaluates the performance of six ML models, each with a different degree of explainability, in classifying the conditions of circulating water pumps (CWPs) by utilizing sensor data from nuclear power plants. To determine the drivers behind the trade-offs presented by this array of models, this work also tests different combinations of CWP units as the training and testing data, degrees of data imbalance, and objective functions for hyperparameter tuning. It was found that black-box models tend to afford superior performance in cases where there are far more instances of one type of labeled data than of any other type. It is recommended that a guided procedure be followed for designing and delivering an ML system that is sufficiently explainable to all involved stakeholders.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗

NASA NTRS2001

Learning Sequences of Actions in Collectives of Autonomous Agents

In this paper we focus on the problem of designing a collective of autonomous agents that individually learn sequences of actions such that the resultant sequence of joint actions achieves a predetermined global objective. We are particularly interested in instances of this problem where centralized control is either impossible or impractical. For single agent systems in similar domains, machine learning methods (e.g., reinforcement learners) have been successfully used. However, applying such solutions directly to multi-agent systems often proves problematic, as agents may work at cross-purposes, or have difficulty in evaluating their contribution to achievement of the global objective, or both. Accordingly, the crucial design step in multiagent systems centers on determining the private objectives of each agent so that as the agents strive for those objectives, the system reaches a good global solution. In this work we consider a version of this problem involving multiple autonomous agents in a grid world. We use concepts from collective intelligence to design goals for the agents that are 'aligned' with the global goal, and are 'learnable' in that agents can readily see how their behavior affects their utility. We show that reinforcement learning agents using those goals outperform both 'natural' extensions of single agent algorithms and global reinforcement, learning solutions based on 'team games'.

Turner, Kagan↗

DOE OSTI2024

The ReSWARM microgravity flight experiments: Planning, control, and model estimation for on‐orbit close proximity operations

Abstract On‐orbit close proximity operations involve robotic spacecraft maneuvering and making decisions for a growing number of mission scenarios demanding autonomy, including on‐orbit assembly, repair, and astronaut assistance. Of these scenarios, on‐orbit assembly is an enabling technology that will allow large space structures to be built in situ, using smaller building block modules. However, like many of these scenarios, robotic on‐orbit assembly involves several technical hurdles, such as changing system models. For instance, grappled modules moved by a free‐flying “assembler” robot can cause significant changes in the combined system inertia, which have cascading impacts on motion planning and control portions of the autonomy stack. Further, on‐orbit assembly and other scenarios require collision‐avoiding motion planning, particularly when operating in a “construction site” scenario of multiple assembler robots and structures. Multiple key technologies that address these complicating factors for autonomous microgravity close proximity operations are detailed in this work, in particular: (1) application of global long‐horizon planning, accomplished using offline and online sampling‐based planner options that consider the system dynamics; (2) adaptation of the recently proposed RATTLE information‐aware planning framework for on‐orbit reconfiguration model learning; and (3) connection with robust control tools to provide low‐level control robustness using current system knowledge. These approaches were demonstrated for an autonomous on‐orbit assembly use case by the RElative Satellite sWarming and Robotic Maneuvering (ReSWARM) experiments using NASA's Astrobee robots on the International Space Station. Results of the ReSWARM experiments are provided along with significant operational and implementation detail discussing the practicalities of hardware implementation and unique aspects of working with the Astrobee free‐flyer robots in microgravity. ReSWARM provides a base set of planning and control tools for robotic close proximity operations, demonstrates them in microgravity, and outlines some of the important hardware aspects that future autonomous free‐flyers will need to consider.

Robotics↗

DOE OSTI2024

ReLIC: Full-Scale Realization of Reinforcement Learning for Infrastructure Control

Prior efforts have shown that deep reinforcement learning (DRL) may provide a new method for controlling networked power systems. Though successful, prior approaches have not yet demonstrated their behavior on systems of realistic scale. This effort examined multiple theoretical and technical approaches to allow a DRL model to operate over a system of 2,000 buses or more. We find that allowing the DRL models to run training episodes in parallel provides near limitless efficiency gains, allowing us to train successful agents to behave on our Kuramoto transmission model of up to 4,000 buses. We further show that we can expand our PowerWorld DRL implementation to systems of up to 25 buses but struggle to go beyond this limit due to PowerWorld’s inability to run multiple instances at once. Finally, we examine a multi-agent approach and find that it performs as well if not better than our existing centralized approach.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1990

Reuse metrics and measurement: A framework

The lessons learned and experience gleaned are described by those who have started to implement the reuse metrics and measurement framework used in controlling the development of common avionics and software for its affiliated aircraft programs. The framework was developed to permit the measurement of the long term cost/benefits resulting from the creation and use of Reusable Software Objects (RSOs). The framework also monitors the efficiency and effectiveness of the Software Reuse Library (SRL). The metrics and measurement framework is defined which was established to allow some determinations and findings to be made relative to software reuse. Seven criteria are discussed which were used to guide the establishment of the proposed reuse framework. Object recapture and creation metrics are explained along with their normalized use in effort, productivity, and quality determination. A single and multiple reuse instance version of a popular cost model is presented which uses these metrics and the measurement scheme proposed to predict the software effort and duration under various reuse assumptions. Studies in using this model to predict actuals taken from the RCI data base of over 1000 completed projects is discussed.

Reifer, Donald J.↗

NASA NTRS2018

Modeling Key Predictors of Airport Runway Configurations Using Learning Algorithms

Advanced traffic flow management automation will need accurate predictions of airport runway configurations. Terminal area weather and traffic demand are generally considered to be the most significant factors in predicting runway configuration. Weather information is forecasted across multiple features, including wind direction, wind speed, gusts, cloud ceilings, visibility, temperature, and precipitation, among many others. We use machine learning techniques on historical weather and runway data to determine weather features that correlate well with runway configurations. We analyze the predictive capability of weather features using different learning models trained on data from four major U.S. airports: Atlanta (ATL), Washington – Dulles (IAD), New York – Kennedy (JFK), and San Francisco (SFO). Wind direction alone is strongly correlated with runway configurations above all other examined factors, as expected. This correlation is the most significant component of the ~80% prediction accuracy in selecting between the two most frequently used runway configurations. However, individual airports show variations on how well the runway configuration decisions correlate with wind direction. While wind direction was identified as the most significant indicator of configuration decisions in ATL, IAD, and JFK, it did not emerge as such at SFO. Traffic demand was not found to be a strong factor in predicting runway configurations at any of the airports analyzed. In rare instances, when high demand cannot be accommodated within the current configuration, temporary changes are likely to be attributable to demand. However, these occurrences are so limited in number that their overall effect is not sufficient to consider traffic demand as a major indicator of runway configuration at the airports analyzed.

Bilimoria, Karl D.↗

NASA NTRSDate not supplied

Medium-Range River Flood Forecasts Using a Long Short-Term Memory Network

River flooding and the impacts are a concern for decision makers throughout the United States. Accurate medium-range forecasts (~3-7 days) are critical for providing advanced outlooks to emergency management officials. Unfortunately, accurately forecasting rainfall-runoff and the subsequent rise and fall within rivers remain a challenge in hydrological modeling. While complex physical modeling systems are the standard for representing the hydrological processes, they are computationally demanding and can require extensive calibration. Further, uncertainties remain in the model parameters and input data. The use of machine learning can reduce some of the computational demand while maintaining high accuracy. Therefore, this project makes use of a Long Short-Term Memory (LSTM) network which explicitly accounts for the time-dependent nature of rainfall-runoff modeling. The developed LSTM was trained to predict river gauge height, or stage height, based on time-lagged input features which include: gauge height to initialize the model, the NASA Short-term Prediction Research and Transition Center’s instance of the Land Information System (SPoRT-LIS) relative soil moisture to describe the rainfall infiltration rate, and 6-hr Multi-Radar Multi-Sensor quantitative precipitation estimate (MRMS QPE). The developed LSTM based system is then used to produce 7-day forecasts with a 6-hr temporal resolution using three different quantitative precipitation forecasts (QPF) from the NWS’s Weather Prediction Center (WPC), the NCEP Global Forecast System (GFS) model and the National Blend of Models (NBM). This trained modeling system has been implemented as an experimental product at over 100 different rivers in collaboration with at multiple National Weather Service (NWS) Forecast Offices and River Forecast Centers (RFC) across the eastern half of the United States. The developed LSTM model achieved average Nash-Sutcliffe efficiency (NSE) 0.89 higher than the equivalent medium-range National Water Model ensemble member forecast over a 7-day forecast. In addition to the initial development and evaluation, this project has continued to expand. While the initial model was developed for precipitation dominated basins, expansion of the project has taken it to basins effected by snow melt. This presentation will provide an overview of the project with focus on recent developments on incorporating snow melt processes into the model.

Andrew T. White↗

NASA NTRSDate not supplied

A Census of Severe Weather as Observed From Aqua: Visible/IR and Passive-Microwave Perspectives of Severe Convection

Severe weather phenomena represent the extreme upper end of the spectrum of convection and precipitation and tend to be highly localized and relatively rare compared to the rest of the distribution, but they can cause damage and loss disproportionate to their scale and frequency. Fortunately, severe convection exhibits distinct signatures in spaceborne remote-sensing datasets (e.g. overshooting cloud tops in visible/IR, or brightness temperature depressions in passive-microwave imagery). Leveraging these signatures individually has become a long-established practice to detect, analyze and establish climatologies of severe thunderstorms, especially in instances where traditional ground-based data may be unavailable. Spaceborne visible/IR and passive-microwave approaches are not without their pitfalls, however: passive-microwave channels have large footprints and exhibit non-uniform beam filling. Visible/IR instruments have fine horizontal resolution but are limited by their insensitivity to processes occurring below cloud top. To address this, we investigate the nearly simultaneous and colocated MODIS (visible/IR) and AMSR-E (passive-microwave) onboard the Aqua satellite to leverage both datasets together and assess the extent to which these datasets can be combined to improve severe thunderstorm detection. We pair AMSR-E and MODIS signatures of severe convection with ground-based weather radar, severe weather reports, and environmental parameters defined by the MERRA-2 reanalysis in six different geographical regimes throughout the Aqua domain. We present a census of potentially severe convective storms and their environments as seen by multiple instruments simultaneously, investigating how MODIS and AMSR-E signatures may be used together to diagnose storm properties and processes, and how the interrelationships between the signatures varies seasonally and geographically. Using statistical machine learning analysis, we aim to quantify the optimal MODIS and AMSR-E parameter sets for discriminating severe from non-severe storm cells and assess what improvement (if any) in detection results from combining the IR, visible, and microwave datasets.

Sarah Bang↗

NASA NTRSDate not supplied

Examining Artifacts from GLOBE Program Research Symposia & Using Network Analysis Techniques to Characterize Students’ Authentic STEM Investigations

For the past several years, the GLOBE Program's International Virtual Science Symposia (IVSS) and Student Research Symposia (SRS) have provided opportunities for U.S. and international students to present their Earth science research investigations to the GLOBE community through online or in-person events. This presentation will share the techniques and findings of an evaluation study that used student posters and written reports to characterize their research investigations through multiple lenses and frameworks. The study began with a list of characteristics drawn from a literature review, an analysis of sample projects, and several reviews by expert stakeholders and scientists, which comprehensively covered diverse relevant frameworks including citizen science, student STEM learning through authentic experiences, and The GLOBE Program model. Once applied to 207 student projects, this list of codes revealed the frequency and prevalence of various qualities and experiences represented by GLOBE student research investigations. An innovative application of social network analysis techniques to the coded dataset revealed frequently cooccurring characteristics. This networking approach identified and conceptually mapped several "clusters" of characteristics that typified student projects, empirically based on the submitted projects themselves. The basic quantitative investigation of frequencies indicates the extent to which various characteristics are present in - or absent from - GLOBE SRS and IVSS projects, while the network analysis provides a descriptive framework for typifying projects. Ultimately, the descriptive framework fostered a suite of assessment tools to help The GLOBE Program's staff, scientists, and research project judges understand the diversity of student research projects. GLOBE can use these tools to identify and respond to areas of need; for instance, the descriptive framework illustrates the potential for further education and training resources related to data analysis, interpretation of data, and credibility of scientific claims. This presentation will share the novel utilization of network analysis techniques to holistically assess and react to student research contributions.

Ann Martin↗

NASA NTRSDate not supplied

Utilization of Machine Learning Techniques for Managing the Tracking and Data Relay Satellite Constellation

National Aeronautics and Space Administration’s (NASA) Goddard Space Flight Center (GSFC) operates a constellation of ten geosynchronous Tracking and Data Relay Satellites (TDRS). The TDRS constellation consists of multiple geosynchronous communication relay satellites located around the equator so they can provide continual coverage of any mission in low earth orbit. The TDRS are located primarily in three oceanic regions around the earth. NASA’s White Sands Complex provides the ground communication support for TDRS located over the Atlantic and Pacific Oceans. Another TDRS ground station in Guam supports the TDRS over the Indian Ocean. With these satellites the TDRS network can provide continuous coverage of satellites in low-earth orbit. The NASA Space Network (SN) project office at GSFC manages the constellation of spacecraft. Major customers of the TDRS constellation include, but are not limited to, the International Space Station and the Hubble Space Telescope. The TDRS constellation has three generations of satellites and has been active for over 30 years providing reliable communication links between customer satellites and corresponding ground stations. However, one of the major concerns for TDRS, and in any space mission, is to ensure the health and safety of the spacecraft. Generally, engineers use telemetry data to monitor and analyze the performance and state of health of the spacecraft. Telemetry data contains hundreds of parameters that monitor each important component in the spacecraft, which can be utilized to recognize and characterize the behavior of the spacecraft. Each parameter contains considerable information to represent time-dependent properties of each spacecraft subsystem and component. During the entire life of a TDRS spacecraft, thousands of gigabytes of telemetry data are transmitted in real-time from the spacecraft to the ground station at the White Sands Complex in Las Cruces, New Mexico, and recorded as historical data sets for engineers to process and analyze the events that occurred on-orbit. These parameters contain the function of multiple spacecraft subsystems, such as the attitude control system (ACS), Thermal, Electrical Power Subsystem (EPS), etc. . The first and second generations have exceeded their required lifetime and NASA is keen to manage these spacecrafts carefully in order to maximize the remaining life using the spacecraft telemetry. The challenge is to know when the risk of losing a spacecraft in geosynchronous orbit exceeds the benefit of continued operations for customer support. In the TDRS fleet, the EPS is the most critical subsystem related to spacecraft operations. Failure of the EPS would strand a spacecraft in geosynchronous orbit. Since EPS provides power to the spacecraft, component failures ultimately lead to the inability to support the spacecraft loads and the communications payload. For instance, TDRS-8 has several anomalies in EPS including the Bus Voltage Limiter (BVL) shunt current, solar array loss of circuits, and failed battery cells. Any of these anomalies can cause critical issues to the spacecraft. Therefore, developing a system to analyze and perform early detection of a potential anomaly is an important issue in telemetry data analysis. In recent years, Telemetry Mining (TM) has been proposed to process telemetry data by using Data Mining (DM) techniques such as classification, clustering, regression and anomaly detection. Anomaly detection, also known as outlier detection, has been widely used in many data mining areas such as remote sensing, medical data processing and digital image processing. The goal of anomaly detection is to detect abnormal data, which contains a relatively low probability of occurrence among the entire data set. Early detection of anomalies is one of the most significant issues in managing the spacecraft configuration. If anomalies can be detected early enough, then the redundant resources can be used to extend the life of the operational spacecraft. We present an unsupervised anomaly detection method to process the EPS data extracted from TDRS-8. This is different from traditional analytical methods, which use telemetry data to illustrate behavior and physical meaning of each spacecraft component. TM connects multiple parameters as a vector and then conducts data analysis on this high dimension telemetry vector. This method is looking at the properties of a high dimensional vector that is able to consider the relationship between different parameters in the anomaly detection problem. This kind of method performs much better than the traditional limit checking method. In addition, we propose a new approach of real-time anomaly detection to process telemetry data in real-time, which can then be applied to spacecraft monitoring with high reliability, low cost and high accuracy.

Machine Learning (ML)↗

NASA NTRSDate not supplied

Advanced Diagnostics for Megahertz Imaging of Mixing, Fuel Spray, and Combustion Processes for Rotating Detonation Combustors

Recent advancements in megahertz rate, high-power, burst-mode laser technology are leveraged to perform and explore imaging measurements that spatially and temporally resolve the mixing, combustion, and detonation flow field in two laboratory-scale rotating detonation combustors (RDCs). In a non-premixed annular RDC, multiple imaging diagnostics are explored to investigate gaseous and liquid injector behavior, the detonation wave structure, and the propellant refill. In one instance, OH planar laser-induced fluorescence (OH-PLIF) imaging is performed up to a 2 MHz repetition rate to track the combustion products and reaction zone locations. In the same annular RDC, a single liquid fuel jet is injected, and laser-based 355-nm imaging of the fuel spray is performed up to a 1 MHz repetition rate. For this configuration, the annular RDC is used as a detonation driver to impose periodic detonation waves to interact with the fuel spray. Moreover, in this annular RDC, a range of tracer-based laser imaging measurements are explored to time-resolve the unsteady oxidizer air recovery and refill process. In a non-premixed linear RDC, planar imaging measurements of the fuel mixing are performed up to a 200 kHz repetition rate using PLIF of a tracer in the fuel supply. The fuel mixing imaging helps explain the origin of the observed pre and post wave burning, detonation structure, and enables quantifying injector recovery timescales. This paper will provide a high-level broad survey of diagnostics applied in these RDCs, lessons learned, and interesting observations.

Combustion↗

NASA NTRSDate not supplied

Advanced Diagnostics For Megahertz Imaging Of Mixing, Fuel Spray, And Combustion Processes For Rotating Detonation Combustors

Combustion↗

NASA NTRS2011

Orion Entry Handling Qualities Assessments

The Orion Command Module (CM) is a capsule designed to bring crew back from the International Space Station (ISS), the moon and beyond. The atmospheric entry portion of the flight is deigned to be flown in autopilot mode for nominal situations. However, there exists the possibility for the crew to take over manual control in off-nominal situations. In these instances, the spacecraft must meet specific handling qualities criteria. To address these criteria two separate assessments of the Orion CM s entry Handling Qualities (HQ) were conducted at NASA s Johnson Space Center (JSC) using the Cooper-Harper scale (Cooper & Harper, 1969). These assessments were conducted in the summers of 2008 and 2010 using the Advanced NASA Technology Architecture for Exploration Studies (ANTARES) six degree of freedom, high fidelity Guidance, Navigation, and Control (GN&C) simulation. This paper will address the specifics of the handling qualities criteria, the vehicle configuration, the scenarios flown, the simulation background and setup, crew interfaces and displays, piloting techniques, ratings and crew comments, pre- and post-fight briefings, lessons learned and changes made to improve the overall system performance. The data collection tools, methods, data reduction and output reports will also be discussed. The objective of the 2008 entry HQ assessment was to evaluate the handling qualities of the CM during a lunar skip return. A lunar skip entry case was selected because it was considered the most demanding of all bank control scenarios. Even though skip entry is not planned to be flown manually, it was hypothesized that if a pilot could fly the harder skip entry case, then they could also fly a simpler loads managed or ballistic (constant bank rate command) entry scenario. In addition, with the evaluation set-up of multiple tasks within the entry case, handling qualities ratings collected in the evaluation could be used to assess other scenarios such as the constant bank angle maintenance case. The 2008 entry assessment was divided into two sections (see Figure 1). Entry I was the first, high speed portion of a lunar return and Entry II was the second, lower speed portion of a lunar return, which is similar (but not identical) to a typical ISS return.

Bihari, B.↗

DOE OSTI2025

Benchmarking large language models for materials synthesis: The case of atomic layer deposition

In this work, we introduce an open-ended question benchmark, ALDbench, to evaluate the performance of large language models (LLMs) in materials synthesis, and, in particular, in the field of atomic layer deposition, a thin film growth technique used in energy applications and microelectronics. Our benchmark comprises questions with a level of difficulty ranging from the graduate level to domain expert current with the state of the art in the field. Human experts reviewed the questions along the criteria of difficulty and specificity, and the model responses along four different criteria: overall quality, specificity, relevance, and accuracy. We ran this benchmark on an instance of OpenAI’s GPT-4o. The responses from the model received a composite quality score of 3.7 on a 1–5 scale, consistent with a passing grade. However, 36% of the questions received at least one below average score. An in-depth analysis of the responses identified at least five instances of suspected hallucination. Finally, we observed statistically significant correlations between the difficulty of the question and the quality of the response, the difficulty of the question and the relevance of the response, the specificity of the question, and the accuracy of the response as graded by the human experts. Furthermore, this emphasizes the need to evaluate LLMs across multiple criteria beyond difficulty or accuracy.

Artificial intelligence↗