Results for “online learning” — Page 6

NASA NTRS2022

Scheduling For Urban Air Mobility Using Safe Learning

This work considers the scheduling problem for Urban Air Mobility (UAM) vehicles travelling between origin-destination pairs with both hard and soft trip deadlines. Each route is described by a discrete probability distribution over trip completion times (or delay) and over interarrival times of requests (or demand) for the route along with a fixed hard or soft deadline. Soft deadlines carry a cost that is incurred when the deadline is missed. An online, safe scheduler is developed that ensures that hard deadlines are never missed and that average cost of missing soft deadlines is minimized. The system is modelled as a Markov Decision Process (MDP) and safe model based learning is used to find the probabilistic distributions over route delays and demand. Monte Carlo Tree Search (MCTS) Earliest Deadline First (EDF) is used to safely explore the learned models in an online fashion and develop a near-optimal non-preemptive scheduling policy. These results are compared with Value Iteration (VI) and MCTS (Random) scheduling solutions.

Urban Air Mobility↗

NASA NTRSDate not supplied

Scheduling for Urban Air Mobility using Safe Learning

This work considers the scheduling problem for Urban Air Mobility (UAM) vehicles travelling between origin-destination pairs with both hard and soft trip deadlines. Each route is described by a discrete probability distribution over trip completion times (or delay) and over interarrival times of requests (or demand) for the route along with a fixed hard or soft deadline. Soft deadlines carry a cost that is incurred when the deadline is missed. An online, safe scheduler is developed that ensures that hard deadlines are never missed and that average cost of missing soft deadlines is minimized. The system is modelled as a Markov Decision Process (MDP) and safe model based learning is used to find the probabilistic distributions over route delays and demand. Monte Carlo Tree Search (MCTS) Earliest Deadline First (EDF) is used to safely explore the learned models in an online fashion and develop a near-optimal non-preemptive scheduling policy. These results are compared with Value Iteration (VI) and MCTS (Random) scheduling solutions.

Urban Air Mobility↗

NASA NTRS2007

Exploring NASA and ESA Atmospheric Data Using GIOVANNI, the Online Visualization and Analysis Tool

Giovanni, the NASA Goddard online visualization and analysis tool (http://giovanni.gsfc.nasa.gov) allows users explore various atmospheric phenomena without learning remote sensing data formats and downloading voluminous data. Using NASA MODIS (Terra and Aqua) and ESA MERIS (ENVISAT) aerosol data as an example, we demonstrate Giovanni usage for online multi-sensor remote sensing data comparison and analysis.

Leptoukh, Gregory↗

DOE OSTI2024

A Tutorial Set to Prepare for Science with the Vera C. Rubin Observatory

In this poster the Rubin Observatory's Community Science team (CST) presents its current suite of tutorials, which are designed to help people make use of simulated data sets in preparation for the upcoming Legacy Survey of Space and Time (LSST). We will show examples of the tutorial contents, provide custom learning modules for different astronomical fields, and describe the online environment for data analysis (the Rubin Science Platform; RSP). We will also supply a checklist for how to obtain an RSP account and access the tutorials. All are welcome to drop by the poster or the Rubin booth in the exhibit hall with questions.

79 ASTRONOMY AND ASTROPHYSICS↗

NASA NTRSDate not supplied

Online Small-Sat Knowledge Repositories and Modeling Tools for Risk Reduction and Enhanced Mission Success

Developers of small satellites face the challenges of a short development schedule, incomplete information about parts from vendors, and limited budgets. The situation is complicated by the need to use commercial electronic parts rather than radiation hardened parts to meet performance and budget constraints. Fortunately, a number of organizations are standing up new modeling and data resources that are available for free online. Motivation for sharing models, part data, and lessons learned include shorter development times, increased mission success likelihood, and training the next generation of space professionals. The purpose of this paper is to make these resources and their complementary capabilities known to the small satellite community.

Small Satellites↗

DOE OSTI2024

A Tutorial Set to Prepare for Science with the Vera C. Rubin Observatory

In this poster the Rubin Observatory's Community Science team (CST) presents its current suite of tutorials, which are designed to help people make use of simulated data sets in preparation for the upcoming Legacy Survey of Space and Time (LSST). We will show examples of the tutorial contents, provide custom learning modules for different astronomical fields, and describe the online environment for data analysis (the Rubin Science Platform; RSP). We will also supply a checklist for how to obtain an RSP account and access the tutorials. All are welcome to drop by the poster or the Rubin booth in the exhibit hall with questions.

79 ASTRONOMY AND ASTROPHYSICS↗

NASA NTRS2005

Online Bagging and Boosting

Bagging and boosting are two of the most well-known ensemble learning methods due to their theoretical performance guarantees and strong experimental results. However, these algorithms have been used mainly in batch mode, i.e., they require the entire training set to be available at once and, in some cases, require random access to the data. In this paper, we present online versions of bagging and boosting that require only one pass through the training data. We build on previously presented work by presenting some theoretical results. We also compare the online and batch algorithms experimentally in terms of accuracy and running time.

Oza, Nikunji C.↗

DOE OSTI2024

PMU Data Quality and Sensor Health Monitoring

Phasor Measurement Units (PMUs) play a critical role in the evolution of the electric power industry by providing high-precision, real-time monitoring of essential power system metrics. However, effectively detecting abnormalities and critical events from PMU data is a complex task, complicated by intricate temporal patterns, a scarcity of labeled data for training algo- rithms, and constraints on online computational power. In this study, we apply TranAD, an innovative algorithm that combines transformer architectures with the refinement of adversarial learning, to both synthetic and real-world PMU datasets for developing a data quality and sensor online health monitoring platform for utilities. Our findings reveal that TranAD not only provides efficient detection and localization but also enhances the detail with which abnormalities are detected, marking a a significant step forward in the field of clean data acquisition processes for power system monitoring

deep neural network, machine learning (ML)↗

DOE OSTI2026

Learning to Trigger: Reinforcement Learning at the Large Hadron Collider

High-throughput scientific facilities such as the Large Hadron Collider depend on real-time event filtering (\textit{triggering}) under tight constraints on bandwidth, latency, and storage. In practice, trigger menus are largely static and hand-tuned and can become suboptimal as detector conditions, pileup, and background composition drift over time. We cast online threshold tuning as a sequential decision-making problem: a reinforcement learning agent ingests streaming summaries of recent rates and signal-sensitive features and updates trigger thresholds to maximize signal efficiency while tracking a target background rate within a tolerance band. We adapt Group-Filtered Policy Optimization (GFPO) to streaming control and introduce two variants (GFPO-F, GFPO-FR) that enforce background rate feasibility during training. On a benchmark that emulates realistic collider operation, we study two representative triggers: a total transverse energy ($H_{T}$) trigger sensitive to pileup variation, and an anomaly-detection (AD) trigger based on reconstruction loss for rare or non-standard signatures. On Monte Carlo streams, our agent increases the fraction of in-tolerance time intervals by 48% ($H_T$) and 28% (AD), with a cumulative gain of up to 2% in signal efficiency on those in-tolerance intervals. Transferring from simulation to \emph{real} collision data (CMS Run 283408), the same agent, without fine-tuning, achieves a 56% ($H_T$) and 28% (AD) in-tolerance improvement over baselines, with further signal-efficiency gain on both triggers. To our knowledge, this is the \emph{first} demonstration of RL-based trigger control on real Large Hadron Collider collision data. Code is available at https://github.com/Zixind/GFPO_LHC (see repo for details).

Ding, Zixin [Chicago U.]↗

DOE OSTI2025

Machine Learning for Scalable and Optimal Load Shedding Under Power System Contingency

Prompt and effective corrective actions in response to unexpected contingencies are crucial for improving power system resilience and preventing cascading blackouts. The optimal load shedding (OLS) accounting for network limits has the potential to address the diverse system-wide impacts of contingency scenarios as compared to traditional local schemes. However, due to the fast cascading propagation of initial contingencies, real-time OLS solutions are challenging to attain in large systems with high computation and communication needs. In this paper, we propose a decentralized design that leverages offline training of a neural network (NN) model for individual load centers to autonomously construct the OLS solutions from locally available measurements. Our learning-for-OLS approach can greatly reduce the computation and communication needs during online emergency responses, thus preventing the cascading propagation of contingencies for enhanced power grid resilience. Numerical studies on both the IEEE 118-bus system and a synthetic Texas 2000-bus system have demonstrated the efficiency and effectiveness of our scalable OLS learning design for timely power system emergency operations.

24 POWER TRANSMISSION AND DISTRIBUTION↗

NASA NTRSDate not supplied

Game Based Learning For Earth Science Applications Training

Current NASA Earth capacity development programs employ mechanisms ranging from online resource sharing, and virtual and in-person trainings to share knowledge. While these programs are highly successful at engaging individuals around the world – in 2018, over 8000 individuals and over 2000 institutions from all 50 US states and over 140 countries were engaged through over 150 projects and trainings – user feedback has highlighted the desire for expanded hands-on, practical experiences in incorporating NASA EO insights with localized data and actions. We aim to address this gap by leveraging the benefits of game-based learning to build user skills in integrating NASA and local EO data to guide decisions for climate resiliency and hazard planning. This project is being executed as a two-phase crowdsourced challenge: 1) Phase 1 will require a well-researched product concept that reflects an understanding of NASA’s Earth data and tools and user needs, and proposes an innovative and interactive game or extended reality experience to train users in identifying relevant NASA data and applying insights to their climate resiliency decisions; 2) Winners of Phase 1 will be provided seed funding to develop a working prototype of the product. We aim to award 1-3 final winners to support the development of more than one game, thereby ensuring that NASA's diverse audiences around the world can access training games that best suit their needs and capabilities. This EO training game project fits in the NASA Earth Science Applied Sciences Program’s Capacity Development Program, contributing to the program mission of “helping people around the world better understand [NASA’s Earth] data and find ways to use them” (https://appliedsciences.nasa.gov/what-we-do/capacity-building). The final training game will complement existing programmatic activities of workforce development, trainings, and collaborative projects, while providing the unique value of providing interactive experiences to users and collecting real-time data and feedback to improve NASA’s Earth applications’ products and services related to climate resilience.

Human centered design↗

NASA NTRSDate not supplied

The Virginia Earth System Science Scholars (VESSS) Experience: Inspiring High School Students in Stem Through Mentorship and Collaborative Skills Development

Since 2016, the Virginia Earth System Science Scholars (VESSS) program offered by the Virginia Space Grant Consortium (VSGC) has provided an interactive science, technology, engineering, and mathematics learning experience for Virginia high school juniors and seniors. The VESSS program was created through a partnership with VSGC, NASA Langley Research Center (LaRC), and Hampton University to engage students in learning about Earth’s climate change and its global impacts. High school students from all areas of Virginia learn about Earth science and climate change in a 16 week online course focused on the Earth’s surface and interior, atmospheric composition, weather, water energy cycles, and climate variability. Students that complete the online course are invited to a no-cost seven-day residential summer academy at NASA LaRC where students are divided into teams based on four of the major spheres of study for Earth: atmosphere, biosphere, hydrosphere, and lithosphere. Each team is guided by a Master Teacher and teams of undergraduate student interns who were also previous VESSS scholars. Students interview for team positions and are given a role reflecting an actual satellite design team position such as project scientist, research scientist, budget analyst, launch and orbit analyst, contamination analyst, and education and public outreach specialist. The students spend a week mentored by NASA and industry scientists and engineers as they learn how to select science goals and objectives, select science instruments, balance a mission budget, determine the mission launch location and orbit, and communicate the importance of the mission to the public. Each team produces a mission proposal, and at the end of the week, each team presents their portion of the mission to a panel of experts, followed by a question and answer period. Through this week students learn teamwork, communication, time management, meeting management, and research skills. Project work is supplemented by team-building activities, NASA facility tours, and lectures by acclaimed guest speakers such as Dr. Christine Darden, a NASA “Hidden Figure.” Students have conveyed through course evaluations and testimonials not only the vast personal learning and growth they achieved, but also the new perspectives they gained on future STEM career fields and opportunities to explore.

Elaine Seasly↗

NASA NTRS2005

Experiments on Supervised Learning Algorithms for Text Categorization

Modern information society is facing the challenge of handling massive volume of online documents, news, intelligence reports, and so on. How to use the information accurately and in a timely manner becomes a major concern in many areas. While the general information may also include images and voice, we focus on the categorization of text data in this paper. We provide a brief overview of the information processing flow for text categorization, and discuss two supervised learning algorithms, viz., support vector machines (SVM) and partial least squares (PLS), which have been successfully applied in other domains, e.g., fault diagnosis [9]. While SVM has been well explored for binary classification and was reported as an efficient algorithm for text categorization, PLS has not yet been applied to text categorization. Our experiments are conducted on three data sets: Reuter's- 21578 dataset about corporate mergers and data acquisitions (ACQ), WebKB and the 20-Newsgroups. Results show that the performance of PLS is comparable to SVM in text categorization. A major drawback of SVM for multi-class categorization is that it requires a voting scheme based on the results of pair-wise classification. PLS does not have this drawback and could be a better candidate for multi-class text categorization.

Namburu, Setu Madhavi↗

DOE OSTI2024

HTESP (High-throughput electronic structure package): A package for high-throughput ab initio calculations

High-throughput ab initio calculations are the indispensable parts of data-driven discovery of new materials with desirable properties, as reflected in the establishment of several online material databases. The accumulation of extensive theoretical data through computations enables data-driven discovery by constructing machine learning and artificial intelligence models to predict novel compounds and forecast their properties. Efficient usage and extraction of data from these existing online material databases can accelerate the next stage materials discovery that targets different and more advanced properties, such as electron–phonon coupling for phonon-mediated superconductivity. However, extracting data from these databases, generating tailored input files for different ab initio calculations, performing such calculations, and analyzing new results can be demanding tasks. Here, in this work, we introduce a software package named “HTESP” (High-Throughput Electronic Structure Package) written in Python and Bash languages, which automates the entire workflow including data extraction, input file generation, calculation submission, result collection and plotting. Our HTESP will help speed up future computational materials discovery processes.

36 MATERIALS SCIENCE↗

NASA NTRS2024

Transitioning Sixty Years of NASA Spacesuit Knowledge Capture Lessons Learned to Searchable Knowledge Transfer Databases

Sixty years of spacesuit knowledge capture and lessons learned by spacesuit subject matter experts are documented on videos and presentations and archived with the NASA Engineering and Safety Center (NESC) Academy. The NESC Academy is a web-based platform that hosts online courses by technical experts. A process is underway to transition these decades of lessons learned from the U.S. Spacesuit Knowledge Capture Program Library into more quicky searchable databases and to proactively provide this information to those working on spacesuit projects. Hundreds of lessons learned from Project Mercury, Gemini, Apollo, Apollo- Soyuz Test Project, Skylab, Space Shuttle, International Space Station, and Artemis have been captured in a format that can be quickly searched, enabling users to find information directly applicable to their needs. Transitioning this information to a NASA wiki page will further enhance search and retrieval of data immediately useful to users. This paper provides information about how the lessons learned were determined and how to access them.

spacesuit↗

NASA NTRSDate not supplied

Transitioning Sixty Years of NASA Spacesuit Knowledge Capture Lessons Learned to Searchable Knowledge Transfer Databases

Sixty years of spacesuit knowledge capture and lessons learned by spacesuit subject matter experts are documented on videos and presentations and archived with the NASA Engineering and Safety Center (NESC) Academy. The NESC Academy is a web-based platform that hosts online courses by technical experts. A process is underway to transition these decades of lessons learned from the U.S. Spacesuit Knowledge Capture Program Library into more quicky searchable databases and to proactively provide this information to those working on spacesuit projects. Hundreds of lessons learned from Project Mercury, Gemini, Apollo, Apollo- Soyuz Test Project, Skylab, Space Shuttle, International Space Station, and Artemis have been captured in a format that can be quickly searched, enabling users to find information directly applicable to their needs. Transitioning this information to a NASA wiki page will further enhance search and retrieval of data immediately useful to users. This paper provides information about how the lessons learned were determined and how to access them.

spacesuit↗

NASA NTRS1995

Locomotion training of legged robots using hybrid machine learning techniques

In this study artificial neural networks and fuzzy logic are used to control the jumping behavior of a three-link uniped robot. The biped locomotion control problem is an increment of the uniped locomotion control. Study of legged locomotion dynamics indicates that a hierarchical controller is required to control the behavior of a legged robot. A structured control strategy is suggested which includes navigator, motion planner, biped coordinator and uniped controllers. A three-link uniped robot simulation is developed to be used as the plant. Neurocontrollers were trained both online and offline. In the case of on-line training, a reinforcement learning technique was used to train the neurocontroller to make the robot jump to a specified height. After several hundred iterations of training, the plant output achieved an accuracy of 7.4%. However, when jump distance and body angular momentum were also included in the control objectives, training time became impractically long. In the case of off-line training, a three-layered backpropagation (BP) network was first used with three inputs, three outputs and 15 to 40 hidden nodes. Pre-generated data were presented to the network with a learning rate as low as 0.003 in order to reach convergence. The low learning rate required for convergence resulted in a very slow training process which took weeks to learn 460 examples. After training, performance of the neurocontroller was rather poor. Consequently, the BP network was replaced by a Cerebeller Model Articulation Controller (CMAC) network. Subsequent experiments described in this document show that the CMAC network is more suitable to the solution of uniped locomotion control problems in terms of both learning efficiency and performance. A new approach is introduced in this report, viz., a self-organizing multiagent cerebeller model for fuzzy-neural control of uniped locomotion is suggested to improve training efficiency. This is currently being evaluated for a possible patent by NASA, Johnson Space Center. An alternative modular approach is also developed which uses separate controllers for each stage of the running stride. A self-organizing fuzzy-neural controller controls the height, distance and angular momentum of the stride. A CMAC-based controller controls the movement of the leg from the time the foot leaves the ground to the time of landing. Because the leg joints are controlled at each time step during flight, movement is smooth and obstacles can be avoided. Initial results indicate that this approach can yield fast, accurate results.

Simon, William E.↗

DOE OSTI2024

RAP: Resource-aware Automated GPU Sharing for Multi-GPU Recommendation Model Training and Input Preprocessing

Ensuring high-quality recommendations for newly onboarded users requires the continuous retraining of Deep Learning Recommendation Models (DLRMs) with freshly generated data. To serve the online DLRM retraining, existing solutions use hundreds of CPU computing nodes designated for input preprocessing, causing significant power consumption that surpasses even the power usage of GPU trainers. To this end, we propose RAP, an end-to-end DLRM training framework that supports Resource-aware Automated GPU sharing for DLRM input Preprocessing and Training. The core idea of RAP is to accurately capture the remaining GPU computing resources during DLRM training for input preprocessing, achieving superior training efficiency without requiring additional resources. Specifically, RAP utilizes a co-running cost model to efficiently assess the costs of various input preprocessing operations, and it implements a resource-aware horizontal fusion technique that adaptively merges smaller kernels according to GPU availability, circumventing any interference with DLRM training. In addition, RAP leverages a heuristic searching algorithm that jointly optimizes both the input preprocessing graph mapping and the co-running schedule to maximize the end-to-end DLRM training throughput. The comprehensive evaluation shows that RAP achieves 78.3× speedup on average over CPU-based DLRM input preprocessing frameworks. In addition, the end-to-end training throughput of RAP is only 2.04% lower than the ideal case, which has no input preprocessing overhead.

Wang, Zheng↗