Search NASA⌕ Search

SEARCH · Search NASA

Results for “learning systems”

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

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

At least 91 records · Page 5
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Human-in-the-loop Sensing and Control for Commercial Building Energy Efficiency and Occupant Comfort

Most of the existing heating, ventilation and air conditioning (HVAC) systems in commercial buildings operate in a conservative manner by assuming maximum occupancy in each room during pre-specified periods of the week, leading to significant energy being wasted as rooms are over-conditioned compared to the actual requirements of the occupants. Though critical, our understanding of occupancy patterns and thermal comfort needs of the occupants in commercial buildings is lacking and it is well known that both of these quantities are stochastic and time-varying, thus requiring sensing solutions to estimate them. This project had the goal of designing, implementing and evaluating a hardware and software solution to ameliorate this challenge. In particular, a depth camera (one whose pixels reveal distance from the camera as opposed to color values) placed on doorways is used to detect entrance and exit events from thermal zones in the building, and thereby estimate their occupancy levels. This information is then fed to a novel control algorithm that can, through interactions with the HVAC system, learn how to provide control inputs that maximize comfort and minimize energy waste. The resulting system represents a significant improvement over existing controllers for commercial HVAC systems and allowed us to improve our understanding of the design of future human-in-the-loop control solutions. For this solution to be feasible, the project had target metrics for its performance and cost. In particular, entrance and exit events for occupants moving about the building would need to be detected with an accuracy higher than 97%; and the resulting control inputs derived from this information would need to lead to approximately 10% energy savings compared to a schedule-based controller. Furthermore, regarding the final hardware design, the project had a target bill of materials (BOM) cost for the sensing solution of less than US$200 per unit while using less than 25W of power on average. All of these target metrics were met or exceeded by our final proposed solution. We performed evaluations by deploying the system in over 20 rooms of different types across 6 commercial buildings in Pittsburgh, PA over the course of three years, and performing targeted controlled experiments to test its performance along the different metrics. The human-in-the-loop control solutions (both hardware and software) developed through this project are expected to lead to significant improvements in the comfort and energy efficiency of HVAC systems used in commercial buildings. The insights we developed through the project pave the way to HVAC systems that can condition interior spaces according to their real-time utilization and the thermal comfort needs of the occupants, thereby reducing energy use. They also open up a new learning-based way of configuring HVAC controllers without having to manually fine-tune them for each building. These innovations can significantly increase the adoption of novel control solutions by the industry and thereby save resources and reduce costs of operation.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Design principles for lifelong learning AI accelerators

Lifelong learning-an agent's ability to learn throughout its lifetime-is a hallmark of biological learning systems and a central challenge for artificial intelligence (AI). The development of lifelong learning algorithms could lead to a range of novel AI applications, but this will also require the development of appropriate hardware accelerators, particularly if the models are to be deployed on edge platforms, which have strict size, weight and power constraints. Here we explore the design of lifelong learning AI accelerators that are intended for deployment in untethered environments. We identify key desirable capabilities for lifelong learning accelerators and highlight metrics to evaluate such accelerators. We then discuss current edge AI accelerators and explore the future design of lifelong learning accelerators, considering the role that different emerging technologies could play. The Perspective explores the future design of lifelong learning artificial intelligence (AI) accelerators that are intended for deployment in untethered environments, identifying key desirable capabilities for such edge AI accelerators and guidance on metrics to evaluate them.

42 ENGINEERING↗
DOE OSTI2025

Federated Learning with Frequency Estimation for Smart Meter Systems

Federated learning (FL) is a powerful framework that enables multiple distributed clients to collaborate without the need to transfer their data to a central server. However, FL does not inherently guarantee the level of privacy that clients often require. In our review of recent studies on privacy-enhancing techniques in FL, we found that frequency estimation (FE) methods remain underexplored. To address this gap, we developed and integrated FE techniques on the client side, further examining the effects of incorporating an adaptive range and a shuffled model. We also analyzed the impact of varying hyper-parameters on privacy preservation. Our results provide clear guidance on the algorithms and configurations that are most effective for enhancing privacy in FL, particularly when using long short-term memory (LSTM) architectures.

Kotevska, Olivera [ORNL] (ORCID:0000000316772243)↗
DOE OSTI2021

Evolutionary NAS in Light of Model Stability for Accurate Continual Learning

Continual learning, the capability to learn new knowledge from streaming data without forgetting the previous knowledge, is a critical requirement for dynamic learning systems, especially for emerging edge devices such as self-driving cars and drones. However, continual learning is still facing the catastrophic forgetting problem. Previous work illustrate that model performance on continual learning is not only related to the learning algorithms but also strongly dependent on the inherited model, i.e., the model where continual learning starts. The better stability of the inherited model, the less catastrophic forgetting and thus, the inherited model should be elaborately selected. Inspired by this finding, we develop an evolutionary neural architecture search (ENAS) algorithm that emphasizes the Stability of the inherited model, namely ENAS-S. ENAS-S aims to find optimal architectures for accurate continual learning on edge devices. On CIFAR-10 and CIFAR-100, we present that ENAS-S achieves competitive architectures with lower catastrophic forgetting and smaller model size when learning from a data stream, as compared with handcrafted DNNs.

Du, Xiaocong↗
DOE OSTI2024

Divide and conquer: Learning chaotic dynamical systems with multistep penalty neural ordinary differential equations

Forecasting high-dimensional dynamical systems is a fundamental challenge in various fields, such as geosciences and engineering. Neural Ordinary Differential Equations (NODEs), which combine the power of neural networks and numerical solvers, have emerged as a promising algorithm for forecasting complex nonlinear dynamical systems. However, classical techniques used for NODE training are ineffective for learning chaotic dynamical systems. In this work, we propose a novel NODE-training approach that allows for robust learning of chaotic dynamical systems. Here, our method addresses the challenges of non-convexity and exploding gradients associated with underlying chaotic dynamics. Training data trajectories from such systems are split into multiple, non-overlapping time windows. In addition to the deviation from the training data, the optimization loss term further penalizes the discontinuities of the predicted trajectory between the time windows. The window size is selected based on the fastest Lyapunov time scale of the system. Multi-step penalty(MP) method is first demonstrated on Lorenz equation, to illustrate how it improves the loss landscape and thereby accelerates the optimization convergence. MP method can optimize chaotic systems in a manner similar to least-squares shadowing with significantly lower computational costs. Our proposed algorithm, denoted the Multistep Penalty NODE, is applied to chaotic systems such as the Kuramoto-Sivashinsky equation, the two-dimensional Kolmogorov flow, and ERA5 reanalysis data for the atmosphere. It is observed that MP-NODE provide viable performance for such chaotic systems, not only for short-term trajectory predictions but also for invariant statistics that are hallmarks of the chaotic nature of these dynamics.

Chaotic dynamical systems↗
DOE OSTI2025

Sparse-Data Deep Learning Strategies for Radiographic Non-Destructive Testing

Radiography is an imaging technique used in a variety of applications, such as medical diagnosis, airport security, and nondestructive testing. We present a deep learning system for extracting information from radiographic images. We perform various prediction tasks using our system, including material classification and regression on the dimensions of a given object that is being radiographed. Our system is designed to address the sparse-data issue for radiographic nondestructive testing applications. It uses a radiographic simulation tool for synthetic data augmentation, and it uses transfer learning with a pre-trained convolutional neural network model. Using this system, our preliminary results indicate that the object geometry regression task saw an improvement of 70% in the R-squared value when using a multi-regime model. In addition, we increase the performance of the object material classification tasks by utilizing data from different imaging systems. In particular, using neutron imaging improved the material classification accuracy by 20% when compared to x-ray imaging.

convolutional neural networks↗
DOE OSTI2021

Efficient Learning of Accurate Surrogates for Simulations of Complex Systems

Machine learning methods are increasingly deployed to construct surrogate models for complex physical systems at a reduced computational cost. However, the predictive capability of these surrogates degrades in the presence of noisy, sparse, or dynamic data. We introduce an online learning method empowered by optimizer-driven sampling that has two advantages over current approaches: it ensures that all local extrema (including endpoints) of the model response surface are included in the training data and it employs a continuous validation and update process in which surrogates undergo retraining when their performance falls below a validity threshold. We find, using benchmark functions, that optimizer-directed sampling generally outperforms traditional sampling methods in terms of accuracy around local extrema even when the scoring metric is biased towards assessing overall accuracy. Finally, the application to dense nuclear matter demonstrates that highly accurate surrogates for a nuclear equation of state model can be reliably auto-generated from expensive calculations using few model evaluations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Efficient learning of accurate surrogates for simulations of complex systems

Machine learning methods are increasingly deployed to construct surrogate models for complex physical systems at a reduced computational cost. However, the predictive capability of these surrogates degrades in the presence of noisy, sparse or dynamic data. Here, we introduce an online learning method empowered by optimizer-driven sampling that has two advantages over current approaches: it ensures that all local extrema (including endpoints) of the model response surface are included in the training data, and it employs a continuous validation and update process in which surrogates undergo retraining when their performance falls below a validity threshold. We find, using benchmark functions, that optimizer-directed sampling generally outperforms traditional sampling methods in terms of accuracy around local extrema even when the scoring metric is biased towards assessing overall accuracy. Finally, the application to dense nuclear matter demonstrates that highly accurate surrogates for a nuclear equation-of-state model can be reliably autogenerated from expensive calculations using few model evaluations.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Estimating Submicron Aerosol Mixing State at the Global Scale With Machine Learning and Earth System Modeling

Abstract This study integrates machine learning and particle‐resolved aerosol simulations to develop emulators that predict submicron aerosol mixing state indices from the Earth system model (ESM) simulations. The emulators predict aerosol mixing state using only quantities that are predicted by the ESM, including bulk aerosol species concentrations, which do not by themselves carry mixing state information. We used PartMC‐MOSAIC as the particle‐resolved model and NCAR's CESM as the ESM. We trained emulators for three different mixing state indices for submicron aerosol in terms of chemical species abundance ( χ a ), the mixing of optically absorbing and nonabsorbing species ( χ o ), and the mixing of hygroscopic and nonhygroscopic species ( χ h ). Our global mixing state maps show considerable spatial and seasonal variability unique to each mixing state index. Seasonal averages varied spatially between 13% and 94% for χ a , between 38% and 94% for χ o , and between 20% and 87% for χ h with global annual averages of 67%, 68%, and 56%, respectively. High values in one index can be consistent with low values in another index depending on the grouping of species and their relative abundance, meaning that each mixing state index captures different aspects of the population mixing state. Although a direct validation with observational data has not been possible yet, our results are consistent with mixing state index values derived from ambient observations. This work is a prototypical example of using machine learning emulators to add information to ESM simulations.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Physics-guided Deep Learning for Power System State Estimation

In the past decade, dramatic progress has been made in the field of machine learning. This paper explores the possibility of applying deep learning in power system state estimation. Traditionally, physics-based models are used including weighted least square (WLS) or weighted least absolute value (WLAV). These models typically consider a single snapshot of the system without capturing temporal correlations of system states. In this paper, a physics-guided deep learning (PGDL) method is proposed. Specifically, inspired by autoencoders, deep neural networks (DNNs) are used to learn the temporal correlations. The estimated system states from DNNs are then checked against physics laws by running through a set of power flow equations. Hence, the proposed PGDL is both data-driven and physics-guided. The accuracy and robustness of the proposed PGDL method are compared with traditional methods in standard IEEE cases. Simulations show promising results and the applicability is further discussed.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

Distributed Data-Driven Optimization for Voltage Regulation in Distribution Systems

Here, this paper proposes a distributed data-driven optimization framework for voltage regulation in distribution systems. The recursive kernel regression and alternating direction method of multipliers (ADMM) are selected to cover the system learning and distributed optimization tasks. The proposed distributed data-driven framework is capable of having a rapid response to system or load changes while considering the operation optimality. Besides, the distributed algorithm parallels the computation tasks and reduces the computational expense of a single agent. To validate the performance of the proposed method, a hypothetical 7-Bus system and the IEEE 123-Bus system are selected to show the effectiveness of the proposed data-driven framework. According to the numerical study results, the proposed method offers great flexibility for selecting customized kernel models for different regions and can effectively improve the system voltage profile in a distributed manner.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2021

A meta-learning based distribution system load forecasting model selection framework

This paper presents a meta-learning based, automatic distribution system load forecasting model selection framework. Furthermore, the framework includes the following processes: feature extraction, candidate model preparation and labeling, offline training, and online model recommendation. Using load forecasting needs and data characteristics as input features, multiple metalearners are used to rank the candidate load forecast models based on their forecasting accuracy. Then, a scoring-voting mechanism is proposed to weights recommendations from each meta-leaner and make the final recommendations. Heterogeneous load forecasting tasks with different temporal and technical requirements at different load aggregation levels are set up to train, validate, and test the performance of the proposed framework. Simulation results demonstrate that the performance of the meta-learning based approach is satisfactory in both seen and unseen forecasting tasks.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Outage Cause Classification of Power Distribution Systems with Machine Learning and Real-World Data

Power distribution systems are geographically dispersed by nature. It may be affected by various factors, such as vegetation, weather, animal and human behaviors. Present response procedures to an outage event massively rely on expert experience and thus tend to be time-consuming. Automatic outage event detection and classification will help to reduce the responding and restoration time. However, this issue is less addressed with existing research done in this area. In this applied research, a set of waveform pre-processing techniques are first proposed to prepare the waveform data for being used as inputs to the classification algorithm. Further, a machine learning-based algorithm is proposed to classify the outage events according to their root causes, e.g. tree contact, animal contact, lightning, etc. Available data include three phase current & voltage waveforms and contextual information during the distribution system outages. The proposed machine learning algorithm takes the current and voltage waveforms as direct inputs in search of features that humans are unable to capture. Real data provided by a distribution company in the East Tennessee region is used to test the proposed pre-processing techniques and the classification algorithm.

Sun, Haoyuan↗
DOE OSTI2024

Active Learning for Metamaterial Optimization on HPC and QC Integrated Systems

Active learning algorithms, integrating machine learning, quantum computing and optics simulation in an iterative loop, offer a promising approach to optimizing metamaterials. However, these algorithms can face difficulties in optimizing highly complex structures due to computational limitations. High-performance computing (HPC) and quantum computing (QC) integrated systems can address these issues by enabling parallel computing. In this study, we develop an active learning algorithm working on HPC-QC integrated systems. We evaluate the performance of optimization processes within active learning (i.e., training a machine learning model, problem-solving with quantum computing, and evaluating optical properties through wave-optics simulation) for highly complex metamaterial cases. Our results showcase that utilizing multiple cores on the integrated system can significantly reduce computational time, thereby enhancing the efficiency of optimization processes. Therefore, we expect that leveraging HPC-QC integrated systems helps effectively tackle large-scale optimization challenges in general.

Kim, Seongmin↗
DOE OSTI2022

End-to-end learning of multiple sequence alignments with differentiable Smith–Waterman

Abstract Motivation Multiple sequence alignments (MSAs) of homologous sequences contain information on structural and functional constraints and their evolutionary histories. Despite their importance for many downstream tasks, such as structure prediction, MSA generation is often treated as a separate pre-processing step, without any guidance from the application it will be used for. Results Here, we implement a smooth and differentiable version of the Smith–Waterman pairwise alignment algorithm that enables jointly learning an MSA and a downstream machine learning system in an end-to-end fashion. To demonstrate its utility, we introduce SMURF (Smooth Markov Unaligned Random Field), a new method that jointly learns an alignment and the parameters of a Markov Random Field for unsupervised contact prediction. We find that SMURF learns MSAs that mildly improve contact prediction on a diverse set of protein and RNA families. As a proof of concept, we demonstrate that by connecting our differentiable alignment module to AlphaFold2 and maximizing predicted confidence, we can learn MSAs that improve structure predictions over the initial MSAs. Interestingly, the alignments that improve AlphaFold predictions are self-inconsistent and can be viewed as adversarial. This work highlights the potential of differentiable dynamic programming to improve neural network pipelines that rely on an alignment and the potential dangers of optimizing predictions of protein sequences with methods that are not fully understood. Availability and implementation Our code and examples are available at: https://github.com/spetti/SMURF. Supplementary information Supplementary data are available at Bioinformatics online.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Measuring Equality in Machine Learning Security Defenses: A Case Study in Speech Recognition

Over the past decade, the machine learning security community has developed a myriad of defenses for evasion attacks. An understudied question in that community is: for whom do these defenses defend? This work considers common approaches to defending learned systems and how security defenses result in performance inequities across different sub-populations. We outline appropriate parity metrics for analysis and begin to answer this question through empirical results of the fairness implications of machine learning security methods. We find that many methods that have been proposed can cause direct harm, like false rejection and unequal benefits from robustness training. The framework we propose for measuring defense equality can be applied to robustly trained models, preprocessing-based defenses, and rejection methods. We identify a set of datasets with a user-centered application and a reasonable computational cost suitable for case studies in measuring the equality of defenses. In our case study of speech command recognition, we show how such adversarial training and augmentation have non-equal but complex protections for social subgroups across gender, accent, and age in relation to user coverage. We present a comparison of equality between two rejection-based defenses: randomized smoothing and neural rejection, finding randomized smoothing more equitable due to the sampling mechanism for minority groups. This represents the first work examining the disparity in the adversarial robustness in the speech domain and the fairness evaluation of rejection-based defenses.

• Artificial intelligence (AI) / machine learning ↗
DOE OSTI2026

Scalable Federated Learning for Scientific Foundation Models on Leadership-Class Systems

Federated learning (FL) at leadership-class HPC systems remains largely unexplored, despite growing interest in deploying federated workflows on modern HPC systems. This paper provides the first system-level empirical characterization of federated fine-tuning of pretrained foundation models on an exascale supercomputer under a multi-node deployment. Using up to 96 concurrent FL clients deployed across Frontier nodes, we study the impact of client scale, model size, data heterogeneity, partial participation, and differential privacy on runtime, communication overhead, and convergence stability. Our results show that pretrained transformer models remain robust to heterogeneity, client dropout, and privacy noise, while system efficiency degrades rapidly with scale as synchronizat and orchestration dominate runtime. We further demonstrate that system-aware execution strategies, including intra-node aggregation and early aggregation, significantly reduce wall-clock time without degrading model quality. These findings establish a practical performance baseline and inform the design of communication-efficient FL systems on leadership-class HPC platforms.

Kotevska, Olivera [ORNL] (ORCID:0000000316772243)↗
DOE OSTI2021

Searching for fat tails in CRISPR-Cas systems: Data analysis and mathematical modeling

Understanding CRISPR-Cas systems—the adaptive defence mechanism that about half of bacterial species and most of archaea use to neutralise viral attacks—is important for explaining the biodiversity observed in the microbial world as well as for editing animal and plant genomes effectively. The CRISPR-Cas system learns from previous viral infections and integrates small pieces from phage genomes called spacers into the microbial genome. The resulting library of spacers collected in CRISPR arrays is then compared with the DNA of potential invaders. One of the most intriguing and least well understood questions about CRISPR-Cas systems is the distribution of spacers across the microbial population. Here, using empirical data, we show that the global distribution of spacer numbers in CRISPR arrays across multiple biomes worldwide typically exhibits scale-invariant power law behaviour, and the standard deviation is greater than the sample mean. We develop a mathematical model of spacer loss and acquisition dynamics which fits observed data from almost four thousand metagenomes well. In analogy to the classical ‘rich-get-richer’ mechanism of power law emergence, the rate of spacer acquisition is proportional to the CRISPR array size, which allows a small proportion of CRISPRs within the population to possess a significant number of spacers. Our study provides an alternative explanation for the rarity of all-resistant super microbes in nature and why proliferation of phages can be highly successful despite the effectiveness of CRISPR-Cas systems.

59 BASIC BIOLOGICAL SCIENCES↗