SEARCH · Search NASA
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.
This repository implements an iterative magnitude pruning algorithm for pruning neural networks in PyTorch. The pruning method involves gradually removing less significant weights from the model to achieve a specified sparsity, followed by fine-tuning the pruned model to recover performance.
In the past decade, coral reefs worldwide have experienced unprecedented stresses due to climate change, ocean acidification, and anthropomorphic pressures, instigating massive bleaching and die-off of these fragile and diverse ecosystems. Furthermore, remote sensing of these shallow marine habitats is hindered by ocean wave distortion, refraction and optical attenuation, leading invariably to data products that are often of low resolution and signal-to-noise (SNR) ratio. However, recent advances in UAV and Fluid Lensing technology have allowed us to capture multispectral 3D imagery of these systems at sub-cm scales from above the water surface, giving us an unprecedented view of their growth and decay. By combining spatial and spectral information from varying resolutions, we seek to augment and improve the classification accuracy of previously low-resolution datasets at large temporal scales.NeMO-Net, the first open-source deep convolutional neural network (CNN) and interactive learning and training software, currently being developed at NASA Ames, is aimed at assessing the present and past dynamics of coral reef ecosystems through determination of percent living cover and morphology. The latest iteration uses fully convolutional networks to segment and identify coral imagery taken by UAVs and satellites, including WorldView-2 and Sentinel. We present results taken from the Indian Ocean where classification accuracy has exceeded 91% for 24 geomorphological classes given ample training data. In addition, we utilize deep Laplacian Pyramid Super-Resolution Networks (LapSRN) to reconstruct high resolution information from low resolution imagery, trained from various UAV and satellite datasets. Finally, in the case of insufficient training data, we have developed an interactive online platform that allows users to easily segment and submit their classifications, which has been integrated with the current NeMO-Net workflow. Specifically, we present results from the Fiji islands in which preliminary user data has allowed for the accurate identification of 9 separate classes, despite issues such as cloud shadowing and spectral variation. The project is being supported by NASA's Earth Science Technology Office (ESTO) Advanced Information Systems Technology (AIST-16) Program.
In February 2025, the DOE user facility Energy Sciences Network (ESnet) held a three-day Data and AI Workshop in Berkeley, California. The objective of the workshop was to identify challenges within ESnet that could be addressed through data-driven methods, to help define ESnet’s data-analysis requirements, and to shape its AI strategy, guiding data-stewardship efforts and the direction of AI research and AIOps exploration for ESnet7, the next iteration of ESnet’s network. This report summarizes the multi-faceted discussions and findings and presents a set of recommendations for next steps.
To facilitate design optimization of turbine blade shape for reusable launching vehicles, appropriate techniques need to be developed to process and estimate the characteristics of the design variables and the response of the output with respect to the variations of the design variables. The purpose of this report is to offer insight into developing appropriate techniques for supporting such design and optimization needs. Neural network and polynomial-based techniques are applied to process aerodynamic data obtained from computational simulations for flows around a two-dimensional airfoil and a generic three- dimensional wing/blade. For the two-dimensional airfoil, a two-layered radial-basis network is designed and trained. The performances of two different design functions for radial-basis networks, one based on the accuracy requirement, whereas the other one based on the limit on the network size. While the number of neurons needed to satisfactorily reproduce the information depends on the size of the data, the neural network technique is shown to be more accurate for large data set (up to 765 simulations have been used) than the polynomial-based response surface method. For the three-dimensional wing/blade case, smaller aerodynamic data sets (between 9 to 25 simulations) are considered, and both the neural network and the polynomial-based response surface techniques improve their performance as the data size increases. It is found while the relative performance of two different network types, a radial-basis network and a back-propagation network, depends on the number of input data, the number of iterations required for radial-basis network is less than that for the back-propagation network.
Learning and solving governing equations of a physical system, represented by partial differential equations (PDEs), from data is a central challenge in many areas of science and engineering. Traditional numerical methods can be computationally expensive for complex systems and require complete governing equations. Existing data-driven machine learning methods require large datasets to learn a surrogate solution operator, which could be impractical. Here, we propose a solution operator learning method that requires only one PDE solution, i.e., one-shot learning, along with suitable initial and boundary conditions. Leveraging the locality of derivatives, we define a local solution operator in small local domains, train it using a neural network, and use it to predict solutions of new input functions via mesh-based fixed-point iteration or meshfree neural-network based approaches. We test our method on various PDEs, complex geometries, and a practical spatial infection spread application, demonstrating its effectiveness and generalization capabilities.
Traditional control theory is well-developed mainly for linear control situations. In non-linear cases there is no general method of generating a good control, so we have to rely on the ability of the experts (operators) to control them. If we want to automate their control, we must acquire their knowledge and translate it into a precise control strategy. The experts' knowledge is usually represented in non-numeric terms, namely, in terms of uncertain statements of the type 'if the obstacle is straight ahead, the distance to it is small, and the velocity of the car is medium, press the brakes hard'. Fuzzy control is a methodology that translates such statements into precise formulas for control. The necessary first step of this strategy consists of assigning membership functions to all the terms that the expert uses in his rules (in our sample phrase these words are 'small', 'medium', and 'hard'). The appropriate choice of a membership function can drastically improve the quality of a fuzzy control. In the simplest cases, we can take the functions whose domains have equally spaced endpoints. Because of that, many software packages for fuzzy control are based on this choice of membership functions. This choice is not very efficient in more complicated cases. Therefore, methods have been developed that use neural networks or generic algorithms to 'tune' membership functions. But this tuning takes lots of time (for example, several thousands iterations are typical for neural networks). In some cases there are evident physical reasons why equally space domains do not work: e.g., if the control variable u is always positive (i.e., if we control temperature in a reactor), then negative values (that are generated by equal spacing) simply make no sense. In this case it sounds reasonable to choose another scale u' = f(u) to represent u, so that equal spacing will work fine for u'. In the present paper we formulate the problem of finding the best rescaling function, solve this problem, and show (on a real-life example) that after an optimal rescaling, the un-tuned fuzzy control can be as good as the best state-of-art traditional non-linear controls.
A method for calculating the properties of discrete random walks on networks is presented. The method divides complex networks into simpler units whose contribution to the mean first-passage time is calculated. The simplified network is then further iterated. The method is demonstrated by calculating mean first-passage times on a segment, a segment with a single dangling bond, a segment with many dangling bonds, and a looplike structure. The results are analyzed and related to the applicability of the Einstein relation between conductance and diffusion.
NOTE: The manuscript associated with this data package is currently in review. The data may be revised based on reviewer feedback. Upon manuscript acceptance, this data package will be updated with the final dataset and additional metadata. This data package is associated with the manuscript “Artificial intelligence-guided iterations between observations and modeling significantly improve environmental predictions” (Malhotra et al., in prep). This effort was designed following ICON (integrated, coordinated, open, and networked) principles to facilitate a model-experiment (ModEx) iteration approach, leveraging crowdsourced sampling across the contiguous United States (CONUS). New machine learning models were created every month to guide sampling locations. Data from the resulting samples were used to test and rebuild the machine learning models for the next round of sampling guidance. Associated sediment and water geochemistry and in situ sensor data can be found at https://data.ess-dive.lbl.gov/datasets/doi:10.15485/1923689, https://data.ess-dive.lbl.gov/datasets/doi:10.15485/1729719, and https://data.ess-dive.lbl.gov/datasets/doi:10.15485/1603775. This data package is associated with two GitHub repositories found at https://github.com/parallelworks/dynamic-learning-rivers and https://github.com/WHONDRS-Hub/ICON-ModEx_Open_Manuscript. In addition to this readme, this data package also includes two file-level metadata (FLMD) files that describes each file and two data dictionaries (DD) that describe all column/row headers and variable definitions. This data package consists of two main folders (1) dynamic-learning-rivers and (2) ICON-ModEx_Open_Manuscript which contain snapshots of the associated GitHub repositories. The input data, output data, and machine learning models used to guide sampling locations are within dynamic-learning-rivers. The folder is organized into five top-level directories: (1) “input_data” holds the training data for the ML models; (2) “ml_models” holds machine learning (ML) models trained on the data in “input_data”; (3) “examples” contains files for direct experimentation with the machine learning model, including scripts for setting up “hindcast” run; (4) “scripts” contains data preprocessing and postprocessing scripts and intermediate results specific to this data set that bookend the ML workflow; and (5) “output_data” holds the overall results of the ML model on that branch. Each trained ML model resides on its own branch in the repository; this means that inputs and outputs can be different branch-to-branch. There is also one hidden directory “.github/workflows”. This hidden directory contains information for how to run the ML workflow as an end-to-end automated GitHub Action but it is not needed for reusing the ML models archived here. Please see the top-level README.md in the GitHub repository for more details on the automation. The scripts and data used to create figures in the manuscript are within ICON-ModEx_Open_Manuscript. The folder is organized into four folders which contain the scripts, data, and pdf for each figure. Within the “fig-model-score-evolution” folder, there is a folder called “intermediate_branch_data” which contains some intermediate files pulled from dynamic-learning-rivers and reorganized to easily integrate into the workflows. NOTE: THIS FOLDER INCLUDES THE FILES AT THE POINT OF PAPER SUBMISSION. IT WILL BE UPDATED ONCE THE PAPER IS ACCEPTED WITH ANY REVISIONS AND WILL INCLUDE A DD/FLMD AT THAT POINT. We thank the United States Forest Service, Washington Department of Fish and Wildlife, Washington Department of Natural Resources, Cowiche Canyon Conservatory, Washington State Parks and Recreation Commission (Scientific Research Permit #210901), and the Confederated Tribes and Bands of the Yakama Nation for access to field locations where the samples labeled “SSS” were collected. We also thank the Yakama Nation Tribal Council and Yakama Nation Fisheries for working with us to facilitate sample collection and optimization of data usage according to their values and worldview. WHONDRS consortium members were asked to provide any acknowledgments for the collection of samples labeled “CM” and the following is a list of acknowledgments that were submitted with their corresponding Site IDs: (MART) Research activities were conducted in part on the Wind River Experimental Forest within the Gifford Pinchot National Forest; (MP- 100379) Philadelphia is part of Lenapehoking, the ancestral homelands of the Lenape peoples; (MP-102398) Land surveyed is the ancestral homelands of the Nookhose'iinenno (Arapaho), Tsis tsis'tas (Cheyenne), and Nuuchu (Ute); (MP-100749 and MP- 100747) Georgia Coastal Ecosystem LTER, OCE-1832178; (SP-70 and SP-72) Eastern Shoshone, Shoshone-Bannock; (MP- 102944) Funded by Oregon Watershed Enhancement Board. On the traditional lands of the Confederated Tribes of the Siletz, Confederated Tribes of the Grand Rhonde, and the Clatsop-Nehalem Confederated Tribe; (MP- 100607) Holiday Creek is located on the traditional territory of the Monacan Indian Nation; (SP-45) Lafayette Blue Springs State Park; (MP-102420) NSF DEB-2016749; (MP-100019) New Hampshire Agriculture Experiment Station; (SP-35) Rayonier (land owner; https://www.rayonier.com/); (MP- 101276) US Department of Energy, Office of Science, Biological and Environmental Research, Subsurface Biogeochemical Research, Watershed Dynamics and Evolution SFA at ORNL; (MP- 103224) Watershed Dynamics and Evolution SFA at ORNL; (MP- 101584) Traditional lands of the Oceti Sakowin (Dakota, Lakota, Nakoda) and Anishinaabe Peoples.
A closed-loop optimal neural-network controller technique was developed to optimize rotorcraft aeromechanical behaviour. This technique utilities a neural-network scheme to provide a general non-linear model of the rotorcraft. A modem constrained optimisation method is used to determine and update the constants in the neural-network plant model as well as to determine the optimal control vector. Current data is read, weighted, and added to a sliding data window. When the specified maximum number of data sets allowed in the data window is exceeded, the oldest data set is and the remaining data sets are re-weighted. This procedure provides at least four additional degrees-of-freedom in addition to the size and geometry of the neural-network itself with which to optimize the overall operation of the controller. These additional degrees-of-freedom are: 1. the maximum length of the sliding data window, 2. the frequency of neural-network updates, 3. the weighting of the individual data sets within the sliding window, and 4. the maximum number of optimisation iterations used for the neural-network updates.
IABG has been using various servohydraulic test facilities for many years for the reproduction of service loads and environmental loads on all kinds of test objects. For more than 15 years, a multi-axis vibration test facility has been under service, originally designed for earthquake simulation but being upgraded to the demands of space testing. First tests with the DFS/STM showed good reproduction accuracy and demonstrated the feasibility of transient vibration testing of space objects on a multi-axis hydraulic shaker. An approach to structural qualification is possible by using this test philosophy. It will be outlined and its obvious advantages over the state-of-the-art single-axis test will be demonstrated by example results. The new test technique has some special requirements to the test facility exceeding those of earthquake testing. Most important is the high reproduction accuracy demanded for a sophisticated control system. The state-of-the-art approach of analog closed-loop control circuits for each actuator combined with a static decoupling network and an off-line iterative waveform control is not able to meet all the demands. Therefore, the future over-all control system is implemented as hierarchical full digital closed-loop system on a highly parallel transputer network. The innermost layer is the digital actuator controller, the second one is the MDOF-control of the table movement. The outermost layer would be the off-line iterative waveform control, which is dedicated only to deal with the interaction of test table and test object or non-linear effects. The outline of the system will be presented.
The rapid adoption of machine learning in various scientific domains calls for the development of best practices and community agreed-upon benchmarking tasks and metrics. We present Matbench Discovery as an example evaluation framework for machine learning energy models, here applied as pre-filters to first-principles computed data in a high-throughput search for stable inorganic crystals. We address the disconnect between (1) thermodynamic stability and formation energy and (2) retrospective and prospective benchmarking for materials discovery. Alongside this paper, we publish a Python package to aid with future model submissions and a growing online leaderboard with adaptive user-defined weighting of various performance metrics allowing researchers to prioritize the metrics they value most. To answer the question of which machine learning methodology performs best at materials discovery, our initial release includes random forests, graph neural networks, one-shot predictors, iterative Bayesian optimizers and universal interatomic potentials. We highlight a misalignment between commonly used regression metrics and more task-relevant classification metrics for materials discovery. Accurate regressors are susceptible to unexpectedly high false-positive rates if those accurate predictions lie close to the decision boundary at 0 eV per atom above the convex hull. The benchmark results demonstrate that universal interatomic potentials have advanced sufficiently to effectively and cheaply pre-screen thermodynamic stable hypothetical materials in future expansions of high-throughput materials databases.
This research focuses on a new neural network scene classification technique. The task is to identify scene elements in Advanced Very High Resolution Radiometry (AVHRR) data from three scene types: polar, desert and smoke from biomass burning in South America (smoke). The ultimate goal of this research is to design and implement a computer system which will identify the clouds present on a whole-Earth satellite view as a means of tracking global climate changes. Previous research has reported results for rule-based systems (Tovinkere et at 1992, 1993) for standard back propagation (Watters et at. 1993) and for a hierarchical approach (Corwin et al 1994) for polar data. This research uses a hierarchical neural network with don't care conditions and applies this technique to complex scenes. A hierarchical neural network consists of a switching network and a collection of leaf networks. The idea of the hierarchical neural network is that it is a simpler task to classify a certain pattern from a subset of patterns than it is to classify a pattern from the entire set. Therefore, the first task is to cluster the classes into groups. The switching, or decision network, performs an initial classification by selecting a leaf network. The leaf networks contain a reduced set of similar classes, and it is in the various leaf networks that the actual classification takes place. The grouping of classes in the various leaf networks is determined by applying an iterative clustering algorithm. Several clustering algorithms were investigated, but due to the size of the data sets, the exhaustive search algorithms were eliminated. A heuristic approach using a confusion matrix from a lightly trained neural network provided the basis for the clustering algorithm. Once the clusters have been identified, the hierarchical network can be trained. The approach of using don't care nodes results from the difficulty in generating extremely complex surfaces in order to separate one class from all of the others. This approach finds pairwise separating surfaces and forms the more complex separating surface from combinations of simpler surfaces. This technique both reduces training time and improves accuracy over the previously reported results. Accuracies of 97.47%, 95.70%, and 99.05% were achieved for the polar, desert and smoke data sets.
The CeO 2 -based memristor has attracted significant attention due to its intrinsic resistive switching (RS) properties, large on/off ratio, and great plasticity, making it a promising candidate for artificial synapses. However, significant challenges such as high power consumption and poor device reliability hinder its broad application in neuromorphic microchips. To tackle these issues, in this work, we design a novel bilayer (BL) memristor by integrating a CeO 2 -based memristor with a Co-CeO 2 vertically aligned nanocomposite (VAN) layer and compare it with the single layer (SL) memristor. Preliminary electrical testing reveals that the BL memristor offers a reduced set/reset voltage (~67% lower), a higher on/off ratio (~5 × 10 2 ), enhanced device reliability, and improved device-to-device variation compared to the SL memristor. Insight from COMSOL simulation, coupled with microstructural analysis, provides a comprehensive elucidation on how the VAN layer facilitates the selective conductive filament (CF) formation. Subsequently, the plasticity of the BL memristor is evaluated through long-term potentiation/depression (LTP/LTD), paired-pulse facilitation (PPF), and spike-time-dependent plasticity (STDP). The spiking neural network (SNN) built upon the BL memristor achieves remarkable accuracy (~94%) after only 12 iterations, underscoring its potential for high-performance neural networks.
Conventionally programmed digital computers can process numbers with great speed and precision, but do not easily recognize patterns or imprecise or contradictory data. Instead of being programmed in the conventional sense, artificial neural networks are capable of self-learning through exposure to repeated examples. However, the training of an ANN can be a time consuming and unpredictable process. A general method is being developed to mate the adaptability of the ANN with the speed and precision of the digital computer. This method was successful in building feedforward networks that can approximate functions and their partial derivatives from examples in a single iteration. The general method also allows the formation of feedforward networks that can approximate the solution to nonlinear ordinary and partial differential equations to desired accuracy without the need of examples. It is believed that continued research will produce artificial neural networks that can be used with confidence in practical scientific computing and engineering applications.
Conventionally programmed digital computers can process numbers with great speed and precision, but do not easily recognize patterns or imprecise or contradictory data. Instead of being programmed in the conventional sense, artificial neural networks (ANN's) are capable of self-learning through exposure to repeated examples. However, the training of an ANN can be a time consuming and unpredictable process. A general method is being developed by the author to mate the adaptability of the ANN with the speed and precision of the digital computer. This method has been successful in building feedforward networks that can approximate functions and their partial derivatives from examples in a single iteration. The general method also allows the formation of feedforward networks that can approximate the solution to nonlinear ordinary and partial differential equations to desired accuracy without the need of examples. It is believed that continued research will produce artificial neural networks that can be used with confidence in practical scientific computing and engineering applications.
This report describes the development and evaluation of mathematical models for predicting human performance from discrete wavelet transforms (DWT) of event-related potentials (ERP) elicited by task-relevant stimuli. The DWT was compared to principal components analysis (PCA) for representation of ERPs in linear regression and neural network models developed to predict a composite measure of human signal detection performance. Linear regression models based on coefficients of the decimated DWT predicted signal detection performance with half as many f ree parameters as comparable models based on PCA scores. In addition, the DWT-based models were more resistant to model degradation due to over-fitting than PCA-based models. Feed-forward neural networks were trained using the backpropagation,-, algorithm to predict signal detection performance based on raw ERPs, PCA scores, or high-power coefficients of the DWT. Neural networks based on high-power DWT coefficients trained with fewer iterations, generalized to new data better, and were more resistant to overfitting than networks based on raw ERPs. Networks based on PCA scores did not generalize to new data as well as either the DWT network or the raw ERP network. The results show that wavelet expansions represent the ERP efficiently and extract behaviorally important features for use in linear regression or neural network models of human performance. The efficiency of the DWT is discussed in terms of its decorrelation and energy compaction properties. In addition, the DWT models provided evidence that a pattern of low-frequency activity (1 to 3.5 Hz) occurring at specific times and scalp locations is a reliable correlate of human signal detection performance.
This report describes the development and evaluation of mathematical models for predicting human performance from discrete wavelet transforms (DWT) of event-related potentials (ERP) elicited by task-relevant stimuli. The DWT was compared to principal components analysis (PCA) for representation of ERPs in linear regression and neural network models developed to predict a composite measure of human signal detection performance. Linear regression models based on coefficients of the decimated DWT predicted signal detection performance with half as many free parameters as comparable models based on PCA scores. In addition, the DWT-based models were more resistant to model degradation due to over-fitting than PCA-based models. Feed-forward neural networks were trained using the backpropagation algorithm to predict signal detection performance based on raw ERPs, PCA scores, or high-power coefficients of the DWT. Neural networks based on high-power DWT coefficients trained with fewer iterations, generalized to new data better, and were more resistant to overfitting than networks based on raw ERPs. Networks based on PCA scores did not generalize to new data as well as either the DWT network or the raw ERP network. The results show that wavelet expansions represent the ERP efficiently and extract behaviorally important features for use in linear regression or neural network models of human performance. The efficiency of the DWT is discussed in terms of its decorrelation and energy compaction properties. In addition, the DWT models provided evidence that a pattern of low-frequency activity (1 to 3.5 Hz) occurring at specific times and scalp locations is a reliable correlate of human signal detection performance. Copyright 1999 Academic Press.
This paper studies the shallow Ritz method for solving the one-dimensional diffusion problem. It is shown that the shallow Ritz method improves the order of approximation dramatically for non-smooth problems. To realize this optimal or nearly optimal order of the shallow Ritz approximation, we develop a damped block Newton (dBN) method that alternates between updates of the linear and non-linear parameters. Per each iteration, the linear and the non-linear parameters are updated by exact inversion and one step of a modified, damped Newton method applied to a reduced non-linear system, respectively. The computational cost of each dBN iteration is $\mathcal{O}$(n). Starting with the non-linear parameters as a uniform partition of the interval, numerical experiments show that the dBN is capable of efficiently moving mesh points to nearly optimal locations. In conclusion, to improve the efficiency of the dBN further, we propose an adaptive damped block Newton (AdBN) method by combining the dBN with the adaptive neuron enhancement (ANE) method [28].