Search NASA⌕ Search

SEARCH · Search NASA

Results for “sparse 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.

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

At least 595 records · Page 33

AWB-GCN: A Graph Convolutional Network Accelerator with Runtime Workload Rebalancing

The recent development of deep learning has been mostly focusing on Euclidean data, such as images, videos, audios, etc. However, most real-world information and relation are often expressed as graphs. To efficiently learn from graph data, graph convolutional networks (GCNs) emerge as a promising approach, showing advantages in several practical applications such as social network analysis, knowledge discovery, 3D modeling, motion capturing, etc. Real-world graphs are usually extremely large and imbalanced, posting significant performance demand and design challenges on the hardware dedicated for GCN inference. In this paper, we propose an architecture design called UW-GCN to accelerate graph convolutional network inference. To tackle the major performance bottleneck from workload imbalance, we propose dynamic neighborhood stealing and remote chunk shuffling techniques, relying on hardware flexibility to achieve hardware auto-tuning under negligible area or delay overhead. Specifically, UW-GCN is able to smartly profile the sparse graph pattern while continuously adjusting the workload distribution via routing reconfiguration among parallel processing elements (PEs). The ideal configuration is then reused in the remaining iterations. To the best of our knowledge, this is the first accelerator design particularly for GCN and the first work relying on hardware auto-tuning, which is normally based on software, to achieve near-optimal workload balance in processing sparse structures.

Geng, Tong↗

The Kokkos OpenMPTarget Backend: Implementation and Lessons Learned

As the supercomputing landscape diversifies, solutions such as Kokkos to write vendor agnostic applications and libraries have risen in popularity. Kokkos provides a programming model designed for performance portability, which allows developers to write a single source implementation that can run efficiently on various architectures. At its heart, Kokkos maps parallel algorithms to architecture and vendor specific backends written in lower level programming models such as CUDA and HIP. Another approach to writing vendor agnostic parallel code is using OpenMP’s directives based approach, which lets developers annotate code to express parallelism. It is implemented at the compiler level and is supported by all major high performance computing vendors, as well as the primary Open Source toolchains GNU and LLVM. Since its inception, Kokkos has used OpenMP to parallelize on CPU architectures. In this paper, we explore leveraging OpenMP for a GPU backend and discuss the challenges we encountered when mapping the Kokkos APIs and semantics to OpenMP target constructs. As an exemplar workload we chose a simple conjugate gradient solver for sparse matrices. We find that performance on NVIDIA and AMD GPUs varies widely based on details of the implementation strategy and the chosen compiler. Furthermore, the performance of the OpenMP implementations decreases with increasing complexity of the investigated algorithms.

Gayatri, Rahulkumar↗

Improving Reliability of Large Language Models for Nuclear Power Plant Diagnostics [Poster]

Large Language Models (LLMs) struggle out of the box when answering factually about detailed questions, especially in domains that are sparsely represented in their training data. This causes hallucinations and reduces reliability making it difficult for them to be used in practice. This work shows that using RAG techniques can improve factual accuracy and reliability, allowing for the application of LLMs in specialized areas, even when those areas that aren’t extensively covered in their initial training.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Improved Subseasonal Forecasting of Extreme Polar Vortices Using Machine Learning

Our research was focused on forecasting the position and shape of the winter stratospheric polar vortex at a subseasonal timescale of 15 days in advance. To achieve this, we employed both statistical and neural network machine learning techniques. The analysis was performed on 42 winter seasons of reanalysis data provided by NASA giving us a total of 6,342 days of data. The state of the polar vortex for determined by using geometric moments to calculate the centroid latitude and the aspect ratio of an ellipse fit onto the vortex. Timeseries for thirty additional precursors were calculated to help improve the predictive capabilities of the algorithm. Feature importance of these precursors was performed using random forest to measure the predictive importance and the ideal number of precursors. Then, using the precursors identified as important, various statistical methods were tested for predictive accuracy with random forest and nearest neighbor performing the best. An echo state network, a type of recurrent neural network that features sparsely connected hidden layer and a reduced number of trainable parameters that allows for rapid training and testing, was also implemented for the forecasting problem. Hyperparameter tuning was performed for each methods using a subset of the training data. The algorithms were trained and tuned on the first 41 years of data, then tested for accuracy on the final year. In general, the centroid latitude of the polar vortex proved easier to predict than the aspect ratio across all algorithms. Random forest outperformed other statistical forecasting algorithms overall but struggled to predict extreme values. Forecasting from echo state network suggested a strong predictive capability past 15 days, but further work is required to fully realize the potential of recurrent neural network approaches.

54 ENVIRONMENTAL SCIENCES↗

Regularized machine learning on molecular graph model explains systematic error in DFT enthalpies

Abstract A major goal of materials research is the discovery of novel and efficient heterogeneous catalysts for various chemical processes. In such studies, the candidate catalyst material is modeled using tens to thousands of chemical species and elementary reactions. Density Functional Theory (DFT) is widely used to calculate the thermochemistry of these species which might be surface species or gas-phase molecules. The use of an approximate exchange correlation functional in the DFT framework introduces an important source of error in such models. This is especially true in the calculation of gas phase molecules whose thermochemistry is calculated using the same planewave basis set as the rest of the surface mechanism. Unfortunately, the nature and magnitude of these errors is unknown for most practical molecules. Here, we investigate the error in the enthalpy of formation for 1676 gaseous species using two different DFT levels of theory and the ‘ground truth values’ obtained from the NIST database. We featurize molecules using graph theory. We use a regularized algorithm to discover a sparse model of the error and identify important molecular fragments that drive this error. The model is robust to rigorous statistical tests and is used to correct DFT thermochemistry, achieving more than an order of magnitude improvement.

36 MATERIALS SCIENCE↗

Simulation-trained machine learning models for Lorentz transmission electron microscopy

Understanding the collective behavior of complex spin textures, such as lattices of magnetic skyrmions, is of fundamental importance for exploring and controlling the emergent ordering of these spin textures and inducing phase transitions. It is also critical to understand the skyrmion–skyrmion interactions for applications such as magnetic skyrmion-enabled reservoir or neuromorphic computing. Magnetic skyrmion lattices can be studied using in situ Lorentz transmission electron microscopy (LTEM), but quantitative and statistically robust analysis of the skyrmion lattices from LTEM images can be difficult. In this work, we show that a convolutional neural network, trained on simulated data, can be applied to perform segmentation of spin textures and to extract quantitative data, such as spin texture size and location, from experimental LTEM images, which cannot be obtained manually. This includes quantitative information about skyrmion size, position, and shape, which can, in turn, be used to calculate skyrmion–skyrmion interactions and lattice ordering. We apply this approach to segmenting images of Néel skyrmion lattices so that we can accurately identify skyrmion size and deformation in both dense and sparse lattices. The model is trained using a large set of micromagnetic simulations as well as simulated LTEM images. This entirely open-source training pipeline can be applied to a wide variety of magnetic features and materials, enabling large-scale statistical studies of spin textures using LTEM.

McCray, Arthur R. C. (ORCID:0000000160774698)↗

Accurate and uncertainty-aware multi-task prediction of HEA properties using prior-guided deep Gaussian processes

Surrogate modeling techniques have become indispensable in accelerating the discovery and optimization of high-entropy alloys (HEAs), especially when integrating computational predictions with sparse experimental observations. This study systematically evaluates the training and testing performance of four prominent surrogate models—conventional Gaussian processes (cGP), Deep Gaussian processes (DGP), encoder-decoder neural networks for multi-output regression and eXtreme Gradient Boosting (XGBoost)—applied to a hybrid dataset of experimental and computational properties of the 8-component HEA system Al-Co-Cr-Cu-Fe-Mn-Ni-V. We specifically assess their capabilities in predicting correlated material properties, including yield strength, hardness, modulus, ultimate tensile strength, elongation, and average hardness under dynamic/quasi-static conditions, alongside auxiliary computational properties. The comparison highlights the strengths of hierarchical deep modeling approaches in handling heteroscedastic, heterotopic, and incomplete data commonly encountered in materials science. Our findings illustrate that combined surrogate models such as DGPs infused with machine-learned priors outperform other surrogates by effectively capturing inter-property correlations and by assimilating prior knowledge. This enhanced predictive accuracy positions the combined surrogate models as powerful tools for robust and data-efficient materials design.

36 MATERIALS SCIENCE↗

Identifying Entangled Physics Relationships Through Sparse Matrix Decomposition to Inform Plasma Fusion Design

We report a sustainable burn platform through inertial confinement fusion (ICF) has been an ongoing challenge for over 50 years. Mitigating engineering limitations and improving the current design involves an understanding of the complex coupling of physical processes. While sophisticated simulation codes are used to model ICF implosions, these tools contain necessary numerical approximation but miss physical processes that limit predictive capability. Identification of relationships between controllable design inputs to ICF experiments and measurable outcomes (e.g., neutron yield, neutron velocity, areal density) from performed experiments can help guide the future design of experiments and development of simulation codes, to potentially improve the accuracy of the computational models used to simulate ICF experiments. We use sparse matrix decomposition methods to identify clusters of a few related design variables. Sparse principal component analysis (SPCA) identifies groupings that are related to the physical origin of the variables (laser, hohlraum, and capsule). A variable importance analysis finds that in addition to variables highly correlated with neutron yield, such as picket power and laser energy, variables that represent a dramatic change of the ICF design, such as number of pulse steps, are also very important. The obtained sparse components are then used to train a random forest (RF) regression surrogate for predicting total yield. The RF performance on the training and testing data compares with the performance of the RF trained using all the design variables considered. This work is intended to inform design changes in future ICF experiments by augmenting the expert intuition and simulation results.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Sparse distributed memory

Sparse distributed memory was proposed be Pentti Kanerva as a realizable architecture that could store large patterns and retrieve them based on partial matches with patterns representing current sensory inputs. This memory exhibits behaviors, both in theory and in experiment, that resemble those previously unapproached by machines - e.g., rapid recognition of faces or odors, discovery of new connections between seemingly unrelated ideas, continuation of a sequence of events when given a cue from the middle, knowing that one doesn't know, or getting stuck with an answer on the tip of one's tongue. These behaviors are now within reach of machines that can be incorporated into the computing systems of robots capable of seeing, talking, and manipulating. Kanerva's theory is a break with the Western rationalistic tradition, allowing a new interpretation of learning and cognition that respects biology and the mysteries of individual human beings.

Denning, Peter J.↗

Los Alamos works on a biologically realistic computer network

Brain neuroscientists and computer scientists call this field neuromimetic computing—building computers inspired by how the cerebral cortex of the brain works. Its cortical processes rely on billions of small biological “switches,” called neurons. To learn, they store and process information. Using an approach called neural networks, researchers are developing computers that simulate neurons and their interconnections. Then computers can learn about their surroundings, interpret data, and make predictions based on it. In practice, however, researchers attempting to simulate neural processing at anything close to the scale and complexity of the brain’s cortical circuits have been stymied by limitations on computer memory and computational power. All that has changed with the new Trinity supercomputer at Los Alamos, which became fully operational in mid-2017. Trinity has unique capabilities designed for the stockpile stewardship mission, which includes highly complex nuclear simulations that have replaced the testing of nuclear weapons. All this capability means Trinity allows a fundamentally different approach to large-scale cortical simulations, enabling an unprecedented leap in the ability to model neural processing. To test that capability on a limited-scale problem, computer scientists and neuroscientists at Los Alamos created a “sparse prediction machine” that runs on a neural network on Trinity. A sparse prediction machine is designed to work like the brain: researchers expose it to data—in this case, various videos of a car driving down a road—without labeling the data in any way. Then the program sorts through that data frame by frame, focuses on the important information, and develops a prediction about the car’s motion. With Trinity’s power, the Los Alamos team simulates the way a brain handles information in its neurons but uses the fewest neurons at any given moment to explain the information at hand. That’s the “sparse” part, and it makes the brain very efficient—and, hopefully, a computer more efficient, too.

97 MATHEMATICS AND COMPUTING↗

Measurement and applications: Exploring the challenges and opportunities of hierarchical federated learning in sensor applications

Sensor applications have become ubiquitous in modern society as the digital age continues to advance. AI-based techniques (e.g., machine learning) are effective at extracting actionable information from large amounts of data. An example would be an automated water irrigation system that uses AI-based techniques on soil quality data to decide how to best distribute water. However, these AI-based techniques are costly in terms of hardware resources, and Internet-of-Things (IoT) sensors are resource-constrained with respect to processing power, energy, and storage capacity. These limitations can compromise the security, performance, and reliability of sensor-driven applications. To address these concerns, cloud computing services can be used by sensor applications for data storage and processing. Unfortunately, cloud-based sensor applications that require real-time processing, such as medical applications (e.g., fall detection and stroke prediction), are vulnerable to issues such as network latency due to the sparse and unreliable networks between the sensor nodes and the cloud server [1]. As users approach the edge of the communications network, latency issues become more severe and frequent. A promising alternative is edge computing, which provides cloud-like capabilities at the edge of the network by pushing storage and processing capabilities from centralized nodes to edge devices that are closer to where the data are gathered, resulting in reduced network delays [2], [3].

Po-Leen Ooi, Melanie↗

DeepM&Mnet: Inferring the electroconvection multiphysics fields based on operator approximation by neural networks

Electroconvection is a multiphysics problem involving coupling of the flow field with the electric field as well as the cation and anion concentration fields. Here, we use electroconvection as a benchmark problem to put forward a new data assimilation framework, the DeepM&Mnet, for simulating multiphysics and multiscale problems at speeds much faster than standard numerical methods using pre-trained neural networks. We first pre-train DeepONets that can predict independently each field, given general inputs from the rest of the fields of the coupled system. DeepONets can approximate nonlinear operators and are composed of two sub-networks, a branch net for the input fields and a trunk net for the locations of the output field. DeepONets, which are extremely fast, are used as building blocks in the DeepM&Mnet and form constraints for the multiphysics solution along with some sparse available measurements of any of the fields. We demonstrate the new methodology and document the accuracy of each individual DeepONet, and subsequently we present two different DeepM&Mnet architectures that infer accurately and efficiently 2D electroconvection fields for unseen electric potentials. Furthermore, the DeepM&Mnet framework is general and can be applied for building any complex multiphysics and multiscale models based on very few measurements using pre-trained DeepONets in a “plug-and-play” mode.

97 MATHEMATICS AND COMPUTING↗

Benchmarking the Performance of Neuromorphic and Spiking Neural Network Simulators

Software simulators play a critical role in the development of new algorithms and system architectures in any field of engineering. Neuromorphic computing, which has shown potential in building brain-inspired energy-efficient hardware, suffers a slow-down in the development cycle due to a lack of flexible and easy-to-use simulators of either neuromorphic hardware itself or of spiking neural networks (SNNs), the type of neural network computation executed on most neuromorphic systems. While there are several openly available neuromorphic or SNN simulation packages developed by a variety of research groups, they have mostly targeted computational neuroscience simulations, and only a few have targeted small-scale machine learning tasks with SNNs. Evaluations or comparisons of these simulators have often targeted computational neuroscience-style workloads. In this work, we seek to evaluate the performance of several publicly available SNN simulators with respect to non-computational neuroscience workloads, in terms of speed, flexibility, and scalability. We evaluate the performance of the NEST, Brian2, Brian2GeNN, BindsNET and Nengo packages under a common front-end neuromorphic framework. Our evaluation tasks include a variety of different network architectures and workload types to mimic the computation common in different algorithms, including feed-forward network inference, genetic algorithms, and reservoir computing. We also study the scalability of each of these simulators when running on different computing hardware, from single core CPU workstations to multi-node supercomputers. Our results show that the BindsNET simulator has the best speed and scalability for most of the SNN workloads (sparse, dense, and layered SNN architectures) on a single core CPU. However, when comparing the simulators leveraging the GPU capabilities, Brian2GeNN outperforms the others for these workloads in terms of scalability. NEST performs the best for small sparse networks and is also the most flexible simulator in terms of reconfiguration capability NEST shows a speedup of at least 2x compared to the other packages when running evolutionary algorithms for SNNs. The multi-node and multi-thread capabilities of NEST show at least 2x speedup compared to the rest of the simulators (single core CPU or GPU based simulators) for large and sparse networks. We conclude our work by providing a set of recommendations on the suitability of employing these simulators for different tasks and scales of operations. We also present the characteristics for a future generic ideal SNN simulator for different neuromorphic computing workloads.

97 MATHEMATICS AND COMPUTING↗

Linear and nonlinear trending and prediction for AVHRR time series data

The variability of AVHRR calibration coefficient in time was analyzed using algorithms of linear and non-linear time series analysis. Specifically we have used the spline trend modeling, autoregressive process analysis, incremental neural network learning algorithm and redundancy functional testing. The analysis performed on available AVHRR data sets revealed that (1) the calibration data have nonlinear dependencies, (2) the calibration data depend strongly on the target temperature, (3) both calibration coefficients and the temperature time series can be modeled, in the first approximation, as autonomous dynamical systems, (4) the high frequency residuals of the analyzed data sets can be best modeled as an autoregressive process of the 10th degree. We have dealt with a nonlinear identification problem and the problem of noise filtering (data smoothing). The system identification and filtering are significant problems for AVHRR data sets. The algorithms outlined in this study can be used for the future EOS missions. Prediction and smoothing algorithms for time series of calibration data provide a functional characterization of the data. Those algorithms can be particularly useful when calibration data are incomplete or sparse.

Smid, J.↗

Invertible Neural Networks for Aerodynamic Design of Wind Turbine Blades

The state-of-the-practice methods for aerodynamic design of wind turbine blads use linearized blade element momentum theory (BEM) to optimize the twist and chord profiles from a pre-selected set of 2D airfoil shapes. In this work, we apply invertible neural network (INN) tools to enable the rapid inverse aerodynamic design of wind turbine blades including component airfoils. The INN is trained on data obtained through the use of robust automated mesh generation and the HAMSTR computational fluid dynamics solver with advanced turbulence and transition models validated for turbine applications. Our design technique is a significant improvement over the state-of-the-practice linearized blade element momentum (BEM) techniques in capturing 3D nonlinear aerodynamic effects that are critical for optimal design of the rotors. This is made possible by developing sparse, invertible neural networks (INNs) for inverse design and optimization that realize a 100x cost reduction compared to adjoint-based computational fluid dynamics (CFD) approaches, while enabling increased robustness of the final design. We demonstrate the INN tool for design of a section of the NREL 5-MW blade. All generated shapes satisfy the desired aerodynamic characteristics, demonstrating the success of the INN approach for inverse design of wind turbine blades.

aerodynamic design↗

Assessing entropy for catalytic processes at complex reactive interfaces

When chemical reactions are accelerated by a catalyst, entropy differences between reactants and their transient intermediates can be the driving force behind the promotion or inhibition of desired and parasitic chemical pathways. Understanding and controlling catalytic processes therefore requires both a fundamental and practicable understanding of entropy in addition to enthalpy. In unstructured media such as the vapor phase equilibrated with sparsely covered surfaces, entropy can be adequately accounted for by well-established approaches based on translational, rotational, and harmonic vibrational partition functions. However, these approximations become inadequate in more complex condensed phase environments, e.g., solid liquid interfaces of confined reaction spaces. In this chapter, we provide an overview of the state-of-art in the computational quantification of entropy and its known ramifications on catalysis. The fundamental roles of thermodynamics and kinetics in catalysis are covered in enough detail to appreciate and contextualize the computational methods employed to compute chemically accurate estimates of entropy. These methods are discussed in appropriate detail and range from the ubiquitous harmonic oscillator approximation where entropy unrelated to high frequency oscillations is typically underestimated, to enhanced free energy sampling with molecular dynamics where the desired accuracy must be weighed against the associated computational cost of obtaining it. The rising importance of machine learning and artificial intelligence in accelerating methodological progress in this field is touched upon, as well. Finally, applications, successes, and pitfalls of using these methods are provided to showcase past and present accomplishments while clarifying where improvements in both understanding and methodology are still needed.

Kollias, Loukas↗

Multimodal super-resolution: discovering hidden physics and its application to fusion plasmas

Understanding complex physical systems often requires integrating data from multiple diagnostics, each with limited resolution or coverage. We present a machine learning framework that reconstructs synthetic high-temporal-resolution data for a target diagnostic using information from other diagnostics, without direct target measurements during the inference. This multimodal super-resolution technique improves diagnostic robustness and enables monitoring even in case of measurement failures or degradation. Applied to fusion plasmas, our method targets edge-localized modes (ELMs), which can damage plasma-facing materials. By reconstructing super-resolution Thomson Scattering data from complementary diagnostics, we uncover fine-scale plasma dynamics and validate the role of resonant magnetic perturbations (RMPs) in ELM suppression through magnetic island formation. The approach provides new observation supporting the plasma profile flattening due to these islands. Our results demonstrate the framework’s ability to generate high-fidelity synthetic diagnostics, offering a powerful tool for ELM control development in future reactors like ITER. The approach is broadly transferable to other domains facing sparse, incomplete, or degraded diagnostic data, opening new avenues for discovery.

Jalalvand, Azarakhsh [Princeton Univ., NJ (United ↗

brulilo, Version 0.x

brulilo is a Python package for building and evolving thermonuclear reaction networks, such as those used in astrophysical explosion calculations. It is very much a current work in progress, and much of the intended functionality is still under development. The main goal of brulilo is an easy-to-use code for quick calculations or post-processing, as well as a learning tool to test various implementations of solvers on both CPUs and GPUs. Currently, brulilo parses an input text file that specifies the reactions to be included in the network. The notation of the inputs are in the typical astrophysical notation; i.e. hydrogen burning to deuterium in the sun, "p + p + e --> d + neutrino", is given as "p(p,)d." Each reaction is added to a list comprising the reaction network, and once completed, the network contains a list of all isotopes and reactions involved. brulilo then parses publicly available nuclear data (binding energy, spin, partition function data, etc.) for each isotope from the WebNucleo group's data at nucleo.ces.clemson.edu. Fits to reaction rate data for each reaction are obtained from the publicly available data provided by the Joint Institute for Nuclear Astrophysics' Reaclib database. Reverse reactions are calculated from detailed balance. The reaction network then consists of a sparse system of mathematically stiff ordinary differential equations (ODEs). Physically, "stiffness" here means that there are some reaction rates that are occurring much faster than others; the ratio of the fast to slow often being greater than 10^15 in astrophysics. Numerical integration of such systems typically involves implicit integration schemes, which brulilo will leverage from the scientific computing Python package, Numpy, as well as implement its own. There has recently been work from researchers at Oak Ridge National Lab who have found good results from applying quasi-equilibrium and quasi-asymptotic conditions to the stiff system of ODEs, which effectively remove some of the stiffness and allow for efficient explicit integration techniques to be used. The original intent of brulilo was to implement these stiffness-alleviating techniques with explicit integrators and compare the performance to traditional implicit integrations of the full stiff system. This is still underway, as the code is very much in an alpha-release state. Furthermore, explicit integrators are often much easier to parallelize than their implicit counterparts. brulilo will implement parallelization of these techniques, leveraging both the Python implementation of MPI, mpi4py, as well as highly parallelized versions targeted at GPUs with PyOpenCL and/or PyCUDA.

Malone, Chris↗