Search NASA⌕ Search

SEARCH · Search NASA

Results for “network function”

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 199 records · Page 11
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Improving the Transportability of a Deep Learning Denoising Model Using Transfer Learning Techniques

The adoption of machine learning techniques in the seismology community has led to great performance improvements in several areas, including signal processing. Specifically, the development of deep learning–based seismic waveform denoising models has the potential to yield improvements in signal detection capabilities for networks operating in particularly noisy environments. Recent advancements in the design of these deep learning denoising models have included the incorporation of continuous and discrete wavelet transform functions into the network architecture to improve the learning capabilities and efficiency of said models. These wavelet transform–based seismic denoising models have shown improved denoising capabilities in regions where there is good agreement between the data features present in the training and evaluation datasets. However, questions remain about the overall transportability of these models to other monitoring regions. Here, in this study, we will determine the baseline transportability of a newly developed multilevel wavelet‐transform convolutional neural network (MWCNN) seismic denoising model. We accomplish this by taking a version of the MWCNN denoising model trained on data collected from the Utah region and evaluating its denoising performance on datasets collected from the neighboring Nevada region, which differ with regard to monitoring sensor types and event histories. We find that there is a notable variability in denoising performance related to the degree of similarity between the initial and new target datasets. The most notable difference in denoising performance is the ability of the denoising model to preserve accurate amplitude information associated with the signal energy present in the waveform data. Finally, we evaluate the ability of transfer learning techniques to improve the transportability of the MWCNN denoising model. We find that although there is still a performance gap present in the denoising results of the MWCNN model, transfer learning did yield improved results.

Quinones, Louis [Sandia National Laboratories (SNL↗
DOE OSTI2020

A Look Inside the Black Box: Using graph-theoretical descriptors to interpret a Continuous-Filter Convolutional Neural Network (CF-CNN) trained on the global and local minimum energy structures of neutral water clusters

A Continuous Filter Convolutional Neural Network (CF-CNN) was trained to predict the potential energy of water cluster networks \ce{(H2O)_{\textit{N}}}, \textit{N}=10--30, corresponding to local minima lying within 5 kcal/mol from the putative minima taken from a newly published database containing over 5 million unique networks. The chemical sampling space of the database was characterized using chemical descriptors derived from graph theory, which led to the identification of important trends in the topology, connectivity, polygon structures associated with the various networks as a function of cluster size. The resulting graphs are available alongside the original database at \url{https://sites.uw.edu/wdbase/}. The CF-CNN trained on a subset of 500,000 networks for (\textit{N}=10, 30) yielded a mean absolute error of 0.002$\pm$0.002 kcal/mol per water molecule, giving the trained CF-CNN the highest accuracy of any neural network-based surrogate model to date. In addition, clusters of sizes not included in the training set exhibited errors of the same magnitude, indicating that the CF-CNN ptotocol is general enough to accurately predict energies of networks for both smaller and larger sizes than those used during training. The graph-theoretical descriptors were developed in order to analyze the properties of the full database and interpret the predictive power of the CF-CNN. Using topology measures, such as the Wiener index and the average shortest path length along with two similarity measures, we showed that all networks from the test set were within the range of the ones from the training set, suggesting that the training set covered the chemical space of interest quite well. Our graph analysis suggests that the mean degree and number of polygons for networks with larger errors tend to lie further from the mean than those with lower errors. The generality of the used CF-CNN was thus demonstrated, while the use of the graph-theoretical descriptors assisted in interpreting the predicted results.

Bilbrey, Jenna A.↗
DOE OSTI2024

LossLens: Diagnostics for Machine Learning Through Loss Landscape Visual Analytics

Modern machine learning often relies on optimizing a neural network's parameters using a loss function to learn complex features. Beyond training, examining the loss function with respect to a network's parameters (i.e., as a loss landscape) can reveal insights into the architecture and learning process. While the local structure of the loss landscape surrounding an individual solution can be characterized using a variety of approaches, the global structure of a loss landscape, which includes potentially many local minima corresponding to different solutions, remains far more difficult to conceptualize and visualize. To address this difficulty, we introduce LossLens, a visual analytics framework that explores loss landscapes at multiple scales. LossLens integrates metrics from global and local scales into a comprehensive visual representation, enhancing model diagnostics. Here we demonstrate LossLens through two case studies: visualizing how residual connections influence a ResNet-20, and visualizing how physical parameters influence a physics-informed neural network (PINN) solving a simple convection problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Kinetics-Informed Neural Networks

Chemical kinetics and reaction engineering consists of the phenomenological framework for the disentanglement of reaction mechanisms, optimization of reaction performance and the rational design of chemical processes. Here, we utilize feed-forward artificial neural networks as basis functions to solve ordinary differential equations (ODEs) constrained by differential algebraic equations (DAEs) that describe microkinetic models (MKMs). We present an algebraic framework for the mathematical description and classification of reaction networks, types of elementary reaction, and chemical species. Under this framework, we demonstrate that the simultaneous training of neural nets and kinetic model parameters in a regularized multi-objective optimization setting leads to the solution of the inverse problem through the estimation of kinetic parameters from synthetic experimental data. We analyze a set of scenarios to establish the extent to which kinetic parameters can be retrieved from transient kinetic data, and assess the robustness of the methodology with respect to statistical noise. Furthermore, this approach to inverse kinetic ODEs can assist in the elucidation of reaction mechanisms based on transient data.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Gradient-enhanced physics-informed neural networks for forward and inverse PDE problems

Deep learning has been shown to be an effective tool in solving partial differential equations (PDEs) through physics-informed neural networks (PINNs). PINNs embed the PDE residual into the loss function of the neural network, and have been successfully employed to solve diverse forward and inverse PDE problems. However, one disadvantage of the first generation of PINNs is that they usually have limited accuracy even with many training points. Here, we propose a new method, gradient-enhanced physics-informed neural networks (gPINNs), for improving the accuracy of PINNs. gPINNs leverage gradient information of the PDE residual and embed the gradient into the loss function. We tested gPINNs extensively and demonstrated the effectiveness of gPINNs in both forward and inverse PDE problems. Our numerical results show that gPINN performs better than PINN with fewer training points. Additionally, we combined gPINN with the method of residual-based adaptive refinement (RAR), a method for improving the distribution of training points adaptively during training, to further improve the performance of gPINN, especially in PDEs with solutions that have steep gradients.

42 ENGINEERING↗
DOE OSTI2022

Throughput Analytics of Cloud Networks

A network of virtual machines at cloud server sites connected over virtual IO connections is a flexible, easily deployable, and cost-effective alternative to a physical network infrastructure with dedicated servers connected over leased fiber lines. We study the throughput performance of such a cloud network by collecting measurements over Google Cloud infrastructure spanning multiple continents. To study its ideal performance and impact of packet losses, we utilize its emulation using dedicated servers and connection hardware emulation devices. We compare the measurements over the cloud network to those over its emulation on a testbed. We examine the throughput profiles of both networks as a function of the round trip time and their utilization-concavity coefficients, estimated using measurements for common TCP versions. The throughput profile's concave-convex shape and its coefficient are critical indicators of the network performance, qualitatively and quantitatively, respectively. The results indicate their overall agreement between the production cloud network and its emulation using dedicated connections, and a near optimal throughput performance of the former except for a few under-performing connections. Also, the number of parallel flows is found to be a dominant factor in optimizing the throughput across various conditions and TCP versions.

Phanekham, Derek↗
DOE OSTI2020

Structural conserved moiety splitting of a stoichiometric matrix

Characterising biochemical reaction network structure in mathematical terms enables the inference of functional biochemical consequences from network structure with existing mathematical techniques and spurs the development of new mathematics that exploits the peculiarities of biochemical network structure. The structure of a biochemical network may be specified by reaction stoichiometry, that is, the relative quantities of each molecule produced and consumed in each reaction of the network. A biochemical network may also be specified at a higher level of resolution in terms of the internal structure of each molecule and how molecular structures are transformed by each reaction in a network. The stoichiometry for a set of reactions can be compiled into a stoichiometric matrix N ϵ Z mxn where each row corresponds to a molecule and each column corresponds to a reaction. We demonstrate that a stoichiometric matrix may be split into the sum of m - rank(N) moiety transition matrices, each of which corresponds to a subnetwork accessible to a structurally identifiable conserved moiety. The existence of this moiety matrix splitting is a property that distinguishes a stoichiometric matrix from an arbitrary rectangular matrix.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Functional Data Analysis in Wearable Body Sensor Networks

Improving response time of indirect room-size calorimeters is still an outstanding problem in metabolic research. Accurate estimates of instantaneous rates of gaseous exchange require numerical differentiation of measured gaseousgas concentrations. We propose a new method to estimate the instantaneous gaseousgas exchange rates in indirect calorimetry. In contrast to the previously developed techniques, the method addresses the problem of differentiation of gaseous concentrations as an ill-posed problem. By applying the method of regularization, the problem of differentiation is converted into a well-posed problem resulting in smooth and consistent gaseous exchange rates. The validity of the method is tested on a large dataset of calorimeter experiments which included 313 human experiments along with 231 alcohol combustion experiments. It is demonstrated that the method is able to reliably differentiate between the “unphysiological” process of alcohol combustion and physiological variations produced by human metabolism. The method also allowed unraveling the previously unreported relative kinetics of O2 consumption and Respiratory Quotient (RQ) in humans. It was found that the kinetics of oxidative fuel selection lags behind the energy expenditure in humans exhibiting some sort of oxidative inertia. The time lag varies from 2-3 min up to 30 min, depending on particular individual. No such lag was found in alcohol combustion experiments. In addition to the relative kinetics of substrate oxidation, two statistical indexes reflecting variability of minute-by-minute RQ were estimated. The indexes were the RQ’s standard deviation and RQ’s first-order derivative. Both indexes showed statistically significant difference between human experiments and alcohol combustion experiments. We conclude that the proposed method can consistently extract physiologically-relevant information from noisy calorimetry data and the aforesaid information can provide additional insights into the mechanism of metabolic fuel selection in humans.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Functional Data Analysis in Wearable Body Sensor Networks

Improving response time of indirect room-size calorimeters is still an outstanding problem in metabolic research. Accurate estimates of instantaneous rates of gaseous exchange require numerical differentiation of measured gaseousgas concentrations. We propose a new method to estimate the instantaneous gaseousgas exchange rates in indirect calorimetry. In contrast to the previously developed techniques, the method addresses the problem of differentiation of gaseous concentrations as an ill-posed problem. By applying the method of regularization, the problem of differentiation is converted into a well-posed problem resulting in smooth and consistent gaseous exchange rates. The validity of the method is tested on a large dataset of calorimeter experiments which included 313 human experiments along with 231 alcohol combustion experiments. It is demonstrated that the method is able to reliably differentiate between the “unphysiological” process of alcohol combustion and physiological variations produced by human metabolism. The method also allowed unraveling the previously unreported relative kinetics of O2 consumption and Respiratory Quotient (RQ) in humans. It was found that the kinetics of oxidative fuel selection lags behind the energy expenditure in humans exhibiting some sort of oxidative inertia. The time lag varies from 2-3 min up to 30 min, depending on particular individual. No such lag was found in alcohol combustion experiments. In addition to the relative kinetics of substrate oxidation, two statistical indexes reflecting variability of minute-by-minute RQ were estimated. The indexes were the RQ’s standard deviation and RQ’s first-order derivative. Both indexes showed statistically significant difference between human experiments and alcohol combustion experiments. We conclude that the proposed method can consistently extract physiologically-relevant information from noisy calorimetry data and the aforesaid information can provide additional insights into the mechanism of metabolic fuel selection in humans.

60 - APPLIED LIFE SCIENCES↗
DOE OSTI2022

Faster network disruption from layered oscillatory dynamics

Nonlinear complex network-coupled systems typically have multiple stable equilibrium states. Following perturbations or due to ambient noise, the system is pushed away from its initial equilibrium, and, depending on the direction and the amplitude of the excursion, it might undergo a transition to another equilibrium. It was recently demonstrated [M. Tyloo, J. Phys. Complex. 3 03LT01 (2022)] that layered complex networks may exhibit amplified fluctuations. Here, I investigate how noise with system-specific correlations impacts the first escape time of nonlinearly coupled oscillators. Interestingly, I show that, not only the strong amplification of the fluctuations is a threat to the good functioning of the network but also the spatial and temporal correlations of the noise along the lowest-lying eigenmodes of the Laplacian matrix. Finally, I analyze first escape times on synthetic networks and compare noise originating from layered dynamics to uncorrelated noise.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Deep Learning without Global Optimization by Random Fourier Neural Networks

Here we introduce a new training algorithm for deep neural networks that utilize random complex exponential activation functions. Our approach employs a Markov chain Monte Carlo sampling procedure to iteratively train network layers, avoiding global and gradient-based optimization while maintaining error control. It consistently attains the theoretical approximation rate for residual networks with complex exponential activation functions, determined by network complexity. Additionally, it enables efficient learning of multiscale and high-frequency features, producing interpretable parameter distributions. Despite using sinusoidal basis functions, we do not observe Gibbs phenomena in approximating discontinuous target functions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Changes in microbial community and network structure precede shrub degradation in a desert ecosystem

Large-scale restoration is intended to promote ecological recovery. Improvements in plant and microbial conditions, however, may slow or even reverse in late succession. To better understand long-term restoration outcomes and underlying drivers of successional pathways, we tracked plant, bacterial and fungal, and soil conditions across a 40-year shrub plantation that was intended to stabilize desertified land in northern China. Here, we found that planted Haloxylon ammodendron shrubs developed and then subsequently became degraded after 30–40 years. Bacterial abundance and α-diversity were much higher than those of fungi, but no significant differences in composition and structure were found in different plantation ages. In contrast, the dominant taxa of fungal communities shifted from symbiotroph and saprotroph species towards pathotroph species with increased soil nutrients in the plantation chronosequence after two decades. The changes in fungal dominant species led to a transition in microbial network structure and function, with an increase in negative linkages among taxa that began in the middle stages of succession. Changes in fungal community structure had direct and indirect negative effects on shrub leaf physiology, root activity, and biomass. Our results highlight the preceding role of a breakdown in soil microbial community composition and network structure on the degradation of shrub performance in long-term desert succession. Our study emphasizes the importance of understanding soil-microbial-plant linkages on restoration outcomes, and mechanisms that can slow or reverse the recovery of ecosystems.

Bacterial and fungal community composition↗
DOE OSTI2020

Stress effects on flow and transport in three-dimensional fracture networks

We investigate the effects of various external stress regimes on fracture apertures, fluid flow, and solute transport in three-dimensional fracture networks. We use well-established geomechanics equations coupled with discrete fracture network modeling to characterize changes in primary flow paths within a complex network as a function of stress magnitude and orientation. These changes manifest in the alterations of the fluid flow field and are measured in terms of Eulerian and Lagrangian flow observables including solute transport, which is a key problem in many hydrologic applications. Changes in primary flow paths affect the solute transport in the network by promoting anomalously early arrival or long tailing behavior. However, early time arrival is not ubiquitous in anisotropically stressed networks and in most cases there is a delayed arrival of solute, which is attributed to (i) the presence of low-velocity zones normal to the flow direction, (ii) the angle between flow direction and major compressive principal stress directions, and (iii) changes in the primary flow paths (i.e., increases in tortuosity and active network structure). Overall, flows become more channelized in anisotropically stressed fracture networks than in unstressed and isotropically stressed networks.

58 GEOSCIENCES↗
DOE OSTI2020

Comparing the performance of density functionals in describing the adsorption of atoms and small molecules on Ni(111)

Reaction networks calculated using density functional theory (DFT) play a crucial role in understanding and predicting heterogeneous catalyst performance. However, results obtained using different density functionals can vary significantly. Accordingly, it is important to assess the accuracy of available functionals in capturing adsorption energetics derived from single-crystal adsorption microcalorimetry experiments. Here, we evaluate how different density functionals describe the molecular and dissociative adsorption of CH 3 I, the adsorption of CH 3 , I, and H, and the dissociative adsorption of CH 4 on Ni(111). First, we identify energetically-preferred adsorbed states for each adsorbate at three surface coverages and calculate the corresponding enthalpies of adsorption at 160 K. Then, we compare the results to experimental measurements published in the literature. For each adsorption process we find at least one functional that is quantitatively accurate within experimental error, but no functional is accurate (within experimental error) for all adsorption processes studied. Even when assuming an additional DFT error of ±20 kJ/mol on top of the experimental error, only the PBE-D3 and RPBE-D3 functionals are accurate for all considered adsorption systems. Overall, these results indicate that quantitative agreement between density functional theory calculations and experimental measurements is both system- and functional-dependent. Importantly, single-crystal adsorption microcalorimetry experiments will continue to play an important role for benchmarking density functionals.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Densely Connected G-invariant Deep Neural Networks with Signed Permutation Representations

We introduce and investigate, for finite groups G, G-invariant deep neural network (GDNN) architectures with ReLU activation that are densely connected- i.e., include all possible skip connections. In contrast to other G-invariant architectures in the literature, the preactivations of theG-DNNs presented here are able to transform by signed permutation representations (signed perm-reps) of G. Moreover, the individual layers of the G-DNNs are not required to be G-equivariant; instead, the preactivations are constrained to be G-equivariant functions of the network input in a way that couples weights across all layers. The result is a richer family of G-invariant architectures never seen previously. We derive an efficient implementation of G-DNNs after a reparameterization of weights, as well as necessary and sufficient conditions for an architecture to be "admissible"- i.e., nondegenerate and inequivalent to smaller architectures. We include code that allows a user to build a G-DNN interactively layer-by-layer, with the final architecture guaranteed to be admissible. We show that there are far more admissible G-DNN architectures than those accessible with the "concatenated ReLU" activation function from the literature. Finally, we apply G-DNNs to two example problems--(1) multiplication in --1, 1} (with theoretical guarantees) and (2) 3D object classification--finding that the inclusion of signed perm-reps significantly boosts predictive performance compared to baselines with only ordinary (i.e., unsigned) perm-reps.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Adaptive time scaling for accelerating explicit finite element analysis

A method for accelerating an explicit finite element analysis (FEA) simulation of a modeled system or process includes performing an initial iteration of the FEA simulation according to a baseline time interval via an FEA computing network, and calculating a criteria ratio of a predetermined set of scaling criteria for the modeled system or process. The method includes determining a time-scaling factor using the criteria ratio via the FEA computing network as a function of the criteria ratio, and then applying the time-scaling factor to the baseline time interval to generate a scaled time interval. The scaled time interval accelerates simulation time of the FEA simulation. The method includes performing a subsequent iteration of the explicit FEA simulation at the scaled time interval using the FEA computing network. The process continues for subsequent iterations, with the time-scaling factor adapting with each iteration.

Chen, Jian↗
DOE OSTI2021

Spectroscopic Signatures of MQ-Resins in Silicone Elastomers

Polysiloxane elastomers have a large application space due to their versatile cross-linking chemistry and highly tunable physical and mechanical properties. One approach for improving the mechanical integrity of commercial polysiloxane “silicone” elastomers while maintaining their optical transparency is the addition of small, silicone-resin molecules to the network. However, both the poly(dimethylsiloxane) (PDMS) network and the silicone-resin particles have an amorphous structure and complex chemistry, which makes the characterization of their structural properties and segmental network dynamics difficult. Here, we report the synthesis and characterization of a series of model silicone networks modified with a specific class of silicone-resin known as MQ-resin using Raman and advanced nuclear magnetic resonance (NMR) spectroscopy methods. Raman spectroscopy was successfully used to quantify the contribution of the MQ-resin to the network, to determine the type of MQ-resin present in the network, and to investigate the completeness of the network cross-linking reaction. Furthermore, solid-state and 1H double-quantum (DQ) NMR spectroscopies were used not only as a detection method for the MQ-resins but also to quantify changes in the segmental dynamics of the network as a function of MQ-resin concentration. The combination of Raman and NMR spectroscopies describes a series of samples where the MQ-resin particles and PDMS chains maintain their independent segmental dynamics up to high concentrations of MQ-resins (40–50% MQ), where the physical properties of the resin dominate the physical properties of the overall network. The results from our spectroscopic analyses are consistent with the results from macroscopic characterization techniques such as solvent uptake and mechanical testing. The spectroscopic insights into the structure–property relationships of PDMS-MQ composites presented in this study are a valuable tool not only for the synthesis and reverse engineering of future generations of commercial silicone elastomers but also for understanding the mechanisms of aging and degradation over the material lifetime.

36 MATERIALS SCIENCE↗