Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “loss functions”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Analyzing and Exploring Training Recipes for Large-Scale Transformer-Based Weather Prediction

Abstract The rapid rise of deep learning (DL) in numerical weather prediction (NWP) has led to a proliferation of models which forecast atmospheric variables with comparable or superior skill than traditional physics-based NWP. However, among these leading DL models, there is a wide variance in both the training settings and architecture used. Further, the lack of thorough ablation studies makes it hard to discern which components are most critical to success. In this work, we show that it is possible to attain high forecast skill even with relatively off-the-shelf architectures, simple training procedures, and moderate compute budgets. Specifically, we train a minimally modified Swin Transformer V2 (SwinV2) on ERA5 data and find that it attains superior skill in terms of mean-square errors of deterministic forecasts when compared against the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS). Almost all DL–NWP systems share a core set of hyperparameters and design decisions. To aid and expedite future DL–NWP research, we present an in-depth, systematic exploration of different loss functions, model sizes and depths, patch sizes, and multistep training objectives. We also examine the model performance with metrics beyond the typical accuracy (ACC) and RMSE and investigate how the performance scales with model size. Through our open-source code, scoring pipelines, and models, we share our findings on key aspects of the training pipeline. These ablations reduce the necessity for expensive hyperparameter tuning and lower the barrier to entry for future DL–NWP research. Significance Statement This study investigates the potential of using large-scale transformer-based models for weather prediction, showing that it is possible to achieve high forecast accuracy with simpler, off-the-shelf architectures. By training a minimally modified SwinV2 transformer on ERA5 data, we show that the model achieves competitive forecast skill in terms of mean-square error for key variables, outperforming the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS) at all lead times. Our findings suggest that effective training strategies, such as multistep fine-tuning and channel-weighted losses, significantly enhance the model’s performance. However, we also highlight that these improvements come with trade-offs in other areas, such as ensemble spread and high-frequency spatial detail. This work highlights the promise of deep learning in improving weather forecasts, which could lead to better preparedness and response to weather events, ultimately benefiting society by providing more reliable weather predictions.

Willard, Jared D. [Lawrence Berkeley National Labo↗
NASA NTRS2015

Rao-Blackwellization for Adaptive Gaussian Sum Nonlinear Model Propagation

When dealing with imperfect data and general models of dynamic systems, the best estimate is always sought in the presence of uncertainty or unknown parameters. In many cases, as the first attempt, the Extended Kalman filter (EKF) provides sufficient solutions to handling issues arising from nonlinear and non-Gaussian estimation problems. But these issues may lead unacceptable performance and even divergence. In order to accurately capture the nonlinearities of most real-world dynamic systems, advanced filtering methods have been created to reduce filter divergence while enhancing performance. Approaches, such as Gaussian sum filtering, grid based Bayesian methods and particle filters are well-known examples of advanced methods used to represent and recursively reproduce an approximation to the state probability density function (pdf). Some of these filtering methods were conceptually developed years before their widespread uses were realized. Advanced nonlinear filtering methods currently benefit from the computing advancements in computational speeds, memory, and parallel processing. Grid based methods, multiple-model approaches and Gaussian sum filtering are numerical solutions that take advantage of different state coordinates or multiple-model methods that reduced the amount of approximations used. Choosing an efficient grid is very difficult for multi-dimensional state spaces, and oftentimes expensive computations must be done at each point. For the original Gaussian sum filter, a weighted sum of Gaussian density functions approximates the pdf but suffers at the update step for the individual component weight selections. In order to improve upon the original Gaussian sum filter, Ref. [2] introduces a weight update approach at the filter propagation stage instead of the measurement update stage. This weight update is performed by minimizing the integral square difference between the true forecast pdf and its Gaussian sum approximation. By adaptively updating each component weight during the nonlinear propagation stage an approximation of the true pdf can be successfully reconstructed. Particle filtering (PF) methods have gained popularity recently for solving nonlinear estimation problems due to their straightforward approach and the processing capabilities mentioned above. The basic concept behind PF is to represent any pdf as a set of random samples. As the number of samples increases, they will theoretically converge to the exact, equivalent representation of the desired pdf. When the estimated qth moment is needed, the samples are used for its construction allowing further analysis of the pdf characteristics. However, filter performance deteriorates as the dimension of the state vector increases. To overcome this problem Ref. [5] applies a marginalization technique for PF methods, decreasing complexity of the system to one linear and another nonlinear state estimation problem. The marginalization theory was originally developed by Rao and Blackwell independently. According to Ref. [6] it improves any given estimator under every convex loss function. The improvement comes from calculating a conditional expected value, often involving integrating out a supportive statistic. In other words, Rao-Blackwellization allows for smaller but separate computations to be carried out while reaching the main objective of the estimator. In the case of improving an estimator's variance, any supporting statistic can be removed and its variance determined. Next, any other information that dependents on the supporting statistic is found along with its respective variance. A new approach is developed here by utilizing the strengths of the adaptive Gaussian sum propagation in Ref. [2] and a marginalization approach used for PF methods found in Ref. [7]. In the following sections a modified filtering approach is presented based on a special state-space model within nonlinear systems to reduce the dimensionality of the optimization problem in Ref. [2]. First, the adaptive Gaussian sum propagation is explained and then the new marginalized adaptive Gaussian sum propagation is derived. Finally, an example simulation is presented.

state estimation↗
DOE OSTI2024

Accurate and Data‐Efficient Micro X‐ray Diffraction Phase Identification Using Multitask Learning: Application to Hydrothermal Fluids

Traditional analysis of highly distorted micro X‐ray diffraction (μ‐XRD) patterns from hydrothermal fluid environments is a time‐consuming process, often requiring substantial data preprocessing and labeled experimental data. Herein, the potential of deep learning with a multitask learning (MTL) architecture to overcome these limitations is demonstrated. MTL models are trained to identify phase information in μ‐XRD patterns, minimizing the need for labeled experimental data and masking preprocessing steps. Notably, MTL models show superior accuracy compared to binary classification convolutional neural networks. Additionally, introducing a tailored cross‐entropy loss function improves MTL model performance. Most significantly, MTL models tuned to analyze raw and unmasked XRD patterns achieve close performance to models analyzing preprocessed data, with minimal accuracy differences. This work indicates that advanced deep learning architectures like MTL can automate arduous data handling tasks, streamline the analysis of distorted XRD patterns, and reduce the reliance on labor‐intensive experimental datasets.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Data‐driven variational method for discrepancy modeling: Dynamics with small‐strain nonlinear elasticity and viscoelasticity

Abstract The effective inclusion of a priori knowledge when embedding known data in physics‐based models of dynamical systems can ensure that the reconstructed model respects physical principles, while simultaneously improving the accuracy of the solution in the previously unseen regions of state space. This paper presents a physics‐constrained data‐driven discrepancy modeling method that variationally embeds known data in the modeling framework. The hierarchical structure of the method yields fine scale variational equations that facilitate the derivation of residuals which are comprised of the first‐principles theory and sensor‐based data from the dynamical system. The embedding of the sensor data via residual terms leads to discrepancy‐informed closure models that yield a method which is driven not only by boundary and initial conditions, but also by measurements that are taken at only a few observation points in the target system. Specifically, the data‐embedding term serves as residual‐based least‐squares loss function, thus retaining variational consistency. Another important relation arises from the interpretation of the stabilization tensor as a kernel function, thereby incorporating a priori knowledge of the problem and adding computational intelligence to the modeling framework. Numerical test cases show that when known data is taken into account, the data driven variational (DDV) method can correctly predict the system response in the presence of several types of discrepancies. Specifically, the damped solution and correct energy time histories are recovered by including known data in the undamped situation. Morlet wavelet analyses reveal that the surrogate problem with embedded data recovers the fundamental frequency band of the target system. The enhanced stability and accuracy of the DDV method is manifested via reconstructed displacement and velocity fields that yield time histories of strain and kinetic energies which match the target systems. The proposed DDV method also serves as a procedure for restoring eigenvalues and eigenvectors of a deficient dynamical system when known data is taken into account, as shown in the numerical test cases presented here.

Masud, Arif↗
DOE OSTI2024

Learning likelihood ratios with neural network classifiers

The likelihood ratio is a crucial quantity for statistical inference in science that enables hypothesis testing, construction of confidence intervals, reweighting of distributions, and more. Many modern scientific applications, however, make use of data- or simulation-driven models for which computing the likelihood ratio can be very difficult or even impossible. By applying the so-called “likelihood ratio trick,” approximations of the likelihood ratio may be computed using clever parametrizations of neural network-based classifiers. A number of different neural network setups can be defined to satisfy this procedure, each with varying performance in approximating the likelihood ratio when using finite training data. We present a series of empirical studies detailing the performance of several common loss functionals and parametrizations of the classifier output in approximating the likelihood ratio of two univariate and multivariate Gaussian distributions as well as simulated high-energy particle physics datasets.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2021

Enhancing searches for resonances with machine learning and moment decomposition

A key challenge in searches for resonant new physics is that classifiers trained to enhance potential signals must not induce localized structures. Such structures could result in a false signal when the background is estimated from data using sideband methods. A variety of techniques have been developed to construct classifiers which are independent from the resonant feature (often a mass). Such strategies are sufficient to avoid localized structures, but are not necessary. We develop a new set of tools using a novel moment loss function (Moment Decomposition or MoDe) which relax the assumption of independence without creating structures in the background. Here, by allowing classifiers to be more flexible, we enhance the sensitivity to new physics without compromising the fidelity of the background estimation.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Characterizing 4-string contact interaction using machine learning

Abstract The geometry of 4-string contact interaction of closed string field theory is characterized using machine learning. We obtain Strebel quadratic differentials on 4-punctured spheres as a neural network by performing unsupervised learning with a custom-built loss function. This allows us to solve for local coordinates and compute their associated mapping radii numerically. We also train a neural network distinguishing vertex from Feynman region. As a check, 4-tachyon contact term in the tachyon potential is computed and a good agreement with the results in the literature is observed. We argue that our algorithm is manifestly independent of number of punctures and scaling it to characterize the geometry ofn-string contact interaction is feasible.

Physics↗
DOE OSTI2024

Physical signatures of fermion-coupled axion dark matter

In the presence of axion dark matter, fermion spins experience an “axion wind” torque and an “axioelectric” force. We investigate new experimental probes of these effects and find that magnetized analogs of multilayer dielectric haloscopes can explore orders of magnitude of new parameter space for the axion-electron coupling. We also revisit the calculation of axion absorption into in-medium excitations, showing that axioelectric absorption is screened in spin-polarized targets, and axion wind absorption can be characterized in terms of a magnetic energy loss function. Finally, our detailed theoretical treatment allows us to critically examine recent claims in the literature. We find that axioelectric corrections to electronic energy levels are smaller than previously estimated and that the purported electron electric dipole moment due to a constant axion field is entirely spurious.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Enhancing Interpretability in Generative Modeling: Statistically Disentangled Latent Spaces Guided by Generative Factors in Scientific Datasets

This study addresses the challenge of statistically extracting generative factors from complex, high-dimensional datasets in unsupervised or semi-supervised settings. We investigate encoder-decoder-based generative models for nonlinear dimensionality reduction, focusing on disentangling low-dimensional latent variables corresponding to independent physical factors. Introducing Aux-VAE, a novel architecture within the classical Variational Autoencoder framework, we achieve disentanglement with minimal modifications to the standard VAE loss function by leveraging prior statistical knowledge through auxiliary variables. These variables guide the shaping of the latent space by aligning latent factors with learned auxiliary variables. We validate the efficacy of Aux-VAE through comparative assessments on multiple datasets, including astronomical simulations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Error-mitigated data-driven circuit learning on noisy quantum hardware

Application-level benchmarks measure how well a quantum device performs meaningful calculations. In the case of parameterized circuit training, the computational task is the preparation of a target quantum state via optimization over a loss landscape. This is complicated by various sources of noise, fixed hardware connectivity, and generative modeling, the choice of target distribution. Gradient-based training has become a useful benchmarking task for noisy intermediate-scale quantum computers because of the additional requirement that the optimization step uses the quantum device to estimate the loss function gradient. In this work, we use gradient-based data-driven circuit learning to qualitatively evaluate the performance of several superconducting platform devices and present results that show how error mitigation can improve the training of quantum circuit Born machines with 28 tunable parameters.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2025

Physics-Informed Neural Networks for PDE-Constrained Optimization and Control

The goal of optimal control is to determine a sequence of inputs for maximizing or minimizing a given performance criterion subject to the dynamics and constraints of the system under observation. This work introduces Control Physics-Informed Neural Networks (PINNs), which simultaneously learn both the system states and the optimal control signal in a single-stage framework that leverages the system’s underlying physical laws. While prior approaches often follow a two-stage process-modeling, the system first and then devising its control—the presented novel framework embeds the necessary optimality conditions directly into the network architecture and loss function. We demonstrate the effectiveness of the novel methodology by solving various open-loop optimal control problems governed by analytical, one-dimensional, and two-dimensional partial differential equations (PDEs).

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

GW-PINN: A deep learning algorithm for solving groundwater flow equations

Machine learning methods provide new perspective for more convenient and efficient prediction of groundwater flow. In this study, a deep learning method “GW-PINN” without labeled data for solving groundwater flow equations with wells was proposed. GW-PINN takes the physics inform neural network (PINN) as the backbone and uses either the hard or soft constraint in the loss function for training. A locally refined sampling strategy (LRS) is adopted to generate the consistent spatial sampling points for problems with strong hydraulic head change, and then combined with an appropriate temporal sampling scheme to obtain the final spatial-temporal sampling points. A snowball-style two-stage training strategy by dividing the temporal domain into two subdomains is designed to decrease the sampling points. Five cases were designed to test the training performance of GW-PINN under different sampling strategies and two constraints. The predicted results of GW-PINN were compared with MODFLOW and the analytical solution. The results demonstrate that GW-PINN possesses strong ability in capturing the hydraulic head change for both confined and un-confined aquifers. The hard constraint owns more robust learning ability than the soft constraint. The LRS strategy can generate more accurate results with much fewer sampling points than traditional sampling strategies, and the snowball-style two-stage training strategy is significantly efficient for problems with the drastic change of hydraulic head. Additionally, the application of GW-PINN as a surrogate model for parameterized groundwater flow equations is illustrated. This study provides an option tool for efficient groundwater flow simulation, especially for those with local refinements are needed.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

ResNet and CycleGAN for pulse shape discrimination of He-4 detector pulses: Recovering pulses conventional algorithms fail to label unanimously

Pulse shape discrimination (PSD) capable detectors, such as He-4, that respond to neutron and gamma-ray 7 interactions have a threshold deposited energy value below which n/γ discrimination vanishes when using 8 conventional PSD algorithms. Recent attempts in applying supervised learning based artificial neural 9 networks for PSD use the pulses in the separated regions to train the networks so they can be used to classify 10 another set of separated pulses. In doing so, pulses previously indistinguishable are not recovered for 11 classification, which would have increased the number of neutron and gamma-ray pulses that could be used 12 for further analysis. Assuming the reason why conventional PSD algorithms have unseparated regions is 13 because the parameter space of the algorithms fail to capture the intrinsic (but subtle) distinguishing 14 behavior of some of the neutron and gamma-ray pulses, a cycle-consistent generative adversarial network 15 (CycleGAN) was trained to amplify those differences and extract well separated neutron and gamma-ray 16 clusters. Results show that, once the network is trained with pulses from separated and unseparated regions, 17 it was able to transform the pulses in the unseparated region to improve the PSD. Subsequent n/γ 18 classification was performed using deep residual network (ResNet) that takes pulses with 512 data points 19 as an input. Two different ResNets were explored – simple ResNet and modified ResNet which takes 20 segmented pulse inputs in the first layer and the corresponding time axis values in the last hidden layer. 21 The later approach enables the network to extract time correlated pulse features to enhance its ability to 22 capture the pulse behaviors relevant for PSD. Although it achieves slightly lower accuracy, 99.41% versus 23 99.89%, based on simply counting the number of correct n/γ labels assigned, compared to the simple 24 ResNet, the modified ResNets architecture was able to decreases the cross-entropy loss function by half, 25 which implies that the correct n/γ labels assigned are less likely to be accidental. PSD parameter 26 distributions based on n/γ classification by ResNet before and after transforming unseparated pulses using 27 CycleGAN show that by enhancing the separation between neutrons and gamma-rays, the transformation 28 helps improve the performance of classifier networks that are trained using labeled dataset. The 29 enhancement of neutron and gamma-ray separation by the CycleGAN increased the PSD figure of merit 30 (FOM) by up to 70% in some regions. Here, the results show that, if a given detector achieves clear separation 31 between neutron and gamma-ray pulses in any energy region, such neural network approaches can help 32 lower the energy threshold for the separation and increasing the number of neutron and gamma-ray pulses 33 that can be used for further analysis.

4He↗
DOE OSTI2023

GMT: A deep learning approach to generalized multivariate translation for scientific data analysis and visualization

In scientific visualization, despite the significant advances of deep learning for data generation, researchers have not thoroughly investigated the issue of data translation. We present a new deep learning approach called generalized multivariate translation (GMT) for multivariate time-varying data analysis and visualization. Like V2V, GMT assumes a preprocessing step that selects suitable variables for translation. However, unlike V2V, which only handles one-to-one variable translation during training and inference, GMT enables one-to-many and many-to-many variable translation in the same framework. We leverage the recent StarGAN design from multi-domain image-to-image translation to achieve this generalization capability. We experiment with different loss functions and injection strategies to explore the best choices and leverage pre-training for performance improvement. We compare GMT with other state-of-the-art methods (i.e., Pix2Pix, V2V, StarGAN). Furthermore, the results demonstrate the overall advantage of GMT in translation quality and generalization ability.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Machine learning accelerated discrete element modeling of granular flows

Granular flows are widely encountered in many industrial processes and natural phenomena. Discrete Element Modeling (DEM) is a useful tool for understanding and troubleshooting devices, which handle granular materials. However, its applicability is significantly limited by the huge computational cost associated with detecting and computing collisions. In this research, the computation speed of DEM was accelerated by orders of magnitude using a convolutional neural network to replace the direct calculation of particle-particle and particle-boundary collisions. The MFiX software was used to generate the training and testing dataset. Additionally, a GPU accelerated TensorFlow model was used to train the neural network and test the results. The model fluctuations caused by different training steps were reduced with a multi-scale loss function. The accuracy was improved with more frames within one training step. The modeling of a rotating drum and a hopper demonstrated the accuracy and efficiency of this machine learning accelerated DEM in the simulation of granular flows.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

COVID-19 dynamics across the US: A deep learning study of human mobility and social behavior

This paper presents a deep learning framework for epidemiology system identification from noisy and sparse observations with quantified uncertainty. The proposed approach employs an ensemble of deep neural networks to infer the time-dependent reproduction number of an infectious disease by formulating a tensor-based multi-step loss function that allows us to efficiently calibrate the model on multiple observed trajectories. The method is applied to a mobility and social behavior-based SEIR model of COVID-19 spread. The model is trained on Google and Unacast mobility data spanning a period of 66 days, and is able to yield accurate future forecasts of COVID-19 spread in 203 US counties within a time-window of 15 days. Interestingly, a sensitivity analysis that assesses the importance of different mobility and social behavior parameters reveals that attendance of close places, including workplaces, residential, and retail and recreational locations, has the largest impact on the effective reproduction number. Furthermore, the model enables us to rapidly probe and quantify the effects of government interventions, such as lock-down and re-opening strategies. Taken together, the proposed framework provides a robust workflow for data-driven epidemiology model discovery under uncertainty and produces probabilistic forecasts for the evolution of a pandemic that can judiciously provide information for policy and decision making. All codes and data accompanying this manuscript are available at https://github.com/PredictiveIntelligenceLab/DeepCOVID19.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Respecting causality for training physics-informed neural networks

While the popularity of physics-informed neural networks (PINNs) is steadily rising, to this date PINNs have not been successful in simulating dynamical systems whose solution exhibits multi-scale, chaotic or turbulent behavior. In this work we attribute this shortcoming to the inability of existing PINNs formulations to respect the spatio-temporal causal structure that is inherent to the evolution of physical systems. We argue that this is a fundamental limitation and a key source of error that can ultimately steer PINN models to converge towards erroneous solutions. We address this pathology by proposing a simple re-formulation of PINNs loss functions that can explicitly account for physical causality during model training. We demonstrate that this simple modification alone is enough to introduce significant accuracy improvements, as well as a practical quantitative mechanism for assessing the convergence of a PINNs model. Here we provide state-of-the-art numerical results across a series of benchmarks for which existing PINNs formulations fail, including the chaotic Lorenz system, the Kuramoto–Sivashinsky equation in the chaotic regime, and the Navier–Stokes equations. To the best of our knowledge, this is the first time that PINNs have been successful in simulating such systems, introducing new opportunities for their applicability to problems of industrial complexity.

97 MATHEMATICS AND COMPUTING↗