Search NASA⌕ Search

SEARCH · Search NASA

Results for “overfitting”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Epistatic Net allows the sparse spectral regularization of deep neural networks for inferring fitness functions

Abstract Despite recent advances in high-throughput combinatorial mutagenesis assays, the number of labeled sequences available to predict molecular functions has remained small for the vastness of the sequence space combined with the ruggedness of many fitness functions. While deep neural networks (DNNs) can capture high-order epistatic interactions among the mutational sites, they tend to overfit to the small number of labeled sequences available for training. Here, we developed Epistatic Net (EN), a method for spectral regularization of DNNs that exploits evidence that epistatic interactions in many fitness functions are sparse. We built a scalable extension of EN, usable for larger sequences, which enables spectral regularization using fast sparse recovery algorithms informed by coding theory. Results on several biological landscapes show that EN consistently improves the prediction accuracy of DNNs and enables them to outperform competing models which assume other priors. EN estimates the higher-order epistatic interactions of DNNs trained on massive sequence spaces-a computational problem that otherwise takes years to solve.

Aghazadeh, Amirali (ORCID:0000000302230873)↗
DOE OSTI2024

Unveiling the effect of composition on nuclear waste immobilization glasses’ durability by nonparametric machine learning

Abstract Ensuring the long-term chemical durability of glasses is critical for nuclear waste immobilization operations. Durable glasses usually undergo qualification for disposal based on their response to standardized tests such as the product consistency test or the vapor hydration test (VHT). The VHT uses elevated temperature and water vapor to accelerate glass alteration and the formation of secondary phases. Understanding the relationship between glass composition and VHT response is of fundamental and practical interest. However, this relationship is complex, non-linear, and sometimes fairly variable, posing challenges in identifying the distinct effect of individual oxides on VHT response. Here, we leverage a dataset comprising 654 Hanford low-activity waste (LAW) glasses across a wide compositional envelope and employ various machine learning techniques to explore this relationship. We find that Gaussian process regression (GPR), a nonparametric regression method, yields the highest predictive accuracy. By utilizing the trained model, we discern the influence of each oxide on the glasses’ VHT response. Moreover, we discuss the trade-off between underfitting and overfitting for extrapolating the material performance in the context of sparse and heterogeneous datasets.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Comparison of machine learning techniques to optimize the analysis of plutonium surrogate material via a portable LIBS device

The utilization of machine learning techniques has become commonplace in the analysis of optical emission spectra. These methods are often limited to variants of principal components analysis (PCA), partial-least squares (PLS), and artificial neural networks (ANNs). A plethora of other techniques exist and are well established in the world of data science, yet are seldom investigated for their use in spectroscopic problems. In this study, machine learning techniques were used to analyze optical emission spectra of laser-induced plasma from ceria pellets doped with silicon in order to predict silicon content. Additionally, a boosted regression ensemble model was created, and its predictive accuracy was compared to that of traditional PCA, PLS, and ANN regression models. Boosted regression tree ensembles yielded fits with R-squared (R2) values as high as 0.964 and mean-squared errors of prediction (MSEPs) as low as 0.074, providing the most accurate predictive model. Neural networks performed with slightly lower R2 values and higher MSEPs compared to the ensemble methods, thus indicating susceptibility to overfitting.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Characterizing magnetized plasmas with dynamic mode decomposition

Accurate and efficient plasma models are essential to understand and control experimental devices. Existing magnetohydrodynamic or kinetic models are nonlinear and computationally intensive and can be difficult to interpret, while often only approximating the true dynamics. In this work, data-driven techniques recently developed in the field of fluid dynamics are leveraged to develop interpretable reduced-order models of plasmas that strike a balance between accuracy and efficiency. In particular, dynamic mode decomposition (DMD) is used to extract spatio-temporal magnetic coherent structures from the experimental and simulation datasets of the helicity injected torus with steady inductive (HIT-SI) experiment. Three-dimensional magnetic surface probes from the HIT-SI experiment are analyzed, along with companion simulations with synthetic internal magnetic probes. A number of leading variants of the DMD algorithm are compared, including the sparsity-promoting and optimized DMD. Optimized DMD results in the highest overall prediction accuracy, while sparsity-promoting DMD yields physically interpretable models that avoid overfitting. These DMD algorithms uncover several coherent magnetic modes that provide new physical insights into the inner plasma structure. These modes were subsequently used to discover a previously unobserved three-dimensional structure in the simulation, rotating at the second injector harmonic. Finally, using data from probes at experimentally accessible locations, DMD identifies a resistive kink mode, a ubiquitous instability seen in magnetized plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2024

Enhancing high-fidelity neural network potentials through low-fidelity sampling

The efficacy of neural network potentials (NNPs) critically depends on the quality of the configurational datasets used for training. Prior research using empirical potentials has shown that well-selected liquid–solid transitional configurations of a metallic system can be translated to other metallic systems. This study demonstrates that such validated configurations can be relabeled using density functional theory (DFT) calculations, thereby enhancing the development of high-fidelity NNPs. Training strategies and sampling approaches are efficiently assessed using empirical potentials and subsequently relabeled via DFT in a highly parallelized fashion for high-fidelity NNP training. Our results reveal that relying solely on energy and force for NNP training is inadequate to prevent overfitting, highlighting the necessity of incorporating stress terms into the loss functions. To optimize training involving force and stress terms, we propose employing transfer learning to fine-tune the weights, ensuring that the potential surface is smooth for these quantities composed of energy derivatives. This approach markedly improves the accuracy of elastic constants derived from simulations in both empirical potential-based NNPs and relabeled DFT-based NNPs. Overall, this study offers significant insights into leveraging empirical potentials to expedite the development of reliable and robust NNPs at the DFT level.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Machine Learning–Augmented Laser-Induced Breakdown Spectroscopy for Spectral Discrimination of Iron Oxalates

Enhanced characterization and phase identification of post-PUREX Pu Oxalates (PuOXA) are pivotal for nonproliferation and pre-detonation nuclear forensics. Despite significant advances in the characterization of PuO 2 samples, little is known about the impact of both the chemical structure and oxidation states of PuOXA (i.e., Pu(III) and Pu(IV)) have on optical emission signatures. Here, we demonstrate the analytical capabilities of laser-induced breakdown spectroscopy (LIBS) applied to Fe(II) and Fe(III) oxalate samples as surrogates for PuOXA, highlighting the discriminating features in the LIBS emission spectra arising from differences in the oxidation states within mixed FeOXA samples. We report the enhancement of spectral feature selection using Principal Component Analysis (PCA), which enables the analytical superiority of machine learning algorithms such as Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), Partial Least Squares Regression (PLSR), Support Vector Regression (SVR), and Random Forest Regression (RFR) over conventional univariate techniques for phase discrimination and chemometric analysis. Cluster analysis revealed how both matrix effects and laser ablation influence cluster separability by introducing spectral artifacts that misdirect the maximization of variance. PCA-selected emission lines were used in the regression models, demonstrating that both univariate and multivariate linear regression models (i.e., PLSR and SVR) can achieve acceptable performance, with machine learning models outperforming conventional calibration regressions. Furthermore, the application of non-linearly activated PCA-selected emission lines illustrates how simplifying the data while retaining captured variance enables the use of less complex and more computationally efficient models. Furthermore, this is particularly evident in the underperformance of RFR, which suffers from increased computational costs and overfitting owing to its high complexity.

Oxalates↗
DOE OSTI2024

Constraining primordial non-Gaussianity from DESI quasar targets and Planck CMB lensing

We detect the cross-correlation between 2.7 million DESI quasar targets across 14,700 deg 2 (180 quasars deg -2 ) and Planck 2018 CMB lensing at ~30σ. We use the cross-correlation on very large scales to constrain local primordial non-Gaussianity via the scale dependence of quasar bias. The DESI quasar targets lie at an effective redshift of 1.51 and are separated into four imaging regions of varying depth and image quality. We select quasar targets from Legacy Survey DR9 imaging, apply additional flux and photometric redshift cuts to improve the purity and reduce the fraction of unclassified redshifts, and use early DESI spectroscopy of 194,000 quasar targets to determine their redshift distribution and stellar contamination fraction (2.6%). Due to significant excess large-scale power in the quasar autocorrelation, we apply weights to mitigate contamination from imaging systematics such as depth, extinction, and stellar density. We use realistic contaminated mocks to determine the greatest number of systematic modes that we can fit, before we are biased by overfitting and spuriously remove real power. We find that linear regression with one to seven imaging templates removed per region accurately recovers the input cross-power, f NL and linear bias. As in previous analyses, our fNL constraint depends on the linear primordial non-Gaussianity bias parameter, b Φ = 2(b - p)δ c assuming universality of the halo mass function. We measure f NL = -26 +45 -40 with p = 1.6 (f NL = -18 +29 -27 with p = 1.0), and find that this result is robust under several systematics tests. Future spectroscopic quasar cross-correlations with Planck lensing can tighten the f NL constraints by a factor of 2 if they can remove the excess power on large scales in the quasar auto power spectrum.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Trigonometric Parallaxes of Two T Dwarfs With Keck and ShaneAO Astrometry

We present trigonometric parallax and proper motion measurements for two T-type brown dwarfs. We derive our measurements from infrared laser guide star adaptive optics observations spanning five years from the ShaneAO/SHARCS and NIRC2/medium-cam instruments on the Shane and Keck telescopes, respectively. To improve our astrometric precision, we measure and apply a distortion correction to our fields for both instruments. We also transform the Keck and ShaneAO astrometric reference frames onto the ICRS using five-parameter parallax and proper motion solutions for background reference stars from Gaia DR2. Fitting for parallax and proper motion, we measure parallaxes of 73.5 ± 9.2 mas and 70.1 ± 6.7 mas for WISEJ19010703+47181688 (WISE1901) and WISEJ21543294+59421370 (WISE2154), respectively. We utilize Monte Carlo methods to estimate the error in our sparse field methods, taking into account overfitting and differential atmospheric refraction. Comparing to previous measurements in the literature, all of our parallax and proper motion values fall within 2σ of the published measurements, and 4 of 6 measurements are within 1σ. These data are among the first parallax measurements of these T dwarfs and serve as precise measurements for calibrating stellar formation models. These two objects are the first results of an ongoing survey of T dwarfs with Keck/NIRC2 and the Shane Adaptive Optics system at Lick Observatory.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Recipes for when physics fails: recovering robust learning of physics informed neural networks

Abstract Physics-informed neural networks (PINNs) have been shown to be effective in solving partial differential equations by capturing the physics induced constraints as a part of the training loss function. This paper shows that a PINN can be sensitive to errors in training data and overfit itself in dynamically propagating these errors over the domain of the solution of the PDE. It also shows how physical regularizations based on continuity criteria and conservation laws fail to address this issue and rather introduce problems of their own causing the deep network to converge to a physics-obeying local minimum instead of the global minimum. We introduce Gaussian process (GP) based smoothing that recovers the performance of a PINN and promises a robust architecture against noise/errors in measurements. Additionally, we illustrate an inexpensive method of quantifying the evolution of uncertainty based on the variance estimation of GPs on boundary data. Robust PINN performance is also shown to be achievable by choice of sparse sets of inducing points based on sparsely induced GPs. We demonstrate the performance of our proposed methods and compare the results from existing benchmark models in literature for time-dependent Schrödinger and Burgers’ equations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Deep energy-pressure regression for a thermodynamically consistent EOS model

Abstract In this paper, we aim to explore novel machine learning (ML) techniques to facilitate and accelerate the construction of universal equation-Of-State (EOS) models with a high accuracy while ensuring important thermodynamic consistency. When applying ML to fit a universal EOS model, there are two key requirements: (1) a high prediction accuracy to ensure precise estimation of relevant physics properties and (2) physical interpretability to support important physics-related downstream applications. We first identify a set of fundamental challenges from the accuracy perspective, including an extremely wide range of input/output space and highly sparse training data. We demonstrate that while a neural network (NN) model may fit the EOS data well, the black-box nature makes it difficult to provide physically interpretable results, leading to weak accountability of prediction results outside the training range and lack of guarantee to meet important thermodynamic consistency constraints. To this end, we propose a principled deep regression model that can be trained following a meta-learning style to predict the desired quantities with a high accuracy using scarce training data. We further introduce a uniquely designed kernel-based regularizer for accurate uncertainty quantification. An ensemble technique is leveraged to battle model overfitting with improved prediction stability. Auto-differentiation is conducted to verify that necessary thermodynamic consistency conditions are maintained. Our evaluation results show an excellent fit of the EOS table and the predicted values are ready to use for important physics-related tasks.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Comparison of linear regression, k-nearest neighbour and random forest methods in airborne laser-scanning-based prediction of growing stock

Abstract In this study, for five sites around the world, we look at the effects of different model types and variable selection approaches on forest yield modelling performances in an area-based approach (ABA). We compared ordinary least squares regression (OLS), k-nearest neighbours (kNN) and random forest (RF). Our objective was to test if there are systematic differences in accuracy between OLS, kNN and RF in ABA predictions of growing stock volume. The analyses are based on a 5-fold cross-validation at five study sites: an eucalyptus plantation, a temperate forest and three different boreal forests. Two completely independent validation datasets were also available for two of the boreal sites. For the kNN, we evaluated multiple measures of distance including Euclidean, Mahalanobis, most similar neighbour (MSN) and an RF-based distance metric. The variable selection approaches we examined included a heuristic approach (for OLS, kNN and RF), exhaustive search among all combinations (OLS only) and all variables together (RF only). Performances varied by model type and variable selection approaches among sites. OLS and RF had similar accuracies and were more efficient than any of the kNN variants. Variable selection did not affect RF performance. Heuristic and exhaustive variable selection performed similarly for OLS. kNN fared the poorest amongst model types, and kNN with RF distance was prone to overfitting when compared with a validation dataset. Additional caution is therefore required when building kNN models for volume prediction though ABA, being preferable instead to opt for models based on OLS with some variable selection, or RF with all variables together.

Cosenza, Diogo N.↗
DOE OSTI2022

Nonlinear dimension reduction for surrogate modeling using gradient information

We introduce a method for the nonlinear dimension reduction of a high-dimensional function $u:{\mathbb{R}}^d\rightarrow{\mathbb{R}}$, $d\gg 1$. Our objective is to identify a nonlinear feature map $g:{\mathbb{R}}^d\rightarrow{\mathbb{R}}^m$, with a prescribed intermediate dimension $m\ll d$, so that $u$ can be well approximated by $f\circ g$ for some profile function $f:{\mathbb{R}}^m\rightarrow{\mathbb{R}}$. We propose to build the feature map by aligning the Jacobian $\nabla g$ with the gradient $\nabla u$, and we theoretically analyze the properties of the resulting $g$. Once $g$ is built, we construct $f$ by solving a gradient-enhanced least squares problem. Our practical algorithm uses a sample $\{{\textbf{x}}^{(i)},u({\textbf{x}}^{(i)}),\nabla u({\textbf{x}}^{(i)})\}_{i=1}^N$ and builds both $g$ and $f$ on adaptive downward-closed polynomial spaces, using cross validation to avoid overfitting. We numerically evaluate the performance of our algorithm across different benchmarks, and explore the impact of the intermediate dimension $m$. We show that building a nonlinear feature map $g$ can permit more accurate approximation of $u$ than a linear $g$, for the same input data set.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Improving galaxy clustering measurements with deep learning: analysis of the DECaLS DR7 data

ABSTRACT Robust measurements of cosmological parameters from galaxy surveys rely on our understanding of systematic effects that impact the observed galaxy density field. In this paper, we present, validate, and implement the idea of adopting the systematics mitigation method of artificial neural networks for modelling the relationship between the target galaxy density field and various observational realities including but not limited to Galactic extinction, seeing, and stellar density. Our method by construction allows a wide class of models and alleviates overtraining by performing k-fold cross-validation and dimensionality reduction via backward feature elimination. By permuting the choice of the training, validation, and test sets, we construct a selection mask for the entire footprint. We apply our method on the extended Baryon Oscillation Spectroscopic Survey (eBOSS) Emission Line Galaxies (ELGs) selection from the Dark Energy Camera Legacy Survey (DECaLS) Data Release 7 and show that the spurious large-scale contamination due to imaging systematics can be significantly reduced by up-weighting the observed galaxy density using the selection mask from the neural network and that our method is more effective than the conventional linear and quadratic polynomial functions. We perform extensive analyses on simulated mock data sets with and without systematic effects. Our analyses indicate that our methodology is more robust to overfitting compared to the conventional methods. This method can be utilized in the catalogue generation of future spectroscopic galaxy surveys such as eBOSS and Dark Energy Spectroscopic Instrument (DESI) to better mitigate observational systematics.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

A general framework for removing point-spread function additive systematics in cosmological weak lensing analysis

ABSTRACT Cosmological weak lensing measurements rely on a precise measurement of the shear two-point correlation function (2PCF) along with a deep understanding of systematics that affect it. In this work, we demonstrate a general framework for detecting and modelling the impact of PSF systematics on the cosmic shear 2PCF and mitigating its impact on cosmological analysis. Our framework can detect PSF leakage and modelling error from all spin-2 quantities contributed by the PSF second and higher moments, rather than just the second moments, using the cross-correlations between galaxy shapes and PSF moments. We interpret null tests using the HSC Year 3 (Y3) catalogs with this formalism and find that leakage from the spin-2 combination of PSF fourth moments is the leading contributor to additive shear systematics, with total contamination that is an order-of-magnitude higher than that contributed by PSF second moments alone. We conducted a mock cosmic shear analysis for HSC Y3 and find that, if uncorrected, PSF systematics can bias the cosmological parameters Ωm and S8 by ∼0.3σ. The traditional second moment-based model can only correct for a 0.1σ bias, leaving the contamination largely uncorrected. We conclude it is necessary to model both PSF second and fourth moment contaminations for HSC Y3 cosmic shear analysis. We also reanalyse the HSC Y1 cosmic shear analysis with our updated systematics model and identify a 0.07σ bias on Ωm when using the more restricted second moment model from the original analysis. We demonstrate how to self-consistently use the method in both real space and Fourier space, assess shear systematics in tomographic bins, and test for PSF model overfitting.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2022

Importance of kernel bandwidth in quantum machine learning

Quantum kernel methods are considered a promising avenue for applying quantum computers to machine learning problems. Identifying hyperparameters controlling the inductive bias of quantum machine learning models is expected to be crucial given the central role hyperparameters play in determining the performance of classical machine learning methods. In this work we introduce the hyperparameter controlling the bandwidth of a quantum kernel and show that it controls the expressivity of the resulting model. We use extensive numerical experiments with multiple quantum kernels and classical data sets to show consistent change in the model behavior from underfitting (bandwidth too large) to overfitting (bandwidth too small), with optimal generalization in between. We draw a connection between the bandwidth of classical and quantum kernels and show analogous behavior in both cases. Furthermore, we show that optimizing the bandwidth can help mitigate the exponential decay of kernel values with qubit count, which is the cause behind recent observations that the performance of quantum kernel methods decreases with qubit count. Here, we reproduce these negative results and show that if the kernel bandwidth is optimized, the performance instead improves with growing qubit count and becomes competitive with the best classical methods.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Effective bias expansion for 21-cm cosmology in redshift space

A near-future detection of the 21-cm signal from the epoch of reionization will provide unique opportunities to probe the underlying cosmology, provided that such cosmological information can be extracted with precision. To this end, we further develop effective field theory (EFT) inspired techniques for the 21-cm brightness temperature field during the epoch of reionization, incorporating renormalized bias and a treatment of redshift space distortions. Notably, we confirm that in redshift space, measures of the 21-cm brightness, e.g., the power spectrum, should have irreducible contributions that lack a bias coefficient and therefore contain direct, astrophysics-free information about the cosmological density field; in this work, we study this effect beyond linear order. To validate our theoretical treatment, we fit the predicted EFT Fourier-space shapes to the thesan suite of hydrodynamical simulations of reionization at the field level, where the considerable number of modes prevents overfitting. We find agreement at the level of a few percent between the 21-cm power spectrum from the EFT fits and simulations over the wave number range k ≲ 0.8 h/Mpc and neutral fraction x HI ≳ 0.4, which is imminently measurable by the Hydrogen Epoch of Reionization Array and future experiments. Furthermore, the ability of the EFT to describe the 21-cm signal extends to simulations that have different astrophysical prescriptions for reionization as well as simulations with interacting dark matter.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Modeling Intercalation Chemistry with Multiredox Reactions by Sparse Lattice Models in Disordered Rocksalt Cathodes

Modern battery materials can contain many elements with substantial site disorder, and their configurational state has been shown to be critical for their performance. The intercalation voltage profile is a critical parameter to evaluate the performance of energy storage. The application of commonly used cluster expansion techniques to model the intercalation thermodynamics of such systems ab initio is challenged by the combinatorial increase in configurational degrees of freedom as the number of species grows. Such challenges necessitate the efficient generation of lattice models without overfitting and proper sampling of the configurational space under the requirement of charge balance in ionic systems. In this work, we introduce a combined approach that addresses these challenges by (1) constructing a robust cluster expansion Hamiltonian using the sparse regression technique, including -norm regularization and structural hierarchy; and (2) implementing semigrand-canonical Monte Carlo to sample charge-balanced ionic configurations using the table-exchange method and an ensemble average approach. These techniques are applied to a disordered rocksalt oxyfluoride (LMNOF) that is part of a family of promising earth-abundant cathode materials. The simulated voltage profile is found to be in good agreement with experimental data and particularly provides a clear demonstration of the and oxygen contributions to the redox potential as a function of content.

25 ENERGY STORAGE↗
DOE OSTI2024

Counter Data Paucity through Adversarial Invariance Encoding: A Case Study on Modeling Battery Thermal Runaway

Lithium-ion batteries, widely used for their durability and high energy storage, face the risk of internal short circuits leading to catastrophic thermal runaway events. These events, triggered by external stimuli like mechanical loads, pose safety concerns in applications such as electric vehicles. Detecting and understanding thermal runaway events is crucial, but physics-driven models struggle to explain the non-linear evolution of battery temperature during these events, considering factors like material composition and state-of-charge. Due to the rarity of these events and the cost of data collection, we propose a deep learning (DL) model to predict battery temperature responses during thermal runaway. The challenge lies in the scarcity of data, making traditional DL models prone to overfitting and learning low-quality representations of the complex process.Our approach introduces a novel few-shot architecture that incorporates an adversarially governed invariant encoding process. This architecture aims to distill "invariant" relationships by addressing distributional shifts in data across various battery properties, facilitating the detection of thermal runaway events. Specifically, our results demonstrate that deep learning models conditioned on these "invariant" representations outperform state-of-the-art baselines, achieving a remarkable 96.8% performance improvement in terms of the popular metric MAPE. This framework presents a promising direction for enhancing battery safety modeling, particularly in the context of rare and complex events like thermal runaway. Our code and code and dataset used for the paper are public1.

Tabassum, Anika [ORNL] (ORCID:0000000254600955)↗