SEARCH · Search NASA

Results for “Level set methods”

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.

SourceResearch connectionFrom yearThrough year

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

At least 145 records · Page 8

All sources NASA NTRS DOE OSTI

DOE OSTI2025

The latent variable proximal point algorithm for variational problems with inequality constraints

The latent variable proximal point (LVPP) algorithm is a framework for solving infinite-dimensional variational problems with pointwise inequality constraints. The algorithm is a saddle point reformulation of the Bregman proximal point algorithm. At the continuous level, the two formulations are equivalent, but the saddle point formulation is more amenable to discretization because it introduces a structure-preserving transformation between a latent function space and the feasible set. Working in this latent space is much more convenient for enforcing inequality constraints than the feasible set, as discretizations can employ general linear combinations of suitable basis functions, and nonlinear solvers can involve general additive updates. LVPP yields numerical methods with observed mesh-independence for obstacle problems, contact, fracture, plasticity, and others besides; in many cases, for the first time. The framework also extends to more complex constraints, providing means to enforce convexity in the Monge–Ampère equation and handling quasi-variational inequalities, where the underlying constraint depends implicitly on the unknown solution. Here, in this paper, we describe the LVPP algorithm in a general form and apply it to ten problems from across mathematics.

Inequality constraints↗

DOE OSTI2025

Search for a light charged Higgs boson in $t \rightarrow H^{\pm } b$ decays, with $H^{\pm } \rightarrow cs$, in $pp$ collisions at $\sqrt{s}={13}\hbox { TeV}$ with the ATLAS detector

A search for a light charged Higgs boson produced in decays of the top quark, $t \rightarrow H^{\pm } b$ with $H^{\pm } \rightarrow cs$, is presented. This search targets the production of top-quark pairs $t\bar{t} \rightarrow WbH^{\pm } b$, with $W \rightarrow ℓv(ℓ = e, μ)$, resulting in a lepton-plus-jets final state characterised by an isolated electron or muon and at least four jets. The search exploits b-quark and c-quark identification techniques as well as multivariate methods to suppress the dominant $t\bar{t}$ background. The data analysed correspond to 140 fb -1 of $pp$ collisions at $\sqrt{s}$ = 13 TeV recorded with the ATLAS detector at the LHC between 2015 and 2018. Observed (expected) 95% confidence-level upper limits on the branching fraction $\mathscr{B}(t \rightarrow H^{\pm } b)$, assuming $\mathscr{B}(t \rightarrow Wb) + \mathscr{B}(t \rightarrow H^{\pm }(\rightarrow cs)b$, are set between 0.066% (0.077%) and 3.6% (2.3%) for a charged Higgs boson with a mass between 60 and 168 GeV.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2024

Integrate Latimer Controls' Solution into RTAC (CRADA Final Report, CRD-23-24672)

Latimer Controls, Inc. was awarded two vouchers under the Department of Energy's American-Made Solar Prize Round 6 to conduct collaborative research at a national laboratory. The National Renewable Energy Laboratory (NREL) was selected as a partner to assist Latimer Controls in the performance evaluation of its photovoltaic (PV) control software. This collaboration focuses on developing a hardware-in-the-loop (HIL) testbed at NREL, which will be used to test and validate the Latimer PV control technology in a realistic yet de-risked environment. Both Latimer and NREL teams will work together to analyze the collected test data, derive insights, and disseminate the scientific findings. Recent studies underscore the potential of solar energy as a zero-marginal-cost and zero-emission flexibility resource within the bulk power system, particularly when integrated with advanced control systems. To enhance the performance of such systems, Latimer Controls has developed leading-edge technologies, including machine learning (ML) algorithms and hierarchical inverter set-point allocation methods. These innovations are designed to estimate the operational headroom of large PV plants for grid integration and control. However, comprehensive validation under real-world conditions remains necessary. To address this gap, the concurrent CRADA project proposes the real-world application and validation of the Latimer Control solution within a HIL environment. Initially, the Latimer algorithm was developed and tested within MATLAB Simulink, a platform suitable for research-level simulations and iterative development. However, transitioning this technology to a real solar site as an industry-ready solution necessitates implementation in a format compatible with widely used solar power plant controllers. In this additional CRADA work, the MATLAB Simulink-based logic will be translated into Structured Text, a programming language compliant with IEC 61131 standards, which is commonly used for custom logic implementations in industry-leading programmable logic controllers (PLCs), such as the Schweitzer SEL real-time automation controller (RTAC). This transition will facilitate the deployment of the Latimer Control solution in real-world solar power plants, thereby advancing the technology towards commercialization.

14 SOLAR ENERGY↗

DOE OSTI2025

Similarity Metric for Data Optimization and Efficient Training of Reactive Machine Learning Force Fields for Hydrocarbon Radiolysis

Radiolysis is a common approach to sterilize polymers, chemically modify them for upcycling, and accelerate their decomposition for recycling purposes. Reactive molecular dynamics (MD) simulations provide a powerful tool to generate atomic-level trajectories of the reactive processes and quantify radiolytic chemical degradation pathways. For this, machine learning (ML) surrogate models for reactive force fields with quantum mechanical accuracy are now widely used, which require ML training data sets that can provide information on atomic environments for target chemical systems. However, radiolysis chemistry can be highly complex and diverse, which poses significant challenges for generating training data to parametrize ML models. In this regard, we developed a method for optimizing the training data set using a cosine similarity metric to help guide training set selection for radiolysis of polyethylene, a model hydrocarbon polymer, as well as to enhance the transferability of our reactive ML force field (MLFF) to a variety of molecular and polymeric systems. Our approach performs atom-by-atom comparisons between local atomic environments to pinpoint important data points associated with rare and localized events, such as radiolysis damage within structures. We apply this approach to train the Chebyshev Interaction Model for Efficient Simulation (ChIMES) MLFF model, which expresses the atomic interaction potentials in terms of linear combinations of many-body Chebyshev polynomials. We first show that our method can reduce our training set size by ∼70% while improving overall accuracy compared to more standard MD model fitting approaches. We then validate our optimum model against diverse hydrocarbon simulation data, including simple alkanes and systems with unsaturated carbon bonds, over a wide range of thermodynamic conditions. Finally, we use our ChIMES model to perform MD simulations of radiolytic damage with large-scale systems that help avoid system size effects. Overall, our approach yields an MD force field that retains most of the accuracy of the underlying quantum method while yielding many orders of improvement in computational efficiency. In conclusion, our efforts will have impact on future hydrocarbon polymer radiolysis studies, where the chemical details of the polymer–radiation interactions can have a strong effect on the resulting products observed in experiments.

Hydrocarbons↗

DOE OSTI2025

An MLCommons Scientific Benchmarks Ontology

Scientific machine learning research spans diverse domains and data modalities, yet existing benchmark efforts remain siloed and lack standardization. This makes novel and transformative applications of machine learning to critical scientific use-cases more fragmented and less clear in pathways to impact. This paper introduces an ontology for scientific benchmarking developed through a unified, community-driven effort that extends the MLCommons ecosystem to cover physics, chemistry, materials science, biology, climate science, and more. Building on prior initiatives such as XAI-BENCH, FastML Science Benchmarks, PDEBench, and the SciMLBench framework, our effort consolidates a large set of disparate benchmarks and frameworks into a single taxonomy of scientific, application, and system-level benchmarks. New benchmarks can be added through an open submission workflow coordinated by the MLCommons Science Working Group and evaluated against a six-category rating rubric that promotes and identifies high-quality benchmarks, enabling stakeholders to select benchmarks that meet their specific needs. The architecture is extensible, supporting future scientific and AI/ML motifs, and we discuss methods for identifying emerging computing patterns for unique scientific workloads. The MLCommons Science Benchmarks Ontology provides a standardized, scalable foundation for reproducible, cross-domain benchmarking in scientific machine learning. A companion webpage for this work has also been developed as the effort evolves: https://mlcommons-science.github.io/benchmark/

Hawks, Ben [Fermilab] (ORCID:0000000157000288)↗

DOE OSTI2025

On the Sampling-Based Computation of Nash Equilibria Under Uncertainty via the Nikaido–Isoda Function

We consider the computation of an equilibrium of a stochastic Nash equilibrium problem, where the player objectives are assumed to be L 0 -Lipschitz continuous and convex, given rival decisions with convex and closed player-specific feasibility sets. To address this problem, we consider minimizing a suitably defined value function defined using the Nikaido–Isoda function. Such an avenue does not necessitate either monotonicity properties of the concatenated gradient map or potentiality requirements on the game but does require a suitable regularity requirement under which a stationary point is a Nash equilibrium. We design and analyze a sampling-enabled projected-gradient-response method, reliant on inexact resolution of a player-level best-response subproblem. Here, by deriving suitable Lipschitzian guarantees on the value function, we derive both asymptotic guarantees for the sequence of generated iterates as well as rate and complexity guarantees for computing a stationary point by appropriate choices of the sampling rate and inexactness sequence.

Nikaido-Isoda function↗

DOE OSTI2025

Prioritizing Nuclear Materials for SAM-3 Neutron Irradiation Campaign: Structural and Cladding Materials Candidates

This report outlines a framework for selecting structural and cladding materials for the Nuclear Science User Facilities (NSUF) SAM-3 neutron irradiation campaign to support the advancement of nuclear energy technologies. The document begins with an introduction that provides background context, highlights the motivations for launching a new irradiation campaign, and defines the overall objectives. The core of the report describes the design considerations for the irradiation campaign, including capsule configurations, irradiation temperature ranges, and target dose levels (defined by displacements per atom, or dpa). The material recommendation was guided by the Specimen Identification and Prioritization (SIP) Working Group, a multidisciplinary team of experts representing national laboratories, academia, industry, federal government and agency. This group played a central role in identifying candidate materials, evaluating technical justifications, and ensuring alignment with boarder programmatic goals. A detailed set of criteria for material prioritization is then presented, taking into account reactor relevance, performance gaps, advanced manufacturing methods, and emerging material classes. Based on the input of SIP working group, specific materials were selected and justified for inclusion in the irradiation campaign by the NSUF leadership and its U.S. Department of Energy (DOE)-Office of Nuclear Energy (NE) management. The final section provides recommended capsule designs, summarizing critical parameters such as material type, fabrication method, sample geometry, irradiation conditions, and specimen quantities. This report serves as a foundation for executing a focused and high-impact neutron irradiation campaign aimed at addressing key materials challenges for both existing and advanced nuclear reactors.

36 - MATERIALS SCIENCE↗

DOE OSTI2025

Estimating soybean yields from high-temporal-resolution multi-source data using deep learning

Accurate and timely crop yield prediction is crucial for ensuring food security and maintaining stable agricultural markets. In recent years, there has been a surge in interest in leveraging high-temporal-resolution, multi-source data for effective crop growth monitoring and yield estimation. A notable challenge arises from the difficulty in capturing the intricate interactions between variables across different time steps within these high-temporal-resolution time series datasets. This complexity hinders the reliable extraction of yield information from voluminous and often noisy datasets, especially during periods of extreme weather events. Here, in this study, we propose an Attention and Graph Isomorphism Network-enhanced Bi-directional Long Short-Term Memory network (AGB-LSTM) for estimating county-level soybean yield in the United States. This model integrates a diverse set of remote sensing data, including Near-Infrared Reflectance of Vegetation (NIRv), Sun-Induced chlorophyll Fluorescence (SIF), and Gross Primary Productivity (GPP), along with environmental covariates. The AGB-LSTM effectively leverages information related to crop yield from high-temporal-resolution time series data (5-days), achieving an accuracy of R²= 0.67 and rRMSE = 14.46%. This approach significantly outperforms traditional machine learning methods such as Random Forest (RF) (R²= 0.52, rRMSE = 17.36%) and Bi-LSTM (R²= 0.58, rRMSE = 16.17%). Sensitivity experiments with different time steps and ranges demonstrated that our model could accurately and stably predict yields 1 to 2 months before harvest. Moreover, data with a finer temporal resolution consistently improved prediction performance, resulting in an approximately 20% increase in and an approximately 20% decrease in rRMSE compared to using monthly composites. We also evaluated the robustness of the model under extreme climate events and observed strong performance (R²= 0.50, rRMSE = 21.32%). Finally, yield mapping for major soybean-producing regions in North America in 2023 revealed spatial patterns that closely matched USDA yield reports. Our findings suggest that the AGB-LSTM model is a promising and effective method for estimating yield and has notable potential for global crop yield forecasting.

Deep learning↗

DOE OSTI2025

Integrated Metrics for County-Level Resilience Ranking Using Entropy and TOPSIS

In the face of atypical weather events, power infrastructure failures, and limited resources for resilience investment, energy decision-makers need data-driven metrics to allocate resilience investments and maximize the reduction of power outage impacts. For state-level planning, for instance, ranking the resilience of each county is key to ensuring effective distribution of resources. In such cases, resilience for each spatial unit is multifaceted and is captured by a set of indicators (i.e., metrics) that can be combined into an overall score that reduces the complexity of power outage dynamics to a single decision metric. However, weighting of these indicators is often addressed by simplifying assumptions (i.e., equal weights) or semi-subjective methods that rely on user-defined weights that can introduce biases (e.g., weighted average score). Within the disaster risk reduction and resilience engineering community, a recurring challenge in multicriteria decision-making is the objective weighting of indicators for composite indices. To address this issue, we have leveraged a Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) combined with an entropy-based weighting approach to calculated the integrated scores. This method objectively determines the importance of each metric, better discerns between spatial units (i.e., counties), and offers a more reliable ranking of counties according to their relative resilience attributes. By improving methods for integrating resilience indicators, our approach helps planners and decision-makers prioritize resources more effectively for more efficient resilience investments.

Bhusal, Narayan [Oak Ridge National Laboratory (OR↗

DOE OSTI2025

Cheaper and more noise-resilient quantum state preparation using eigenvector continuation

Subspace methods are powerful, noise-resilient methods that can effectively prepare ground states on quantum computers. The challenge is to get a subspace with a small condition number that spans the states of interest using minimal quantum resources. In this work, we will use eigenvector continuation to build a subspace from the low-lying states of a set of Hamiltonians. The basis vectors are prepared using truncated versions of standard state preparation methods such as imaginary-time evolution (ITE), adiabatic state preparation (ASP), and variational quantum eigensolver. By using these truncated methods combined with eigenvector continuation, we can directly improve upon them, obtaining more accurate ground-state energies at a reduced cost. We use several spin systems to demonstrate convergence even when methods like ITE and ASP fail, such as ASP in the presence of level crossings and ITE with vanishing energy gaps. We also showcase the noise resilience of this approach beyond the gains already made by having shallower quantum circuits. Furthermore, our findings suggest that eigenvector continuation can be used to improve existing state preparation methods in the near term.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2024

How Accurate Are Simulations and Experiments for the Lattice Energies of Molecular Crystals?

Molecular crystals play a central role in a wide range of scientific fields, including pharmaceuticals and organic semiconductor devices. However, they are challenging systems to model accurately with computational approaches because of a delicate interplay of intermolecular interactions such as hydrogen bonding and Van der Waals dispersion forces. Here, by exploiting recent algorithmic developments, we report the first set of diffusion Monte Carlo lattice energies for all 23 molecular crystals in the popular and widely used X23 dataset. Comparisons with previous state-of-the-art lattice energy predictions (on a subset of the dataset) and a careful analysis of experimental sublimation enthalpies reveals that high-accuracy computational methods are now at least as reliable as (computationally derived) experiments for the lattice energies of molecular crystals. Overall, this work demonstrates the feasibility of high-level explicitly correlated electronic structure methods for broad benchmarking studies in complex condensed phase systems, and signposts a route towards closer agreement between experiment and simulation. Published by the American Physical Society 2024

Physics↗

DOE OSTI2026

Identifying Differential Equations in Fourier Domain (FourierIdent)

We investigate identifying differential equations in the frequency domain. Fourier analysis is an important tool in theoretical analysis and numerical solvers of differential equations, yet there is limited work in exploring this connection in the identification of differential equations. This paper aims to identify the underlying differential equation in the frequency domain, from a given single realization of the differential equation perturbed by noise. Such setting imposes difficulties which are different from other identification methods where computation is carried out in the physical domain. We propose several ways to mitigate the challenges arising from noise in data and large differences in the magnitudes of frequency responses. The main takeaways are that identifying differential equations solely in the frequency domain is challenging, the method we propose is based on a form of domain partitions in the frequency domain, and this method shows benefits for complex data even with high level of noise. We introduce a Fourier feature denoising, and define the meaningful data region and the core regions of features to reduce the effect of noise in the frequency domain and to enhance the accuracy in coefficient identification. The proposed method is tested on various differential equations with linear, nonlinear, and high-order derivative feature terms, and shows advantages on complex data with many frequency modes, even under high level of noise.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2026

EC-Bench: A Benchmark for Enzyme Commission Number Prediction

Enzymes are proteins that catalyze specific biochemical reactions in cells. Enzyme Commission (EC) numbers are used to annotate enzymes in a four-level hierarchy that classifies enzymes based on the specific chemical reactions they catalyze. Accurate EC number prediction is essential for understanding enzyme functions. Despite the availability of numerous methods for predicting EC numbers from protein sequences, there is no unified framework for evaluating and studying such methods systematically. This gap limits the ability of the community to identify the most effective approaches for enzyme annotation. We introduce EC-Bench, a benchmark for EC number prediction, consisting of 1) an initial representative set of existing methods (including homology-based, deep learning, contrastive learning, and language model methods), 2) existing and novel accuracy and efficiency performance metrics, and 3) selected datasets to allow for comprehensive comparative study. EC-Bench is open-source and provides a framework for researchers to not only compare among existing methods objectively under uniform conditions, but also to introduce and effectively evaluate performance of new methods in a comparative framework. To demonstrate the utility of EC-Bench, we perform extensive experimentation to compare the existing EC number prediction methods and establish their advantages and disadvantages in a variety of prediction tasks, namely “exact EC number prediction”, “EC number completion” and (partial or additional) “EC number recommendation”. We find wide variation in the performance of different methods, but also subtle but potentially useful differences in the performance of different methods across tasks and for different parts of the EC hierarchy.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2025

Exploring Data Set Bias and Decision Support with Predictive Uncertainty Through Bayesian Approximations and Convolutional Neural Networks

Individual seismic catalogs can contain multiscale observations from fault level to global scales and associated waveforms from discrete events reflect crustal structure across many different scales and locations. Seismic network aperture, geographic location, and observation distance may not provide informative guidance or intuition on how different catalogs will behave across models trained under different conditions. We rely on uncertainty to provide guardrails for when to trust model decisions, but understanding when our uncertainty is trustworthy is an open challenge. Here, in this work, we explore Bayesian approximation methods for assigning predictive uncertainty in seismic event classification problems. We find that computationally expensive Bayesian approximations do not outperform simple ensemble methods. We also find that when exploiting multiple seismic event catalogs, joint training with data from all the catalogs combined with Bayesian approximations and supervised training for classification can obscure bias and result in less robust uncertainty while also not providing substantial performance benefits compared to training individual models for each catalog.

58 GEOSCIENCES↗

DOE OSTI2024

Path integrals, complex probabilities and the discrete Weyl representation

Abstract A discrete formulation of the real-time path integral as the expectation value of a functional of paths with respect to a complex probability on a sample space of discrete valued paths is explored. The formulation in terms of complex probabilities is motivated by a recent reinterpretation of the real-time path integral as the expectation value of a potential functional with respect to a complex probability distribution on cylinder sets of paths. The discrete formulation in this work is based on a discrete version of the Weyl algebra that can be applied to any observable with a finite number of outcomes. The origin of the complex probability in this work is the completeness relation. In the discrete formulation the complex probability exactly factors into products of conditional probabilities and exact unitarity is maintained at each level of approximation. The approximation of infinite dimensional quantum systems by discrete systems is discussed. The method is illustrated by applying it to scattering theory and quantum field theory. The implications of these applications for quantum computing is discussed.

Physics↗

DOE OSTI2025

Fermilab Booster loss modelling and rebalancing using Bayesian methods

Fermilab Booster is being upgraded for the PIP-II project to support 20Hz ramp rate at higher intensities. Loss trip limits determine the achievable peak power. To meet PIP-II requirements, losses need to be halved as compared to current levels. Losses primarily occur at injection and transition crossing, with both gradually increasing and threshold-like intensity-dependent behaviors. The existing simulation models are not yet good enough for quantitative loss predictions. In practice, it will be necessary to tune up the Booster using iterative methods and operator intuition. In this paper we present an effort to systematically model Booster losses using active learning (Bayesian exploration) techniques, and subsequently to rebalance them for higher trip limit margins. We first created several sets of spatially and temporally isolated orbit and optics knobs, and trained Gaussian process models for each beam loss monitor as well as beam current. This is a complex task due to safety and timing requirements – we discuss mitigations such as uncertainty constraints and approximate fitting. Once models are stable, we perform large-scale single and multi-objective tuning using scalarized objectives made up of critical beam loss locations. Our results demonstrate significant rebalancing of losses, increasing trip margins, as well as an overall improvement in beam transmission efficiency. We are exploring how to combine existing simulations with experimental data and automate the collection procedure so that more advanced surrogate models can be created over time.

Kuklev, Nikita [Fermilab]↗

DOE OSTI2024

Circulating levels of micronutrients and risk of osteomyelitis: a Mendelian randomization study

Background Few observational studies have investigated the effect of micronutrients on osteomyelitis, and these findings are limited by confounding and conflicting results. Therefore, we conducted Mendelian randomization (MR) analyses to evaluate the association between blood levels of eight micronutrients (copper, selenium, zinc, vitamin B12, vitamin C, and vitamin D, vitamin B6, vitamin E) and the risk of osteomyelitis. Methods We performed the two-sample and multivariable Mendelian randomization (MVMR) to investigate causation, where instrument variables for the predictor (micronutrients) were derived from the summary data of micronutrients from independent cohorts of European ancestry. The outcome instrumental variables were used from the summary data of European-ancestry individuals ( n = 486,484). The threshold of statistical significance was set at p < 0.00625. Results We found a significant causal association that elevated zinc heightens the risk of developing osteomyelitis in European ancestry individuals OR = 1.23 [95% confidence interval (CI) [1.07, 1.43]; p = 4.26E-03]. Similarly, vitamin B6 showed a similar significant causal effect on osteomyelitis as a risk factor OR = 2.78 (95% CI [1.34, 5.76]; p = 6.04E-03; in the secondary analysis). Post-hoc analysis suggested this result (vitamin B6). However, the multivariable Mendelian randomization (MVMR) provides evidence against the causal association between zinc and osteomyelitis OR = 0.98(95% CI [−0.11, 0.07]; p = 7.20E-1). After searching in PhenoScanner, no SNP with confounding factors was found in the analysis of vitamin B6. There was no evidence of a reverse causal impact of osteomyelitis on zinc and vitamin B6. Conclusion This study supported a strong causal association between vitamin B6 and osteomyelitis while reporting a dubious causal association between zinc and osteomyelitis.

Zhang, Xu↗

DOE OSTI2026

Leveraging large language models to address data scarcity in machine learning for graphene synthesis

Machine learning in experimental materials science faces significant challenges due to the scarcity of data, which are costly and time-consuming to generate, particularly when relying on in-house experiments. Literature data mining offers a potential solution but introduces issues like mixed data quality, inconsistent formats, and non-uniform reporting of synthesis parameters, resulting in partially missing and heterogeneous features across the dataset. Here, we propose data imputation and feature engineering methods that employ pre-trained large language models (LLMs) to enhance machine learning performance on scarce, heterogeneous datasets, demonstrated on graphene CVD synthesis data and the ML-HydPARK hydrogen storage dataset. GPT models perform data imputation via tailored prompting and semantic normalization of inconsistently reported features through embeddings, for example, to harmonize the complex nomenclature of CVD substrates. Beyond yielding more diverse and richer feature representations than traditional methods such as K-nearest neighbors (KNN) and Multivariate Imputation by Chained Equations (MICE), LLM-based data imputation is evaluated against dataset characteristics and prompting strategies. We vary the level of autonomy granted to the LLM, from generic prompting that leverages pre-trained knowledge for autonomous data generation to data-informed prompting that constrains outputs using target-specific information, and demonstrate which level of autonomy yields superior imputation performance across datasets and feature types. The proposed data engineering methods markedly improve downstream performance; for example, in graphene layer number classification using a support vector machine (SVM), binary accuracy increases from 39% to 65% and ternary accuracy from 52% to 72%. Fine-tuning experiments on both datasets show that combining our proposed LLM-based data imputation and feature encoding methods with numerical machine learning predictors outperforms standalone fine-tuned LLM predictors in data-scarce settings. The proposed strategies emphasize data enhancement techniques rather than refining learning architectures or regularizing loss functions, offering a broadly applicable framework for improving machine learning performance on scarce, inhomogeneous datasets.

Chemical vapor deposition↗