Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “machine learning potentials”

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 1,027 records · Page 57

Autonomous Molecular Structure Imaging with High-Resolution Atomic Force Microscopy for Molecular Mixture Discovery

Due to its single-molecule sensitivity, high-resolution atomic force microscopy (HR-AFM) has proved to be a valuable and uniquely advantageous tool to study complex molecular mixtures, which hold promise for developing clean energy and achieving environmental sustainability. However, significant challenges remain to achieve the full potential of the sophisticated and time-consuming experiments. Automation combined with machine learning (ML) and artificial intelligence (AI) is key to overcoming these challenges. Here we present Auto-HR-AFM, an AI tool to automatically collect HR-AFM images of petroleum-based mixtures. In this study, we trained an instance segmentation model to teach Auto-HR-AFM how to recognize features in HR-AFM images. Auto-HR-AFM then uses that information to optimize the imaging by adjusting the probe-molecule distance for each molecule in the run. Auto-HR-AFM is the initial tool that will lead to fully automated scanning probe microscopy (SPM) experiments, from start to finish. This automation will allow SPM to become a mainstream characterization technique for complex mixtures, an otherwise unattainable target.

36 MATERIALS SCIENCE↗

The Value of Forecasters‐in‐the‐Loop in Real‐Time Flood Forecasting in the Age of Machine Learning

Machine learning (ML) applications in hydrological forecasting are increasingly prevalent and show great potential. However, many previous studies have only evaluated performance through reanalysis or retrospective simulations compared to simplified baselines. This study provides the first assessment of ML performance against actual operational forecasting systems operated by the California Nevada River Forecast Center (CNRFC), which combines the Community Hydrologic Prediction System (CHPS) with forecasters-in-the-loop. Results demonstrate that forecasters-in-the-loop systems consistently outperform ML models in both general forecasts and flood alerting across lead times up to 96 hr, even when ML models use observed forcings, while CNRFC operational process relies on biased weather forecasts. Our analysis reveals that forecaster expertise maintains forecast reliability despite inaccurate precipitation inputs, with human-guided systems showing superior performance degradation characteristics at extended lead times. These findings highlight the irreplaceable value of human expertise in operational forecasting and caution against overstating current ML capabilities in real-world applications.

Tran, Vinh Ngoc [Univ. of Michigan, Ann Arbor, MI ↗

Exciting DeePMD: Learning excited-state energies, forces, and non-adiabatic couplings

We extend the DeePMD neural network architecture to predict electronic structure properties necessary to perform non-adiabatic dynamics simulations. While learning the excited state energies and forces follows a straightforward extension of the DeePMD approach for ground-state energies and forces, how to learn the map between the non-adiabatic coupling vectors (NACV) and the local chemical environment descriptors of DeePMD is less trivial. Most implementations of machine-learning-based non-adiabatic dynamics inherently approximate the NACVs, with an underlying assumption that the energy-difference-scaled NACVs are conservative fields. We overcome this approximation, implementing the method recently introduced by Richardson [J. Chem. Phys. 158, 011102 (2023)], which learns the symmetric dyad of the energy-difference-scaled NACV. Furthermore, the efficiency and accuracy of our neural network architecture are demonstrated through the example of the methaniminium cation CH 2 NH 2 + .

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Deceptive Infusion of Data: A Novel Data Masking Paradigm for High-Valued Systems

This work addresses how analysts of a high-valued system (e.g., nuclear reactor, aircraft turbine designs) can extract findable, accessible, interoperable, and reusable scientific data for public dissemination to artificial intelligence and machine-learning (AI/ML) researchers in a manner that cannot be reverse-engineered, potentially compromising sensitive or proprietary information. State-of-the-art methods address this problem through data masking techniques, which allow access to a subset of the information while obfuscating private and potentially identifying information (e.g., personally identifying medical data). These methods are unsuitable for industrial engineering processes, where AI/ML tools need explicit access to all the data available to draw the best inference about the system to help optimize its performance and identify its vulnerabilities, etc. Our novel deceptive infusion of data paradigm provides a solution to this conundrum by developing a mathematical approach capable of concealing the identity of the system while providing full access to all the features employed by AI/ML tools to ensure their optimal performance.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Accelerate microstructure evolution simulation using graph neural networks with adaptive spatiotemporal resolution

Abstract Surrogate models driven by sizeable datasets and scientific machine-learning methods have emerged as an attractive microstructure simulation tool with the potential to deliver predictive microstructure evolution dynamics with huge savings in computational costs. Taking 2D and 3D grain growth simulations as an example, we present a completely overhauled computational framework based on graph neural networks with not only excellent agreement to both the ground truth phase-field methods and theoretical predictions, but enhanced accuracy and efficiency compared to previous works based on convolutional neural networks. These improvements can be attributed to the graph representation, both improved predictive power and a more flexible data structure amenable to adaptive mesh refinement. As the simulated microstructures coarsen, our method can adaptively adopt remeshed grids and larger timesteps to achieve further speedup. The data-to-model pipeline with training procedures together with the source codes are provided.

36 MATERIALS SCIENCE↗

Shortest Path and Neighborhood Subgraph Extraction on a Spiking Memristive Neuromorphic Implementation

Spiking neuromorphic computers (SNCs) are promising as a post Moore's law technology partly because of their potential for very low power computation. SNCs have primarily been demonstrated on machine learning and neural network applications, but they can also be used for applications beyond machine learning that can leverage SNC properties such as massively parallel computation and collocated processing and memory. Here, we demonstrate two graph problems (shortest path and neighborhood subgraph extraction) that can be solved using SNCs. We discuss the approach for mapping these applications to an SNC. We also estimate the performance of a memristive SNC for these applications on three real-world graphs.

Schuman, Catherine↗

The Arctic LDRD project: What happened?

This presentation overviews results from the 2019 LDRD Reserve project “Diagnosing near-future changes in Arctic sea ice and ocean conditions” and a follow-up 2020 CSES Rapid Response project “Multi- Model Risk Assessment for Arctic Navigability.” We analyzed Arctic simulations from 4 CMIP6 model ensembles (E3SM, CESM-CAM, CESM-WACCM and CESM-DPLE), comparing them with observations. We then applied a machine-learning approach using copulas to produce joint probabilities of adverse conditions along potential shipping transit routes through the Arctic. Here we compare and contrast the different models and examine the utility of the copula approach applied to these data sets.

58 GEOSCIENCES↗

Neutron-antineutron oscillation search with MicroBooNE and DUNE

The Deep Underground Neutrino Experiment (DUNE) is an international project aiming at neutrino physics and astrophysics and a search for phenomena predicted by theories beyond the standard model (BSM). The excellent imaging capability of Liquid Argon Time Projection Chamber (LArTPC) technology, particle tracking and identification utilized in the Far Detector, as well as the Far Detector size and underground placement, allow the experiment to achieve high sensitivity to various rare processes. BSM theories predict the existence of baryon number non-conservation effects, in particular when the baryon number changes by 2. Here we discuss the sensitivity of DUNE to neutron-antineutron oscillation. With full event simulation and reconstruction using the LArSoft package, we have investigated the background to potential signal events from atmospheric neutrino interactions and particle misidentification, and utilized machine learning techniques to enhance the discrimination between signal and background. The methodologies being developed for a high-sensitivity search for neutron-antineutron oscillation in DUNE can also be demonstrated with the currently running MicroBooNE LArTPC. We discuss progress on demonstrating the developed techniques with the first-ever search for neutron-antineutron oscillation in a LArTPC using MicroBooNE data.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

A machine learning approach to predict metabolic pathway dynamics from time-series multiomics data

New synthetic biology capabilities hold the promise of dramatically improving our ability to engineer biological systems. However, a fundamental hurdle in realizing this potential is our inability to accurately predict biological behavior after modifying the corresponding genotype. Kinetic models have traditionally been used to predict pathway dynamics in bioengineered systems, but they take significant time to develop, and rely heavily on domain expertise. Here, we show that the combination of machine learning and abundant multiomics data (proteomics and metabolomics) can be used to effectively predict pathway dynamics in an automated fashion. The new method outperforms a classical kinetic model, and produces qualitative and quantitative predictions that can be used to productively guide bioengineering efforts. This method systematically leverages arbitrary amounts of new data to improve predictions, and does not assume any particular interactions, but rather implicitly chooses the most predictive ones.

59 BASIC BIOLOGICAL SCIENCES↗

Considerations regarding the Use of Computer Vision Machine Learning in Safety-Related or Risk-Significant Applications in Nuclear Power Plants

With the advancements made to date in the field of artificial intelligence (AI), significant potential exists to utilize AI capabilities for nuclear power plant (NPP) applications. AI can replicate human decision making and it is usually faster and more accurate than humans. For implementations that impact critical NPP applications (e.g., safety-related or non-safety systems that potentially affect overall plant risk), a deeper safety analysis of the AI methods is necessary. AI applied to NPP operations could resemble the use of digital I&C (DI&C) because such applications involve digital computer hardware and custom-designed software that input plant data, execute complex software algorithms, and output the results to a system or licensed human operator to potentially provoke an action. For AI methods to be compliant with current safety requirements for DI&C, AI compatibility must be evaluated, and AI-related gaps may exist that prevent the prompt deployment of AI in NPPs. This effort aims to evaluate how example AI technologies align with the DI&C safety framework, and discusses how they could be analyzed, modeled, tested, and validated in a manner similar to typical DI&C technologies. Because AI is a broad field that encompasses areas such as machine learning (ML), natural language processing, and computer vision, this research focused on a subset of methods categorized as the computer vision ML (CVML) methods. This report explores two CVML use cases, gauge reading and fire watch, considered relevant to the DI&C standards, as they could play a safety-critical role. For the gauge reading use case, a CVML-enabled technology that can read gauges at oblique angles is utilized. For the fire watch use case, a CVML-enabled technology is utilized that migrates fire watch from a manual (human) approach to automated fire detection. These use cases are mainly intended to give context to the CVML system discussion. This effort assumes the worst-case scenario, with the CVML system being used to replace a safety-related or risk-significant system, thus requiring evaluation. Evaluating CVML against most of the relevant safety requirements for DI&C yielded several CVML-specific considerations due to the uniqueness of its characteristics in comparison with typical DI&C systems. For example, CVML models often employ commonly used (open-source) datasets, and it is not always possible to determine the level of overlap among open-source datasets. Therefore, the independence of the developed CVML models when demonstrating diversity is questionable, therefore creating vulnerability to common cause failure (CCF). The design verification process is also impacted since the data overlap could result in overestimation of the software validation and verification (V&V) performance results. Section 2 of this report evaluates a list of the identified CVML-specific characteristics and discusses the resulting considerations and potential solutions in the context of each referenced requirement. A summation is provided in Section 3. This report is not to be used as a guideline. It was developed to identify and consider issues in the implementation of ML technologies used to augment activities that may have a bearing on plant operation. The report draws parallels to the use of DI&C technologies, for which many standards are available to guide their use in nuclear plant operation. It considers the technologies and some of the potential implications of their use in safety-related applications but is not intended to address regulatory or licensing related issues.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

MO-F-CAMPUS-J-02: Automatic Recognition of Patient Treatment Site in Portal Images Using Machine Learning

Purpose: To investigate the method to automatically recognize the treatment site in the X-Ray portal images. It could be useful to detect potential treatment errors, and to provide guidance to sequential tasks, e.g. automatically verify the patient daily setup. Methods: The portal images were exported from MOSAIQ as DICOM files, and were 1) processed with a threshold based intensity transformation algorithm to enhance contrast, and 2) where then down-sampled (from 1024×768 to 128×96) by using bi-cubic interpolation algorithm. An appearance-based vector space model (VSM) was used to rearrange the images into vectors. A principal component analysis (PCA) method was used to reduce the vector dimensions. A multi-class support vector machine (SVM), with radial basis function kernel, was used to build the treatment site recognition models. These models were then used to recognize the treatment sites in the portal image. Portal images of 120 patients were included in the study. The images were selected to cover six treatment sites: brain, head and neck, breast, lung, abdomen and pelvis. Each site had images of the twenty patients. Cross-validation experiments were performed to evaluate the performance. Results: MATLAB image processing Toolbox and scikit-learn (a machine learning library in python) were used to implement the proposed method. The average accuracies using the AP and RT images separately were 95% and 94% respectively. The average accuracy using AP and RT images together was 98%. Computation time was ∼0.16 seconds per patient with AP or RT image, ∼0.33 seconds per patient with both of AP and RT images. Conclusion: The proposed method of treatment site recognition is efficient and accurate. It is not sensitive to the differences of image intensity, size and positions of patients in the portal images. It could be useful for the patient safety assurance. The work was partially supported by a research grant from Varian Medical System.

60 APPLIED LIFE SCIENCES↗

Machine Learning Applied to a Variable Charge Atomistic Model for Cu/Hf Binary Alloy Oxide Heterostructures [Machine Learning a Variable Charge Atomistic Model for Cu/Hf Binary Alloy Oxide Heterostructures]

Alloy oxide heterostructures are essential for a wide variety of energy applications, including anticorrosion coatings, bifunctional catalysis, energy storage, as well as emerging platforms for multilevel nonvolatile memory and neuromorphic devices. The key role played by oxide composition, density, and stoichiometry in governing electrochemical reactions, interface structure, and ionic transport in alloy oxides remains largely unknown. This is primarily due to the lack of variable charge interatomic potential models that can adequately describe the various atomic/ionic interactions in alloy oxides within a unified framework. Here, we introduce a charge transfer ionic potential (CTIP) model for Cu/Hf/O alloy system and demonstrate its ability to accurately capture the complex potential energy surface owing to dynamically varying interactions, including metallic (Cu/Hf), ionic (Cu/Hf/Ox), mixed environment (interfaces). We leverage supervised machine learning methods powered by genetic algorithms coupled with local simplex optimization to efficiently navigate through a high-dimensional parameter space and identify an optimum set of an independent set of CTIP parameters. We train our model against an extensive first-principles based data set that includes lattice constants, cohesive energies, equations of state, and elastic constants of various experimentally observed polymorphs of hafnia, copper oxide, hafnium, as well as copper. Our machine-learned CTIP model captures the structure, elastic properties, thermodynamics, and energetic ordering of various polymorphs. It also accurately predicts the surface properties of both oxides and their metals. To demonstrate the suitability of our CTIP model for investigating dynamic processes, we employed it to identify the atomistic scale mechanisms associated with the initial nanoscale oxidation and surface oxide growth on Cu-doped hafnia surfaces. Here, our machine-learned CTIP model can be used to probe the dynamic response of Cu/Hf/O alloy interfaces subjected to external stimuli (e.g., electric field, pressure, temperature, strain, etc.) and a variety of atomistic phenomena including the dynamics of switching in emerging neuromorphic platforms.

36 MATERIALS SCIENCE↗

Machine Learning Analysis of Impact of Western US Fires on Central US Hailstorms

Fires, including wildfires, harm air quality and essential public services like transportation, communication, and utilities. These fires can also influence atmospheric conditions, including temperature and aerosols, potentially affecting severe convective storms. Here, we investigate the remote impacts of fires in the western United States (WUS) on the occurrence of large hail (size: $\geqslant$ 2.54 cm) in the central US (CUS) over the 20-year period of 2001–20 using the machine learning (ML), Random Forest (RF), and Extreme Gradient Boosting (XGB) methods. The developed RF and XGB models demonstrate high accuracy (> 90%) and F1 scores of up to 0.78 in predicting large hail occurrences when WUS fires and CUS hailstorms coincide, particularly in four states (Wyoming, South Dakota, Nebraska, and Kansas). The key contributing variables identified from both ML models include the meteorological variables in the fire region (temperature and moisture), the westerly wind over the plume transport path, and the fire features (i.e., the maximum fire power and burned area). Importantly, the results confirm a linkage between WUS fires and severe weather in the CUS, corroborating the findings of our previous modeling study conducted on case simulations with a detailed physics model.

54 ENVIRONMENTAL SCIENCES↗

Failure Analysis–Informed Risk Assessment Framework for Geological Carbon Storage Using Numerical Simulation and Machine Learning

Geological carbon storage (GCS) is recognized as a critical technology for achieving large-scale reductions in anthropogenic carbon dioxide (CO 2 ) emissions. Ensuring long-term containment and safety requires robust risk assessment frameworks that account for geological uncertainty and identify potential failure scenarios. Among various indicators, the area of review (AoR) serves as a key metric for evaluating storage performance, regulatory compliance, and monitoring design, as it delineates the spatial extent impacted by pressure buildup and plume migration. However, conventional AoR-based risk assessments typically perturb parameters within narrow uncertainty bounds, potentially overlooking rare but high-impact events arising from extreme geological conditions. In this study, we present a failure analysis–informed risk assessment framework for large-scale GCS projects to improve site prescreening and monitoring design. A suite of 300 numerical simulations was generated using stochastic geological models that vary five key parameters: net-to-gross ratio, anisotropy azimuth, porosity multiplier, permeability multiplier, and vertical-to-horizontal permeability ratio. Among these, 200 realizations represent normal geological uncertainty, while 100 additional cases explore extreme yet plausible conditions for failure-case analysis. The AoR was simulated and computed from pressure and CO 2 saturation fields, where the baseline AoR boundary, representing the extent predicted under typical geological uncertainty, was defined as the union of 200 normal-range simulations, and failure was identified when extreme-range cases exceeded this baseline. Results show that incorporating broader parameter uncertainty produces significantly larger AoR extents, underscoring the potential underestimation of risk under conventional uncertainty ranges. Furthermore, spatial probability maps derived from failure-induced AoR exceedance identify regions requiring enhanced monitoring attention. Various machine learning (ML)–based classifiers were developed to predict failure occurrence from geological parameters, with the random forest model achieving the highest performance (F1-score of 0.986). Consistent findings from correlation coefficient, feature importance, and Sobol sensitivity analyses reveal that low net-to-gross ratios and permeability multipliers are the dominant risk drivers, reflecting reduced reservoir connectivity and limited pressure dissipation. Altogether, these results provide a novel framework for risk-informed site prescreening and monitoring design that explicitly considers rare but high-impact geological scenarios in GCS projects.

25 ENERGY STORAGE↗

Predicting Intensive Care Unit Length of Stay and Mortality Using Patient Vital Signs: Machine Learning Model Development and Validation

Background: Patient monitoring is vital in all stages of care. In particular, intensive care unit (ICU) patient monitoring has the potential to reduce complications and morbidity, and to increase the quality of care by enabling hospitals to deliver higher-quality, cost-effective patient care, and improve the quality of medical services in the ICU. Objective: We here report the development and validation of ICU length of stay and mortality prediction models. The models will be used in an intelligent ICU patient monitoring module of an Intelligent Remote Patient Monitoring (IRPM) framework that monitors the health status of patients, and generates timely alerts, maneuver guidance, or reports when adverse medical conditions are predicted. Methods: We utilized the publicly available Medical Information Mart for Intensive Care (MIMIC) database to extract ICU stay data for adult patients to build two prediction models: one for mortality prediction and another for ICU length of stay. For the mortality model, we applied six commonly used machine learning (ML) binary classification algorithms for predicting the discharge status (survived or not). For the length of stay model, we applied the same six ML algorithms for binary classification using the median patient population ICU stay of 2.64 days. For the regression-based classification, we used two ML algorithms for predicting the number of days. We built two variations of each prediction model: one using 12 baseline demographic and vital sign features, and the other based on our proposed quantiles approach, in which we use 21 extra features engineered from the baseline vital sign features, including their modified means, standard deviations, and quantile percentages. Results: We could perform predictive modeling with minimal features while maintaining reasonable performance using the quantiles approach. The best accuracy achieved in the mortality model was approximately 89% using the random forest algorithm. The highest accuracy achieved in the length of stay model, based on the population median ICU stay (2.64 days), was approximately 65% using the random forest algorithm. Conclusions: The novelty in our approach is that we built models to predict ICU length of stay and mortality with reasonable accuracy based on a combination of ML and the quantiles approach that utilizes only vital signs available from the patient’s profile without the need to use any external features. This approach is based on feature engineering of the vital signs by including their modified means, standard deviations, and quantile percentages of the original features, which provided a richer dataset to achieve better predictive power in our models.

59 BASIC BIOLOGICAL SCIENCES↗

A combined experimental and machine learning exploration of Ti 2-x Zr x MnCrFeNi high entropy Laves hydrides

A series of high entropy AB 2 -type Ti 2-x Zr x MnCrFeNi alloys (x = 0.6, 0.7, 0.8, 0.9, 1.0, 1.1 and 1.2) were synthesized to investigate their potential for hydrogen storage and chemical compression. The influence of the Ti/Zr ratio was explored in terms of structural, microstructural and thermodynamic properties. The storage capacity together with the reaction enthalpy and entropy changes of the synthesized high entropy alloys were compared to predictions from Machine Learning (ML) to investigate changes in these properties across the explored composition space. The results revealed that a decreasing Zr content consistently lowered the hydride formation enthalpy and increased the plateau pressure from 8 to >90 bar H 2 at 25 °C, in good agreement with ML predictions. Selected compositions (x = 1.0 and 1.2) demonstrated reversible hydrogen storage capability over 150 cycles, with capacities of 1.34–1.40 wt % H 2 and remarkable reaction kinetics (<4 min) at ambient temperature. These experimental and computational findings highlight the potential of this Laves-HEA system as tuneable, stable, and cost-effective materials suitable for long-term operations in stationary hydrogen storage and compression applications.

36 MATERIALS SCIENCE↗

Deep learning for electron and scanning probe microscopy: From materials design to atomic fabrication

Machine learning and artificial intelligence (ML/AI) are rapidly becoming an indispensable part of physics research, with applications ranging from theory and materials prediction to high-throughput data analysis. In parallel, the recent successes in applying ML/AI methods for autonomous systems from robotics through self-driving cars to organic and inorganic synthesis are generating enthusiasm for the potential of these techniques to enable automated and autonomous experiment in imaging. Here, we discuss recent progress in application of machine learning methods in scanning transmission electron microscopy and scanning probe microscopy, from applications such as data compression and exploratory data analysis to physics learning to atomic fabrication.

36 MATERIALS SCIENCE↗

Unraveling Network-induced Memory Contention: Deeper Insights with Machine Learning

Remote Direct Memory Access (RDMA) is expected to be an integral communication mechanism for future exascale systems enabling asynchronous data transfers, so that applications may fully utilize CPU resources while simultaneously sharing data amongst remote nodes. We examine Network-induced Memory Contention (NiMC) on Infiniband networks. We expose the interactions between RDMA, main-memory and cache, when applications and out-of-band services compete for memory resources. We then explore NiMCs resulting impact on application-level performance. For a range of hardware technologies and HPC workloads, we quantify NiMC and show that NiMCs impact grows with scale resulting in up to 3X performance degradation at scales as small as 8K processes even in applications that previously have been shown to be performance resilient in the presence of noise. In addition, this work examines the problem of predicting NiMC's impact on applications by leveraging machine learning and easily accessible performance counters. This approach provides additional insights about the root cause of NiMC and facilitates dynamic selection of potential solutions. Finally, we evaluated three potential techniques to reduce NiMCs impact, namely hardware offloading, core reservation and network throttling.

97 MATHEMATICS AND COMPUTING↗