Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “learned”

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 1,063 records · Page 59

Machine learning-led semi-automated medium optimization reveals salt as key for flaviolin production in Pseudomonas putida

Although synthetic biology can produce valuable chemicals in a renewable manner, its progress is still hindered by a lack of predictive capabilities. Media optimization is a critical, and often overlooked, process which is essential to obtain the titers, rates and yields needed for commercial viability. Here, we present a molecule- and host-agnostic active learning process for media optimization that is enabled by a fast and highly repeatable semi-automated pipeline. Its application yielded 60% and 70% increases in titer, and 350% increase in process yield in three different campaigns for flaviolin production in Pseudomonas putida KT2440. Explainable Artificial Intelligence techniques pinpointed that, surprisingly, common salt (NaCl) is the most important component influencing production. The optimal salt concentration is very high, comparable to seawater and close to the limits that P. putida can tolerate. The availability of fast Design-Build-Test-Learn (DBTL) cycles allowed us to show that performance improvements for active learning are rarely monotonous. This work illustrates how machine learning and automation can change the paradigm of current synthetic biology research to make it more effective and informative, and suggests a cost-effective and underexploited strategy to facilitate the high titers, rates and yields essential for commercial viability.

59 BASIC BIOLOGICAL SCIENCES↗

Machine learning inversion of interatomic force constants from single-crystal inelastic neutron scattering

Atomic vibrations govern many macroscopic properties of materials, but experiments to comprehensively probe them remain challenging. Inelastic neutron scattering (INS) is a powerful technique to map phonon dispersions in crystals, especially when leveraging modern time-of-flight (ToF) spectrometers with large detectors. However, efficiently and robustly extracting interatomic force constants (FCs) parameterizing phonon dynamics from experimental spectra remains a bottleneck due to the complexity and high dimensionality of ToF INS datasets. Here, we present a machine learning approach for the direct inversion of FCs from single-crystal INS measurements. The framework leverages synthetic training data generated using universal machine-learned force fields and an efficient physics-based forward model. We benchmark two neural architectures–one emphasizing structured latent representation learning and the other direct, supervised spectral regression–across simulated datasets for two materials under idealized and noisy conditions. The latent-representation model is subsequently applied to experimental single-crystal INS data on germanium. The model is shown to reproduce FCs derived from both first-principles simulations and from iterative optimization, and furthermore achieves reliable inference even from sparse, single-orientation measurements representing short data acquisitions. Analysis of the learned latent space reveals semantically continuous and physically interpretable encodings that support strong cross-domain generalization. By bridging theoretical and experimental domains, we establish a path toward rapid inversion of experimental spectra and data-driven interpretation of temperature-dependent lattice dynamics.

42 ENGINEERING↗

Machine learning approaches for structural and thermodynamic properties of a Lennard-Jones fluid

Predicting the functional properties of many molecular systems relies on understanding how atomistic interactions give rise to macroscale observables. However, current attempts to develop predictive models for the structural and thermodynamic properties of condensed-phase systems often rely on extensive parameter fitting to empirically selected functional forms whose effectiveness is limited to a narrow range of physical conditions. Here, we illustrate how these traditional fitting paradigms can be superseded using machine learning. Specifically, we use the results of molecular dynamics simulations to train machine learning protocols that are able to produce the radial distribution function, pressure, and internal energy of a Lennard-Jones fluid with increased accuracy in comparison to previous theoretical methods. The radial distribution function is determined using a variant of the segmented linear regression with the multivariate function decomposition approach developed by Craven et al. [J. Phys. Chem. Lett. 11, 4372 (2020)]. The pressure and internal energy are determined using expressions containing the learned radial distribution function and also a kernel ridge regression process that is trained directly on thermodynamic properties measured in simulation. The presented results suggest that the structural and thermodynamic properties of fluids may be determined more accurately through machine learning than through human-guided functional forms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

When do short-range atomistic machine-learning models fall short?

In this work, we explore the role of long-range interactions in atomistic machine-learning models by analyzing the effects on fitting accuracy, isolated cluster properties, and bulk thermodynamic properties. Such models have become increasingly popular in molecular simulations given their ability to learn highly complex and multi-dimensional interactions within a local environment; however, many of them fundamentally lack a description of explicit long-range interactions. In order to provide a well-defined benchmark system with precisely known pairwise interactions, we chose as the reference model a flexible version of the Extended Simple Point Charge (SPC/E) water model. Our analysis shows that while local representations are sufficient for predictions of the condensed liquid phase, the short-range nature of machine-learning models falls short in representing cluster and vapor phase properties. These findings provide an improved understanding of the role of long-range interactions in machine learning models and the regimes where they are necessary.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Improved accuracy and transferability of molecular-orbital-based machine learning: Organics, transition-metal complexes, non-covalent interactions, and transition states

We report molecular-orbital-based machine learning (MOB-ML) provides a general framework for the prediction of accurate correlation energies at the cost of obtaining molecular orbitals. The application of Nesbet’s theorem makes it possible to recast a typical extrapolation task, training on correlation energies for small molecules and predicting correlation energies for large molecules, into an interpolation task based on the properties of orbital pairs. We demonstrate the importance of preserving physical constraints, including invariance conditions and size consistency, when generating the input for the machine learning model. Numerical improvements are demonstrated for different datasets covering total and relative energies for thermally accessible organic and transition-metal containing molecules, non-covalent interactions, and transition-state energies. MOB-ML requires training data from only 1% of the QM7b-T dataset (i.e., only 70 organic molecules with seven and fewer heavy atoms) to predict the total energy of the remaining 99% of this dataset with sub-kcal/mol accuracy. This MOB-ML model is significantly more accurate than other methods when transferred to a dataset comprising of 13 heavy atom molecules, exhibiting no loss of accuracy on a size intensive (i.e., per-electron) basis. It is shown that MOB-ML also works well for extrapolating to transition-state structures, predicting the barrier region for malonaldehyde intramolecular proton-transfer to within 0.35 kcal/mol when only trained on reactant/product-like structures. Finally, the use of the Gaussian process variance enables an active learning strategy for extending the MOB-ML model to new regions of chemical space with minimal effort. We demonstrate this active learning strategy by extending a QM7b-T model to describe non-covalent interactions in the protein backbone–backbone interaction dataset to an accuracy of 0.28 kcal/mol.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Beyond optimization—supervised learning applications in relativistic laser-plasma experiments

We explore the applications of machine learning techniques in relativistic laser-plasma experiments beyond optimization purposes. We predict the beam charge of electrons produced in a laser wakefield accelerator given the laser wavefront change caused by a deformable mirror. Machine learning enables feature analysis beyond merely searching for an optimal beam charge, showing that specific aberrations in the laser wavefront are favored in generating higher beam charges. Supervised learning models allow characterizing the measured data quality as well as recognizing irreproducible data and potential outliers. Furthermore, we also include virtual measurement errors in the experimental data to examine the model robustness under these conditions. This work demonstrates how machine learning methods can benefit data analysis and physics interpretation in a highly nonlinear problem of relativistic laser-plasma interaction.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Learning the temporal evolution of multivariate densities via normalizing flows

In this work, we propose a method to learn multivariate probability distributions using sample path data from stochastic differential equations. Specifically, we consider temporally evolving probability distributions (e.g., those produced by integrating local or nonlocal Fokker–Planck equations). Here, we analyze this evolution through machine learning assisted construction of a time-dependent mapping that takes a reference distribution (say, a Gaussian) to each and every instance of our evolving distribution. If the reference distribution is the initial condition of a Fokker–Planck equation, what we learn is the time-T map of the corresponding solution. Specifically, the learned map is a multivariate normalizing flow that deforms the support of the reference density to the support of each and every density snapshot in time. We demonstrate that this approach can approximate probability density function evolutions in time from observed sampled data for systems driven by both Brownian and Lévy noise. We present examples with two- and three-dimensional, uni- and multimodal distributions to validate the method.

97 MATHEMATICS AND COMPUTING↗

Transfer learning driven design optimization for inertial confinement fusion

Transfer learning is a promising approach to create predictive models that incorporate simulation and experimental data into a common framework. In this technique, a neural network is first trained on a large database of simulations and then partially retrained on sparse sets of experimental data to adjust predictions to be more consistent with reality. Previously, this technique has been used to create predictive models of Omega [Humbird et al., IEEE Trans. Plasma Sci. 48, 61–70 (2019)] and NIF [Humbird et al., Phys. Plasmas 28, 042709 (2021); Kustowski et al., Mach. Learn. 3, 015035 (2022)] inertial confinement fusion (ICF) experiments that are more accurate than simulations alone. Here, in this work, we conduct a transfer learning driven hypothetical ICF campaign in which the goal is to maximize experimental neutron yield via Bayesian optimization. The transfer learning model achieves yields within 5% of the maximum achievable yield in a modest-sized design space in fewer than 20 experiments. Furthermore, we demonstrate that this method is more efficient at optimizing designs than traditional model calibration techniques commonly employed in ICF design. Such an approach to ICF design could enable robust optimization of experimental performance under uncertainty.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Unraveling the impact of initial choices and in-loop interventions on learning dynamics in autonomous scanning probe microscopy

The current focus in Autonomous Experimentation (AE) is on developing robust workflows to conduct the AE effectively. This entails the need for well-defined approaches to guide the AE process, including strategies for hyperparameter tuning and high-level human interventions within the workflow loop. This paper presents a comprehensive analysis of the influence of initial experimental conditions and in-loop interventions on the learning dynamics of Deep Kernel Learning (DKL) within the realm of AE in scanning probe microscopy. We explore the concept of the “seed effect,” where the initial experiment setup has a substantial impact on the subsequent learning trajectory. Additionally, we introduce an approach of the seed point interventions in AE allowing the operator to influence the exploration process. Using a dataset from Piezoresponse Force Microscopy on PbTiO 3 thin films, we illustrate the impact of the “seed effect” and in-loop seed interventions on the effectiveness of DKL in predicting material properties. The study highlights the importance of initial choices and adaptive interventions in optimizing learning rates and enhancing the efficiency of automated material characterization. This work offers valuable insights into designing more robust and effective AE workflows in microscopy with potential applications across various characterization techniques.

47 OTHER INSTRUMENTATION↗

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↗

Scalable learning of potentials to predict time-dependent Hartree–Fock dynamics

We propose a framework to learn the time-dependent Hartree–Fock (TDHF) inter-electronic potential of a molecule from its electron density dynamics. Although the entire TDHF Hamiltonian, including the inter-electronic potential, can be computed from first principles, we use this problem as a testbed to develop strategies that can be applied to learn a priori unknown terms that arise in other methods/approaches to quantum dynamics, e.g., emerging problems such as learning exchange–correlation potentials for time-dependent density functional theory. We develop, train, and test three models of the TDHF inter-electronic potential, each parameterized by a four-index tensor of size up to 60 × 60 × 60 × 60. Two of the models preserve Hermitian symmetry, while one model preserves an eight-fold permutation symmetry that implies Hermitian symmetry. Across seven different molecular systems, we find that accounting for the deeper eight-fold symmetry leads to the best-performing model across three metrics: training efficiency, test set predictive power, and direct comparison of true and learned inter-electronic potentials. All three models, when trained on ensembles of field-free trajectories, generate accurate electron dynamics predictions even in a field-on regime that lies outside the training set. To enable our models to scale to large molecular systems, we derive expressions for Jacobian-vector products that enable iterative, matrix-free training.

97 MATHEMATICS AND COMPUTING↗

Transfer learning for analysis of collective and non-collective Thomson scattering spectra

Thomson scattering (TS) diagnostics provide reliable, minimally perturbative measurements of fundamental plasma parameters, such as electron density (⁠n e ) and electron temperature (⁠T e ⁠). Deep neural networks can provide accurate estimates of ⁠n e and T e when conventional fitting algorithms may fail, such as when TS spectra are dominated by noise, or when fast analysis is required for real-time operation. Although deep neural networks typically require large training sets, transfer learning can improve model performance on a target task with limited data by leveraging pre-trained models from related source tasks, where select hidden layers are further trained using target data. We present five architecturally diverse deep neural networks, pre-trained on synthetic TS data and adapted for experimentally measured TS data, to evaluate the efficacy of transfer learning in estimating n e and T e in both the collective and non-collective scattering regimes. We evaluate errors in n e and T e estimates as a function of training set size for models trained with and without transfer learning, and we observe decreases in model error from transfer learning when the training set contains ≲ 200 experimentally measured spectra.

Artificial neural networks↗

Definitions, methods, and applications in interpretable machine learning

Machine-learning models have demonstrated great success in learning complex patterns that enable them to make predictions about unobserved data. In addition to using models for prediction, the ability to interpret what a model has learned is receiving an increasing amount of attention. However, this increased focus has led to considerable confusion about the notion of interpretability. In particular, it is unclear how the wide array of proposed interpretation methods are related and what common concepts can be used to evaluate them. We aim to address these concerns by defining interpretability in the context of machine learning and introducing the predictive, descriptive, relevant (PDR) framework for discussing interpretations. The PDR framework provides 3 overarching desiderata for evaluation: predictive accuracy, descriptive accuracy, and relevancy, with relevancy judged relative to a human audience. Moreover, to help manage the deluge of interpretation methods, we introduce a categorization of existing techniques into model-based and post hoc categories, with subgroups including sparsity, modularity, and simulatability. To demonstrate how practitioners can use the PDR framework to evaluate and understand interpretations, we provide numerous real-world examples. These examples highlight the often underappreciated role played by human audiences in discussions of interpretability. Finally, based on our framework, we discuss limitations of existing methods and directions for future work. We hope that this work will provide a common vocabulary that will make it easier for both practitioners and researchers to discuss and choose from the full range of interpretation methods.

97 MATHEMATICS AND COMPUTING↗

A Reinforcement Learning Approach to Augment Conventional PID Control in Nuclear Power Plant Transient Operation

The ability of nuclear reactors to operate their power conversion cycles more flexibly will enhance their value to energy grids with variable pricing. Current nuclear control systems are typically classical controllers that are often based on proportional-integral-derivative (PID) control. This paper presents a method of augmenting the existing PID control for difficult transient operations in nuclear power plants using a reinforcement learning–derived feedforward signal applied in real time. The agents, which are trained on a test thermal load-following problem, are designed to improve steam generator outlet temperature control for a range of fast load-following scenarios covering ramp rates from 9%/min to 15%/min. Several reinforcement learning algorithms were initially investigated for the training of the feedforward agents with deep Q-learning (DQN) and proximal policy optimization (PPO) networks, which were found to be the most promising. The DQN controllers utilize discrete actions, giving them a better disturbance rejection at steady state but inconsistent response to initial temperature deviations. In contrast, PPO-trained agents, which take continuous actions except for a dead zone around zero, were shown to have the best combination of high disturbance rejection at steady state and good tracking of the desired temperature value. The ability of the PPO agent was also examined, with the average time of decision making found to be on the order of 1 ms. The fault properties of the controller under the loss of the reinforcement learning agent feedforward signal were also examined. The controller showed strong performance in situations of “no-signal” faults. but was less good at handling “stuck-at” faults, where the feedforward signal remains at a set value. In both cases, however, the PID was able to successfully maintain stability, eventually returning the system to a steady state. It is hoped that this work will allow for the proposed control architecture to be examined for more difficult control problems such that it may eventually be used to adapt existing nuclear plants for more aggressive load-following on grids of the future.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Reinforcement Learning for Anomaly Detection in Nuclear Power Plant Operation and Maintenance

In nuclear power plants (NPPs), timely identification of sensor and human errors is critical to ensure safe and efficient plant operations. Anomaly detection models can be employed for this task. However, traditional anomaly detection approaches may have high dependency on labeled datasets and struggle with adaptability in complex, dynamic environments. Reinforcement learning (RL) has demonstrated significant potential in fault diagnosis and anomaly detection; however, its application to anomaly detection in NPPs remains a relatively underexplored research direction. Hence, to address this gap, in this study, we present a novel physics-informed reinforcement learning model, PIRL-AD: Physics-Informed Reinforcement Learning for Anomaly Detection, that integrates domain knowledge from calorimetric equations into the RL framework for enhanced sensor and human error anomaly detection. We evaluate the performance of PIRL-AD against a non-physics informed RL benchmark and a support vector machine (SVM) on data collected from a forced flow loop testbed. Experimental results suggest that PIRL-AD outperforms other baselines on a range of anomalous datasets that include both sensor and human-induced anomalies across key performance metrics, statistically outperforming the RL and SVM benchmarks with respect to geometric mean (respectively, 92.96% vs. 91.06% vs. 83.01%) and F1-score (respectively, 89.23% vs. 86.98% vs. 77.01%). Furthermore, the findings suggest the potential of physics-integrated reinforcement learning models for enhanced anomaly detection performance in NPPs.

Reinforcement learning↗

Subspace-Driven Learning for Anomaly Detection in Process Transients

Nuclear power plant (NPP) monitoring and diagnostic centers are actively investigating and implementing automated anomaly detection algorithms to help plants catch anomalies sooner, thereby preventing or reducing the duration of unexpected shutdowns. Current machine learning-based anomaly detection methods are expected to be highly effective during stable, full-power operations because NPPs typically operate as baseload power generators, meaning there are extensive operating data available from plant equipment. However, it is expected that anomaly detection methods will face significant challenges during transient conditions (i.e., when power output falls below full power) because plants only occasionally operate at these lower power levels, generating sparse transient operational data, and resulting in false alarms or missed detections. Here, to address this issue, transfer learning is used, which for this problem leverages knowledge (in the form of learned features) from stable, full-power operations to improve detection accuracy during transient conditions, even with limited data. In this effort, a novel subspace approach is developed to transfer a subset of the data features from full power operation to transients. This approach is validated through experiments using synthetic data and was found to outperform two baseline transfer learning approaches in anomaly detection performance across a range of amounts of transient data used in the training process.

46 - INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AN↗

Exploring electron beam induced atomic assembly via reinforcement learning in a molecular dynamics environment

We report atom-by-atom assembly of functional materials and devices is perceived as one of the ultimate targets of nanotechnology. Recently it has been shown that the beam of a scanning transmission electron microscope can be used for targeted manipulation of individual atoms. However, the process is highly dynamic in nature rendering control difficult. One possible solution is to instead train artificial agents to perform the atomic manipulation in an automated manner without need for human intervention. As a first step to realizing this goal, we explore how artificial agents can be trained for atomic manipulation in a simplified molecular dynamics environment of graphene with Si dopants, using reinforcement learning. We find that it is possible to engineer the reward function of the agent in such a way as to encourage formation of local clusters of dopants under different constraints. This study shows the potential for reinforcement learning in nanoscale fabrication, and crucially, that the dynamics learned by agents encode specific elements of important physics that can be learned.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

How accurate is a machine learning-based wind speed extrapolation under a round-robin approach?

As the size of commercial wind turbines keeps increasing, having accurate ways to vertically extrapolate wind speed is essential to obtain a precise characterization of the wind resource for wind energy production. Recently, machine learning has been proposed and applied to extrapolate wind speed to hub heights. However, previous studies trained and tested the machine learning methods at the same site, giving them an unfair advantage over the conventional extrapolation techniques, which are instead more universal. Here, we use data from four sites in Oklahoma to test a round-robin validation approach for machine learning, under which we train a random forest at a site, and test it at a different site, where the model has no prior knowledge of the wind resource. We quantify how the accuracy of this technique varies with distance from the training site, and we find that it outperforms conventional techniques for wind extrapolation at all the considered spatial separations. We then assess how the accuracy of the machine-learning based approach varies when it is used to predict wind speed in a wind farm far wake. Finally, we explore as case study the performance of the random forest in extrapolating winds during a low-level jet event.

17 WIND ENERGY↗