Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Neural Networks (Computer)”

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 1,657 records · Page 92

Intelligent Experiments Through Real-time AI: Fast Data Processing and Autonomous Detector Control for sPHENIX and Future EIC Detectors (Final Report)

The overall vision of this project was to integrate real-time artificial intelligence (AI) directly into the data acquisition and detector-control systems of nuclear physics experiments, including both fast online event selection and an autonomous detector-control feedback loop. The work carried out under the award focused on the fast online event-selection half of that vision: the efficient recording of low-momentum heavy-flavor (HF) hadron decays in proton-proton collisions at the sPHENIX experiment at the Relativistic Heavy Ion Collider (RHIC)—an observable that requires fast tracking and topological trigger selection not previously demonstrated at RHIC, and that is essential for QCD studies at future facilities such as the Electron-Ion Collider (EIC). The autonomous detector-control (GPU-based feedback) component named in the project title remained a design concept and was not implemented under this award. The Massachusetts Institute of Technology (MIT) group led the offline simulation and data processing needed to train the machine-learning (ML) models, the translation of trained models to Field-Programmable Gate Array (FPGA) firmware using the hls4ml framework, and the physics validation of heavy-flavor reconstruction. Over the award period, the team developed and hardware-tested the principal components of an AI-based heavy-flavor trigger on simulated and recorded sPHENIX tracker data: a software Bipartite Graph Attention Network (BiGAT) trigger model reaching > 95% signal efficiency at 99% background rejection; an FPGA-native hit clusterizer matching the offline clustering; smaller networks synthesized to FPGA within the required sub-10 µs latency; and an assembled decoder–clusterizer–inference firmware chain exercised on the FELIX readout board. A complete, fully integrated hardware demonstrator was not finished within the award period. This report documents the project goals, the MIT group’s contributions, the technical accomplishments, and the outlook toward applications at the future EIC ePIC detector.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Physics-informed neural networks for heterogeneous poroelastic media

This study presents a novel physics-informed neural network (PINN) framework for modeling poroelasticity in heterogeneous media with material interfaces. The approach introduces a composite neural network (CoNN) where separate neural networks predict displacement and pressure variables for each material. While sharing identical activation functions, these networks are independently trained for all other parameters. To address challenges posed by heterogeneous material interfaces, the CoNN is integrated with the Interface-PINNs (I-PINNs) framework (Sarma et al., Comput. Methods Appl. Mech. Eng. 429: 117135, 2024), allowing different activation functions across material interfaces. Further, this ensures accurate approximation of discontinuous solution fields and gradients. Performance and accuracy of this combined architecture were evaluated against the conventional PINNs approach, a single neural network (SNN) architecture, and the eXtended PINNs (XPINNs) framework through two one-dimensional benchmark examples with discontinuous material properties. The results show that the proposed CoNN with I-PINNs architecture achieves an RMSE that is two orders of magnitude better than the conventional PINNs approach and is at least 40 times faster than the SNN framework. Compared to XPINNs, the proposed method achieves an RMSE at least one order of magnitude better and is 40% faster.

42 ENGINEERING↗

Adaptive Optimization of Aircraft Engine Performance Using Neural Networks

Preliminary results are presented on the development of an adaptive neural network based control algorithm to enhance aircraft engine performance. This work builds upon a previous National Aeronautics and Space Administration (NASA) effort known as Performance Seeking Control (PSC). PSC is an adaptive control algorithm which contains a model of the aircraft's propulsion system which is updated on-line to match the operation of the aircraft's actual propulsion system. Information from the on-line model is used to adapt the control system during flight to allow optimal operation of the aircraft's propulsion system (inlet, engine, and nozzle) to improve aircraft engine performance without compromising reliability or operability. Performance Seeking Control has been shown to yield reductions in fuel flow, increases in thrust, and reductions in engine fan turbine inlet temperature. The neural network based adaptive control, like PSC, will contain a model of the propulsion system which will be used to calculate optimal control commands on-line. Hopes are that it will be able to provide some additional benefits above and beyond those of PSC. The PSC algorithm is computationally intensive, it is valid only at near steady-state flight conditions, and it has no way to adapt or learn on-line. These issues are being addressed in the development of the optimal neural controller. Specialized neural network processing hardware is being developed to run the software, the algorithm will be valid at steady-state and transient conditions, and will take advantage of the on-line learning capability of neural networks. Future plans include testing the neural network software and hardware prototype against an aircraft engine simulation. In this paper, the proposed neural network software and hardware is described and preliminary neural network training results are presented.

Simon, Donald L.↗

HEPOM: Using Graph Neural Networks for the Accelerated Predictions of Hydrolysis Free Energies in Different pH Conditions

Hydrolysis is a fundamental family of chemical reactions where water facilitates the cleavage of bonds. The process is ubiquitous in biological and chemical systems, owing to water’s remarkable versatility as a solvent. However, accurately predicting the feasibility of hydrolysis through computational techniques is a difficult task, as subtle changes in reactant structure like heteroatom substitutions or neighboring functional groups can influence the reaction outcome. Furthermore, hydrolysis is sensitive to the pH of the aqueous medium, and the same reaction can have different reaction properties at different pH conditions. In this work, we have combined reaction templates and high-throughput ab initio calculations to construct a diverse data set of hydrolysis free energies. The developed framework automatically identifies reaction centers, generates hydrolysis products, and utilizes a trained graph neural network (GNN) model to predict ΔG values for all potential hydrolysis reactions in a given molecule. The long-term goal of the work is to develop a data-driven, computational tool for high-throughput screening of pH-specific hydrolytic stability and the rapid prediction of reaction products, which can then be applied in a wide array of applications including chemical recycling of polymers and ion-conducting membranes for clean energy generation and storage.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

SENTRA: A Modular Computational Graph Framework for Critical Mineral and Materials Supply Chains: Part I: Network Construction Latent-Quantity Estimation, and Temporal Graph Forecasting

Global supply chains for critical minerals and materials are complex, evolving networks of countries, products, production stages, and trade relationships. Existing analytical approaches are limited by fragmented data and static network representations that do not capture the dynamic production dependencies linking raw materials, intermediate products, and final goods across multiple countries. Trade and production statistics provide only a partial view of domestic production, inventories, and material flows, making it difficult to identify indirect sourcing pathways, hidden dependencies, and embedded foreign exposures. This paper introduces the Supply Chain Exposure Network Tracking and Risk Assessment (SENTRA) framework, a modular graph-based computational framework for constructing, analyzing, and forecasting dynamic supply chain networks. As the first paper in a three-part methodological series, it establishes the computational foundation of SENTRA by constructing a temporal attributed multi-relational graph whose nodes represent product–country pairs and whose edges encode observed trade and within-country value-chain relationships. Statistical estimation and constrained optimization recover latent production, final demand, and product input dependency coefficients while enforcing economic accounting constraints. Graph-derived exposure measures quantify direct, transshipment, value-chain, and multi-hop supply chain dependencies independently of the forecasting model. A temporal graph forecasting architecture based on a relational graph neural network then forecasts the evolution of the graph under mass-balance constraints with distribution-free conformal uncertainty quantification. Validation on the global aluminum supply chain shows that the learned graph representations recover economically meaningful supply chain structure, accurately forecast out-of-sample trade relationships, and produce well-calibrated prediction intervals. Subsequent papers apply this computational foundation to exposure assessment, disruption analysis, and scenario-based policy analysis, and extend the framework to multimaterial supply chain modeling and decision support.

36 MATERIALS SCIENCE↗

When and why PINNs fail to train: A neural tangent kernel perspective

Physics-informed neural networks (PINNs) have lately received great attention thanks to their flexibility in tackling a wide range of forward and inverse problems involving partial differential equations. However, despite their noticeable empirical success, little is known about how such constrained neural networks behave during their training via gradient descent. More importantly, even less is known about why such models sometimes fail to train at all. Here in this work, we aim to investigate these questions through the lens of the Neural Tangent Kernel (NTK); a kernel that captures the behavior of fully-connected neural networks in the infinite width limit during training via gradient descent. Specifically, we derive the NTK of PINNs and prove that, under appropriate conditions, it converges to a deterministic kernel that stays constant during training in the infinite-width limit. This allows us to analyze the training dynamics of PINNs through the lens of their limiting NTK and find a remarkable discrepancy in the convergence rate of the different loss components contributing to the total training error. To address this fundamental pathology, we propose a novel gradient descent algorithm that utilizes the eigenvalues of the NTK to adaptively calibrate the convergence rate of the total training error. Finally, we perform a series of numerical experiments to verify the correctness of our theory and the practical effectiveness of the proposed algorithms.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Real-time signal detection for Cyclotron Radiation Emission Spectroscopy measurements using antenna arrays

Cyclotron Radiation Emission Spectroscopy (CRES) is a technique for precision measurement of the energies of charged particles, which is being developed by the Project 8 Collaboration to measure the neutrino mass using tritium beta-decay spectroscopy. Project 8 seeks to use the CRES technique to measure the neutrino mass with a sensitivity of 40 meV, requiring a large supply of tritium atoms stored in a multi-cubic meter detector volume. Antenna arrays are one potential technology compatible with an experiment of this scale, but the capability of an antenna-based CRES experiment to measure the neutrino mass depends on the efficiency of the signal detection algorithms. Here, in this paper, we develop efficiency models for three signal detection algorithms and compare them using simulations from a prototype antenna-based CRES experiment as a case-study. The algorithms include a power threshold, a matched filter template bank, and a neural network based machine learning approach, which are analyzed in terms of their average detection efficiency and relative computational cost. It is found that significant improvements in detection efficiency and, therefore, neutrino mass sensitivity are achievable, with only a moderate increase in computation cost, by utilizing either the matched filter or machine learning approach in place of a power threshold, which is the baseline signal detection algorithm used in previous CRES experiments by Project 8.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

FPDeep: Scalable Acceleration of CNN Training on Deeply-Pipelined FPGA Clusters

In this paper, we propose a framework called FPDeep, which uses a hybrid of model and layer paral- lelism to configure distributed reconfigurable clusters to train DNNs. This approach has numerous benefits. First, the design does not suffer from batch size growth. Second, novel workload and weight partitioning leads to balanced loads of both among nodes. And third, the entire system is fine-grained pipeline. This leads to high parallelism and utilization and also minimizes the time features need to be cached while waiting for back-propagation.

Wang, Tianqi↗

Beam Hardening Artifact Reduction in X-Ray CT Reconstruction of 3D Printed Metal Parts Leveraging Deep Learning and CAD Models

Nondestructive evaluation (NDE) of additively manufactured (AM) parts is important for understanding the impacts of various process parameters and qualifying the built part. X-ray computed tomography (XCT) has played a critical role in rapid NDE and characterization of AM parts. However, XCT of metal AM parts can be challenging because of artifacts produced by standard reconstruction algorithms as a result of a confounding effect called “beam hardening.” Beam hardening artifacts complicate the analysis of XCT images and adversely impact the process of detecting defects, such as pores and cracks, which is key to ensuring the quality of the parts being printed. In this work, we propose a novel framework based on using available computer-aided design (CAD) models for parts to be manufactured, accurate XCT simulations, and a deep-neural network to produce high-quality XCT reconstructions from data that are affected by noise and beam hardening. Using extensive experiments with simulated data sets, we demonstrate that our method can significantly improve the reconstruction quality, thereby enabling better detection of defects compared with the state of the art. We also present promising preliminary results of applying the deep networks trained using CAD models to experimental data obtained from XCT of an AM jet-engine turbine blade.

Ziabari, Amir↗

Toward fusion plasma scenario planning for NSTX-U using machine-learning-accelerated models

One of the most promising devices for realizing power production through nuclear fusion is the tokamak. To maximize performance, it is preferable that tokamak reactors achieve advanced operating scenarios characterized by good plasma confinement, improved magnetohydrodynamic (MHD) stability, and a largely non-inductively driven plasma current. Such scenarios could enable steady-state reactor operation with high \emph{fusion gain} --- the ratio of produced fusion power to the external power provided through the plasma boundary. Precise and robust control of the evolution of the plasma boundary shape as well as the spatial distribution of the plasma current, density, temperature, and rotation will be essential to achieving and maintaining such scenarios. The complexity of the evolution of tokamak plasmas, arising due to nonlinearities and coupling between various parameters, motivates the use of model-based control algorithms that can account for the system dynamics. In this work, a learning-based accelerated model trained on data from the National Spherical Torus Experiment Upgrade (NSTX-U) is employed to develop planning and control strategies for regulating the density and temperature profile evolution around desired trajectories. The proposed model combines empirical scaling laws developed across multiple devices with neural networks trained on empirical data from NSTX-U and a database of first-principles-based computationally intensive simulations. The reduced execution time of the accelerated model will enable practical application of optimization algorithms and reinforcement learning approaches for scenario planning and control development. An initial demonstration of applying optimization approaches to the learning-based model is presented, including a strategy for mitigating the effect of leaving the finite validity range of the accelerated model. The approach shows promise for actuator planning between experiments and in real-time.

machine learning↗

Explainable Neural Architecture Search (XNAS)

Code for the paper Learning Interpretable Models Through Multi-Objective Neural Architecture Search by Zachariah Carmichael, Tim Moon, and Sam Ade Jacobs. Monumental advances in deep learning have led to unprecedented achievements across a multitude of domains. While the performance of deep neural networks is indubitable, the architectural design and interpretability of such models are nontrivial. Research has been introduced to automate the design of neural network architectures through neural architecture search (NAS). Recent progress has made these methods more pragmatic by exploiting distributed computation and novel optimization algorithms. However, there is little work in optimizing architectures for interpretability. To this end, we propose a multiobjective distributed NAS framework that optimizes for both task performance and introspection. We leverage the non-dominated sorting genetic algorithm (NSGA-II) and explainable AI (XAI) techniques to reward architectures that can be better comprehended by humans. The framework is evaluated on several image classification datasets. We demonstrate that jointly optimizing for introspection ability and task error leads to more disentangled architectures that perform within tolerable error.

Carmichael, ZachariahJ↗

Monitoring the Microgravity Environment Quality On-Board the International Space Station Using Soft Computing Techniques

This paper presents an artificial intelligence monitoring system developed by the NASA Glenn Principal Investigator Microgravity Services project to help the principal investigator teams identify the primary vibratory disturbance sources that are active, at any moment in time, on-board the International Space Station, which might impact the microgravity environment their experiments are exposed to. From the Principal Investigator Microgravity Services' web site, the principal investigator teams can monitor via a graphical display, in near real time, which event(s) is/are on, such as crew activities, pumps, fans, centrifuges, compressor, crew exercise, platform structural modes, etc., and decide whether or not to run their experiments based on the acceleration environment associated with a specific event. This monitoring system is focused primarily on detecting the vibratory disturbance sources, but could be used as well to detect some of the transient disturbance sources, depending on the events duration. The system has built-in capability to detect both known and unknown vibratory disturbance sources. Several soft computing techniques such as Kohonen's Self-Organizing Feature Map, Learning Vector Quantization, Back-Propagation Neural Networks, and Fuzzy Logic were used to design the system.

Jules, Kenol↗

Monitoring the Microgravity Environment Quality On-board the International Space Station Using Soft Computing Techniques: Preliminary System Performance Results - Part 2

This paper presents the preliminary performance results of the artificial intelligence monitoring system in full operational mode using near real time acceleration data downlinked from the International Space Station. Preliminary microgravity environment characterization analysis result for the International Space Station (Increment-2), using the monitoring system is presented. Also, comparison between the system predicted performance based on ground test data for the US laboratory "Destiny" module and actual on-orbit performance, using measured acceleration data from the U.S. laboratory module of the International Space Station is presented. Finally, preliminary on-orbit disturbance magnitude levels are presented for the Experiment of Physics of Colloids in Space, which are compared with on ground test data. The ground test data for the Experiment of Physics of Colloids in Space were acquired from the Microgravity Emission Laboratory, located at the NASA Glenn Research Center, Cleveland, Ohio. The artificial intelligence was developed by the NASA Glenn Principal Investigator Microgravity Services Project to help the principal investigator teams identify the primary vibratory disturbance sources that are active, at any moment of time, on-board the International Space Station, which might impact the microgravity environment their experiments are exposed to. From the Principal Investigator Microgravity Services' web site, the principal investigator teams can monitor via a dynamic graphical display, implemented in Java, in near real time, which event(s) is/are on, such as crew activities, pumps, fans, centrifuges, compressor, crew exercise, structural modes, etc., and decide whether or not to run their experiments, whenever that is an option, based on the acceleration magnitude and frequency sensitivity associated with that experiment. This monitoring system detects primarily the vibratory disturbance sources. The system has built-in capability to detect both known and unknown vibratory disturbance sources. Several soft computing techniques such as Kohonen's Self-Organizing Feature Map, Learning Vector Quantization, Back-Propagation Neural Networks, and Fuzzy Logic were used to design the system.

Jules, Kenol↗

Real Time On-line Space Research Laboratory Environment Monitoring with Off-line Trend and Prediction Analysis

One of the responsibilities of the NASA Glenn Principal Investigator Microgravity Services is to support NASA sponsored investigators in the area of reduced-gravity acceleration data analysis, interpretation and the monitoring of the reduced-gravity environment on-board various carriers. With the International Space Station currently operational, a significant amount of acceleration data is being down-linked and processed on ground for both the space station onboard environment characterization (and verification) and scientific experiments. Therefore, to help principal investigator teams monitor the acceleration level on-board the International Space Station to avoid undesirable impact on their experiment, when possible, the NASA Glenn Principal Investigator Microgravity Services developed an artificial intelligence monitoring system, which detects in near real time any change in the environment susceptible to affect onboard experiments. The main objective of the monitoring system is to help research teams identify the vibratory disturbances that are active at any instant of time onboard the International Space Station that might impact the environment in which their experiment is being conducted. The monitoring system allows any space research scientist, at any location and at any time, to see the current acceleration level on-board the Space Station via the World Wide Web. From the NASA Glenn s Exploration Systems Division web site, research scientists can see in near real time the active disturbances, such as pumps, fans, compressor, crew exercise, re-boost, extra-vehicular activity, etc., and decide whether or not to continue operating or stopping (or making note of such activity for later correlation with science results) their experiments based on the g-level associated with that specific event. A dynamic graphical display accessible via the World Wide Web shows the status of all the vibratory disturbance activities with their degree of confidence as well as their g-level contribution to the environment. The system can detect both known and unknown vibratory disturbance activities. It can also perform trend analysis and prediction by analyzing past data over many Increments of the space station for selected disturbance activities. This feature can be used to monitor the health of onboard mechanical systems to detect and prevent potential system failure as well as for use by research scientists during their science results analysis. Examples of both real time on-line vibratory disturbance detection and off-line trend analysis are presented in this paper. Several soft computing techniques such as Kohonen s Self-Organizing Feature Map, Learning Vector Quantization, Back-Propagation Neural Networks, and Fuzzy Logic were used to design the system.

Jules, Kenol↗

High precision computing with charge domain devices and a pseudo-spectral method therefor

The present invention enhances the bit resolution of a CCD/CID MVM processor by storing each bit of each matrix element as a separate CCD charge packet. The bits of each input vector are separately multiplied by each bit of each matrix element in massive parallelism and the resulting products are combined appropriately to synthesize the correct product. In another aspect of the invention, such arrays are employed in a pseudo-spectral method of the invention, in which partial differential equations are solved by expressing each derivative analytically as matrices, and the state function is updated at each computation cycle by multiplying it by the matrices. The matrices are treated as synaptic arrays of a neural network and the state function vector elements are treated as neurons. In a further aspect of the invention, moving target detection is performed by driving the soliton equation with a vector of detector outputs. The neural architecture consists of two synaptic arrays corresponding to the two differential terms of the soliton-equation and an adder connected to the output thereof and to the output of the detector array to drive the soliton equation.

Barhen, Jacob↗

Deep Interacting Multiple Model Filtering

In this paper, a deep learning-based multiple model estimation framework is presented for the state estimation of hybrid dynamical systems from high dimensional observations such as camera images. A low dimensional vector which represents the measurement of the latent dynamical system and its corresponding variance are learned using a deep encoder neural network. An Interacting Multiple Model (IMM) filter is used to generate the latent state estimates and covariances using multiple dynamical models, which can be learned using backpropagation through time. The state estimates of the dynamical system and the corresponding covariance matrix are generated from the latent state estimates and covariance using a deep decoder neural network. The whole network is trained in an end-to-end manner using a loss function which minimizes the negative log-likelihood of the neural network parameters. Simulation results are presented using a 2D bouncing ball example and estimation error statistics are computed which demonstrates the accuracy and consistency of the estimation.

Ghananeel Rotithor↗

GeoLoRA: Geometric integration for parameter efficient fine-tuning

Low-Rank Adaptation (LoRA) has become a widely used method for parameter-efficient fine-tuning of large-scale, pre-trained neural networks. However, LoRA and its extensions face several challenges, including the need for rank adaptivity, robustness, and computational efficiency during the fine-tuning process. We introduce GeoLoRA, a novel approach that addresses these limitations by leveraging dynamical low-rank approximation theory. GeoLoRA requires only a single backpropagation pass over the small-rank adapters, significantly reducing computational cost as compared to similar dynamical low-rank training methods and making it faster than popular baselines such as AdaLoRA. This allows GeoLoRA to efficiently adapt the allocated parameter budget across the model, achieving smaller low-rank adapters compared to heuristic methods like AdaLoRA and LoRA, while maintaining critical convergence, descent, and error-bound theoretical guarantees. The resulting method is not only more efficient but also more robust to varying hyperparameter settings. We demonstrate the effectiveness of GeoLoRA on several state-of-the-art benchmarks, showing that it outperforms existing methods in both accuracy and computational efficiency.

Schotthoefer, Steffen [ORNL] (ORCID:00000002156965↗

Optimal Parameter for the Training of Multilayer Perceptron Neural Networks by Using Hierarchical Genetic Algorithm

This paper deals with the controversial topic of the selection of the parameters of a genetic algorithm, in this case hierarchical, used for training of multilayer perceptron neural networks for the binary classification. The parameters to select are the crossover and mutation probabilities of the control and parametric genes and the permanency percent. The results can be considered as a guide for using this kind of algorithm.

97 MATHEMATICS AND COMPUTING↗