Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Learning with errors”

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 1,675 records · Page 93
All sourcesNASA NTRSDOE OSTI
NASA NTRS1990

Intelligent control and adaptive systems; Proceedings of the Meeting, Philadelphia, PA, Nov. 7, 8, 1989

Various papers on intelligent control and adaptive systems are presented. Individual topics addressed include: control architecture for a Mars walking vehicle, representation for error detection and recovery in robot task plans, real-time operating system for robots, execution monitoring of a mobile robot system, statistical mechanics models for motion and force planning, global kinematics for manipulator planning and control, exploration of unknown mechanical assemblies through manipulation, low-level representations for robot vision, harmonic functions for robot path construction, simulation of dual behavior of an autonomous system. Also discussed are: control framework for hand-arm coordination, neural network approach to multivehicle navigation, electronic neural networks for global optimization, neural network for L1 norm linear regression, planning for assembly with robot hands, neural networks in dynamical systems, control design with iterative learning, improved fuzzy process control of spacecraft autonomous rendezvous using a genetic algorithm.

Rodriguez, Guillermo↗
NASA NTRSDate not supplied

Leveraging Human Performance Data to Change the Narrative that People are the Safety Problem

The study of errors and failure has a long and productive history in the behavioral sciences. By studying how systems fail, we rule out various mechanisms for how those systems might work, thereby refining our theories of how they actually work. Human performance, however, includes more than errors; human performance comprises both failures and successes. A systematic bias to collect and analyze data only on error affects the decisions we make as a community by promoting the narrative that “people are the safety problem.” This narrative manifests in both obvious and subtle ways in the design of systems intended for human use. When the only safety data that are available are about human failure, then “data-driven” designs can only consider that humans fail. Changing this narrative will depend on new data and new ways to examine data – specifically, data on the processes by which human create and contribute to safety. An alternate narrative is that people represent a primary source of safety, through their capability to anticipate, monitor for, respond to, and learn from expected and unexpected change. This presentation will describe research efforts to expand the range of safety-relevant events to include not just rare safety failures but frequent safety successes. These efforts include use of data from both operations and simulations to develop methods and metrics for learning from structured observation, self-report, and system data.

Jon Holbrook↗
DOE OSTI2022

WinnowML: Stable feature selection for maximizing prediction accuracy of time-based system modeling

Online deep learning (ODL) has become an important methodology for modeling time-based performance of computer systems. An open problem is the intelligent selection of features from raw workload traces of computer systems. The best methods are overly sensitive to noisy data, causing frequent feature changes and re-training. Using all available features inflates training time and introduces model artifacts if some features should have been dropped. We present WinnowML, a method for automatically determining the most relevant feature subset for a predictive time-series model. WinnowML combines existing feature ranking algorithms and a history of each feature's ranking to iteratively rank a feature set to lower prediction error and maximize long term relevance. From this ranked feature set, the most relevant and stable subset is selected to train a model. Experimentally, we show how WinnowML can lower a model's mean absolute relative error up to 42% on average compared to the closest performing approach. Additionally, we lower the fluctuation in feature ranking and selection up to 65%. We also demonstrate how to combine WinnowML and a model search tool to provide improvements in performance of up to 14.5% when compared to using all the feature available.

Bel, Oceane MS↗
DOE OSTI2024

Understanding GPU Memory Corruption at Extreme Scale: The Summit Case Study

GPU memory corruption and in particular double-bit errors (DBEs) remain one of the least understood aspects of HPC system reliability. Albeit rare, their occurrences always lead to job termination and can potentially cost thousands of node-hours, either from wasted computations or as the overhead from regular checkpointing needed to minimize the losses. As supercomputers and their components simultaneously grow in scale, density, failure rates, and environmental footprint, the efficiency of HPC operations becomes both an imperative and a challenge. We examine DBEs using system telemetry data and logs collected from the Summit supercomputer, equipped with 27,648 Tesla V100 GPUs with 2nd-generation high-bandwidth memory (HBM2). Using exploratory data analysis and statistical learning, we extract several insights about memory reliability in such GPUs. We find that GPUs with prior DBE occurrences are prone to experience them again due to otherwise harmless factors, correlate this phenomenon with GPU placement, and suggest manufacturing variability as a factor. On the general population of GPUs, we link DBEs to short- and long-term high power consumption modes while finding no significant correlation with higher temperatures. We also show that the workload type can be a factor in memory’s propensity to corruption.

Oles, Vlad↗
DOE OSTI2021

A daily, 250 m and real-time gross primary productivity product (2000–present) covering the contiguous United States

Abstract. Gross primary productivity (GPP) quantifies the amount of carbon dioxide (CO2) fixed by plants through photosynthesis. Although as a key quantity of terrestrial ecosystems, there is a lack of high-spatial-and-temporal-resolution, real-time and observation-based GPP products. To address this critical gap, here we leverage a state-of-the-art vegetation index, near-infrared reflectance of vegetation (NIRV), along with accurate photosynthetically active radiation (PAR), to produce a SatelLite Only Photosynthesis Estimation (SLOPE) GPP product for the contiguous United States (CONUS). Compared to existing GPP products, the proposed SLOPE product is advanced in its spatial resolution (250 m versus >500 m), temporal resolution (daily versus 8 d), instantaneity (latency of 1 d versus >2 weeks) and quantitative uncertainty (on a per-pixel and daily basis versus no uncertainty information available). These characteristics are achieved because of several technical innovations employed in this study: (1) SLOPE couples machine learning models with MODIS atmosphere and land products to accurately estimate PAR. (2) SLOPE couples highly efficient and pragmatic gap-filling and filtering algorithms with surface reflectance acquired by both Terra and Aqua MODIS satellites to derive a soil-adjusted NIRV (SANIRV) dataset. (3) SLOPE couples a temporal pattern recognition approach with a long-term Cropland Data Layer (CDL) product to predict dynamic C4 crop fraction. Through developing a parsimonious model with only two slope parameters, the proposed SLOPE product explains 85 % of the spatial and temporal variations in GPP acquired from 49 AmeriFlux eddy-covariance sites (324 site years), with a root-mean-square error (RMSE) of 1.63 gC m−2 d−1. The median R2 over C3 and C4 crop sites reaches 0.87 and 0.94, respectively, indicating great potentials for monitoring crops, in particular bioenergy crops, at the field level. With such a satisfactory performance and its distinct characteristics in spatiotemporal resolution and instantaneity, the proposed SLOPE GPP product is promising for biological and environmental research, carbon cycle research, and a broad range of real-time applications at the regional scale. The archived dataset is available at https://doi.org/10.3334/ORNLDAAC/1786 (download page: https://daac.ornl.gov/daacdata/cms/SLOPE_GPP_CONUS/data/, last access: 20 January 2021) (Jiang and Guan, 2020), and the real-time dataset is available upon request.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1992

Disturbance rejection in repetitive controllers

The p-integrator is a digital controller that learns from previous experience executing a repetitive command in order to improve its performance in future repetitions. This class of repetitive controller is an analog of the integral controller of classical control theory. This paper addresses several issues related to the performance of the p-integrator. This is followed by a derivation that will demonstrate the repetitive disturbance rejection capabilities of the controller. The effect of the tracking performance in practical implementations with prefiltering and postfiltering devices are studied - using an exact and approximate formulation. Relations showing the actual system response during steady state operation with these filters are derived. The results are then extended to include the tracking error performance of systems subjected to repetitive disturbances. Finally, the system response between sample times is developed.

Solcz, Edward J.↗
DOE OSTI2021

Generalization Equations for Machine Learners Based on Physical and Abstract Laws

The physical and abstract laws derived from the first principles have been recently exploited to customize and sharpen machine learning (ML) methods and also derive their generalization equations. These laws often encapsulate knowledge that complements datasets and ML models. We present a generic framework that uses these laws to provide ML codes that are transferable across multiple areas, including data transport infrastructures and thermal hydraulics analytics of nuclear reactors. By anchoring on datasets from these areas and the statistical generalization theory, we present a rigorous approach to co-develop ML solutions and the generalization equations that characterize them, by exploiting the structure and constraints from the laws. We present illustrative examples using practical problems from existing literature on the performance characterization of data transport infrastructures, and the sensor error and power level estimation in nuclear reactor systems using sensor measurements of primary and secondary coolant systems, respectively.

Rao, Nageswara↗
NASA NTRS1990

Investigation of automated task learning, decomposition and scheduling

The details and results of research conducted in the application of neural networks to task planning and decomposition are presented. Task planning and decomposition are operations that humans perform in a reasonably efficient manner. Without the use of good heuristics and usually much human interaction, automatic planners and decomposers generally do not perform well due to the intractable nature of the problems under consideration. The human-like performance of neural networks has shown promise for generating acceptable solutions to intractable problems such as planning and decomposition. This was the primary reasoning behind attempting the study. The basis for the work is the use of state machines to model tasks. State machine models provide a useful means for examining the structure of tasks since many formal techniques have been developed for their analysis and synthesis. It is the approach to integrate the strong algebraic foundations of state machines with the heretofore trial-and-error approach to neural network synthesis.

Livingston, David L.↗
DOE OSTI2025

Systematic softening in universal machine learning interatomic potentials

Machine learning interatomic potentials (MLIPs) have introduced a new paradigm for atomic simulations. Recent advancements have led to universal MLIPs (uMLIPs) that are pre-trained on diverse datasets, providing opportunities for universal force fields and foundational machine learning models. However, their performance in extrapolating to out-of-distribution complex atomic environments remains unclear. In this study, we highlight a consistent potential energy surface (PES) softening effect in three uMLIPs: M3GNet, CHGNet, and MACE-MP-0, which is characterized by energy and force underprediction in atomic-modeling benchmarks including surfaces, defects, solid-solution energetics, ion migration barriers, phonon vibration modes, and general high-energy states. The PES softening behavior originates primarily from the systematically underpredicted PES curvature, which derives from the biased sampling of near-equilibrium atomic arrangements in uMLIP pre-training datasets. Our findings suggest that a considerable fraction of uMLIP errors are highly systematic, and can therefore be efficiently corrected. We argue for the importance of a comprehensive materials dataset with improved PES sampling for next-generation foundational MLIPs.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Rapid quantitative analysis of trace elements in plutonium alloys using a handheld laser-induced breakdown spectroscopy (LIBS) device coupled with chemometrics and machine learning

Here, we present the first reported quantification of trace elements in plutonium via a portable laser-induced breakdown spectroscopy (LIBS) device and demonstrate the use of chemometric analysis to enhance the handheld device's sensitivity and precision. Quantification of trace elements such as iron and nickel in plutonium metal via LIBS is a challenging problem due to the complex nature of the plutonium optical emission spectra. While rapid analysis of plutonium alloys has been demonstrated using portable LIBS devices, such as the SciAps Z300, their detection limits for trace elements are severely constrained by their achievable pulse power and length, light collection optics, and detectors. In this paper, analytical methods are evaluated as a means to circumvent the detection constraints. Three chemometric methods often used in analytical spectroscopy are evaluated; principal component regression, partial least-squares regression, and artificial neural networks. These models are evaluated based on goodness-of-fit metrics, root mean-squared error, and their achievable limits of detection (LoDs). Partial least squares proved superior for determining content of iron and nickel in plutonium metal, yielding LoDs of 15 and 20 ppm, respectively. These results of identifying the undesirable trace elements in plutonium components are critical for applications such as fabricating radioisotope thermoelectric generators or nuclear fuel.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2025

Variational AutoEncoders Reveal Intensifying GPP Extremes in Continental United States based on CESM2 Simulations

Climate extremes significantly impact terrestrial carbon cycle dynamics, necessitating robust methods for detecting and analyzing anomalous behavior in plant productivity. This study presents a novel application of variational autoencoders (VAE) for identifying extreme events in gross primary productivity (GPP) from Community Earth System Model version 2 simulations across four AR6 regions in the Continental United States. We compare VAE-based anomaly detection with traditional singular spectral analysis (SSA) methods across three time periods: 1850-80, 1950-80, and 2050-80 under SSP5-8.5 scenario. The VAE architecture employs three dense layers and a latent space with input sequence length of 12 months, training on normalized GPP time series to reconstruct the GPP and identify anomalies based on reconstruction errors. Extreme events are defined using 5th percentile thresholds applied to both VAE and SSA anomalies. Results demonstrate strong regional agreement between VAE and SSA methods in spatial patterns of extreme event frequencies, despite VAE consistently producing higher threshold values (179-756 GgC for VAE vs. 100-784 GgC for SSA across regions and periods). Both methods reveal increasing magnitudes and frequencies of negative carbon cycle extremes toward 2050-80, particularly in Western and Central North America. The VAE approach shows comparable performance to established SSA techniques while offering computational advantages and enhanced capability for capturing non-linear temporal dependencies in carbon cycle variability. This research demonstrates the potential of deep learning approaches for extremes detection and provides a foundation for improved understanding of future carbon cycle risks under future conditions.

Sharma, Bharat [ORNL] (ORCID:0000000266982487)↗
DOE OSTI2025

Predicting Dynamic-to-Static Correction Factor from Petrophysical Data and Chemostratigraphy using Unsupervised Machine Learning

Estimating static mechanical properties of stratigraphic layers is critical for optimizing subsurface engineering applications. To estimate dynamic-to-static correction factor F ds (static-to-dynamic Young’s modulus ratio) across the Caney shale interval in Oklahoma, USA, we integrated triaxial test measurements and petrophysical data, including well logs and X-ray fluorescence (XRF) using unsupervised machine learning (ML). We used a novel workflow that includes principal component analysis (PCA) to reduce data set dimensionality of well logs and XRF data sets—both separately and combined—creating three scenarios, and later applied inverse distance weighting (IDW) to derive F ds profiles for these scenarios. Furthermore, we applied K-means clustering on each scenario to predict depositional facies, and built a stiffness zonation profile through chemostratigraphic analysis of the terrigenous elements to validate the predicted F ds . The predicted F ds profile from each scenario using the PCA-IDW method was compared with the constant F ds approach from our previous study by calculating the root mean square error (RMSE). The combined data sets scenario yielded the lowest RMSE value of 0.113, while the RMSE values for the well logs and XRF scenarios were 0.131 and 0.129, respectively. In addition, the predicted F ds from the XRF scenario well-matched the stiffness zonation from the chemostratigraphic analysis that was built using the optimized K-means clustering of nine clusters for that scenario. These methods and findings offer a valuable tool for refining lithological classification and improving the F ds profile, potentially enhancing drilling and stimulation strategies for subsurface energy engineering applications.

clastic rock↗
DOE OSTI2024

Machine learning approaches for integrating multi-omics data to expand microbiome annotation

Preliminary: This final report corresponds to a grant (DE-SC0021216) that was awarded to the University of Montana. Mid-way through the grant period, I relocated from the University of Montana to the University of Arizona. The grant was ended at University of Montana in late 2022, with all efforts concluding on 08/26/22; the remaining funds supporting the project were relinquished by University of Montana, and were later awarded to University of Arizona under a new grant, with start date 04/01/23. This report focuses on results of research efforts at UMontana through 08/26/22. Results: We made progress in each of the three aims of the proposal. We released software that identifies and fills gaps in the annotation of metabolic proteins within bacterial genomes. We made substantial progress in developing software for alignment-based annotation of protein coding DNA, allowing for coding frameshifts caused by sequencing error. Finally, we made notable progress in developing AI methods (specifically: a neural embedding model) for identifying similarities between protein sequences based on amino-wise latent vectors. These efforts were supplemented by development of methods for protein modeling in support of predicting protein-drug binding activity, and by my leadership of a team in the NIH/DOE 2021 Petabyte-Scale Sequence Search hack-a-thon.

59 BASIC BIOLOGICAL SCIENCES↗
NASA NTRS1989

Prototype part task trainer: A remote manipulator system simulator

The Part Task Trainer program (PTT) is a kinematic simulation of the Remote Manipulator System (RMS) for the orbiter. The purpose of the PTT is to supply a low cost man-in-the-loop simulator, allowing the student to learn operational procedures which then can be used in the more expensive full scale simulators. PTT will allow the crew members to work on their arm operation skills without the need for other people running the simulation. The controlling algorithms for the arm were coded out of the Functional Subsystem Requirements Document to ensure realistic operation of the simulation. Relying on the hardware of the workstation to provide fast refresh rates for full shaded images allows the simulation to be run on small low cost stand alone work stations, removing the need to be tied into a multi-million dollar computer for the simulation. PTT will allow the student to make errors which in full scale mock up simulators might cause failures or damage hardware. On the screen the user is shown a graphical representation of the RMS control panel in the aft cockpit of the orbiter, along with a main view window and up to six trunion and guide windows. The dials drawn on the panel may be turned to select the desired mode of operation. The inputs controlling the arm are read from a chair with a Translational Hand Controller (THC) and a Rotational Hand Controller (RHC) attached to it.

Shores, David↗
NASA NTRS2014

Using a Support Vector Machine and a Land Surface Model to Estimate Large-Scale Passive Microwave Temperatures over Snow-Covered Land in North America

A support vector machine (SVM), a machine learning technique developed from statistical learning theory, is employed for the purpose of estimating passive microwave (PMW) brightness temperatures over snow-covered land in North America as observed by the Advanced Microwave Scanning Radiometer (AMSR-E) satellite sensor. The capability of the trained SVM is compared relative to the artificial neural network (ANN) estimates originally presented in [14]. The results suggest the SVM outperforms the ANN at 10.65 GHz, 18.7 GHz, and 36.5 GHz for both vertically and horizontally-polarized PMW radiation. When compared against daily AMSR-E measurements not used during the training procedure and subsequently averaged across the North American domain over the 9-year study period, the root mean squared error in the SVM output is 8 K or less while the anomaly correlation coefficient is 0.7 or greater. When compared relative to the results from the ANN at any of the six frequency and polarization combinations tested, the root mean squared error was reduced by more than 18 percent while the anomaly correlation coefficient was increased by more than 52 percent. Further, the temporal and spatial variability in the modeled brightness temperatures via the SVM more closely agrees with that found in the original AMSR-E measurements. These findings suggest the SVM is a superior alternative to the ANN for eventual use as a measurement operator within a data assimilation framework.

Modeling↗
DOE OSTI2024

Interface PINNs (I-PINNs): A physics-informed neural networks framework for interface problems

Here, we present a novel physics-informed neural networks (PINNs) framework for modeling interface problems, termed Interface PINNs (I-PINNs). I-PINNs uses different neural networks for any two subdomains separated by a sharp interface such that the neural networks differ only through their activation functions while the other parameters remain identical. The performance of I-PINNs, conventional PINNs, and other existing domain-decomposition PINNs methods such as extended PINNs (XPINNs) and multi-domain PINN (M-PINN) is compared through several one-dimensional, two-dimensional, and three-dimensional benchmark elliptic interface problems. The results demonstrate that I-PINNs provides a root-mean-square-error accuracy, at least two orders of magnitude better than conventional PINNs and XPINNs at approximately one-tenth of the computational cost of conventional PINNs and half the cost of XPINNs. Additionally, while I-PINNs and M-PINN provide comparable accuracies, M-PINN is found to be approximately 50% more expensive.

42 ENGINEERING↗
DOE OSTI2020

Variational fast forwarding for quantum simulation beyond the coherence time

Abstract Trotterization-based, iterative approaches to quantum simulation (QS) are restricted to simulation times less than the coherence time of the quantum computer (QC), which limits their utility in the near term. Here, we present a hybrid quantum-classical algorithm, called variational fast forwarding (VFF), for decreasing the quantum circuit depth of QSs. VFF seeks an approximate diagonalization of a short-time simulation to enable longer-time simulations using a constant number of gates. Our error analysis provides two results: (1) the simulation error of VFF scales at worst linearly in the fast-forwarded simulation time, and (2) our cost function’s operational meaning as an upper bound on average-case simulation error provides a natural termination condition for VFF. We implement VFF for the Hubbard, Ising, and Heisenberg models on a simulator. In addition, we implement VFF on Rigetti’s QC to demonstrate simulation beyond the coherence time. Finally, we show how to estimate energy eigenvalues using VFF.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

The PAU Survey: narrow-band photometric redshifts using Gaussian processes

Here, we study the performance of the hybrid template machine learning photometric redshift (photo- z ) algorithm delight , which uses Gaussian processes, on a subset of the early data release of the Physics of the Accelerating Universe Survey (PAUS). We calibrate the fluxes of the 40 PAUS narrow bands with six broad-band fluxes ( uBVriz ) in the Cosmic Evolution Survey (COSMOS) field using three different methods, including a new method that utilizes the correlation between the apparent size and overall flux of the galaxy. We use a rich set of empirically derived galaxy spectral templates as guides to train the Gaussian process, and we show that our results are competitive with other standard photometric redshift algorithms. delight achieves a photo- z 68th percentile error of σ 68 = 0.0081(1 + z ) without any quality cut for galaxies with i auto < 22.5 as compared to 0.0089(1 + z ) and 0.0202(1 + z ) for the bpz and annz 2 codes, respectively. delight is also shown to produce more accurate probability distribution functions for individual redshift estimates than bpz and annz 2. Common photo- z outliers of delight and bcnz 2 (previously applied to PAUS) are found to be primarily caused by outliers in the narrow-band fluxes, with a small number of cases potentially indicating spectroscopic redshift failures in the reference sample. In the process, we introduce performance metrics derived from the results of bcnz 2 and delight , allowing us to achieve a photo- z quality of σ 68 < 0.0035(1 + z ) at a magnitude of i auto < 22.5 while keeping 50 per cent objects of the galaxy sample.

79 ASTRONOMY AND ASTROPHYSICS↗