SEARCH · Engineering Papers

Results for “constrained learning”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

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

At least 379 records · Page 21

All sources NASA NTRS DOE OSTI

NASA NTRS2016

Probabilistic Modeling of Aircraft Trajectories for Dynamic Separation Volumes

With a proliferation of new and unconventional vehicles and operations expected in the future, the ab initio airspace design will require new approaches to trajectory prediction for separation assurance and other air traffic management functions. This paper presents an approach to probabilistic modeling of the trajectory of an aircraft when its intent is unknown. The approach uses a set of feature functions to constrain a maximum entropy probability distribution based on a set of observed aircraft trajectories. This model can be used to sample new aircraft trajectories to form an ensemble reflecting the variability in an aircraft's intent. The model learning process ensures that the variability in this ensemble reflects the behavior observed in the original data set. Computational examples are presented.

Lewis, Timothy A.↗

DOE OSTI2023

IP Protection in TinyML

Tiny machine learning (TinyML) is an essential component of emerging smart microcontrollers (MCUs). How- ever, the protection of the intellectual property (IP) of the model is an increasing concern due to the lack of desktop/server-grade resources on these power-constrained devices. In this paper, we propose STML, a system and algorithm co-design to Secure IP of TinyML on MCUs with ARM TrustZone. Our design jointly optimizes memory utilization and latency while ensuring the security and accuracy of emerging models. We implemented a prototype and benchmarked with 7 models, demonstrating STML reduces 40% of model protection runtime overhead on average.

42 ENGINEERING↗

DOE OSTI2025

Considerations for Distributed Edge Data Centers and Use of Building Loads to Support Large Interconnections

The rapid expansion of artificial intelligence (AI) and machine learning is driving unprecedented electricity demand from data centers. It is predicted that by 2030, 90% of AI workloads will be inference-based, requiring interconnection of multiple low-latency edge data centers (<20 MW) sited closer to end users - often on already constrained distribution feeders. Although individually small, these loads can aggregate to large loads per feeder, straining infrastructure, creating multi-year interconnection delays, and driving up customer costs. This paper proposes a data center-focused grid-integration framework that combines feeder hosting capacity analysis with building energy efficiency, building load flexibility, and waste heat reuse to expand effective feeder and substation headroom. Such approaches can reduce interconnection delays, lower costs for ratepayers, and accelerate AI-ready infrastructure deployment.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2023

Application of unsupervised deep learning to image segmentation and in-situ contact angle measurements in a CO 2 -water-rock system

Rock surface wettability is a critical property that regulates multiphase flows in porous media, which can be quantified using the surface contact angle (CA). X-ray micro-computed tomography (μCT) provides an effective approach to in-situ measurements of surface CAs. However, the CA measurement accuracy depends significantly on the quality of CT image segmentation, which is the clustering of CT pixels into separate phases. Inspired by this, we developed a deep learning (DL)-based CA measurement workflow. Motivated by the recent tremendous progress in unsupervised learning techniques and aiming to avoid expensive manual data annotations, an unsupervised DL pipeline for CT image segmentation was proposed and implemented, which includes unsupervised model training and post-processing. The unsupervised model training was driven by a novel loss function constrained with feature similarity and spatial continuity and implemented by iterative forward and backward paths; the former clustered the pixel-wise feature vectors extracted by convolution neural networks, whereas the latter updated the parameters using gradient descent. An over-segmentation strategy was adopted for model training. The post-processing steps based on agglomerative hierarchical clustering (AHC) were implemented to further merge the over-segmented model output to the desired cluster number, which is intended to improve the efficiency of image segmentation. The developed unsupervised DL pipeline was compared with other commonly-used image segmentation methods using pixel-wise and physics-based evaluation metrics on a synthetic raw-image dataset, which had a known ground truth. The unsupervised DL pipeline showed the best performance. Next, the segmented images were input to an automatic CA measurement tool, and the results were validated by comparisons with manual measurements. The CA values from the manual and automatic measurements showed similar distributions and statistical properties. The automatic measurement demonstrated a wider spectrum because of the much larger number of measurement data points. The primary novelty of the unsupervised DL pipeline developed in this study lies in the novel loss function and the over-segmentation strategy associated with AHC post-processing. Finally, the workflow has been proven an efficient tool for pore-scale wettability characterization, which has a wide range of applications in fundamental studies of multiphase flows in natural porous media, which have critical implications to geological carbon sequestration, hydrocarbon energy recovery, and contaminant transport in groundwater.

42 ENGINEERING↗

DOE OSTI2023

Feature-adjacent multi-fidelity physics-informed machine learning for partial differential equations

Physics-informed neural networks have emerged as an alternative method for solving partial differential equations. However, for complex problems, the training of such networks can still require high-fidelity data which can be expensive to generate. To reduce or even eliminate the dependency on high-fidelity data, we propose a novel multi-fidelity architecture which is based on a feature space shared by the low- and high-fidelity solutions. In the feature space, the projections of the low-fidelity and high-fidelity solutions are adjacent by constraining their relative distance. The feature space is represented with an encoder and its mapping to the original solution space is effected through a decoder. In conclusion, the proposed multi-fidelity approach is validated on forward and inverse problems for steady and unsteady problems described by partial differential equations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

A high-throughput approach for statistical process optimization in Laser Powder Bed Fusion

Process variability is inherent in metal additive manufacturing (AM). However, it is often overlooked in process optimization frameworks, constraining the understanding of process uncertainties and their influence on parameter selection. To address this, we present an integrated framework that combines high-throughput single-track experiments, GAN-based melt pool geometry extraction, robust statistical and machine learning modeling, and uncertainty-quantified process mapping. Process variability is characterized through single-track melt pool behaviors, and its influence on defect formation is systematically quantified to enable statistically guided process parameter optimization. This approach is demonstrated on Laser Powder Bed Fusion (L-PBF) of stainless steel 316L, effectively capturing the interplay between process parameters, melt pool variability, and defect probability. By integrating uncertainty quantification into process optimization, this study provides a structured methodology for addressing variability challenges in AM quality control, ultimately contributing to enhanced manufacturing reliability.

Laser Powder Bed Fusion↗

DOE OSTI2020

Incorporating Electronic Information into Machine Learning Potential Energy Surfaces via Approaching the Ground-State Electronic Energy as a Function of Atom-Based Electronic Populations

Machine learning (ML) approximations to density functional theory (DFT) potential energy surfaces (PESs) are showing great promise for reducing the computational cost of accurate molecular simulations, but at present, they are not applicable to varying electronic states, and in particular, they are not well suited for molecular systems in which the local electronic structure is sensitive to the medium to long-range electronic environment. With this issue as the focal point, we present a new machine learning approach called "BpopNN" for obtaining efficient approximations to DFT PESs. Conceptually, the methodology is based on approaching the true DFT energy as a function of electron populations on atoms; in practice, this is realized with available density functionals and constrained DFT (CDFT). Additionally, the new approach creates approximations to this function with neural networks. These approximations thereby incorporate electronic information naturally into a ML approach, and optimizing the model energy with respect to populations allows the electronic terms to self-consistently adapt to the environment, as in DFT. We confirm the effectiveness of this approach with a variety of calculations on LinHn clusters.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2025

Active oversight and quality control in standard Bayesian optimization for autonomous experiments

The fusion of experimental automation and machine learning has catalyzed a new era in materials research, prominently featuring Gaussian Process (GP) Bayesian Optimization (BO) driven autonomous experiments. Here we introduce a Dual-GP approach that enhances traditional GPBO by adding a secondary surrogate model to dynamically constrain the experimental space based on real-time assessments of the raw experimental data. This Dual-GP approach enhances the optimization efficiency of traditional GPBO by isolating more promising space for BO sampling and more valuable experimental data for primary GP training. We also incorporate a flexible, human-in-the-loop intervention method in the Dual-GP workflow to adjust for unanticipated results. We demonstrate the effectiveness of the Dual-GP model with synthetic model data and implement this approach in autonomous pulsed laser deposition experimental data. This Dual-GP approach has broad applicability in diverse GPBO-driven experimental settings, providing a more adaptable and precise framework for refining autonomous experimentation for more efficient optimization.

36 MATERIALS SCIENCE↗

DOE OSTI2026

Electron Neutrino Selection for Neutrino Oscillation Searches in the SBN Program

The Short- Baseline Neutrino (SBN) Pprogram at Fermilab employs a near detector, SBND, and a far detector, ICARUS, to search for baseline-dependent changes to the measure muon neutrino disappearance and electron neutrino appearance and disappearance from the muon neutrino-dominated Booster Neutrino Beam flux. The use of the same liquid argon time projection chamber (LArTPC) technology in both detectors and the same neutrino beam enables a systematically well-constrained oscillation measurement, by leveraging the high statistics of O(10 million) neutrino interactions expected at SBND through 2027 to constrain flux and cross-section systematics. In this poster, we present high-statistics, high-purity electron neutrino selections in SBND and ICARUS and related electron shower energy calibration measurements in SBND enabled by the SPINE deep learning-based particle reconstruction framework. Furthermore, we discuss how these samples can enable high sensitivity to '3+1' sterile neutrino oscillations in a combined fit accounting for systematic uncertainties and correlations.

Oza, Nupur [Nevis Labs, Columbia U.; Columbia U.]↗

DOE OSTI2021

Physics-constrained, low-dimensional models for magnetohydrodynamics: First-principles and data-driven approaches

Plasmas are highly nonlinear and multiscale, motivating a hierarchy of models to understand and describe their behavior. However, there is a scarcity of plasma models of lower fidelity than magnetohydrodynamics (MHD), although these reduced models hold promise for understanding key physical mechanisms, efficient computation, and real-time optimization and control. Galerkin models, obtained by projection of the MHD equations onto a truncated modal basis, and data-driven models, obtained by modern machine learning and system identification, can furnish this gap in the lower levels of the model hierarchy. This work develops a reduced-order modeling framework for compressible plasmas, leveraging decades of progress in projection-based and data-driven modeling of fluids. We begin by formalizing projection-based model reduction for nonlinear MHD systems. To avoid separate modal decompositions for the magnetic, velocity, and pressure fields, we introduce an energy inner product to synthesize all of the fields into a dimensionally consistent, reduced-order basis. Next, we obtain an analytic model by Galerkin projection of the Hall-MHD equations onto these modes. We illustrate how global conservation laws constrain the model parameters, revealing symmetries that can be enforced in data-driven models, directly connecting these models to the underlying physics. We demonstrate the effectiveness of this approach on data from high-fidelity numerical simulations of a three-dimensional spheromak experiment. Finally, this manuscript builds a bridge to the extensive Galerkin literature in fluid mechanics and facilitates future principled development of projection-based and data-driven models for plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

NASA NTRS2012

What We Have Learned About Clusters From a Decade of Arcsecond Resolution X-ray Observations

This talk will briefly review the main findings from Chandra high angular resolution observations of galaxy clusters, emphasizing results on cluster astrophysics. Chandra has discovered shock fronts in merging systems, providing information on the shock Mach number and velocity, and for best-observed shocks, constraining the microphysical properties of the intracluster medium (ICM). Cold fronts, a Chandra discovery, are ubiquitous both in merging clusters and in the cool ccres of relaxed systems. They reveal the structure and strength of the intracluster magnetic fields and constrain the ICM viscosity a combined with radio data, these observations also shed light on the production of ultra-relativistic particles that are known to coexist with thermal plasma. Finally, in nearly all cool cores, Chandra observes cavities in the ICM that are produced by the central AGN. All these phenomena will be extremely interesting for high-resolution SZ studies.

Markevitch, Maxim↗

DOE OSTI2023

Knowledge Distillation for Anomaly Detection

Unsupervised deep learning techniques are widely used to identify anomalous behaviour. The performance of such methods is a product of the amount of training data and the model size. However, the size is often a limiting factor for the deployment on resource-constrained devices. Here, we present a novel procedure based on knowledge distillation for compressing an unsupervised anomaly detection model into a supervised deployable one and we suggest a set of techniques to improve the detection sensitivity. Compressed models perform comparably to their larger counterparts while significantly reducing the size and memory footprint.

Pol, Adrian Alan↗

DOE OSTI2024

What Can We Learn about the Unstable Equation-of-state Branch from Neutron Star Mergers?

Abstract The equation of state (EOS) of dense strongly interacting matter can be probed by astrophysical observations of neutron stars (NS), such as X-ray detections of pulsars or the measurement of the tidal deformability of NSs during the inspiral stage of NS mergers. These observations constrain the EOS at most up to the density of the maximum-mass configuration, n TOV , which is the highest density that can be explored by stable NSs for a given EOS. However, under the right circumstances, binary neutron star (BNS) mergers can create a postmerger remnant that explores densities above n TOV . In this work, we explore whether the EOS above n TOV can be measured from gravitational-wave or electromagnetic observations of the postmerger remnant. We perform a total of 25 numerical-relativity simulations of BNS mergers for a range of EOSs and find no case in which different descriptions of the matter above n TOV have a detectable impact on postmerger observables. Hence, we conclude that the EOS above n TOV can likely not be probed through BNS merger observations for the current and next generation of detectors.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2025

ORBIT-2: Scaling Exascale Vision Foundation Models for Weather and Climate Downscaling

Sparse observations and coarse-resolution climate models limit effective regional decision-making, underscoring the need for robust downscaling. However, existing AI methods struggle with generalization across variables and geographies and are constrained by the quadratic complexity of Vision Transformer (ViT) self-attention. We introduce ORBIT-2, a scalable foundation model for global, hyper-resolution climate downscaling. ORBIT-2 incorporates two key innovations: (1) Residual Slim ViT (Reslim), a lightweight architecture with residual learning and Bayesian regularization for efficient, robust prediction; and (2) TILES, a tile-wise sequence scaling algorithm that reduces self-attention complexity from quadratic to linear, enabling long-sequence processing and massive parallelism. ORBIT-2 scales to 10 billion parameters across 65,536 GPUs, achieving up to 4.1 ExaFLOPS sustained throughput and 74–98% strong scaling efficiency. It supports downscaling to 0.9 km global resolution and processes sequences up to 4.2 billion tokens. On 7 km resolution benchmarks, ORBIT-2 achieves high accuracy with R2 scores in range of 0.98–0.99 against observation data.

Wang, Xiao [ORNL] (ORCID:0000000165451943)↗

DOE OSTI2025

Data-Driven Kinetic Reaction Networks for Separation Chemistry

Understanding complex, multistep chemical reactions at the molecular level is a major challenge whose solution would greatly benefit the design and optimization of numerous chemical processes. The separation of rare-earth (4f) and actinide (5f) elements is an example where improving our chemical understanding is important for designing and optimizing new chemistries, even with a limited number of observations. Here, in this work, we leverage data-driven artificial intelligence and machine-learning approaches to develop kinetic reaction networks that describe the liquid–liquid extraction mechanism of uranium using N,N-di-2-ethylhexyl-isobutyramide (DEHiBA). Specifically, we compare and contrast the properties of two classes of models: (1) purely data-driven models that are regularized using chemistry-agnostic, L1 regression and (2) chemistry-informed models that are regularized using relative reaction energies provided by quantum mechanical calculations. We observe that purely data-driven models are unbiased, simple, and accurate in their predictions of experimental measurements when provided with sufficient data but are difficult to fully constrain and interpret. In contrast, chemistry-informed models exhibit significantly improved chemical interpretability and consistency, providing a detailed description of the separation process while achieving high accuracy through ensemble averaging. Overall, the dominant species predicted to be extracted into the organic phase is UO 2 (NO 3 ) 2 (DEHiBA) 2 , agreeing with experimental slope analysis, thermodynamic modeling, EXAFS, and crystal structures. This work demonstrates that leveraging the fundamental structure of the problem can lead to efficient learning schemes that provide both accurate predictions and chemical insights at a low computational cost.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2020

Approximate Inverse Chain Preconditioner: Iteration Count Case Study for Spectral Support Solvers

As the growing availability of computational power slows, there has been an increasing reliance on algorithmic advances. However, faster algorithms alone will not necessarily bridge the gap in allowing computational scientists to study problems at the edge of scientific discovery in the next several decades. Often, it is necessary to simplify or precondition solvers to accelerate the study of large systems of linear equations commonly seen in a number of scientific fields. Preconditioning a problem to increase efficiency is often seen as the best approach; yet, preconditioners which are fast, smart, and efficient do not always exist. Following the progress of [1], we present a new preconditioner for symmetric diagonally dominant (SDD) systems of linear equations. These systems are common in certain PDEs, network science, and supervised learning among others. Based on spectral support graph theory, this new preconditioner builds off of the work of [2], computing and applying a V-cycle chain of approximate inverse matrices. This preconditioner approach is both algebraic in nature as well as hierarchically-constrained depending on the condition number of the system to be solved. Due to its generation of an Approximate Inverse Chain of matrices, we refer to this as the AIC preconditioner. We further accelerate the AIC preconditioner by utilizing precomputations to simplify setup and multiplications in the con-text of an iterative Krylov-subspace solver. While these iterative solvers can greatly reduce solution time, the number of iterations can grow large quickly in the absence of good preconditioners. Initial results for the AIC preconditioner have shown a very large reduction in iteration counts for SDD systems as compared to standard preconditioners such as Incomplete Cholesky (ICC) and Multigrid (MG). We further show significant reduction in iteration counts against the more advanced Combinatorial Multigrid (CMG) preconditioner. We have further developed no-fill sparsification techniques to ensure that the computational cost of applying the AIC preconditioner does not grow prohibitively large as the depth of the V-cycle grows for systems with larger condition numbers. Our numerical results have shown that these sparsifiers maintain the sparsity structure of our system while also displaying significant reductions in iteration counts.1 2

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Approximate Inverse Chain Preconditioner: Iteration Count Case Study for Spectral Support Solvers

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Exploratory analysis of machine learning techniques in the Nevada geothermal play fairway analysis

Play fairway analysis (PFA) is commonly used to generate geothermal potential maps and guide exploration studies, with a particular focus on locating and characterizing blind geothermal systems. This study evaluates the application of machine learning techniques to PFA in the Great Basin region of Nevada. Following the evaluation of various techniques, we identified two approaches to PFA that produced promising results, 1) supervised Bayesian probabilistic neural networks to generate geothermal potential maps with confidence intervals, and 2) unsupervised principal component analysis paired with k-means clustering to generate both cluster maps to help identify spatial patterns, as well as new combined feature inputs. We applied these techniques to perform a comparative analysis between two principal sets of geological and geophysical features related to permeability and heat and a set of positive (known geothermal resources) and negative training sites (known drill sites with unsuitable geothermal conditions). We found that these methods constrain previously unrecognized feature controls on geothermal favorability, many of which are spatially organized within the extent of cluster groups and the major structural-hydrologic domains of the study area. Furthermore, we utilized exploratory unsupervised modeling to highlight spatial relationships between input data and predictive output results of our supervised modeling. As a result, we demonstrate how our models compare to the previous Nevada PFA and how the rapid insights these machine learning techniques offer may support future assessments of both known and undiscovered blind geothermal systems in the Great Basin region of Nevada and beyond.

15 GEOTHERMAL ENERGY↗