Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Distributed training”

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 397 records · Page 22
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Extending Conformal Prediction to Hidden Markov Models with Exact Validity via de Finetti’s Theorem for Markov Chains

Conformal prediction is a widely used method to quantify uncertainty in settings where the data is independent and identically distributed (IID), or more generally, exchangeable. Conformal prediction takes in a pre-trained classifier and a calibration dataset as inputs, and returns a function which maps feature vectors to subsets of classes. The output of the returned function for a new feature vector is guaranteed to contain the true class with a pre-specified confidence. Despite its success and usefulness in IID settings, extending conformal prediction to non-exchangeable (e.g., Markovian) data in a manner that provably preserves all desirable theoretical properties has largely remained an open problem. As a solution, we extend conformal prediction to the setting of a Hidden Markov Model (HMM) with unknown parameters. The key idea behind the proposed method is to partition the non-exchangeable Markovian data from the HMM into exchangeable blocks by exploiting the de Finetti’s Theorem for Markov Chains discovered by Diaconis and Freedman (1980). The permutations of the exchangeable blocks are then viewed as randomizations of the observed Markovian data from the HMM. The proposed method provably retains all desirable theoretical guarantees offered by the classical conformal prediction framework and is general enough to be useful in many sequential prediction problems.

Nettasinghe, Don Buddhika Wijayantha↗
DOE OSTI2024

Deep learning-based temporal deconvolution for photon time-of-flight distribution retrieval

The acquisition of the time of flight (ToF) of photons has found numerous applications in the biomedical field. Over the last decades, a few strategies have been proposed to deconvolve the temporal instrument response function (IRF) that distorts the experimental time-resolved data. However, these methods require burdensome computational strategies and regularization terms to mitigate noise contributions. Herein, we propose a deep learning model specifically to perform the deconvolution task in fluorescence lifetime imaging (FLI). The model is trained and validated with representative simulated FLI data with the goal of retrieving the true photon ToF distribution. Its performance and robustness are validated with well-controlled in vitro experiments using three time-resolved imaging modalities with markedly different temporal IRFs. The model aptitude is further established with in vivo preclinical investigation. Overall, these in vitro and in vivo validations demonstrate the flexibility and accuracy of deep learning model-based deconvolution in time-resolved FLI and diffuse optical imaging.

Pandey, Vikas (ORCID:0000000154771095)↗
DOE OSTI2026

D$e^+e^-$ffusion: Capturing the Beam-Beam Physics of $e^+e^-$ Collisions with Diffusion Models

Beam-induced backgrounds at high-luminosity $e^+e^-$ colliders, such as the FCC-ee, are dominated by incoherent pair creation (IPC), and require computationally expensive simulations with dedicated Monte Carlo (MC) event generators. Reliable detector and machine-detector interface studies necessitate event samples that are several orders of magnitude larger than what is practically attainable with existing MC. To alleviate this bottleneck, we present D$e^+e^-$ffusion, a denoising diffusion probabilistic model that operates as a permutation-equivariant, set-valued surrogate for fast IPC simulation. Trained on a small GuineaPig++ sample, D$e^+e^-$ffusion faithfully reproduces the marginal and joint kinematic, angular, and positional distributions of all three IPC production processes. In addition, we assess the fidelity at the detector level by propagating both Geant4 and D$e^+e^-$ffusion events through a Geant4 simulation of the CLD vertex detector and by training a transformer-based two-sample classifier; the classifier achieves an area under the ROC curve of $0.553 \pm 0.016$. The trained model generates events nearly four orders of magnitude faster than Geant4, paving the way for a fast-simulation surrogate for FCC-ee design studies.

Chahine, Antonio [Imperial Coll., London]↗
DOE OSTI2024

Projective Integral Updates for High-Dimensional Variational Inference

Variational inference is an approximation framework for Bayesian inference that seeks to improve quantified uncertainty in predictions by optimizing a simplified distribution over parameters to stand in for the full posterior. Capturing model variations that remain consistent with training data enables more robust predictions by reducing parameter sensitivity. This work introduces a fixed-point optimization for variational inference that is applicable when every feasible log density can be expressed as a linear combination of functions from a given basis. In such cases, the optimizer becomes a fixed-point of projective integral updates. When the basis spans univariate quadratics in each parameter, the feasible distributions are Gaussian mean-fields and the projective integral updates yield quasi-Newton variational Bayes (QNVB). Other bases and updates are also possible. Since these updates require high-dimensional integration, this work begins by proposing an efficient quasirandom sequence of quadratures for mean-field distributions. Each iterate of the sequence contains two evaluation points that combine to correctly integrate all univariate quadratic functions and, if the mean-field factors are symmetric, all univariate cubics. More importantly, averaging results over short subsequences achieves periodic exactness on a much larger space of multivariate polynomials of quadratic total degree. The corresponding variational updates require four loss evaluations with standard (not second-order) backpropagation to eliminate error terms from over half of all multivariate quadratic basis functions. Furthermore, this integration technique is motivated by first proposing stochastic blocked mean-field quadratures, which may be useful in other contexts. A PyTorch implementation of QNVB allows for better control over model uncertainty during training than competing methods. Experiments demonstrate superior generalizability for multiple learning problems and architectures.

Gaussian mean-field↗
DOE OSTI2023

Combining synchrotron X-ray diffraction, mechanistic modeling and machine learning for in situ subsurface temperature quantification during laser melting

Laser melting, such as that encountered during additive manufacturing, produces extreme gradients of temperature in both space and time, which in turn influence microstructural development in the material. Qualification and model validation of the process itself and the resulting material necessitate the ability to characterize these temperature fields. However, well established means to directly probe the material temperature below the surface of an alloy while it is being processed are limited. To address this gap in characterization capabilities, a novel means is presented to extract subsurface temperature-distribution metrics, with uncertainty, from in situ synchrotron X-ray diffraction measurements to provide quantitative temperature evolution data during laser melting. Temperature-distribution metrics are determined using Gaussian process regression supervised machine-learning surrogate models trained with a combination of mechanistic modeling (heat transfer and fluid flow) and X-ray diffraction simulation. The trained surrogate model uncertainties are found to range from 5 to 15% depending on the metric and current temperature. The surrogate models are then applied to experimental data to extract temperature metrics from an Inconel 625 nickel superalloy wall specimen during laser melting. The maximum temperatures of the solid phase in the diffraction volume through melting and cooling are found to reach the solidus temperature as expected, with the mean and minimum temperatures found to be several hundred degrees less. The extracted temperature metrics near melting are determined to be more accurate because of the lower relative levels of mechanical elastic strains. However, uncertainties for temperature metrics during cooling are increased due to the effects of thermomechanical stress.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Blockchain Enabled Intelligence of Federated Systems (BELIEFS): An attack-tolerant trustable distributed intelligence paradigm

In this article, a Blockchain Enabled Intelligence of Federated Systems (BELIEFS) is proposed to conduct cooperative control for the multi-regional large-scale power system with a multi-agents system (MAS). By establishing a two levels blockchain, each regional AI agent can simultaneously manage intra-regional controllers and cooperate with other AI agents. Under the consensus mechanism, the agents, which respectively conducted distributed deep reinforcement learning (DDRL) algorithm in multi-regions, can have the tolerant capability of malicious attacks in their training process. The demonstration of the proposed approach is within a multi-regional large-scale interconnected power system. Under the mode of "centralized dispatching and hierarchical management", this article aims to definite a mathematical model to deal with the control problem of the power systems. With the comparison experiments, the effectiveness and efficiency of our proposed method in the training process are verified. In addition, malicious attacks are set on the main chain and shard chains to verify the attack-tolerant capability. We expect that such approach and results can suggest a new paradigm of attack-tolerant trustable distributed AI deployment.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Embedded, Real-Time, and Distributed Traveling Wave Fault Location Method Using Graph Convolutional Neural Networks

This work proposes and develops an implementation of a fault location method to provide a fast and resilient protection scheme for power distribution systems. The method analyzes the transient dynamics of traveling waves (TWs) to generate features using the discrete wavelet transform (DWT), which are then used to train several graph convolutional network (GCN) models. Faults are simulated in the IEEE 34-node system, which is divided into three protection zones (PZs). The goal is to identify the PZ in which the fault occurs. The GCN models create a distributed protection scheme, as all nodes are able to retrieve a prediction. Given that message-passing between nodes occurs both during training and in the execution of the model, the resiliency of such schemes to communication losses was analyzed and demonstrated. One of the models, which only uses voltage measurements, was implemented on a Texas Instruments F28379D development board. The execution times were monitored to assess the speed of the protection scheme. It is shown that the proposed method can be executed in approximately a millisecond, which is comparable to existing TW protection in the transmission system. For experimental purposes, a DWT-based detection method is employed. A design of a setup to playback TWs using two development boards is also addressed.

Jiménez-Aparicio, Miguel (ORCID:000000016864461X)↗
DOE OSTI2026

Multi-task Parallelism for Robust Pre-training of Graph Foundation Models on Multi-source, Multi-fidelity Atomistic Modeling Data

Graph foundation models using graph neural networks promise sustainable, efficient atomistic modeling. To tackle challenges of processing multi-source, multi-fidelity data during pre-training, recent studies employ multi-task learning, in which shared message passing layers initially process input atomistic structures regardless of source, then route them to multiple decoding heads that predict data-specific outputs. This approach stabilizes pre-training and enhances a model’s transferability to unexplored chemical regions. Preliminary results on approximately four million structures are encouraging, yet questions remain about generalizability to larger, more diverse datasets and scalability on supercomputers. We propose a multi-task parallelism method that distributes each head across computing resources with GPU acceleration. Implemented in the open-source HydraGNN architecture, our method was trained on over 24 million structures from five datasets and tested on the Perlmutter, Aurora, and Frontier supercomputers, demonstrating efficient scaling on all three highly heterogeneous super-computing architectures.

Lupo Pasini, Massimiliano [ORNL] (ORCID:0000000249↗
DOE OSTI2024

Detecting Anomalies in Time Series Using Kernel Density Approaches

This paper introduces a novel anomaly detection approach tailored for time series data with exclusive reliance on normal events during training. Our key innovation lies in the application of kernel-density estimation (KDE) to scrutinize reconstruction errors, providing an empirically derived probability distribution for normal events post-reconstruction. This non-parametric density estimation technique offers a nuanced understanding of anomaly detection, differentiating it from prevalent threshold-based mechanisms in existing methodologies. In post-training, events are encoded, decoded, and evaluated against the estimated density, providing a comprehensive notion of normality. In addition, we propose a data augmentation strategy involving variational autoencoder-generated events and a smoothing step for enhanced model robustness. The significance of our autoencoder-based approach is evident in its capacity to learn normal representation without prior anomaly knowledge. Through the KDE step on reconstruction errors, our method addresses the versatility of anomalies, departing from assumptions tied to larger reconstruction errors for anomalous events. Our proposed likelihood measure then distinguishes normal from anomalous events, providing a concise yet comprehensive anomaly detection solution. The extensive experimental results support the feasibility of our proposed method, yielding significantly improved classification performance by nearly 10% on the UCR benchmark data.

Frehner, Robin↗
DOE OSTI2020

Preliminary results in using Deep Learning to emulate BLOB, a nuclear interaction model

Purpose: A reliable model to simulate nuclear interactions is fundamental for Ion-therapy. We already showed how BLOB (“Boltzmann-Langevin One Body”), a model developed to simulate heavy ion interactions up to few hundreds of MeV/u, could simulate also 12 C reactions in the same energy domain. However, its computation time is too long for any medical application. For this reason we present the possibility of emulating it with a Deep Learning algorithm. Methods: The BLOB final state is a Probability Density Function (PDF) of finding a nucleon in a position of the phase space. We discretised this PDF and trained a Variational Auto-Encoder (VAE) to reproduce such a discrete PDF. As a proof of concept, we developed and trained a VAE to emulate BLOB in simulating the interactions of 12 C with 12 C at 62 MeV/u. To have more control on the generation, we forced the VAE latent space to be organised with respect to the impact parameter (b) training a classifier of b jointly with the VAE. Results: In this work, the distributions obtained from the VAE are similar to the input ones and the computation time needed to use the VAE as a generator is negligible. Conclusions: We show that it is possible to use a Deep Learning approach to emulate a model developed to simulate nuclear reactions in the energy range of interest for Ion-therapy. We foresee the implementation of the generation part in C++ and to interface it with the most used Monte Carlo toolkit: Geant4.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Accelerating the Discovery of New, Single Phase High Entropy Ceramics via Active Learning

High-entropy ceramics have garnered interest due to their remarkable hardness, compressive strength, thermal stability, and fracture toughness; yet the discovery of new high-entropy ceramics (out of a tremendous number of possible elemental permutations) still largely requires costly, inefficient, trial-and-error experimental and computational approaches. The entropy forming ability (EFA) factor was recently proposed as a computational descriptor that positively correlates with the likelihood that a 5-metal high-entropy carbide (HECs) will form the desired single phase, homogeneous solid solution; however, discovery of new compositions is computationally expensive. If you consider 8 candidate metals, the HEC EFA approach uses 49 optimizations for each of the 56 unique 5-metal carbides, requiring a total of 2744 costly density functional theory calculations. Here, we describe an orders-of-magnitude more efficient active learning (AL) approach for identifying novel HECs. To begin, we compared numerous methods for generating composition-based feature vectors (e.g., magpie and mat2vec), deployed an ensemble of machine learning (ML) models to generate an average and distribution of predictions, and then utilized the distribution as an uncertainty. Here we then deployed an AL approach to extract new training data points where the ensemble of ML models predicted a high EFA value or was uncertain of the prediction. Our approach has the combined benefit of decreasing the amount of training data required to reach acceptable prediction qualities and biases the predictions toward identifying HECs with the desired high EFA values, which are tentatively correlated with the formation of single phase HECs. Using this approach, we increased the number of 5-metal carbides screened from 56 to 15,504, revealing 4 compositions with record-high EFA values that were previously unreported in the literature. Our AL framework is also generalizable and could be modified to rationally predict optimized candidate materials/combinations with a wide range of desired properties (e.g., mechanical stability, thermal conductivity).

36 MATERIALS SCIENCE↗
DOE OSTI2020

Minimal Experimental Bias on the Hydrogen Bond Greatly Improves Ab Initio Molecular Dynamics Simulations of Water

Experiment Directed Simulations (EDS) is a method within a class of techniques seeking to improve molecular simulations by minimally biasing the system Hamiltonian to reproduce certain experimental observables. In a previous application of EDS to ab initio molecular dynamics (AIMD) simulation based on electronic density functional theory (DFT), the AIMD simulations of water were biased to reproduce its experimentally derived solvation structure. In particular, by solely biasing the O-O pair correlation functions, other structural and dynamical properties that were not biased were improved. In this work, the hypothesis is tested that directly biasing the OH pair correlation (and hence the H-O∙∙∙H hydrogen bonding), will provide an even better improvement of DFT-based water properties in AIMD simulations. The logic behind this hypothesis is that for most electronic DFT descriptions of water the hydrogen bonding is known to be deficient due to anomalous charge transfer and over polarization in the DFT. Using recent advances to the EDS learning algorithm, we thus train a minimal bias on AIMD water that reproduces the O-H radial distribution function derived from the highly accurate MB-pol model of water. Finally, it is then confirmed that biasing the O-H pair correlation alone can lead to improved AIMD water properties, with structural and dynamical properties in even closer to experiment than the previous EDS-AIMD model.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Data-Driven PMU Noise Emulation Framework using Gradient-Penalty-Based Wasserstein GAN

Availability of phasor measurement unit (PMUs) data has led to research on data-driven algorithms for event monitoring, control and ensuring stability of the grid. Unavailability of infrequent critical event field PMU data with component failures is driving the need to generate realistic synthetic PMU data for research. The synthetic data from power system simulation software often neglect noise profiles of received phasors, thus creating some discrepancies between real PMU data and synthetic ones. To address this issue, this work presents an initial study on the noise characteristics of PMUs, as well as presenting models for recreating their unique noise signatures. The proposed method, utilizing the Wasserstein generative adversarial network with gradient penalty (WGAN-GP) architecture, provides an excellent benchmark for matching the noise distribution. One can use a well-learned GAN model to draw noise signatures from a distribution that seemingly mirrors the real PMU noise distribution, while also being able to be detached from the PMU data once the training is done. Based on the observed results and employed data-driven methodology, it is expected that the proposed methods can be adapted to replicate the behavior of other sensors, providing research and other applications with a tool for data synthesis and sensor characterization.

PMU↗
DOE OSTI2021

Computer-aided Abnormality Detection in Chest Radiographs in a Clinical Setting via Domain-adaptation

Deep learning (DL) models are being deployed at medical centers to aid radiologists for diagnosis of lung conditions from chest radiographs. Such models are often trained on a large volume of publicly available labeled radiographs. These pre-trained DL models’ ability to generalize in clinical settings is poor because of the changes in data distributions between publicly available and privately held radiographs. In chest radiographs, the heterogeneity in distributions arises from the diverse conditions in X-ray equipment and their configurations used for generating the images. In the machine learning community, the challenges posed by the heterogeneity in the data generation source is known as domain shift, which is a mode shift in the generative model. In this work, we introduce a domain-shift detection and removal method to overcome this problem. Our experimental results show the proposed method’s effectiveness in deploying a pre-trained DL model for abnormality detection in chest radiographs in a clinical setting.

Dubey, Abhishek↗
DOE OSTI2024

Distributed-Memory Sparse Deep Neural Network Inference Using Global Arrays

Partitioned Global Address Space (PGAS) models exhibit tremendous promise in developing efficient and productive distributed-memory parallel applications. They have been used extensively in scientific computations due to conveniently offering a ``shared-memory''-like model and convenient interfaces that separate communication with synchronization. Traditionally, PGAS communication models have been applied to dense/contiguously distributed data, but most modern applications depict varied levels of sparsity. Existing PGAS models require certain adaptations to support distributed sparse computations, since associated computations often require matrix arithmetic, in addition to data movement. The Global Arrays toolkit from Pacific Northwest National Laboratory (PNNL) is one of the earliest PGAS models to combine one-sided data communication and distributed matrix operations and is still used in the popular NWChem quantum chemistry suite. Recently, we have expanded the Global Arrays toolkit to support common sparse operations, like sparse matrix-dense matrix multiplies (SpMM), sparse matrix-sparse matrix multiplication (SpGEMM) and Sampled Dense-Dense Matrix Multiplication (SDDMM). As it turns out, these operations are the bedrock of sparse Deep Learning (DL); sparse deep neural networks and Graph Neural Networks (GNNs) have gained increasing attention recently in achieving speedups on training and inference with reduced memory footprints. Unlike scientific applications in High Performance Computing (HPC), modern (distributed-memory capable) DL toolkits often rely on non-standardized and closed-source vendor software optimizations, creating challenges in software-hardware co-design at scale. Our goal is to support a variety of distributed-memory sparse matrix operations and helper functions in the newly created Sparse Global Arrays (SGA), such that it is possible to build portable and productive Machine Learning scenarios for algorithm/software and hardware codesign purposes. Contemporary data-parallel schemes for training/inference are undergoing a major overhaul since model replication limits scalability and causes resource inefficiencies. As such, we have adopted tensor parallelism in decomposing the model and inputs, to mitigate memory issues. Current implementation is built on top of MPI and uses CPUs to maximize the portability across the platforms.

Distributed computing, machine learning↗
DOE OSTI2024

Multi-agent voltage control in distribution systems using GAN-DRL-based approach

Active distribution grids can experience voltage fluctuations and violations due to the high penetration of variable distributed energy resources (DERs). These problems might occur because of the uncertain and variable generation natures of these resources, especially solar photovoltaic resources, during panel shadowing scenarios. Volt-VAR control (VVC) is an efficient method that controls the reactive power set-points of the inverters to regulate the voltage of distribution grids. Although several VVC approaches have been proposed recently, the performance of these approaches degrades significantly if behind-the-meter solar generation data are unobservable/missing. Therefore, it is necessary to impute missing/unobservable PV data accurately to be utilized in VVC approaches. Further, this paper proposes a model-free, data-driven, centrally trained, and decentrally executed multi-agent deep reinforcement learning-based VVC architecture to regulate the voltage of distribution networks. A generative adversarial network (GAN) is incorporated to impute the unobservable PV data accurately, which improves the performance of the proposed control architecture. The proposed multi-agent-soft-actor–critic algorithm (MASAC)-based VVC technique utilizes the actual PV dataset as well as the imputed dataset from the GAN framework to learn the optimal coordinated control policy for controlling the optimal reactive power set-points of PV inverters. The effectiveness of the proposed approach is analyzed on a modified IEEE 34-bus test case with added PV inverters. The results are compared and analyzed with a base case model with no VVC and VVC with a local droop control approach, genetic algorithm optimization, and a centralized soft actor–critic-based approach. Moreover, the performance of the proposed approach is compared with that of a multi-agent VVC framework without using the PV generation data and load information as the system state. The results illustrate that the proposed method with more state input improves the voltage profile and reduces the power loss of the network across various loading and PV generation scenarios.

14 SOLAR ENERGY↗
DOE OSTI2025

Out-of-Distribution Generalization for Learning Quantum Channels with Low-Energy Coherent States

When experimentally learning the action of a continuous-variable quantum process by probing it with inputs, there will often be some restriction on the input states used. One experimentally simple way to probe a quantum channel is to use low-energy coherent states. Learning a quantum channel in this way presents difficulties, due to the fact that two channels may act similarly on low-energy inputs but very differently for high-energy inputs. They may also act similarly on coherent-state inputs but differently on nonclassical inputs. Extrapolating the behavior of a channel for more general input states from its action on the far more limited set of low-energy coherent states is a case of out-of-distribution generalization. To be sure that such generalization gives meaningful results, one needs to relate error bounds for the training set to bounds that are valid for all inputs. We show that for any pair of channels that act sufficiently similarly on low-energy coherent-state inputs, one can bound how different the input-output relations are for any (high-energy or highly nonclassical) input. This proves that out-of-distribution generalization is always possible for learning quantum channels using low-energy coherent states, as long as enough samples are used.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗