Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Resnet”

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 73 records · Page 4
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Accelerating Binarized Neural Networks via Bit-Tensor-Cores in Turing GPUs

Despite foreseeing tremendous speedups over conventional deep neural networks, the performance advantage of binarized neural networks (BNNs) has merely been showcased on general-purpose processors such as CPUs and GPUs. In fact, due to being unable to leverage bit-level-parallelism with a word-based architecture, GPUs have been criticized for extremely low utilization (1%) when executing BNNs. Consequently, the latest tensorcores in NVIDIA Turing GPUs start to experimentally support bit computation. In this work, we look into this brand new bit computation capability and characterize its unique features. We show that the stride of memory access can significantly affect performance delivery and a data-format co-design is highly desired to support the tensorcores for achieving superior performance than existing software solutions without tensorcores. We realize the tensorcore-accelerated BNN design, particularly the major functions for fully-connect and convolution layers — bit matrix multiplication and bit convolution. Evaluations on two NVIDIA Turing GPUs show that, with ResNet-18, our BTC-BNN design can process ImageNet at a rate of 5.6K images per second, 77% faster than state-of-the-art. Our BNN approach is released on https://github.com/pnnl/TCBNN.

Li, Ang↗
DOE OSTI2025

Optimizing transmit field inhomogeneity of parallel RF transmit design in 7T MRI using deep learning

Ultrahigh field (UHF) Magnetic Resonance Imaging (MRI) provides a higher signal-to-noise ratio and, thereby, higher spatial resolution. However, UHF MRI introduces challenges such as transmit radiofrequency (RF) field (B+1) inhomogeneities, leading to uneven flip angles and image intensity anomalies. These issues can significantly degrade imaging quality and its medical applications. This study addresses B+1 field homogeneity through a novel deep learning-based strategy. Traditional methods like Magnitude Least Squares (MLS) optimization have been effective but are time-consuming and dependent on the patient’s presence. Recent machine learning approaches, such as RF Shim Prediction by Iteratively Projected Ridge Regression and deep learning frameworks, have shown promise but face limitations like extensive training times and oversimplified architectures. We propose a two-step deep learning strategy. First, we obtain the desired reference RF shimming weights from multi-channel B+1 fields using random-initialized Adaptive Moment Estimation. Then, we employ Residual Networks (ResNets) to train a model that maps B+1 fields to target RF shimming outputs. Our approach does not rely on pre-calculated reference optimizations for the testing process and efficiently learns residual functions. Comparative studies with traditional MLS optimization demonstrate our method’s advantages in terms of speed and accuracy. The proposed strategy achieves a faster and more efficient RF shimming design, significantly improving imaging quality at UHF. This advancement holds potential for broader applications in medical imaging and diagnostics.

Lu, Zhengyi [Vanderbilt University]↗
DOE OSTI2025

Automated nuclear cloud feature extraction from film

Chemical, biological, radiological, nuclear, and explosives incidents require rapid detection and characterization for appropriate response. For a nuclear detonation, visible-light cameras may be used to locate the cloud and characterize fallout deposition when coupled with numerical models. Films from the United States’ nuclear testing era compose the only sizeable collection of imagery depicting high-yield detonations. These films offer unique insights into characteristics of flows involving scales that are difficult to replicate experimentally, and they are a valuable source of data for the validation of models for nuclear fallout transport, either as part of emergency response or forensic activities. In this work, we implement modern computer vision and machine learning techniques to identify and track the cloud automatically and subsequently determine the time dependence of some of its features. We trained a ResNet-18 image classifier on hundreds of images to categorize nuclear cloud morphology. Each category or cloud regime is determined by early cloud evolution and is associated to constitutive properties of the flow, such as distribution of vorticity. Next, we identified keypoint features using the KAZE algorithm and tracked these keypoints in the images, allowing us to determine the dimensions and velocities of the cloud across film frames. These measurements converted to real-world units provide valuable experimental data that can be used in the development and validation of nuclear cloud models. We compared the results of this method against manual cloud rise measurements from two different films. In one, our automated method accelerated the feature extraction process without sacrificing measurement accuracy.

Khristy, Joel [ORNL] (ORCID:0000000209963060)↗
DOE OSTI2025

Neutrino interaction vertex reconstruction in DUNE with Pandora deep learning

The Pandora Software Development Kit and algorithm libraries perform reconstruction of neutrino interactions in liquid argon time projection chamber detectors. Pandora is the primary event reconstruction software used at the Deep Underground Neutrino Experiment, which will operate four large-scale liquid argon time projection chambers at the far detector site in South Dakota, producing high-resolution images of charged particles emerging from neutrino interactions. While these high-resolution images provide excellent opportunities for physics, the complex topologies require sophisticated pattern recognition capabilities to interpret signals from the detectors as physically meaningful objects that form the inputs to physics analyses. A critical component is the identification of the neutrino interaction vertex. Subsequent reconstruction algorithms use this location to identify the individual primary particles and ensure they each result in a separate reconstructed particle. A new vertex-finding procedure described in this article integrates a U-ResNet neural network performing hit-level classification into the multi-algorithm approach used by Pandora to identify the neutrino interaction vertex. The machine learning solution is seamlessly integrated into a chain of pattern-recognition algorithms. The technique substantially outperforms the previous BDT-based solution, with a more than 20% increase in the efficiency of sub-1 cm vertex reconstruction across all neutrino flavours.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2021

APNN-TC: Accelerating Arbitrary Precision Neural Networks on Ampere GPU Tensor Cores

Over the years, accelerating neural networks with quantization has been widely studied. Unfortunately, prior efforts with diverse precisions (e.g., 1-bit weights and 2-bit activations) are usually restricted by limited precision support on GPUs (e.g., int1 and int4). To break such restrictions, we introduce the first Arbitrary Precision Neural Network framework (APNN-TC) to fully exploit quantization benefits on Ampere GPU Tensor Cores. Specifically, APNN-TC first incorporates a novel emulation algorithm to support arbitrary short bit-width computation with int1 compute primitives and XOR/AND Boolean operations. Second, APNN-TC integrates arbitrary precision layer designs to efficiently map our emulation algorithm to Tensor Cores with novel batching strategies and specialized memory organization. Third, APNN-TC embodies a novel arbitrary precision NN design to minimize memory access across layers and further improve performance. Extensive evaluations show that APNN-TC can achieve significant speedup over CUTLASS kernels and various NN models, such as ResNet and VGG.

Feng, Boyuan↗
DOE OSTI2023

Scaling Resolution of Gigapixel Whole Slide Images Using Spatial Decomposition on Convolutional Neural Networks

Gigapixel images are prevalent in scientific domains ranging from remote sensing, and satellite imagery to microscopy, etc. However, training a deep learning model at the natural resolution of those images has been a challenge in terms of both, overcoming the resource limit (e.g. HBM memory constraints), as well as scaling up to a large number of GPUs. In this paper, we trained Residual neural Networks (ResNet) on 22,528 x 22,528-pixel size images using a distributed spatial decomposition method on 2,304 GPUs on the Summit Supercomputer. We applied our method on a Whole Slide Imaging (WSI) dataset from The Cancer Genome Atlas (TCGA) database. WSI images can be in the size of 100,000 x 100,000 pixels or even larger, and in this work we studied the effect of image resolution on a classification task, while achieving state-of-the-art AUC scores. Moreover, our approach doesn't need pixel-level labels, since we're avoiding patching from the WSI images completely, while adding the capability of training arbitrary large-size images. This is achieved through a distributed spatial decomposition method, by leveraging the non-block fat-tree interconnect network of the Summit architecture, which enabled GPU-to-GPU direct communication. Finally, detailed performance analysis results are shown, as well as a comparison with a data-parallel approach when possible.

Tsaris, Aristeidis (aris)↗
DOE OSTI2021

ROI Hide and Seek Protocol v1

1. Segmentation We provide scripts for the model definition of the U-net architecture adapted from: https://github.com/jvanvugt/pytorch-unet/blob/master/unet.py We developed scripts for preparing the lung segmentation data set. We developed scripts for training the U-Net architecture. We developed scripts for applying the trained U-Net model to perform the ROI Hide and Seek protocol on the classification dataset to create the modified dataset. 2. Classification We provide scripts for the training of the COVID-Net models provided by Linda Wang, this code is adapted from her github repository: https://github.com/lindawangg/COVID-Net/tree/d7b36831d854f57de5bc7557217f5439e86e016f. These scripts were modified to save training log information as well as to load the provided models in their github repo. We developed scripts for training standard Neural Network Models (resnet 50, vgg 11, Alexnet) on the COVID datsets along with the ROI Hide and Seek altered datasets.

Sadre, Robbie↗
DOE OSTI2025

Moltensaltpropnet

MoltenSaltPropnet is a physics-informed machine learning framework that aims to predict the thermophysical properties of molten fluoride and chloride salt mixtures, which are crucial for the design and safety of Generation IV molten salt reactors. The code processes data from the Molten-Salt Thermal Properties Database (MSTDB-TP) and the Janz compendium, converting critically evaluated correlations into fast, differentiable surrogate models for density, viscosity, thermal conductivity, and heat capacity across 448 distinct salt systems. The implementation consists of several key components: 1. Data Curation: The code parses and cleans the raw data, normalizing elemental mole fractions and extracting relevant regression coefficients for various thermophysical properties. 2. Feature Engineering: It generates fixed-length numerical descriptors that encapsulate the composition and temperature, incorporating polynomial interaction terms and dimensionality-reduction techniques to optimize model performance. 3. Coefficient Learning: Four different machine learning architectures are employed: a deep residual network (ResNet), a Kolmogorov–Arnold network (KAN), a sparsity-inducing neural network (SNN), and classical regression models. Each model learns to predict coefficients that define the temperature-dependent correlations for the thermophysical properties. 4. Property Reconstruction: The predicted coefficients are used to compute temperature-dependent property values, ensuring positivity and monotonic trends through a composite loss function that enforces physical constraints. 5. User Interface: An open-source web application enables users to filter the database, train task-specific models, and visualize the results, allowing for rapid exploration of candidate salt mixtures. MoltenSaltPropnet bridges the gap between limited experimental data and high-fidelity reactor simulations, providing a powerful tool for researchers in the field of molten salt reactors and advanced nuclear energy systems.

Retamales, Mauricio Eduardo Tano [Idaho National L↗
DOE OSTI2026

smol_model_benchmark

Machine learning models for protein–small molecule binding prediction using graph (GAT), image-based (ResNet), and transformer-based SMILES approaches (ChemBERTa and MIST). The code includes training pipelines and experiments on the dataset from the BELKA Kaggle competition.

Gibson, Kaetlyn [Los Alamos National Lab]↗
DOE OSTI2024

Forward variable selection enables fast and accurate dynamic system identification with Karhunen-Loève decomposed Gaussian processes

A promising approach for scalable Gaussian processes (GPs) is the Karhunen-Loève (KL) decomposition, in which the GP kernel is represented by a set of basis functions which are the eigenfunctions of the kernel operator. Such decomposed kernels have the potential to be very fast, and do not depend on the selection of a reduced set of inducing points. However KL decompositions lead to high dimensionality, and variable selection thus becomes paramount. This paper reports a new method of forward variable selection, enabled by the ordered nature of the basis functions in the KL expansion of the Bayesian Smoothing Spline ANOVA kernel (BSS-ANOVA), coupled with fast Gibbs sampling in a fully Bayesian approach. It quickly and effectively limits the number of terms, yielding a method with competitive accuracies, training and inference times for tabular datasets of low feature set dimensionality. Theoretical computational complexities are O ( N P 2 ) in training and O ( P ) per point in inference, where N is the number of instances and P the number of expansion terms. The inference speed and accuracy makes the method especially useful for dynamic systems identification, by modeling the dynamics in the tangent space as a static problem, then integrating the learned dynamics using a high-order scheme. The methods are demonstrated on two dynamic datasets: a ‘Susceptible, Infected, Recovered’ (SIR) toy problem, along with the experimental ‘Cascaded Tanks’ benchmark dataset. Comparisons on the static prediction of time derivatives are made with a random forest (RF), a residual neural network (ResNet), and the Orthogonal Additive Kernel (OAK) inducing points scalable GP, while for the timeseries prediction comparisons are made with LSTM and GRU recurrent neural networks (RNNs) along with the SINDy package.

Hayes, Kyle↗
DOE OSTI2021

Deep learning of parameterized equations with applications to uncertainty quantification

We propose a learning algorithm for discovering unknown parameterized dynamical systems by using observational data of the state variables. Our method is built upon and extends the recent work of discovering unknown dynamical systems, in particular those using deep neural network (DNN). We propose a DNN structure, largely based upon the residual network (ResNet), to not only learn the unknown form of the governing equation but also take into account the random effect embedded in the system, which is generated by the random parameters. Once the DNN model is successfully constructed, it is able to produce system prediction over longer term and for arbitrary parameter values. For uncertainty quantification, it allows us to conduct uncertainty analysis by evaluating solution statistics over the parameter space.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

DOE Deep Energy Retrofit Cost Survey

A survey was conducted by the Lawrence Berkeley National Laboratory on deep energy retrofit (DER) market drivers, opportunities, and challenges. The survey was part a research study sponsored by the U.S. Department of Energy to gather information on the costs of DER from home performance contractors and stakeholders. Cost data was gathered from DER projects that use a comprehensive, whole-home approach to drastically reduce energy use and improve performance. DER projects often aim at reducing energy use by 50% or more. In addition, these projects can improve home comfort and potentially benefiting occupant health. Yet, market adoption of DER has been limited. Major limiting factors include complex projects, high costs, perceived risks, extensive disruption, and unfamiliar work scopes to some contractors. In order to better understand what motivates and deters DER projects in today’s market, a survey was conducted to gather this information, and to learn about promising approaches and technologies from the industry perspective. Past surveys on homeowners and home energy performance professionals have studied the motivations and barriers of energy efficiency retrofits. Two surveys of home energy performance professionals were conducted in recent years. The Resources for the Future (RFF) Home Energy Audit and Retrofit Survey was conducted in 2011 by recruiting energy auditors and retrofit installers through members of Efficiency First and Building Performance Institute (BPI) accredited contractors. The survey asked about the business and services that respondents provide, how often homeowners follow their recommendations to retrofit their homes, and the respondents’ opinion on barriers faced by the industry. The survey found that not enough homeowners know about energy audits, but more importantly, it is the high cost of retrofits compared to low energy prices that is responsible for few energy audits and retrofits being completed. The RESNET Deep Retrofit Industry Stakeholder Survey was conducted shortly after launching of the EnergySmart Home Performance Team program. The EnergySmart Team program involves a formal agreement among allied contractors who are engaged in high performance retrofits. Using this allied team approach, teams can pool their expertise and provide each other with customer referrals. The survey asked EnergySmart Team members and outside stakeholders on questions about market and technical barriers in performing home energy retrofits. Survey respondents identified lack of consumer awareness and lack of affordable financing for consumers as the leading market barriers to home energy retrofits. In their written comments, many survey respondents also echoed that the high costs of retrofit compared to low energy prices is a market barrier. Respondents found “certain housing characteristics that prevent effective retrofit” and “energy analysis software inaccuracy or limitations” are the two leading technical barriers. Their choices for technical barriers were reflective of the energy rater/auditor role played by the majority (78%) of the survey respondents. In comparison to past surveys, this work aimed to gather inputs from a broader segment of the home performance industry to identify the opportunities and barriers faced by DERs from all perspectives. The survey asks for project costs to help breakdown the high costs of DERs. This survey is also motivated by a need to better understand the role of DERs in reducing energy use by the residential sectors and meeting climate goals. The survey is designed towards obtaining more substantive inputs from survey respondents by encouraging written comments, rather than setting the goal to reach a large number of respondents. We took this approach because DER is currently still a niche market, so it is more valuable to gather in-depth inputs from individuals who are performing this work rather than getting to the masses.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

TrojAI Alternate Analysis

In this portion of the TrojAI evaluation, we focus on the cyber-network-c2-mar2024 dataset. Recall that in this round ResNet18 and ResNet34 neural networks (NN) were trained on the USTC-TFC2016 dataset with the aim of distinguishing between benign versus botnet command and control (c2) packets. A range of bytes from each packet was reformatted into a 28x28 pixel image, and the collection of reformatted packets served as the training (and testing) data for the two ResNet models. For some of the data a trigger watermark was strategically placed to affect various inputs to the NNs. This watermarked, or poisoned, data in turn created a poisoned, or trojaned NN. The data were poisoned in different ways ultimately creating different trojaned NNs. This collection of trojaned NNs was combined with various versions of not trojaned NNs and served as the training and testing data for the performers. The performers’ task was to construct a classifier to distinguish between the trojaned and not trojaned models. It was previously noted that the performers struggled with the cyber-network-c2-mar2024 dataset, motivating this investigation of potential reasons the performers experienced challenges.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Synthetic Atmospheric River Ensembles Generated by Deep-AR

This dataset contains 35,850 synthetic landfalling atmospheric river (AR) realizations generated by the Deep-AR two-stage deep-learning framework over the Northeast Pacific and U.S. West Coast. The archive contains 25 stochastic ensemble members for each of 1,434 held-out observed seed events. Each synthetic realization is initialized from conditions 48 hours before the corresponding observed AR landfall and is generated autoregressively at 6-hour intervals over a 144-hour period. Deep-AR combines a deterministic residual network (ResNet) that advances the large-scale atmospheric state with a Wasserstein generative adversarial network (WGAN) that produces stochastic, high-resolution fields. Each HDF5 file contains 0.25° gridded synthetic integrated vapor transport components (qu, qv), 10 m wind components (u10, v10), and 6-hour accumulated precipitation on a common 200 × 480 grid. The files also include coordinate and datetime arrays. This dataset supports AR hazard analysis, ensemble-based uncertainty characterization, precipitation-extremes research, and regional stress testing. Synthetic files follow the naming convention deepar.model.YYYYMMDD.HHMMSS.vNN.h5. YYYYMMDD.HHMMSS identifies the UTC initial-condition timestamp, which occurs 48 hours before the diagnosed observed landfall, and vNN identifies the zero-padded ensemble member, ranging from v01 through v25. Each synthetic file can be paired with its corresponding observed file by matching the initial-condition timestamp. The paired observed file follows the naming convention deepar.obs.YYYYMMDD.HHMMSS.h5 and is available in the separately registered oracle/deepar.obs dataset at https://wdh.energy.gov/ds/oracle/deepar.obs (DOI: https://doi.org/10.21947/3377671).

17 WIND ENERGY↗
DOE OSTI2020

Robust Learning with Implicit Residual Networks

In this effort, we propose a new deep architecture utilizing residual blocks inspired by implicit discretization schemes. As opposed to the standard feed-forward networks, the outputs of the proposed implicit residual blocks are defined as the fixed points of the appropriately chosen nonlinear transformations. We show that this choice leads to the improved stability of both forward and backward propagations, has a favorable impact on the generalization power, and allows for control the robustness of the network with only a few hyperparameters. In addition, the proposed reformulation of ResNet does not introduce new parameters and can potentially lead to a reduction in the number of required layers due to improved forward stability. Finally, we derive the memory-efficient training algorithm, propose a stochastic regularization technique, and provide numerical results in support of our findings.

Reshniak, Viktor↗
DOE OSTI2022

Evaluating Image Classification Deep Convolutional Neural Network Architectures for Remaining Useful Life Estimation of Turbofan Engines

Accurate estimation of the remaining useful life (RUL) is a key component of condition-based maintenance (CBM) and prognosis and health management (PHM). Data-based models for the estimation of RUL are of particular interest because expert knowledge of systems is not always available, and physical modeling is often not feasible. Additionally, using data-based models, which make decisions based on raw sensor data, allow features to be learned instead of manually determined. In this work, deep convolutional neural network (CNN) architectures are investigated for their ability to estimate the RUL of turbofan engines. To improve the accuracy of the models, CNN architectures, which have proven successful in image classification, are implemented and tested. Specifically, the blocks used in the Visual Geometry Group (VGG) architecture, inception modules used in the GoogLeNet architecture, and residual blocks used in the ResNet architecture are incorporated. To account for varying flight lengths, the input to the models is a window of time series data collected from the engine under test. Window locations at the climb, cruise, and descent stages are considered. To further improve the RUL estimations, multiple overlapping windows at each location are used. This increases the amount of training data available and is found to increase the accuracy of the resulting RUL estimations by averaging the estimates from all overlapping segments. The model is trained and tested using the new Commercial Modular Aero-Propulsion System Simulation (N-CMAPSS) data set, and high prognosis accuracy was achieved. Furthermore, this work expands on the model developed and used in the 2021 PHM Society Data Challenge, which received second place.

convolutional neural networks↗
DOE OSTI2022

Neural Networks for Nuclear Reactions in MAESTROeX

We demonstrate the use of neural networks to accelerate the reaction steps in the MAESTROeX stellar hydrodynamics code. A traditional MAESTROeX simulation uses a stiff ODE integrator for the reactions; here, we employ a ResNet architecture and describe details relating to the architecture, training, and validation of our networks. Our customized approach includes options for the form of the loss functions, a demonstration that the use of parallel neural networks leads to increased accuracy, and a description of a perturbational approach in the training step that robustifies the model. We test our approach on millimeter-scale flames using a single-step, 3-isotope network describing the first stages of carbon fusion occurring in Type Ia supernovae. We train the neural networks using simulation data from a standard MAESTROeX simulation, and show that the resulting model can be effectively applied to different flame configurations. This work lays the groundwork for more complex networks, and iterative time-integration strategies that can leverage the efficiency of the neural networks.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

RU Net for Automatic Characterization of TRISO Fuel Cross Sections

TRistructural ISOtropic (TRISO) particle fuel is a type of nuclear fuel known for its high-temperature and high-burnup performance. Each sub-millimeter diameter TRISO particle consists of uranium-oxycarbide (UCO) or UO2 fuel kernel, coated with buffer, inner pyrolytic carbon (IPyC), silicon carbide (SiC), and outer pyrolytic carbon (OPyC) layers. The SiC layer acts as the main containment barrier for the TRISO particle to retain the fission products, while the IPyC and OPyC layers provide additional barriers to the release of fission products, especially fission gases. During irradiation, phenomena like kernel swelling, buffer densification, and IPyC fracture may impact fuel performance. Post-irradiation microscopy on entire compact cross sections or samples of individual particles deconsolidated from compacts is often used to identify these irradiation-induced changes in morphology. However, each fuel compact generally contains thousands of TRISO particles. To get statistical information on these phenomena, it is cumbersome work if done manually. For example, to get information about swelling/densification behaviors of different layers or kernels after irradiation, researchers previously manually measured the perimeter of each TRISO layer in hundreds of particles after four rounds of iterative grinding and polishing encompassing more than 2000 cross-section images for a total of four fuel compacts. To attempt to reduce the subjectivity inherent in that process and accelerate data analysis, we conducted a study on the automatic TRISO layer segmentation on cross-sectional microscopic images using Convolutional Neural Networks (CNNs). CNNs are a class of machine learning algorithms specifically designed for processing structured grid data that have gained popularity in recent years due to their remarkable performance in various computer vision tasks, including image classification, object detection, and image segmentation. In this research, we have generated the large irradiated TRISO layer dataset with more than 2000 cross-section TRISO microscopic images and the corresponding annotated images. Based on these annotated images, we have employed different CNNs for automatic segmentation of different TRISO layers. These include RU-Net (developed in this study), as well as three existing architectures: U-Net, Residual Network (ResNet), and Attention U-Net. The preliminary results show that the model based on RU-Net has the best performance in terms of intersection-over-union (IoU). Through the aid of these CNN models, we can expedite the analysis of TRISO particle cross-sections, significantly reducing the manual labor involved and improving the objectivity of the segmentation results.

Convolutional Neural Networks↗