Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Convolution”

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 1,603 records · Page 89
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

FLASH: FPGA-Accelerated Smart Switches with GCN Case Study

Some communication switches, e.g., the Mellanox SHArP and those in the IBM BlueGene clusters, are augmented to process packets at the application level with fixed-function collectives. This approach, however, lacks flexibility, which limits their applicability in diverse and dynamic workloads. Recently, a new type of programmable packet processor, which uses high-level languages, e.g., P4, has emerged as possible candidates. P4-based switches, however, fall short in certain applications, including machine learning, where capabilities not currently supported by P4 are needed. These include more complex calculation, such as sparse computation and fused multiply-accumulate, data-intensive floating point operations, data reuse, and significant memory. The problem addressed here is that such a switch augmentation needs to support: a large amount of state, significant flexible compute capability, and ease of programming, all while maintaining full functionality, including ensuring high throughput, and demonstrating utility. In this work, we propose a programmable look-aside-type accelerator that can be embedded into, or attached to, existing communication switch pipelines and that is capable of processing packets at line-rate. The proposed in-switch accelerator is based on mixing an ISA (subset of RISC-V instructions) with dataflow graphs (found in CGRAs). To augment performance, vector instructions are also supported. To facilitate usability, we have developed a complete toolchain to compile user-provided C/C++ codes to appropriate back-end instructions for configuring the accelerator. While this approach is flexible enough to support various workloads, in this paper, we consider Graph Convolutional Networks (GCNs) as a case study. Experimental results show that this approach considerably improves the performance of distributed GCN applications.

Haghi, Pouya↗
DOE OSTI2023

Architecture-Preserving Provable Repair of Deep Neural Networks

Deep neural networks (DNNs) are becoming increasingly important components of software, and are considered the state-of-the-art solution for a number of problems, such as image recognition. However, DNNs are far from infallible, and incorrect behavior of DNNs can have disastrous real-world consequences. This paper addresses the problem of architecture-preserving V-polytope provable repair of DNNs. A V-polytope defines a convex bounded polytope using its vertex representation. V-polytope provable repair guarantees that the repaired DNN satisfies the given specification on the infinite set of points in the given V-polytope. An architecture-preserving repair only modifies the parameters of the DNN, without modifying its architecture. The repair has the flexibility to modify multiple layers of the DNN, and runs in polynomial time. It supports DNNs with activation functions that have some linear pieces, as well as fully-connected, convolutional, pooling and residual layers. To the best our knowledge, this is the first provable repair approach that has all of these features. We implement our approach in a tool called APRNN. Using MNIST, ImageNet, and ACAS Xu DNNs, we show that it has better efficiency, scalability, and generalization compared to PRDNN and REASSURE, prior provable repair methods that are not architecture preserving.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Self-Supervised T-GCN for Detection of Disturbance and Propagation in Power Grid

Urban power systems increasingly rely on dense sensing to monitor grid reliability, yet disturbance labels are scarce and events are rare. We present a self-supervised spatio-temporal method that detects, localizes, and characterizes grid frequency disturbances across urban areas using only unlabeled data. Our approach trains a tiny Temporal Graph Convolutional Network (T-GCN) to forecast per-site frequency residuals (deviation from 60 Hz). The sensor graph is constructed directly from signals using pre-event Pearson correlation with a cross-correlation lag penalty without geocoding. At inference, node-level anomalies are the model's forecast errors; region-level alarms arise from connected components of high-score nodes. We estimate disturbance propagation by computing per-node arrival times (first persistent exceedance), then fit a planar or time-of-arrival model to obtain direction, speed, and an epicenter proxy. With only three real events collected at decisecond resolution across U.S. cities, we evaluate the T-GCN and report time-to-detect, footprint size, and propagation consistency. We further show that short-window embeddings from the T-GCN's hidden states enable few-shot event-vs-background recognition via a simple prototypical classifier. Despite minimal data and no labels, our system yields fast, spatially coherent detection and interpretable propagation maps, offering a practical, lightweight pathway to city-scale grid resilience analytics.

Niu, Haoran [ORNL] (ORCID:0000000155228297)↗
DOE OSTI2026

wa-hls4ml: A Benchmark and Surrogate Models for hls4ml Resource and Latency Estimation

As machine learning (ML) is increasingly implemented in hardware to address real-time challenges in scientific applications, the development of advanced toolchains has significantly reduced the time required to iterate on various designs. These advancements have solved major obstacles, but also exposed new challenges. For example, processes that were not previously considered bottlenecks, such as hardware synthesis, are becoming limiting factors in the rapid iteration of designs. To mitigate these emerging constraints, multiple efforts have been undertaken to develop an ML-based surrogate model that estimates resource usage of ML accelerator architectures. We introduce wa-hls4ml, a benchmark for ML accelerator resource and latency estimation, and its corresponding initial dataset of over 680,000 fully connected and convolutional neural networks, all synthesized using hls4ml and targeting Xilinx FPGAs. The benchmark evaluates the performance of resource and latency predictors against several common ML model architectures, primarily originating from scientific domains, as exemplar models, and the average performance across a subset of the dataset. Additionally, we introduce GNN- and transformer-based surrogate models that predict latency and resources for ML accelerators. We present the architecture and performance of the models and find that the models generally predict latency and resources for the 75% percentile within several percent of the synthesized resources on the synthetic test dataset.

Hawks, Benjamin [Fermilab] (ORCID:0000000157000288↗
DOE OSTI2026

ObstacleSense: Low-Power Neuromorphic Vision for Corridor Obstacle Awareness in Low-Level ADAS

The automotive industry’s pursuit of Level 5 autonomy is constrained by substantial perception-compute power requirements, often reaching 1, 000 + watts in full autonomy stacks. Reducing this energy burden requires rethinking perception not only at the high-end autonomy level, but also at the foundational Advanced Driver Assistance Systems (ADAS) level where low-power, safety-critical sensing can have broad impact. Neuromorphic vision provides a promising starting point: HD Dynamic Vision Sensors (DVS) can operate below 100 mW at the sensor level by reporting only asynchronous brightness changes. However, low-power sensing alone is insufficient if downstream perception reintroduces dense, energy-intensive computation. In particular, many event-driven object-detection pipelines still rely on CNN backbones, while purely spiking alternatives often trade away accuracy or ignore deployment constraints. We introduce ObstacleSense, a highly compact, CNN-free hybrid ANN–SNN framework for Level 0–1 forward-corridor obstacle awareness. Instead of performing full-scene object detection with a convolutional feature backbone, ObstacleSense targets the safety-critical question of whether the ego corridor is occupied and how far the nearest obstacle is. The architecture combines polarity-conditioned event encoding, lightweight temporal spiking dynamics, axial spatial mixing, and coarse-to-fine range estimation within a regular fixed-grid compute pattern. This design avoids the dense CNN backbone commonly used in event-based detection while maintaining a small state footprint suitable for eventual small-FPGA deployment. Before hardware mapping, we evaluate the software implementation using a model-side power proxy derived from MACs, weight and activation traffic, and spiking state updates under shared FP16 assumptions. On simulated CARLA event corpora, the deployment-oriented model achieves 0.9464 objectness F1, 0.9978 grid-level mAP, and 0.8987 m distance Mean Absolute Error at an estimated 1.92 mW proxy cost, while maintaining performance on unseen generalization test sequences.

Johnson-Scott, Zac [ORNL]↗
DOE OSTI2022

Methods—Kintsugi Imaging of Battery Electrodes: Distinguishing Pores from the Carbon Binder Domain using Pt Deposition

The mesostructure of porous electrodes used in lithium-ion batteries strongly influences cell performance. Accurate imaging of the distribution of phases in these electrodes would allow this relationship to be better understood through simulation. However, imaging the nanoscale features in these components is challenging. While scanning electron microscopy is able to achieve the required resolution, it has well established difficulties imaging porous media. This is because the flat imaging planes prepared using focused ion beam milling will intersect with the pores, which makes the images hard to interpret as the inside walls of the pores are observed. It is common to infiltrate porous media with resin prior to imaging to help resolve this issue, but both the nanoscale porosity and the chemical similarity of the resins to the battery materials undermine the utility of this approach for most electrodes. In this study, a technique is demonstrated which uses in situ infiltration of platinum to fill the pores and thus enhance their contrast during imaging. Reminiscent of the Japanese art of repairing cracked ceramics with precious metals, this technique is referred to as the kintsugi method. The images resulting from applying this technique to a conventional porous cathode are presented and then segmented using a multi-channel convolutional method. We show that while some cracks in active material particles were empty, others appear to be filled (perhaps with the carbon binder phase), which will have implications for the rate performance of the cell. Energy dispersive X-ray spectroscopy was used to validate the distribution of phases resulting from image analysis, which also suggested a graded distribution of the binder relative to the carbon additive. The equipment required to use the kintsugi method is commonly available in major research facilities and so we hope that this method will be rapidly adopted to improve the imaging of electrode materials and porous media in general.

25 ENERGY STORAGE↗
DOE OSTI2023

Machine Learning Benchmarks for the Classification of Equivalent Circuit Models from Electrochemical Impedance Spectra

Analysis of Electrochemical Impedance Spectroscopy (EIS) data for electrochemical systems often consists of defining an Equivalent Circuit Model (ECM) using expert knowledge and then optimizing the model parameters to deconvolute various resistance, capacitive, inductive, or diffusion responses. For small data sets, this procedure can be conducted manually; however, it is not feasible to manually define a proper ECM for extensive data sets with a wide range of EIS responses. Automatic identification of an ECM would substantially accelerate the analysis of large sets of EIS data. We showcase machine learning methods to classify the ECMs of 9,300 impedance spectra provided by QuantumScape for the BatteryDEV hackathon. The best-performing approach is a gradient-boosted tree model utilizing a library to automatically generate features, followed by a random forest model using the raw spectral data. A convolutional neural network using boolean images of Nyquist representations is presented as an alternative, although it achieves a lower accuracy. We publish the data and open source the associated code. The approaches described in this article can serve as benchmarks for further studies. A key remaining challenge is the identifiability of the labels, underlined by the model performances and the comparison of misclassified spectra.

25 ENERGY STORAGE↗
DOE OSTI2021

Carbon Corrosion in Polymer Electrolyte Fuel Cells: A Complex Interplay between Morphological Changes and Electrochemical Performance

Due to the high gravimetric energy density of hydrogen, the focus of implementation of polymer electrolyte fuel cells (PEFCs) has shifted from light duty passenger vehicles to heavy duty vehicles such as buses, trucks, locomotives and marine vessels. A mechanistic understanding of degradation is therefore necessary to improve durability and efficiency. During start-up and shut-down (SUSD) of PEFC systems, the catalyst (Pt nanoparticles embedded on carbon support) undergoes local potentials ~ 1 - 1.5 V caused by a combination of fuel (H 2 ) starvation, mixed fuel region and cell reversal. This leads to a series of degradation phenomenon including reduction in cathode catalyst layer (cCL) thickness and porosity, loss in electrochemical surface area (ECSA), ionomer degradation and loss in electrical contact, therefore resulting in severe performance loss. The convoluted relationship between these individual degradation mechanisms, their chronology and their effects on electrochemical performance are yet unresolved. Here, the complex interplay between morphological changes due to carbon corrosion and its effects on the electrochemical performance were analyzed using a combination of detailed electrochemical characterization, spectroscopy, and electron microscopy techniques.

25 ENERGY STORAGE↗
DOE OSTI2020

Image Search for Scientific Data (camBIR) v1.0

camBIR is the CAMERA based image retrieval, a software that takes several different kinds of imagery, such as 2 dimensional images, both scientific as well as non scientific. camBIR is specially designed for training different kinds of Deep Convolutional Neural Network, including new Deep Learning architectures, such as ResNet, Inception-ResNet, NasNet, DenseNet…. It can be used to automate the classification, sort and rank of X-ray imagery such as microCT, GISAXS, GIWAXS, SAXS and WAXS.

Ushizima, Daniela↗
DOE OSTI2021

Graphene U-Net v1

Graphene U-Net is a library that provides a simplified platform for training deep neural networks for the task of microscopy image segmentation. It contains functions and classes that make the process of training a neural network simple such that non-ML experts can train and evaluate models on their own datasets. It uses Pytorch as a backend and can run on both CUDA-enable GPUs and CPUs. It contains the main library file Microscopy_Unet.py as well as unet.py which contains the UNET model used for this software. This can be replaced with any other fully convolutional deep learning architecture with relative ease.

Sadre, Robbie↗
DOE OSTI2021

CNN-Encoder-Decoder Model

Code and data for training CNN-Encoder-Decoder model described in the publication 'Noise reduction in X-ray photon correlation spectroscopy with convolutional neural networks encoder-decoder models.'

Konstantinova, Tatiana [Brookhaven National Lab. (↗
DOE OSTI2021

TomoEncoders: 3D Autoencoders for feature extraction in X-ray tomography

Real-time steering of time-resolved or in-situ X-ray tomography requires capturing changes in morphological descriptors in a sample (e.g., porosity, particle size, and crack width) during continuous data acquisition. Image segmentation (2D or3D) followed by quantitative measurement is the conventional method for tracking changes in these descriptors with respect to a previous time-step or a 3D search in a volume. However, image segmentation is expensive. As a faster and unsupervised alternative, a feature-extraction approach using a convolutional autoencoders was developed, where the latent space of the encoder responds to relative changes in morphology with-out prior knowledge of the morphological descriptors.

TEKAWADE, ANIKET↗
DOE OSTI2021

HydraGNN

Distributed PyTorch implementation of multi-headed graph convolutional neural networks

Lupo Pasini, Massimiliano [Oak Ridge National Lab.↗
DOE OSTI2021

Attend and Decode: 4D fMRI Task State Decoding Using Attention Models

Source code for Brain Attend and Decode paper. Functional magnetic resonance imaging (fMRI) is a neuroimaging modality that captures the blood oxygen level in a subject's brain while the subject either rests or performs a variety of functional tasks under different conditions. Given fMRI data, the problem of inferring the task, known as task state decoding, is challenging due to the high dimensionality (hundreds of million sampling points per datum) and complex spatio-temporal blood flow patterns inherent in the data. In this work, we propose to tackle the fMRI task state decoding problem by casting it as a 4D spatiotemporal classification problem. We present a novel architecture called Brain Attend and Decode (BAnD), that uses residual convolutional neural networks for spatial feature extraction and self-attention mechanisms for temporal modeling. We achieve significant performance gain compared to previous works on a 7-task benchmark from the large-scale Human Connectome Project-Young Adult (HCP-YA) dataset. We also investigate the transferability of BAnD's extracted features on unseen HCP tasks, either by freezing the spatial feature extraction layers and retraining the temporal model, or finetuning the entire model. The pre-trained features from BAnD are useful on similar tasks while finetuning them yields competitive results on unseen tasks/conditions.

Ng, BrendaM.↗
DOE OSTI2023

AEflow (Autoencoder fluid flow compression network) [SWR-22-29]

As the size of turbulent flow simulations continues to grow, in situ data compression is becoming increasingly important for visualization, analysis, and restart checkpointing. For these applications, single-pass compression techniques with low computational and communication overhead are crucial. In this paper we present a deep-learning approach to in situ compression using an autoencoder architecture that is customized for three-dimensional turbulent flows and is well suited for contemporary heterogeneous computing resources. The autoencoder is compared against a recently introduced randomized single-pass singular value decomposition (SVD) for three different canonical turbulent flows: decaying homogeneous isotropic turbulence, a Taylor-Green vortex, and turbulent channel flow. Our proposed fully convolutional autoencoder architecture compresses turbulent flow snapshots by a factor of 64 with a single pass, allows for arbitrarily sized input fields, is cheaper to compute than the randomized single-pass SVD for typical simulation sizes, performs well on unseen flow configurations, and has been made publicly available. The results reported here show that the autoencoder dramatically outperforms a randomized single-pass SVD with similar compression ratio and yields comparable performance to a higher-rank decomposition with an order of magnitude less compression in regard to preserving a number of important statistical quantities such as turbulent kinetic energy, enstrophy, and Reynolds stresses.

King, Ryan↗
DOE OSTI2022

Tomo2Mesh: Fast Reconstruction and Visualization of Tomography Data in Mesh Format

Tomo2Mesh is an open-source project targeted towards real-time reconstruction, segmentation, and visualization of computed tomography (CT) data in mesh format. The CT reconstruction scheme is based on filtered back-projection of voxel subsets. The segmentation scheme uses a 3D convolutional neural network. To allow for fast, real-time reconstruction, voxel subsets are first identified by coarse reconstruction. The detail in specific regions of interest is improved through full reconstruction of voxel subsets in that region. Data structures are implemented to store and process voxel subsets. Finally, voxel data is labeled using connected components to detect disconnected regions such as voids whose morphological attributes (e.g., Feret diameter, principal axis orientation, local number density, size, etc.) can be measured also in real-time. Finally, a fast marching cubes implementation processes labeled voxel data into a triangular face mesh (vertices and faces) in .ply format for visualization in Paraview or other mesh visualization tools. The code provides a simple programming interface for detecting, classifying, and visualizing regions of interest based on morphology. For example, detected voids can be classified as round pores or extended cracks. Highly porous neighborhoods can be identified based on local number density. The mesh texture (or color) is assigned based these morphological attributes to allow smart visualization scenarios in real-time (e.g., show only long cracks). At the time of first release (July 2022), extraction of face mesh for visualization for raw CT data from a 2 megapixel camera would take between 1-5 minutes for most scenarios.

TEKAWADE, ANIKET↗
DOE OSTI2022

labquake_future_prediction

The labquake_future_prediction code is a collection of python modules and scripts that serves as supporting information for the article “Predicting future laboratory fault friction through deep learning” for publication in the journal of “Geophysical Research Letters”. It is designed to predict laboratory fault slips in the immediate future by scanning continuous acoustic emission (AE) waveforms recorded in laboratory biaxial shear experiments. The predictions are made with a deep learning model based on convolutional encoder-decoder (CED) models and the Transformer model primarily developed for Natural Language Processing (NLP). The deep learning model is trained with the tensorflow package using publicly available laboratory data sets in standard binary file format in numpy. The utility functions for reading data files, configuring model hyperparameters, constructing the CED and Transformer models, training and testing of the models are defined in python module files. The workflow of training the models for labquake future predictions and the multiple GPU’s rapid model hyperparameter optimization as described in the journal article, are demonstrated in accompanying python script files and Jupyter notebooks.

Wang, Kun↗
DOE OSTI2022

InversionNet

InversionNet is a software to solve subsurface imaging problems. It leverages a convolutional neural network with an encoder-decoder structure to model the correspondence from seismic data to subsurface velocity structures.

Lin, Youzuo↗