Search NASA⌕ Search

SEARCH · Search NASA

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 73 records · Page 4

Convolutional neural networks on the HEALPix sphere: a pixel-based algorithm and its application to CMB data analysis

We describe a novel method for the application of convolutional neural networks (CNNs) to fields defined on the sphere, using the Hierarchical Equal Area Latitude Pixelization scheme (HEALPix). Specifically, we have developed a pixel-based approach to implement convolutional and pooling layers on the spherical surface, similarly to what is commonly done for CNNs applied to Euclidean space. The main advantage of our algorithm is to be fully integrable with existing, highly optimized libraries for NNs (e.g., PyTorch, TensorFlow, etc.). We present two applications of our method: (i) recognition of handwritten digits projected on the sphere; (ii) estimation of cosmological parameter from simulated maps of the cosmic microwave background (CMB). The latter represents the main target of this exploratory work, whose goal is to show the applicability of our CNN to CMB parameter estimation. We have built a simple NN architecture, consisting of four convolutional and pooling layers, and we have used it for all the applications explored herein. Concerning the recognition of handwritten digits, our CNN reaches an accuracy of ~95%, comparable with other existing spherical CNNs, and this is true regardless of the position and orientation of the image on the sphere. For CMB-related applications, we tested the CNN on the estimation of a mock cosmological parameter, defining the angular scale at which the power spectrum of a Gaussian field projected on the sphere peaks. We estimated the value of this parameter directly from simulated maps, in several cases: temperature and polarization maps, presence of white noise, and partially covered maps. For temperature maps, the NN performances are comparable with those from standard spectrum-based Bayesian methods. For polarization, CNNs perform about a factor four worse than standard algorithms. Nonetheless, our results demonstrate, for the first time, that CNNs are able to extract information from polarization fields, both in full-sky and masked maps, and to distinguish between E and B -modes in pixel space. Lastly, we have applied our CNN to the estimation of the Thomson scattering optical depth at reionization ( τ ) from simulated CMB maps. Even without any specific optimization of the NN architecture, we reach an accuracy comparable with standard Bayesian methods. This work represents a first step towards the exploitation of NNs in CMB parameter estimation and demonstrates the feasibility of our approach.

79 ASTRONOMY AND ASTROPHYSICS↗

FAST PIXEL SPACE CONVOLUTION FOR COSMIC MICROWAVE BACKGROUND SURVEYS WITH ASYMMETRIC BEAMS AND COMPLEX SCAN STRATEGIES: FEBeCoP

Precise measurement of the angular power spectrum of the cosmic microwave background (CMB) temperature and polarization anisotropy can tightly constrain many cosmological models and parameters. However, accurate measurements can only be realized in practice provided all major systematic effects have been taken into account. Beam asymmetry, coupled with the scan strategy, is a major source of systematic error in scanning CMB experiments such as Planck, the focus of our current interest. We envision Monte Carlo methods to rigorously study and account for the systematic effect of beams in CMB analysis. Toward that goal, we have developed a fast pixel space convolution method that can simulate sky maps observed by a scanning instrument, taking into account real beam shapes and scan strategy. The essence is to pre-compute the 'effective beams' using a computer code, 'Fast Effective Beam Convolution in Pixel space' (FEBeCoP), that we have developed for the Planck mission. The code computes effective beams given the focal plane beam characteristics of the Planck instrument and the full history of actual satellite pointing, and performs very fast convolution of sky signals using the effective beams. In this paper, we describe the algorithm and the computational scheme that has been implemented. We also outline a few applications of the effective beams in the precision analysis of Planck data, for characterizing the CMB anisotropy and for detecting and measuring properties of point sources.

79 ASTRONOMY AND ASTROPHYSICS↗

Information Extraction from Cancer Pathology Reports with Graph Convolution Networks for Natural Language Texts

Graph-of-words is a flexible and efficient text representation which addresses well-known challenges, such as word ordering and variation of expressions, to natural language processing. In this paper, we consider the latest graph-based convolutional neural network technique, the Text Graph Convolutional Network (Text GCN), in the context of performing classification tasks on free-form natural language texts. To do this, we designed a study of multi-task information extraction from medical text documents. We implemented multi-task learning in the Text GCN, performed hyperparameter optimization, and measured the clinical task performances. We evaluated micro and macro-F1 scores of four information extraction tasks, including subsite, laterality, behavior, and histological grades from cancer pathology reports. The scores for the Text GCN significantly outperformed our previous studies with convolutional neural networks, suggesting that the Text GCN model is superior to traditional models in task performance.

Yoon, Hong-Jun↗

Evolving Larger Convolutional Layer Kernel Sizes for a Settlement Detection Deep-Learner on Summit

Deep-learner hyper-parameters, such as kernel sizes, batch sizes, and learning rates, can significantly influence the quality of trained models. The state of the art for finding optimal hyper-parameters generally uses a brute force, grid search approach, random search, or Bayesian-based optimization among other techniques. We applied an evolutionary algorithm to optimize kernel sizes for a convolutional neural network used to detect settlements in satellite imagery. Usually convolutional layer kernel sizes are small - typically one, three, or five - but we found that the system converged at, or near, kernel sizes of nine for the last convolutional layer, and that this occurred for multiple runs using two different datasets. Moreover, the larger kernel sizes had fewer false positives than the 3x3 kernel sizes found as optimal via a brute force uniform grid search. This suggests that this large kernel size may be leveraging patterns found in larger areal features in the source imagery, and that this may be generalized as possible guidance for similar remote sensing deep-learning tasks.

Coletti, Mark↗

Algorithm-Based Fault Tolerance for Convolutional Neural Networks

Convolutional neural networks (CNNs) are becoming more and more important for solving challenging and critical problems in many fields. CNN inference applications have been deployed in safety-critical systems, which may suffer from soft errors caused by high-energy particles, high temperature, or abnormal voltage. Of critical importance is ensuring the stability of the CNN inference process against soft errors. Traditional fault tolerance methods are not suitable for CNN inference because error-correcting code is unable to protect computational components, instruction duplication techniques incur high overhead, and existing algorithm-based fault tolerance (ABFT) techniques cannot protect all convolution implementations. In this paper, we focus on how to protect the CNN inference process against soft errors as efficiently as possible, with the following three contributions. (1) We propose several systematic ABFT schemes based on checksum techniques and analyze their fault protection ability and runtime thoroughly. Unlike traditional ABFT based on matrix-matrix multiplication, our schemes support any convolution implementations. (2) We design a novel workflow integrating all the proposed schemes to obtain a high detection/correction ability with limited total runtime overhead. (3) We perform our evaluation using ImageNet with well-known CNN models including AlexNet, VGG-19, ResNet-18, and YOLOv2. Here, experimental results demonstrate that our implementation can handle soft errors with very limited runtime overhead (4%~8% in both error-free and error-injected situations).

97 MATHEMATICS AND COMPUTING↗

TH-E-BRE-03: A Novel Method to Account for Ion Chamber Volume Averaging Effect in a Commercial Treatment Planning System Through Convolution

Purpose: Fourier-based deconvolution approaches used to eliminate ion chamber volume averaging effect (VAE) suffer from measurement noise. This work aims to investigate a novel method to account for ion chamber VAE through convolution in a commercial treatment planning system (TPS). Methods: Beam profiles of various field sizes and depths of an Elekta Synergy were collected with a finite size ion chamber (CC13) to derive a clinically acceptable beam model for a commercial TPS (Pinnacle{sup 3}), following the vendor-recommended modeling process. The TPS-calculated profiles were then externally convolved with a Gaussian function representing the chamber (σ = chamber radius). The agreement between the convolved profiles and measured profiles was evaluated with a one dimensional Gamma analysis (1%/1mm) as an objective function for optimization. TPS beam model parameters for focal and extra-focal sources were optimized and loaded back into the TPS for new calculation. This process was repeated until the objective function converged using a Simplex optimization method. Planar dose of 30 IMRT beams were calculated with both the clinical and the re-optimized beam models and compared with MapCHEC™ measurements to evaluate the new beam model. Results: After re-optimization, the two orthogonal source sizes for the focal source reduced from 0.20/0.16 cm to 0.01/0.01 cm, which were the minimal allowed values in Pinnacle. No significant change in the parameters for the extra-focal source was observed. With the re-optimized beam model, average Gamma passing rate for the 30 IMRT beams increased from 92.1% to 99.5% with a 3%/3mm criterion and from 82.6% to 97.2% with a 2%/2mm criterion. Conclusion: We proposed a novel method to account for ion chamber VAE in a commercial TPS through convolution. The reoptimized beam model, with VAE accounted for through a reliable and easy-to-implement convolution and optimization approach, outperforms the original beam model in standard IMRT QA verification.

07 ISOTOPE AND RADIATION SOURCES↗

SU-E-T-423: Fast Photon Convolution Calculation with a 3D-Ideal Kernel On the GPU

Purpose: The calculation time is a trade-off for improving the accuracy of convolution dose calculation with fine calculation spacing of the KERMA kernel. We investigated to accelerate the convolution calculation using an ideal kernel on the Graphic Processing Units (GPU). Methods: The calculation was performed on the AMD graphics hardware of Dual FirePro D700 and our algorithm was implemented using the Aparapi that convert Java bytecode to OpenCL. The process of dose calculation was separated with the TERMA and KERMA steps. The dose deposited at the coordinate (x, y, z) was determined in the process. In the dose calculation running on the central processing unit (CPU) of Intel Xeon E5, the calculation loops were performed for all calculation points. On the GPU computation, all of the calculation processes for the points were sent to the GPU and the multi-thread computation was done. In this study, the dose calculation was performed in a water equivalent homogeneous phantom with 150{sup 3} voxels (2 mm calculation grid) and the calculation speed on the GPU to that on the CPU and the accuracy of PDD were compared. Results: The calculation time for the GPU and the CPU were 3.3 sec and 4.4 hour, respectively. The calculation speed for the GPU was 4800 times faster than that for the CPU. The PDD curve for the GPU was perfectly matched to that for the CPU. Conclusion: The convolution calculation with the ideal kernel on the GPU was clinically acceptable for time and may be more accurate in an inhomogeneous region. Intensity modulated arc therapy needs dose calculations for different gantry angles at many control points. Thus, it would be more practical that the kernel uses a coarse spacing technique if the calculation is faster while keeping the similar accuracy to a current treatment planning system.

60 APPLIED LIFE SCIENCES↗

Technical Note: Impact of the geometry dependence of the ion chamber detector response function on a convolution-based method to address the volume averaging effect

Purpose: To investigate the geometry dependence of the detector response function (DRF) of three commonly used scanning ionization chambers and its impact on a convolution-based method to address the volume averaging effect (VAE). Methods: A convolution-based approach has been proposed recently to address the ionization chamber VAE. It simulates the VAE in the treatment planning system (TPS) by iteratively convolving the calculated beam profiles with the DRF while optimizing the beam model. Since the convolved and the measured profiles are subject to the same VAE, the calculated profiles match the implicit “real” ones when the optimization converges. Three DRFs (Gaussian, Lorentzian, and parabolic function) were used for three ionization chambers (CC04, CC13, and SNC125c) in this study. Geometry dependent/independent DRFs were obtained by minimizing the difference between the ionization chamber-measured profiles and the diode-measured profiles convolved with the DRFs. These DRFs were used to obtain eighteen beam models for a commercial TPS. Accuracy of the beam models were evaluated by assessing the 20%–80% penumbra width difference (PWD) between the computed and diode-measured beam profiles. Results: The convolution-based approach was found to be effective for all three ionization chambers with significant improvement for all beam models. Up to 17% geometry dependence of the three DRFs was observed for the studied ionization chambers. With geometry dependent DRFs, the PWD was within 0.80 mm for the parabolic function and CC04 combination and within 0.50 mm for other combinations; with geometry independent DRFs, the PWD was within 1.00 mm for all cases. When using the Gaussian function as the DRF, accounting for geometry dependence led to marginal improvement (PWD < 0.20 mm) for CC04; the improvement ranged from 0.38 to 0.65 mm for CC13; for SNC125c, the improvement was slightly above 0.50 mm. Conclusions: Although all three DRFs were found adequate to represent the response of the studied ionization chambers, the Gaussian function was favored due to its superior overall performance. The geometry dependence of the DRFs can be significant for clinical applications involving small fields such as stereotactic radiotherapy.

60 APPLIED LIFE SCIENCES↗

First- and Second-Order Methods for Online Convolutional Dictionary Learning

Convolutional sparse representations are a form of sparse representation with a structured, translation-invariant dictionary. Most convolutional dictionary learning algorithms to date operate in batch mode, requiring simultaneous access to all training images during the learning process, which results in very high memory usage and severely limits the training data size that can be used. Very recently, however, a number of authors have considered the design of online convolutional dictionary learning algorithms that offer far better scaling of memory and computational cost with training set size than batch methods. This study extends our prior work, improving a number of aspects of our previous algorithm; proposing an entirely new one, with better performance, that supports the inclusion of a spatial mask for learning from incomplete data; and providing a rigorous theoretical analysis of these methods.

97 MATHEMATICS AND COMPUTING↗

PyTorch Implementation of Log-Additive Convolutional Neural Networks

This code is a collection of python code that defines, trains, and tests Log-Additive Convolutional Neural Networks. The model components and training routine are based on the PyTorch python library. The code implements the Log-Additive Convolutional Neural Networks as described in Pagendam et al. 2023. In addition to the Log-Additive Convolutional Neural Networks, this library also defines the Log-Normal Density loss function as described in Pagendam et al. 2023. Code from this paper is not publicly available, so the Pytorch implementation of this type of model is unique to this library.

Callis, Skylar↗

Fully Convolutional Spatio-Temporal Models for Representation Learning in Plasma Science

We have trained a fully convolutional spatio-temporal model for fast and accurate representation learning in the challenging exemplar application area of fusion energy plasma science. The onset of major disruptions is a critically important fusion energy science issue that must be resolved for advanced tokamak plasmas such as the $25B burning plasma international thermonuclear experimental reactor (ITER) experiment. While a variety of statistical methods have been used to address the problem of tokamak disruption prediction and control, recent approaches based on deep learning have proven particularly compelling. In the present paper, we introduce further improvements to the fusion recurrent neural network (FRNN) software suite, which delivered cross-machine disruption predictions with unprecedented accuracy using a large database of experimental signals from two major tokamaks. Up to now, FRNN was based on the long short-term memory (LSTM) variant of recurrent neural networks to leverage the temporal information in the data. Here, we implement and apply the "temporal convolutional neural network (TCN)" architecture to the time-dependent input signals. Furthermore, this allows highly optimized convolution operations to carry the majority of the computational load of training, thus enabling a reduction in training time, and the effective use of high-performance computing resources for hyperparameter tuning. At the same time, the TCN-based architecture achieves better predictive performance when compared with the LSTM architecture for various tasks for a representative fusion database.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Accelerated Time-Domain Modeling of Electromagnetic Pulse Excitation of Finite-Length Dissipative Conductors over a Ground Plane via Function Fitting and Recursive Convolution

In this report we overview the fundamental concepts for a pair of techniques which together greatly hasten computational predictions of electromagnetic pulse (EMP) excitation of finite-length dissipative conductors over a ground plane. In a time- domain, transmission line (TL) model implementation, predictions are computationally bottlenecked time-wise, either for late-time predictions (about 100ns-10000ns range) or predictions concerning EMP excitation of long TLs (order of kilometers or more ). This is because the method requires a temporal convolution to account for the losses in the ground. Addressing this to facilitate practical simulation of EMP excitation of TLs, we first apply a technique to extract an (approximate) complex exponential function basis-fit to the ground/Earth's impedance function, followed by incorporating this into a recursion-based convolution acceleration technique. Because the recursion-based method only requires the evaluation of the most recent voltage history data (versus the entire history in a "brute-force" convolution evaluation), we achieve necessary time speed- ups across a variety of TL/Earth geometry/material scenarios. Intentionally Left Blank

36 MATERIALS SCIENCE↗

Efficient convolutional sparse coding

Computationally efficient algorithms may be applied for fast dictionary learning solving the convolutional sparse coding problem in the Fourier domain. More specifically, efficient convolutional sparse coding may be derived within an alternating direction method of multipliers (ADMM) framework that utilizes fast Fourier transforms (FFT) to solve the main linear system in the frequency domain. Such algorithms may enable a significant reduction in computational cost over conventional approaches by implementing a linear solver for the most critical and computationally expensive component of the conventional iterative algorithm. The theoretical computational cost of the algorithm may be reduced from O(M.sup.3N) to O(MN log N), where N is the dimensionality of the data and M is the number of elements in the dictionary. This significant improvement in efficiency may greatly increase the range of problems that can practically be addressed via convolutional sparse representations.

97 MATHEMATICS AND COMPUTING↗

Harnessing graph convolutional neural networks for identification of glassy states in metallic glasses

Graph Convolutional Neural Networks (GCNNs) have emerged as powerful tools for analyzing materials. In this study, we employ GCNNs to examine structural characteristics of CuZr metallic glasses (MGs) and identify their states. We use molecular dynamics to simulate the quenching process of CuZr, using cooling rates ranging from 10 9 to 10 15 K/s, to produce six unique glassy states. For each state, we create a dataset comprising 1,800 distinct samples. We evaluate the effectiveness of various GCNNs, including Graph Attention Neural Network (GANN), Graph Sample and AggreGatE (GraphSAGE), Graph Isomorphism Network (GIN), and Relational Graph Convolutional Neural Network (RGCN). GANN and GraphSAGE demonstrate comparable performance, achieving an overall accuracy of 81% in classifying the MG states. Furthermore, these results underscore the potential of GCNNs to detect subtle structural variances in disordered materials and point to broader application of deep learning in the analysis of MGs and other amorphous substances.

36 MATERIALS SCIENCE↗

Convoluted filtering for process cycle modeling

Principles of materials science and engineering, physics, mathematics, and information science are used to extract knowledge and insights from the process-structure–property-performance relationships hidden in materials data. The process-structure modeling can be accelerated without loss of interpretability, with artificial intelligence tools that mimic the salient features of the process and process-structure relations. In this work, a novel convoluted model-filtering technique was exploited to build and successfully train the Convoluted Filter (CoFi) artifacts for Fe-based alloy heat treatment cycles. The artifacts were pre-trained to filter out deep models that change the surrogate microstructure state after the heat treatment at ambient conditions. Direct representation of the thermal cycle features within knowledge Graph facilitated development of meaningful data models for microstructure evolution, which reduce overfitting to limited datasets.

36 MATERIALS SCIENCE↗

Multiresolution convolutional autoencoders

Herein we propose a multi-resolution convolutional autoencoder (MrCAE) architecture that integrates and leverages three highly successful mathematical architectures: (i) multigrid methods, (ii) convolutional autoencoders and (iii) transfer learning. The method provides an adaptive, hierarchical architecture that capitalizes on a progressive training approach for multiscale spatio-temporal data. This framework allows for inputs across multiple scales: starting from a compact (small number of weights) network architecture and low-resolution data, our network progressively deepens and widens itself in a principled manner to encode new information in the higher resolution data based on its current performance of reconstruction. Basic transfer learning techniques are applied to ensure information learned from previous training steps can be rapidly transferred to the larger network. As a result, the network can dynamically capture different scaled features at different depths of the network. The performance gains of this adaptive multiscale architecture are illustrated through a sequence of numerical experiments on synthetic examples and real-world spatial-temporal data.

97 MATHEMATICS AND COMPUTING↗

Upsampling Monte Carlo reactor simulation tallies in depleted LWR assemblies fueled with LEU and HALEU using a convolutional neural network

Simulating nuclear reactor cores at the highest achievable spatial and energy resolution is critical in modeling these systems accurately. Increasing the resolution, however, can dramatically increase the memory and central processing unit time required to run simulations. A convolutional neural network was shown previously to accurately upsample tally results of simulated light water reactor assemblies fueled with fresh, low enriched uranium. Here, we show that a convolutional neural network can be used to upsample tally results in assemblies containing fresh and depleted fuel enriched from 1.6 to 19.9 atom percent. The network was trained using neutron flux tallies from simulations of light water reactor assemblies with a range of fuel and coolant temperatures and a diverse selection of geometries. Accurate predictions of flux tallies are possible even on test assemblies with geometries and burnup levels well outside the range of those present in the training and validation data. The network improves the data density by a factor of 8 over a broad range of light water reactor assemblies while incurring insignificant additional computational cost to a Monte Carlo simulation.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Solving the Orszag–Tang vortex magnetohydrodynamics problem with physics-constrained convolutional neural networks

We study the 2D Orszag–Tang vortex magnetohydrodynamics (MHD) problem through the use of physics-constrained convolutional neural networks (PCNNs) for forecasting the density, ρ, and the magnetic field, B, as well as the prediction of B given the velocity field v of the fluid. In addition to translation equivariance from the convolutional architecture, other physics constraints were embedded: absence of magnetic monopoles, non-negativity of ρ, use of only relevant variables, and the periodic boundary conditions of the problem. The use of only relevant variables and the hard constraint of non-negative ρ were found to facilitate learning greatly. The divergenceless condition ∇·B=0 was implemented as a hard constraint up to machine precision through the use of a magnetic potential to define B=∇×A. Residual networks and data augmentation were also used to improve performance. This allowed for some of the residual models to function as surrogate models and provide reasonably accurate simulations. For the prediction task, the PCNNs were evaluated against a physics-informed neural network, which had the ideal MHD induction equation as a soft constraint. Several models were able to generate highly accurate fields, which are visually almost indistinguishable and have low mean squared error. Only methods with built-in hard constraints produced physical fields with ∇·B=0. The use of PCNNs for MHD has the potential to produce physically consistent real-time simulations to serve as virtual diagnostics in cases where inferences must be made with limited observables.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗