Microstructural classification of unirradiated LiAlO2 pellets by deep learning methods
Not Available
Pazdernik, Karl (ORCID:0000000314104006)↗
SEARCH · Search NASA
Results for “classification algorithms”
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.
Efficient Active Learning for Gaussian Process Classification by Error Reduction
Active learning sequentially selects the best instance for labeling by optimizing an acquisition function to enhance data/label efficiency. The selection can be either from a discrete instance set (pool-based scenario) or a continuous instance space (query synthesis scenario). In this work, we study both active learning scenarios for Gaussian Process Classification (GPC). The existing active learning strategies that maximize the Estimated Error Reduction (EER) aim at reducing the classification error after training with the new acquired instance in a onestep-look-ahead manner. The computation of EER-based acquisition functions is typically prohibitive as it requires retraining the GPC with every new query. Moreover, as the EER is not smooth, it can not be combined with gradient-based optimization techniques to efficiently explore the continuous instance space for query synthesis. To overcome these critical limitations, we develop computationally efficient algorithms for EER-based active learning with GPC. Further, we derive the joint predictive distribution of label pairs as a one-dimensional integral, as a result of which the computation of the acquisition function avoids retraining the GPC for each query, remarkably reducing the computational overhead. We also derive the gradient chain rule to efficiently calculate the gradient of the acquisition function, which leads to the first query synthesis active learning algorithm implementing EER-based strategies. Our experiments clearly demonstrate the computational efficiency of the proposed algorithms. We also benchmark our algorithms on both synthetic and real-world datasets, which show superior performance in terms of sampling efficiency compared to the existing state-of-the-art algorithms.
97 MATHEMATICS AND COMPUTING↗
A new computer approach to mixed feature classification for forestry application
A computer approach for mapping mixed forest features (i.e., types, classes) from computer classification maps is discussed. Mixed features such as mixed softwood/hardwood stands are treated as admixtures of softwood and hardwood areas. Large-area mixed features are identified and small-area features neglected when the nominal size of a mixed feature can be specified. The computer program merges small isolated areas into surrounding areas by the iterative manipulation of the postprocessing algorithm that eliminates small connected sets. For a forestry application, computer-classified LANDSAT multispectral scanner data of the Sam Houston National Forest were used to demonstrate the proposed approach. The technique was successful in cleaning the salt-and-pepper appearance of multiclass classification maps and in mapping admixtures of softwood areas and hardwood areas. However, the computer-mapped mixed areas matched very poorly with the ground truth because of inadequate resolution and inappropriate definition of mixed features.
Kan, E. P.↗
Detecting technological maturity from bibliometric patterns
We report the capability to identify emergent technologies based upon easily accessed open-source indicators, such as publications, is important for decision-makers in industry and government. The scientific contribution of this work is the proposition of a machine learning approach to the detection of the maturity of emerging technologies based on publication counts. Time-series of publication counts have universal features that distinguish emerging and growing technologies. We train an artificial neural network classifier, a supervised machine learning algorithm, upon these features to predict the maturity (emergent vs. growth) of an arbitrary technology. With a training set comprised of 22 technologies we obtain a classification accuracy ranging from 58.3% to 100% with an average accuracy of 84.6% for six test technologies. To enhance classifier performance, we augmented the training corpus with synthetic time-series technology life cycle curves, formed by calculating weighted averages of curves in the original training set. Training the classifier on the synthetic data set resulted in improved accuracy, ranging from 83.3% to 100% with an average accuracy of 90.4% for the test technologies. The performance of our classifier exceeds that of competing machine learning approaches in the literature, which report an average classification accuracy of only 85.7% at maximum. Moreover, in contrast to current methods our approach does not require subject matter expertise to generate training labels, and it can be automated and scaled.
97 MATHEMATICS AND COMPUTING↗
Lidar Cloud Detection with Fully Convolutional Networks
In this contribution, we present a novel approach for segmenting laser radar (lidar) imagery into geometric time- height cloud locations with a fully convolutional network (FCN). We describe a semi-supervised learning method to train the FCN by: pre-training the classification layers of the FCN with image-level annotations, pre-training the en- tire FCN with the cloud locations of the MPLCMASK cloud mask algorithm, and fully supervised learning with hand-labeled cloud locations. We show the model achieves higher levels of cloud identification compared to the cloud mask algorithm implementation.
Cromwell, Erol LD↗
Explainable Neural Architecture Search (XNAS)
Code for the paper Learning Interpretable Models Through Multi-Objective Neural Architecture Search by Zachariah Carmichael, Tim Moon, and Sam Ade Jacobs. Monumental advances in deep learning have led to unprecedented achievements across a multitude of domains. While the performance of deep neural networks is indubitable, the architectural design and interpretability of such models are nontrivial. Research has been introduced to automate the design of neural network architectures through neural architecture search (NAS). Recent progress has made these methods more pragmatic by exploiting distributed computation and novel optimization algorithms. However, there is little work in optimizing architectures for interpretability. To this end, we propose a multiobjective distributed NAS framework that optimizes for both task performance and introspection. We leverage the non-dominated sorting genetic algorithm (NSGA-II) and explainable AI (XAI) techniques to reward architectures that can be better comprehended by humans. The framework is evaluated on several image classification datasets. We demonstrate that jointly optimizing for introspection ability and task error leads to more disentangled architectures that perform within tolerable error.
Carmichael, ZachariahJ↗
HD-Bind: Encoding of Molecular Structure with Low Precision, Hyperdimensional Binary Representations
Publicly available collections of drug-like molecules have grown to comprise tens of billions of compounds due to advances in combinatorial chemistry. Traditional methods for identifying "hit" molecules from a large collection of potential drug-like candidates have relied on biophysical theory to compute approximations to the Gibbs free energy of the binding interaction between the drug and its protein target. These approaches have the major drawback that they require exceptional computing capabilities for even relatively small collections of molecules. Hyperdimensional Computing (HDC) is a recently-proposed learning paradigm that represents data with high-dimension binary vectors; this allows the use of low-precision binary vector arithmetic to create models of the data that can be learned without the need for the gradient-based optimization required in many conventional machine learning and deep learning methods. This algorithmic simplicity allows for acceleration in hardware that has been previously demonstrated in a range of application areas. We consider existing HDC approaches for molecular property classification and introduce two novel encodings of a commonly-used molecular representation, the extended connectivity fingerprint (ECFP). We show that HDC-based inference methods are as much as 91 times more efficient than traditional machine learning methods, and achieve an acceleration of nearly nine orders of magnitude compared to molecular docking. Our results show that HDC accelerated methods retain competitive accuracy on a number of well-studied tasks such as molecular property predictions using the MoleculeNet dataset, and bind/no-bind activity classification using the DUD-E and LIT-PCBA datasets. Our work thus motivates further investigation into molecular representation learning to develop ultraefficient pre-screening tools.
Jones, William↗
Passband Signal Detection at the Edge
Algorithms for radio frequency (RF) spectrum awareness need to be compatible with edge hardware to be practical for many applications. We developed a signal detection and classification model for the ZCU111 RF System-on-a-Chip (RFSoC) that operates on the fast Fourier transform of passband RF data. The system can detect and classify multiple signals of interest and display the predictions in real-time. The model consists of a modified ConvNeXt backbone and YOLOv3 head to operate on the Deep Learning Processing Unit on the RFSoC. We gathered datasets for training and testing by using a software defined radio to transmit example signals of Wi-Fi 802.11 b/g, Wi-Fi 802.11 n, FM Radio, LTE and LTE-M. By leveraging multiple inputs on the RFSoC frontend, the datasets span up to 4 GHz of bandwidth. The models showed high performance in classification accuracy, center frequency error, bandwidth error, and detection accuracy for both single and multi-signal datasets.
42 ENGINEERING↗
Classification of multispectral image data by the Binary Diamond neural network and by nonparametric, pixel-by-pixel methods
The classification of multispectral image data obtained from satellites has become an important tool for generating ground cover maps. This study deals with the application of nonparametric pixel-by-pixel classification methods in the classification of pixels, based on their multispectral data. A new neural network, the Binary Diamond, is introduced, and its performance is compared with a nearest neighbor algorithm and a back-propagation network. The Binary Diamond is a multilayer, feed-forward neural network, which learns from examples in unsupervised, 'one-shot' mode. It recruits its neurons according to the actual training set, as it learns. The comparisons of the algorithms were done by using a realistic data base, consisting of approximately 90,000 Landsat 4 Thematic Mapper pixels. The Binary Diamond and the nearest neighbor performances were close, with some advantages to the Binary Diamond. The performance of the back-propagation network lagged behind. An efficient nearest neighbor algorithm, the binned nearest neighbor, is described. Ways for improving the performances, such as merging categories, and analyzing nonboundary pixels, are addressed and evaluated.
Salu, Yehuda↗
Landsat Science: 40 Years of Innovation and Opportunity
Landsat satellites have provided unparalleled Earth-observing data for nearly 40 years, allowing scientists to describe, monitor and model the global environment during a period of time that has seen dramatic changes in population growth, land use, and climate. The success of the Landsat program can be attributed to well-designed instrument specifications, astute engineering, comprehensive global acquisition and calibration strategies, and innovative scientists who have developed analytical techniques and applications to address a wide range of needs at local to global scales (e.g., crop production, water resource management, human health and environmental quality, urbanization, deforestation and biodiversity). Early Landsat contributions included inventories of natural resources and land cover classification maps, which were initially prepared by a visual interpretation of Landsat imagery. Over time, advances in computer technology facilitated the development of sophisticated image processing algorithms and complex ecosystem modeling, enabling scientists to create accurate, reproducible, and more realistic simulations of biogeochemical processes (e.g., plant production and ecosystem dynamics). Today, the Landsat data archive is freely available for download through the USGS, creating new opportunities for scientists to generate global image datasets, develop new change detection algorithms, and provide products in support of operational programs such as Reducing Emissions from Deforestation and Forest Degradation in Developing Countries (REDD). In particular, the use of dense (approximately annual) time series to characterize both rapid and progressive landscape change has yielded new insights into how the land environment is responding to anthropogenic and natural pressures. The launch of the Landsat Data Continuity Mission (LDCM) satellite in 2012 will continue to propel innovative Landsat science.
Cook, Bruce D.↗
Cloud-Based Solutions for Monitoring Coastal Ecosystems and Prioritization of Restoration Efforts Across Belize
In recent years the availability of automated change detection algorithms in Google Earth Engine has permitted cloud-based processing of large time series of satellite imagery. Models such as the Continuous Change Detection and Classification (CCDC), CCDC-Spectral Mixture Analysis (CCDC-SMA), and Landsat-based Detection of Trends in Disturbance and Recovery (LandTrendr) allow users to exploit decades of Earth Observations (EO) , leveraging the Landsat archive and data from other sensors to detect disturbance in forest ecosystems. Despite the wide adoption of these methods, robust documentation and growing community of users, little research has explored their use in mangrove environments. Mangroves are dynamic environments subject to changes due to not only the natural migration of mudflats but also coastal erosion, urban expansion, aquaculture practices, etc. This work aims to identify best practices for the application of these models to identify and monitor changes in the Belizean mangroves, which experienced an estimated 5.4% decrease in extent between 1980 and 2017 (Cherrington et al. 2020). Partnering directly with the Belizean Forest Department, our team will develop a replicable, efficient methodology to annually update the country’s mangrove extent employing EO-based change detection. This collaboration supports Belize’s marine conservation and climate resilience efforts, such as its Blue Bond agreement with the Nature Conservancy, building on national processes to monitor these critical ecosystems.
Mangroves↗
cTULIP: application of a human-based RNA-seq primary tumor classification tool for cross-species primary tumor classification in canine
The domestic dog, Canis familiaris, is quickly gaining traction as an advantageous model for use in the study of cancer, one of the leading causes of death worldwide. Naturally occurring canine cancers share clinical, histological, and molecular characteristics with the corresponding human diseases. In this study, we take a deep-learning approach to test how similar the gene expression profile of canine glioma and bladder cancer (BLCA) tumors are to the corresponding human tumors. We likewise develop a tool for identifying misclassified or outlier samples in large canine oncological datasets, analogous to that which was developed for human datasets. We test a number of machine learning algorithms and found that a convolutional neural network outperformed logistic regression and random forest approaches. We use a recently developed RNA-seq-based convolutional neural network, TULIP, to test the robustness of a human-data-trained primary tumor classification tool on cross-species primary tumor prediction. Our study ultimately highlights the molecular similarities between canine and human BLCA and glioma tumors, showing that protein-coding one-to-one homologs shared between humans and canines, are sufficient to distinguish between BLCA and gliomas. The results of this study indicate that using protein-coding one-to-one homologs as the features in the input layer of TULIP performs good primary tumor prediction in both humans and canines. Furthermore, our analysis shows that our selected features also contain the majority of features with known clinical relevance in BLCA and gliomas. Our success in using a human-data-trained model for cross-species primary tumor prediction also sheds light on the conservation of oncological pathways in humans and canines, further underscoring the importance of the canine model system in the study of human disease.
60 APPLIED LIFE SCIENCES↗
RU-net for automatic characterization of TRISO fuel cross sections
During irradiation, phenomena such as kernel swelling and buffer densification may impact the performance of tristructural isotropic (TRISO) particle fuel. Post-irradiation microscopy is often used to identify these irradiation-induced morphologic changes. However, each fuel compact generally contains thousands of TRISO particles. Manually performing the work to get statistical information on these phenomena is cumbersome and subjective. Here, to reduce the subjectivity inherent in that process and to accelerate data analysis, we used convolutional neural networks (CNNs) to automatically segment cross-sectional images of microscopic TRISO layers. CNNs are a class of machine-learning algorithms specifically designed for processing structured grid data. They 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 generated a large irradiated TRISO layer dataset with more than 2,000 microscopic images of cross-sectional TRISO particles and the corresponding annotated images. Based on these annotated images, we used different CNNs to automatically segment different TRISO layers. These CNNs 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 performs best in terms of Intersection over Union (IoU). Using 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.
11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
Reliable photometric membership (RPM) of galaxies in clusters – I. A machine learning method and its performance in the local universe
ABSTRACT We introduce a new method to determine galaxy cluster membership based solely on photometric properties. We adopt a machine learning approach to recover a cluster membership probability from galaxy photometric parameters and finally derive a membership classification. After testing several machine learning techniques (such as stochastic gradient boosting, model averaged neural network and k-nearest neighbours), we found the support vector machine algorithm to perform better when applied to our data. Our training and validation data are from the Sloan Digital Sky Survey main sample. Hence, to be complete to $M_r^* + 3$, we limit our work to 30 clusters with $z$phot-cl ≤ 0.045. Masses (M200) are larger than $\sim 0.6\times 10^{14} \, \mathrm{M}_{\odot }$ (most above $3\times 10^{14} \, \mathrm{M}_{\odot }$). Our results are derived taking in account all galaxies in the line of sight of each cluster, with no photometric redshift cuts or background corrections. Our method is non-parametric, making no assumptions on the number density or luminosity profiles of galaxies in clusters. Our approach delivers extremely accurate results (completeness, C $\sim 92{\rm{ per\ cent}}$ and purity, P $\sim 87{\rm{ per\ cent}}$) within R200, so that we named our code reliable photometric membership. We discuss possible dependencies on magnitude, colour, and cluster mass. Finally, we present some applications of our method, stressing its impact to galaxy evolution and cosmological studies based on future large-scale surveys, such as eROSITA, EUCLID, and LSST.
Lopes, Paulo A. A.↗
PERSIANN Dynamic Infrared–Rain Rate (PDIR-Now): A Near-Real-Time, Quasi-Global Satellite Precipitation Dataset
This study presents the Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks–Dynamic Infrared Rain Rate (PDIR-Now) near-real-time precipitation dataset. This dataset provides hourly, quasi-global, infrared-based precipitation estimates at 0.04° × 0.04° spatial resolution with a short latency (15–60 min). It is intended to supersede the PERSIANN–Cloud Classification System (PERSIANN-CCS) dataset previously produced as the near-real-time product of the PERSIANN family. We first provide a brief description of the algorithm’s fundamentals and the input data used for deriving precipitation estimates. Second, we provide an extensive evaluation of the PDIR-Now dataset over annual, monthly, daily, and subdaily scales. Last, the article presents information on the dissemination of the dataset through the Center for Hydrometeorology and Remote Sensing (CHRS) web-based interfaces. The evaluation, conducted over the period 2017–18, demonstrates the utility of PDIR-Now and its improvement over PERSIANN-CCS at all temporal scales. Specifically, PDIR-Now improves the estimation of rain/no-rain days as demonstrated by a critical success index (CSI) of 0.53 compared to 0.47 of PERSIANN-CCS. In addition, PDIR-Now improves the estimation of seasonal and diurnal cycles of precipitation as well as regional precipitation patterns erroneously estimated by PERSIANN-CCS. Finally, an evaluation is carried out to examine the performance of PDIR-Now in capturing two extreme events, Hurricane Harvey and a cluster of summer thunderstorms that occurred over the Netherlands, where it is shown that PDIR-Now adequately represents spatial precipitation patterns as well as subdaily precipitation rates with a correlation coefficient (CORR) of 0.64 for Hurricane Harvey and 0.76 for the Netherlands thunderstorms.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
tRNA functional signatures classify plastids as late-branching cyanobacteria
Eukaryotes acquired the trait of oxygenic photosynthesis through endosymbiosis of the cyanobacterial progenitor of plastid organelles. Despite recent advances in the phylogenomics of Cyanobacteria, the phylogenetic root of plastids remains controversial. Although a single origin of plastids by endosymbiosis is broadly supported, recent phylogenomic studies are contradictory on whether plastids branch early or late within Cyanobacteria. One underlying cause may be poor fit of evolutionary models to complex phylogenomic data. Using Posterior Predictive Analysis, we show that recently applied evolutionary models poorly fit three phylogenomic datasets curated from cyanobacteria and plastid genomes because of heterogeneities in both substitution processes across sites and of compositions across lineages. To circumvent these sources of bias, we developed CYANO-MLP, a machine learning algorithm that consistently and accurately phylogenetically classifies (“phyloclassifies”) cyanobacterial genomes to their clade of origin based on bioinformatically predicted function-informative features in tRNA gene complements. Classification of cyanobacterial genomes with CYANO-MLP is accurate and robust to deletion of clades, unbalanced sampling, and compositional heterogeneity in input tRNA data. CYANO-MLP consistently classifies plastid genomes into a late-branching cyanobacterial sub-clade containing single-cell, starch-producing, nitrogen-fixing ecotypes, consistent with metabolic and gene transfer data. Phylogenomic data of cyanobacteria and plastids exhibit both site-process heterogeneities and compositional heterogeneities across lineages. These aspects of the data require careful modeling to avoid bias in phylogenomic estimation. Furthermore, we show that amino acid recoding strategies may be insufficient to mitigate bias from compositional heterogeneities. However, the combination of our novel tRNA-specific strategy with machine learning in CYANO-MLP appears robust to these sources of bias with high accuracy in phyloclassification of cyanobacterial genomes. CYANO-MLP consistently classifies plastids as late-branching Cyanobacteria, consistent with independent evidence from signature-based approaches and some previous phylogenetic studies.
59 BASIC BIOLOGICAL SCIENCES↗
The Spectroscopic Data Processing Pipeline for the Dark Energy Spectroscopic Instrument
Abstract We describe the spectroscopic data processing pipeline of the Dark Energy Spectroscopic Instrument (DESI), which is conducting a redshift survey of about 40 million galaxies and quasars using a purpose-built instrument on the 4 m Mayall Telescope at Kitt Peak National Observatory. The main goal of DESI is to measure with unprecedented precision the expansion history of the universe with the baryon acoustic oscillation technique and the growth rate of structure with redshift space distortions. Ten spectrographs with three cameras each disperse the light from 5000 fibers onto 30 CCDs, covering the near-UV to near-infrared (3600–9800 Å) with a spectral resolution ranging from 2000 to 5000. The DESI data pipeline generates wavelength- and flux-calibrated spectra of all the targets, along with spectroscopic classifications and redshift measurements. Fully processed data from each night are typically available to the DESI collaboration the following morning. We give details about the pipeline’s algorithms, and provide performance results on the stability of the optics, the quality of the sky background subtraction, and the precision and accuracy of the instrumental calibration. This pipeline has been used to process the DESI Survey Validation data set, and has exceeded the project’s requirements for redshift performance, with high efficiency and a purity greater than 99% for all target classes.
79 ASTRONOMY AND ASTROPHYSICS↗
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↗