Search NASA⌕ Search

SEARCH · Search NASA

Results for “unsupervised”

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 847 records · Page 47

Using Dynamic Quantum Clustering to Analyze Hierarchically Heterogeneous Samples on the Nanoscale

Dynamic Quantum Clustering (DQC) is an unsupervised, high visual data mining technique. DQC was tested as an analysis method for X-ray Absorption Near Edge Structure (XANES) data from the Transmission X-ray Microscopy (TXM) group. The TXM group images hierarchically heterogeneous materials with nanoscale resolution and large field of view. XANES data consists of energy spectra for each pixel of an image. It was determined that DQC successfully identifies structure in data of this type without prior knowledge of the components in the sample. Clusters and sub-clusters clearly reflected features of the spectra that identified chemical component, chemical environment, and density in the image. DQC can also be used in conjunction with the established data analysis technique, which does require knowledge of components present.

99 GENERAL AND MISCELLANEOUS↗

Automated UF 6 Cylinder Enrichment Assay: Status of the Hybrid Enrichment Verification Array (HEVA) Project: POTAS Phase II

Pacific Northwest National Laboratory (PNNL) intends to automate the UF 6 cylinder nondestructive assay (NDA) verification currently performed by the International Atomic Energy Agency (IAEA) at enrichment plants. PNNL is proposing the installation of a portal monitor at a key measurement point to positively identify each cylinder, measure its mass and enrichment, store the data along with operator inputs in a secure database, and maintain continuity of knowledge on measured cylinders until inspector arrival. This report summarizes the status of the research and development of an enrichment assay methodology supporting the cylinder verification concept. The enrichment assay approach exploits a hybrid of two passively-detected ionizing-radiation signatures: the traditional enrichment meter signature (186-keV photon peak area) and a non-traditional signature, manifested in the high-energy (3 to 8 MeV) gamma-ray continuum, generated by neutron emission from UF 6 . PNNL has designed, fabricated, and field-tested several prototype assay sensor packages in an effort to demonstrate proof-of-principle for the hybrid assay approach, quantify the expected assay precision for various categories of cylinder contents, and assess the potential for unsupervised deployment of the technology in a portal-monitor form factor. We refer to recent sensor-package prototypes as the Hybrid Enrichment Verification Array (HEVA). The report provides an overview of the assay signatures and summarizes the results of several HEVA field measurement campaigns on populations of Type 30B UF 6 cylinders containing low-enriched uranium (LEU), natural uranium (NU), and depleted uranium (DU). Approaches to performance optimization of the assay technique via radiation transport modeling are briefly described, as are spectroscopic and data-analysis algorithms.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

RolX: Role Extraction and Mining in Large Networks

Given a graph, how can we automatically discover roles for nodes? Roles could be, eg., ’bridges’, or ’periphery-nodes’, etc. Roles are compact summaries of a node’s behavior that generalize across networks. They enable numerous novel and useful network mining tasks, such as sense-making, searching for similar nodes, and node classification. We propose RolX (Role eXtraction), a scalable (linear in the number of edges), unsupervised learning approach for automatically extracting roles from general network data. We demonstrate the effectiveness of RolX on several network mining tasks, from exploratory data analysis to network transfer learning. Moreover, we compare network role discovery with network community discovery. We highlight fundamental differences between the two (e.g., roles generalize across disconnected networks, communities do not).

97 MATHEMATICS AND COMPUTING↗

Expectation-maximization algorithms for learning a finite mixture of univariate survival time distributions from partially specified class values

Heterogeneity exists on a data set when samples from di erent classes are merged into the data set. Finite mixture models can be used to represent a survival time distribution on heterogeneous patient group by the proportions of each class and by the survival time distribution within each class as well. The heterogeneous data set cannot be explicitly decomposed to homogeneous subgroups unless all the samples are precisely labeled by their origin classes; such impossibility of decomposition is a barrier to overcome for estimating nite mixture models. The expectation-maximization (EM) algorithm has been used to obtain maximum likelihood estimates of nite mixture models by soft-decomposition of heterogeneous samples without labels for a subset or the entire set of data. In medical surveillance databases we can find partially labeled data, that is, while not completely unlabeled there is only imprecise information about class values. In this study we propose new EM algorithms that take advantages of using such partial labels, and thus incorporate more information than traditional EM algorithms. We particularly propose four variants of the EM algorithm named EM-OCML, EM-PCML, EM-HCML and EM-CPCML, each of which assumes a specific mechanism of missing class values. We conducted a simulation study on exponential survival trees with five classes and showed that the advantages of incorporating substantial amount of partially labeled data can be highly signi cant. We also showed model selection based on AIC values fairly works to select the best proposed algorithm on each specific data set. A case study on a real-world data set of gastric cancer provided by Surveillance, Epidemiology and End Results (SEER) program showed a superiority of EM-CPCML to not only the other proposed EM algorithms but also conventional supervised, unsupervised and semi-supervised learning algorithms.

97 MATHEMATICS AND COMPUTING↗

Encoding and Analyzing Aerial Imagery Using Geospatial Semantic Graphs

While collection capabilities have yielded an ever-increasing volume of aerial imagery, analytic techniques for identifying patterns in and extracting relevant information from this data have seriously lagged. The vast majority of imagery is never examined, due to a combination of the limited bandwidth of human analysts and limitations of existing analysis tools. In this report, we describe an alternative, novel approach to both encoding and analyzing aerial imagery, using the concept of a geospatial semantic graph. The advantages of our approach are twofold. First, intuitive templates can be easily specified in terms of the domain language in which an analyst converses. These templates can be used to automatically and efficiently search large graph databases, for specific patterns of interest. Second, unsupervised machine learning techniques can be applied to automatically identify patterns in the graph databases, exposing recurring motifs in imagery. We illustrate our approach using real-world data for Anne Arundel County, Maryland, and compare the performance of our approach to that of an expert human analyst.

97 MATHEMATICS AND COMPUTING↗

PANTHER: Trajectory Analysis

We want to organize a body of trajectories in order to identify, search for, classify and predict behavior among objects such as aircraft and ships. Existing comparison functions such as the Fréchet distance are computationally expensive and yield counterintuitive results in some cases. We propose an approach using feature vectors whose components represent succinctly the salient information in trajectories. These features incorporate basic information such as total distance traveled and distance between start/stop points as well as geometric features related to the properties of the convex hull, trajectory curvature and general distance geometry. Additionally, these features can generally be mapped easily to behaviors of interest to humans that are searching large databases. Most of these geometric features are invariant under rigid transformation. We demonstrate the use of different subsets of these features to identify trajectories similar to an exemplar, cluster a database of several hundred thousand trajectories, predict destination and apply unsupervised machine learning algorithms.

42 ENGINEERING↗

Big-Data Multi-Energy Iterative Volumetric Reconstruction Methods for As-Built Validation & Verification Applications

This document archives the results developed by the Lab Directed Research and Development (LDRD) project sponsored by Sandia National Laboratories (SNL). In this work, it is shown that SNL has developed the first known high-energy hyperspectral computed tomography system for industrial and security applications. The main results gained from this work include dramatic beam-hardening artifact reduction by using the hyperspectral reconstruction as a bandpass filter without the need for any other computation or pre-processing; additionally, this work demonstrated the ability to use supervised and unsupervised learning methods on the hyperspectral reconstruction data for the application of materials characterization and identification which is not possible using traditional computed tomography systems or approaches.

47 OTHER INSTRUMENTATION↗

A Scalable HPC Insider Threat Monitoring System

Cyber Equations, Inc. is reporting on the results of its DOE SBIR Phase I project for the development of a Scalable HPC Insider Threat Monitoring System. This final report to the Department of Energy for Task DE-FOA-0001618 covers the period from 02/21/2017 to 02/20/2018 and summarizes the larger research accomplishments, which can be divided in theoretical and laboratory experiments. One of the many challenges for the U.S. Department of Energy (DOE) is the support of the high-performance computing (HPC) centers while providing quality information services, strategic vision, leadership, technical expertise, and capabilities necessary to enable the DOE to deliver effective, efficient, economical, and secure HPC services. In order to realize a secure DOE laboratory enterprise, DOE facilities operations must counter a wide range of cyber insider threats, from unsophisticated adversaries up to advanced, nation state sponsored actors. The DOE laboratories also presents an enormous insider attack surface of an ever-growing number of computers, networks, and devices all running a myriad of software. In addition, the DOE laboratories are connected to a variety of external organizations which multiplies the possibility of insider threat. During this SBIR project Cyber Equations developed a strategic combination of supervised and unsupervised machine learning techniques to control scale and classify whether entities, such as hosts, actors, or documents, are showing suspicious insider behavior. We conducted experiments using real cyber data to show that changes in observable behavior (e.g., changes in outbound traffic patterns) are indicative of attacks executing (e.g., insiders exfiltrating data, bots calling home, or an attack spreading). The features and methods we use, have been used in the anomaly detection research and development community for a significant amount of time and robust implementations are openly available. K-Means clustering, Decision Trees, and Support Vector Machine based classifiers are readily available through the R language and open source H20 machine learning framework. The Cyber Equations approach is novel in its use of these classifiers combined with Hadoop based parallel processing technology. The five stages of processing used by CyberInsider improve scalability and reduce the high number of false positives normally associated with network anomaly detection. In summary, the Cyber Equations has successfully accomplished all of its Phase I tasks and demonstrated the feasibility of the Insider Threat Monitoring System.

97 MATHEMATICS AND COMPUTING↗

Feature Selection and Inferential Procedures for Video Data [Slides]

With the rise of electronic and high-dimensional data, new and innovative feature detection and statistical methods are required to perform accurate and meaningful statistical analysis of these datasets that provide unique statistical challenges. In the area of feature detection, much of the recent feature detection research in the computer vision community has focused on deep learning methods, which require large amounts of labeled training data. However, in many application areas, training data is very limited and often difficult to obtain. We develop methods for fast, unsupervised, precise feature detection for video data based on optical flows, edge detection, and clustering methods. We also use pretrained neural networks and interpretable linear models to extract features using very limited training data. In the area of statistics, while high-dimensional data analysis has been a main focus of recent statistical methodological research, much focus has been on populations of high-dimensional vectors, rather than populations of high-dimensional tensors, which are three-dimensional arrays that can be used to model dependent images, such as images taken of the same person or ripped video frames. Our feature detection method is a non-model-based method that fusses information from dense optical flow, raw image pixels, and frame differences to generate detections. Our hypothesis testing methods are based on the assumption that dependent images are concatenated into a tensor that follows a tensor normal distribution, and from this assumption, we derive likelihood-ratio, score, and regression-based tests for one- and multiple-sample testing problems. Our methods will be illustrated on simulated and real datasets. We conclude this report with comments on the relationship between feature detection and hypothesis testing methods.

97 MATHEMATICS AND COMPUTING↗

Social Media Account Resolution and Verification

In this SAND report, we discuss various authorship attribution techniques for social media platforms to identify separate accounts controlled by one user. We look at three main categories of techniques: stylometric, semantic, and temporal, and we test several algorithms in each category. We then combine these techniques to increase predictive power, and perform cross-platform account verification on Twitter and Stack Exchange datasets. Additionally, we briefly examine graphical techniques, and analyze the predictive power of unsupervised approaches. Finally, we apply our techniques to unbalanced data sets. Our experiments reveal that normalized compression distance shows great promise in analyzing social media.

97 MATHEMATICS AND COMPUTING↗

Automated Particle Classifier Q3 FY18 Quarterly Report

This is the third quarter of the third and final year of this project. Work during this quarter included completing and submitting the unsupervised statistical model paper for the American Statistical Association journal Statistical Analysis and Data Mining , (SADM), and nearly finishing the other paper on the semi-supervised statistical model. We prepared and presented a talk for NSARD 2018. Two papers were published in NIMA this quarter: the comprehensive paper on performance and statistical learning approaches, and the FPGA implementation paper. The team prepared and submitted to NA-22 a description of followon projects for FY19, in part based on discussions at NSARD with the NA-22 Program Manager.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

2018 fourth quarter report to NA-22 on Particle Classifier

This was the final quarter of the third and final year of this project. The team prepared and submitted this project’s final report to NA-22. Work during this quarter included responding to reviewer requests for the unsupervised statistical model paper for the American Statistical Association journal, and completing and submitting the paper about the semi-supervised statistical model to NIMA. We received the report from our North Carolina State University collaborators from the NA-22-sponsored CNEC Consortium, Professor Wilson and student Allen. The report describes their investigation of Dirichlet Process (DP) Gaussian Mixture Models based on pulse shape data from this project. The report is an appendix to this project’s final report, and may be converted into a journal paper in the coming year. Work was performed on the user interface to the Struck firmware. The project team submitted a draft workplan to NA-22 regarding additional work to diffuse improved PSD methods through the NA-22 fast neutron detection community.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

FORTÉ Machine Learning

Of the non-corrupted data collected by the Orbiting Experiment (forté) satellite’s Photo-Diode Detector during the year 2001, I estimate that 7.9% of 914 894 signals are noise. My result differs dramatically from Guillen’s estimate of 96%. To arrive at this estimate, I used Gaussian mixture model (GMM) clustering–unsupervised machine learning–to aggregate the wave forms into groups based on the absolute value of the lowest 25 positive frequency discrete Fourier transform coefficients. Then, I marked several of the groups as noise by inspecting a random sampling of wave forms from each group. Marking groups as either noise or non-noise is a supervised binary classification operation. After removing the signals in noise groups from further consideration, I clustered the remaining signals into families. Again, I used a GMM, but for the familial clustering I used a Non-Negative Matrix Factorization feature vector transform. The result was 9 distinct families of lightning signals, as well as a second stage of noise filtering. To efficiently represent the entirety of the signal space, I broke each family into deciles based on their distance from the family mean. In this case, distance means the log-likelihood based on the GMM. Signals in lower deciles are more similar in shape and amplitude to their family average. I took the top 200 samples from each decile of each group, resulting in 18 000 signals. These signal approximately represent the entirety of the forté observations. To represent outliers, I also kept a zoo of the 1000 signals furthest from any family’s average. All told, the resulting data set represents the forté data with a reduction of about 51:1. To allow synthesis of an arbitrarily large number of test signals, I also captured each family’s average signal and the time-sample covariance matrix over the signals in each family. Using these two pieces of information, I can synthesize new waveforms by using a Gaussian random realization from the family average and covariance matrix. I wrote a program to test the synthesis quality. The program shows me two signals on the screen, one synthesized and one randomly drawn from the data. I attempted to identify the synthesized signal. Although the synthesis is imperfect, in an A/B comparison I only correctly chose the synthesized signal 36% of the time.

97 MATHEMATICS AND COMPUTING↗

Self-Organizing Maps and Their Applications to Data Analysis

Self-Organizing Maps(SOMs) are a form of unsupervised neural network that are used for visualization and exploratory data analysis of high dimensional datasets. Our goal was to understand how we can use a SOM to gain insights about datasets. We do this by first understanding the initialization, training, error metrics, and convergence properties of the SOM. Next we discuss the ways to interpret and visualize a Self-Organizing Map. Finally we used real datasets to understand what the Self-Organizing Map can tell us about labeled and unlabeled data. Based on experiments with our datasets we found that the Self-Organizing Map can tell us about the spacing and position of high dimensional clusters, help us find non-linear patterns, and give us insight into the shape of our data.

97 MATHEMATICS AND COMPUTING↗

Future SDN-HPON Control Plane Architecture and Protocol for On-Demand Terabit End-to-End Extreme-Scale Science Applications

The UC Davis team conducted architecture, algorithm and experimental studies for next generation ultra-high-bandwidth optical networks in support of extreme-scale science applications. In particular, over the course of this 3-year project, the UC Davis team achieved the following main accomplishments: Designed a network architecture for automated multi-domain service provisioning: Designed information exchange schemes between autonomous systems (domains) Designed application programming interfaces to orchestrate resource reservation in data centers, high performance computation facilities, and communication networks belonging to different autonomous systems Experimentally assessed the multi-domain network architecture in a distributed field trial set-up connecting premises in three continents Designed a broker-based orchestrated multi-domain optical networking framework supporting observe-analyze-act cycle based cognitive inter-domain service provisioning: Designed an inter-domain operation mechanism, e.g., topology abstraction, for the broker plane and domain managers to collaboratively accomplish inter-domain service provisioning while respecting the autonomy of each administrative domain Designed an alien wavelength monitoring scheme and an artificial neural network based optical signal-to-noise ratio estimator for quality-of-transmission aware inter-domain service provisioning Developed a deep neural network based multi-domain traffic predictor and a cognitive routing and spectrum assignment algorithm to realize knowledge-based autonomous traffic engineering in multi-domain networks Implemented an optical signal-to-noise ratio estimator and traffic predictor with Google’s Tensor flow and assessed their performance using experimental and synthetic data Demonstrated the benefits of the proposed cognitive schemes with experiments ona two-domain elastic optical network testbed and evaluations on a six-domain simulation framework Built a two-domain software-defined networking control plane testbed using the ONOS platform and demonstrated the workflow of broker and domain manager cooperation, including hierarchical performance monitoring, topology abstraction and cognitive resource allocation. Designed self-learning cognitive service provisioning schemes using advanced machine learning technologies: Designed a self-taught anomaly detection framework for optical networks, including an unsupervised data clustering module for pattern recognition, a supervised data regression and classification module for online anomaly detection, and the system operation principle Experimentally validated the effectiveness of the self-taught anomaly detection framework with various failure/anomaly scenarios. Designed a self-learning autonomic routing, modulation and spectrum assignment (RMSA) provisioning framework for elastic optical networks with deep reinforcement learning. Designed effective modeling and training mechanisms for the self-learning RMSA framework Conducted extensive numerical simulations with various traffic patterns and topologies to verify the advantage of the self-learning RMSA framework.

42 ENGINEERING↗

SDN-NGenIA, a Software Defined Next Generation Integrated Architecture for HEP and Data Intensive Science

This project aims to enable the LHC and other leading programs in high energy physics and other global science domains funded by the DOE to operate with a new level of efficiency and control, through the development of a Next Generation Integrated Architecture (NGenIA) based on intelligent software defined network (SDN)-driven network systems coupled to high throughput applications. While the initial focus is on the challenging LHC use case, the systems and products being developed are general, and apply to many fields of data intensive science ranging from astrophysical sky surveys to bioinformatics and earth observation, as well as other organizations facing the challenges of extracting knowledge from distributed multi-petabyte data stores. A central concept in this development program is a new paradigm ‘consistent network operations’ among widely distributed computing and storage facilities, where stable high throughput flows, at set rates, cross load-balanced network paths, up to flexible high water marks that are adjusted in real time to accommodate other network traffic. The large smooth flows are launched and managed by SDN services that act in concert with the experiments’ site-resident data distribution and management systems, to meet the expanding needs of the science programs. The technical goals include the construction of autonomous, intelligent site-resident services that interact dynamically with network-resident services, and with the science programs’ principal data distribution and management tools, to request or command network resources in support of high throughput petascale to exascale workflow Specific work items include: 1. Developing compact Data Transfer Nodes (DTNs) with auto-tuning functions that support data transfer rates in the 100 Gbps to the 1 Tbps range when used with high throughput data transfer applications 2. Deep site orchestration among virtualized clusters, storage subsystems and subnets to successfully coschedule CPU, storage and network resources 3. Science-program designed site architectures, operational modes, and policy and resource usage priorities, adjudicated across multiple network domains and virtual organizations, using the orchestration functions and methods 4. Seamlessly extending end-to-end operation across both extra-site and intra-site boundaries through the use of Open vSwitch (OVS), FireQoS + next generation Science DMZs 5. Novel methods of system integration that enable granular control of extreme scale long distance transfers through flow matching of scattered source-destination address pairs to multi-domain dynamic circuits 6. Funneling massive sets of streams to DTNs at the site edge hosting petascale buffer pools configured for flows of 100 Gbps and up, exploiting state of the art data transfers 7. Adaptive scheduling based on pervasive end-to-end monitoring, including DTN or compute-node resident agents providing comprehensive end-system profiling 8. Developing unsupervised and supervised machine learning and modeling methods to optimize the workflow involving terabyte to multi-petabyte datasets.

47 OTHER INSTRUMENTATION↗

Event Definition for the Automated Detection of Nuclear Proliferation Activities

In FY2020, Savannah River National Laboratory (SRNL) in collaboration with the Discovery Analytics Center (DAC) at Virginia Polytechnic Institute and State University (VT) began developing a demonstration prototype system that uses multiple machine learning and data analytic methods on largescale open data sources to identify new, developing, or undeclared nuclear programs. One of the most challenging aspects of applying machine learning techniques to such a problem is the high likelihood of extremely sparse data from disparate sources. To overcome this challenge, the current work will use a strategic combination of supervised, semi-supervised, and unsupervised learning techniques to ingest and fuse data streams to make a forecast of nuclear activities in a targeted geospatial location. Identifying potential data sources and training supervised learning algorithms is dependent upon the development of a robust foundation of targeted event domains that fundamentally define the nuclear activities of interest. This report documents the definition of a hierarchical structure for both nuclear activity and event domains that will be used to guide the research team in development or use of existing semantic dictionaries that are instrumental to searching, parsing, and categorizing events for the forecasting system’s use.

97 MATHEMATICS AND COMPUTING↗

Transfer Learning using Denoising Auto-Encoders for Cellular-Level Annotation of Tumor in Pathology Slides

Adversarial examples can produce altered classifications using only seemingly innocuous, imperceptible perturbations to the original image. The imperceptibility of adversarial perturbations suggests that the corresponding classifiers use decision criteria different than those of a human. In a medical setting, inexplicable decision criteria confound a pathologist’s willingness to trust machine-generated annotations. Here, we analyze denoising tumor detection models to see if they are robust to imperceptible adversarial perturbations. Moreover, to be more fully trusted by pathologists, we require tumor detectors that generate interpretable annotations which segment pathology slides into tumorous and normal regions at the cellular level. We therefore compare transfer learning based on two different autoencoder architectures, one derived from a deep denoising bottleneck autoencoder and one from an over-complete sparse autoencoder. Both autoencoders were first trained in an unsupervised manner on a set of pathology slides drawn from the Camelyon16 dataset. The latent representations produced by each autoencoder were then passed to separate neural networks that were trained in a supervised manner on binary tumor-normal masks generated by pathologists at cellular resolution. Both tumor detectors supported better than 90% AUC PR as measured by the area under the precision/recall curve on a held-out pathology slide. To assess the underlying decision criteria used by both tumor detectors, we constructed imperceptible adversarial examples which reduced the AUC PR of both models to less than 70%. Random noise of the same amplitude had almost no effect on the AUC PR of either model. Additionally, each tumor detector was resistant to adversarial “transfer” attacks targeting the other. The adversarial perturbations showed strong characteristic differences: the deep denoising models perturbations were a very diffuse, seemingly unrecognizable pattern while the sparse coding models perturbations showed traces of tissue cells.

47 OTHER INSTRUMENTATION↗