Search NASA⌕ Search

Engineering topics

Park, Byung

Publications and source records attributed to Park, Byung.

DENSECL: Haze Mitigation Using Dense Blocks and Contrastive Loss Regularization

Haze, which occurs as a result of the scattering of light in the atmosphere by small particles, diminishes the visibility of scene objects, inflicting important image applications such as object detection. To address the problem, this paper introduces a new physics-based end-to-end deep learning approach to haze mitigation in outdoor scenes, including those in airborne images. The proposed model named DenseCL is designed with dense blocks and adopts a contrastive loss function as an additional regularization. The model also maintains the cycle consistency by remapping the dehazed outputs into a hazy image using the physics-based light scattering function. DenseCL has been trained with publicly available outdoor images and demonstrates outstanding performance on outdoor, indoor, and remotely sensed nonhomogeneous haze satellite images.

Mitra, Somosmita↗

CPViz: Visualizing clinical pathways represented in higher-order networks

To improve clinical care practice, it is important to understand the variability of clinical pathways executed in different contexts (e.g., pathways in different geographical locations, demographics, and phenotypic groups). A common way of representing clinical pathways is through network-based representations that capture trajectories of treatment steps. However, first-order networks, which are based on the Markovian property and the de facto standard model to represent transitions between steps, often fail to capture real trajectories. This paper introduces a visual analytic tool to explore and compare pathways represented in higher-order networks. Because each higher node in the network is a subtrajectory (i.e., partial or full history of treatment steps), the tool can display true sequences of treatment steps and compute the similarity of the two networks in a space of higher-order nodes. The tool also highlights areas in which the two networks are similar and dissimilar and how a certain subtrajectory is realized differently in different pathways. The paper demonstrates the tool's usefulness by applying it to multiple antidepressant pharmacotherapy pathways for veterans diagnosed with major depressive disorder and by illustrating heterogeneity in prescription patterns across pathways.

Chae, Junghoon↗

Real-time Multi-granular Analytics Framework for HIT Systems

Streaming analytics is the process of ingesting and digesting live data from multiple data sources. In the healthcare domain, as the importance of extracting immediate insights while data are streaming into the system grows, the focus is shifting from batch processing to streaming analytics. With data increasing dramatically at high speeds, many informatics designs have been proposed to adapt healthcare domain into this new environment. In our previous work, we introduced a prototype of health informatics technology (HIT) framework that aims to address challenges in adopting state-of-the-art technologies to enable advanced healthcare analytic tasks in new streaming environments. We recently made major updates to the framework so that anomaly from multiple streaming data sources at different granularity levels can be detected in near real-time. In this paper, we detail the implementation and deployment of the framework in Kubernetes clusters and report its performances when tested on electronic health record (EHR) data of Veterans Affairs.

Park, Byung↗

A Weakly-Supervised, Multitask Deep Learning Framework for Shadow Mitigation in Remote Sensing Imagery

We propose a weakly-supervised, multitask framework for training a convolutional neural network to solve the problem of cloud shadow mitigation given only cloud and shadow masks as labels. The network minimizes the Wasserstein distance between shadows and their proximal sunlit neighborhoods, generating a supervisory signal directly from within the input image. We extract further utility from the shadow mask through multitask learning by introducing an auxiliary task of shadow segmentation. Our approach is advantageous since it performs mitigation in an end-to-end framework which requires only a shadowed image for inference. We apply this process to the Landsat 8 OLI SPARCS validation data set and demonstrate plausible results.

Couwenhoven, Scott↗

A Knowledge Network-Based Approach to Facilitate Annotation of Clinical Pathway Component Clusters

Mining electronic health records (EHRs) to identify contextually related clinical concept clusters that tend to co-occur temporarily and consistently could improve data-driven clinical pathway (CP) construction. However, the automatic extraction of contextually related clinical concept clusters contains a vast amount of irrelevant information. Hence, this paper proposes a knowledge network-enabled literature-based discovery (LBD) approach to remove noise from clusters. The authors used published literature to filter spurious concepts from the clusters and used data from the US Department of Veterans Affairs’s major depressive disorder (MDD) cohort of Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) for their experimentation. The approach was applied to 2,967 clusters extracted from the MDD OEF/OIF database. The experimental results demonstrate that the proposed approach can filter 94% of the irrelevant information. Moreover, the authors applied various network mining algorithms to analyze the clusters and demonstrated that LBD, along with network mining techniques, is a useful method for finding accurate contextually related clinical concept clusters. This could help domain researchers perform advanced analytics in CPs.

Hasan, S M Shamimul↗

Haze Mitigation in High-Resolution Satellite Imagery using Enhanced Style-Transfer Neural Network and Normalization Across Multiple GPUs

Despite recent advances in deep learning approaches, haze mitigation in large satellite images is still a challenging problem. Due to amorphous nature of haze, object detection or image segmentation approaches are not applicable. Also it is practically infeasible to obtain ground truths for training. Bounded memory capacity of GPUs is another constraint that limits the size of image to be processed. In this paper, we propose a style transfer based neural network approach to mitigate haze in a large overhead imagery. The network is trained without paired ground truths; further, perception loss is added to restore vivid colors, enhance contrast and minimize artifacts. The paper also illustrates our use of multiple GPUs in a collective way to produce a single coherent clear image where each GPU dehazes different portions of a large hazy image.

Park, Byung↗

Advanced Health Information Technology Analytic Framework and Application to Hazard Detection

Health Information Technology (HIT) aims to improve healthcare outcomes by organizing and analyzing various health-related data. With data accumulating at a staggering rate, the importance of real-time analytics has been increasing dramatically, shifting the focus of informatics from batch processing to streaming analytics. HIT is also facing unprecedented challenges in adapting to this new requirement and leveraging advanced IT technologies. This paper introduces a HIT data and compute platform that supports multi-granularity real-time analytics from heterogeneous data sources. The paper first identifies functional requirements and proposes a framework that satisfies the requirements using state-of-the-art big data technologies including Apache Kafka, Spark Structured Streaming Engine, and Delta Lake. To demonstrate its capability to support data analytics in multiple time granularities analytics, a statistical process control-based hazard detection algorithm has been implemented on top of the framework to detect unexpected hazards from order cancellation data of the Department of US Veterans Affairs (VA) in near real-time.

Kumar, Mohit↗

Characterizing Sub-Cohorts via Data Normalization and Representation Learning

The process of identifying a cohort of interest is a very challenging task. It requires manually inspecting many patient records of complex structure that might include medical coding errors and missing data. This paper presents a computational pipeline for refining the process of cohort selection based on medical concepts recorded in the electronic health records (EHRs). The pipeline extracts EHR data for a given cohort and normalizes this data using standard vocabularies. Then a stacked denoising autoencoder is used to embed the normalized patient vectors in a low dimensional space, where the patients are subsequently clustered into sub-cohorts. The goal is to represent the cohort in a standard format and abstract variants of sub-populations. As a use-case, we applied the pipeline to 1.8 million Veterans diagnosed with major depressive disorder (MDD), and identified four meaningful sub-cohorts using the features learned by the autoencoder. Then, each sub-cohort was explored using a set of keywords for interpretation.

Rush III, Everett↗

Converting Clinical Pathways to BPM+ Standards: A Case Study in Stable Ischemic Heart Disease

Clinical pathways (CPs) are structured healthcare plans designed to implement evidence-based clinical guidelines, medical algorithms, and protocols. In recent years, a community called BPM+ Health has worked to establish a shareable and computer-consumable representation of CP, leveraging standard notations. These notations, collectively referred to as BPM+, include the Business Process Management and Notation (BPMN), Case Management Model and Notation (CMMN), and Decision Model and Notation (DMN), which aim to support clinical management and standardized communication between different stakeholders. However, the adaptation of these notations for the existing guidelines has largely been left unexplored. This paper introduces procedural steps and criteria considerations to apply components of BPM+ notations to reconstruct a guideline for Stable Ischemic Heart Disease. This paper describes how each of the three different notations is mapped to a medical guideline and discusses the advantages and limitations of representing CPs with BPM+ as compared with paper-based medical guidelines.

Chae, Junghoon↗

High performance Data Driven Agent-based Modeling Framework for Simulation of Commute Mode Choices in Metropolitan Area

Active transportation, human-powered transportation modes such as walking and bicycling, not only reduces the carbon footprint from the transportation sector but also promotes healthy living by offering opportunities for people to build physical activity into their daily routine. To encourage active transportation through urban planning and public campaigns, it is of significant importance to infer factors that substantially influence commuters in their transportation mode choice process. This necessitates a flexible and repeatable tool that can evaluate how a policy is perceived by individual commuters and convert their decisions into macro level understanding. This paper introduces one such effort that is specifically designed for studies of transportation mode choices in metropolitan areas. It provides results from a high-resolution data driven simulation based on high performance computing implementation of the agent-based model framework for home-to-work commute trips. The framework uses a graph-partition based technique that can leverage the interaction structure of agents within a geographic proximity and can boost the simulation execution time. Further, based on a flexible design, it can run ABM with different levels of computing resources-from multi core workstations to an HPC grid. The framework has been tested on the Titan Cray XK7 supercomputer of the Oak Ridge Leadership Computing Facility.

Park, Byung↗

A Big Data Analytics Framework for HPC Log Data: Three Case Studies Using the Titan Supercomputer Log

Reliability, availability and serviceability (RAS) logs of high performance computing (HPC) resources, when closely investigated in spatial and temporal dimensions, can provide invaluable information regarding system status, performance, and resource utilization. These data are often generated from multiple logging systems and sensors that cover many components of the system. The analysis of these data for finding persistent temporal and spatial insights faces two main difficulties: the volume of RAS logs makes manual inspection difficult and the unstructured nature and unique properties of log data produced by each subsystem adds another dimension of difficulty in identifying implicit correlation among recorded events. To address these issues, we recently developed a multi-user Big Data analytics framework for HPC log data at Oak Ridge National Laboratory (ORNL). This paper introduces three in-progress data analytics projects that leverage this framework to assess system status, mine event patterns, and study correlations between user applications and system events. We describe the motivation of each project and detail their workflows using three years of log data collected from ORNL's Titan supercomputer.

Park, Byung↗

A Comprehensive Informative Metric for Analyzing HPC System Status Using the LogSCAN Platform

Log processing by Spark and Cassandra-based ANalytics (LogSCAN) is a newly developed analytical platform that provides flexible and scalable data gathering, transformation and computation. One major challenge is to effectively summarize the status of a complex computer system, such as the Titan supercomputer at the Oak Ridge Leadership Computing Facility (OLCF). Although there is plenty of operational and maintenance information collected and stored in real time, which may yield insights about short- and long-term system status, it is difficult to present this information in a comprehensive form. In this work, we present system information entropy (SIE), a newly developed metric that leverages the powers of traditional machine learning techniques and information theory. By compressing the multivariant multi-dimensional event information recorded during the operation of the targeted system into a single time series of SIE, we demonstrate that the historical system status can be sensitively represented concisely and comprehensively. Given a sharp indicator as SIE, we argue that follow-up analytics based on SIE will reveal in-depth knowledge about system status using other sophisticated approaches, such as pattern recognition in the temporal domain or causality analysis incorporating extra independent metrics of the system.

Hui, Yawei↗

Improving Quality of Observational Streaming Medical Data by Using Long Short-Term Memory Networks (LSTMs)

We present an exploration of the encoder-decoder structured Long Short-Term Memory Network (LSTM) as a detector of the anomalous missing observations in streaming medical data by using the difference between the LSTM-reconstructed and observed values as the anomaly detector. We experiment with time-series data from bedside monitoring devices from the available Medical Information Mart for Intensive Care Database (MIMIC). Our results show that not only encoder-decoder LSTM approach works well for detecting the difference between anomalous and normal missing observations in streaming medical data, but also has an imputation potential for the missing data.

Bowie, Michael B.↗

Big Data Meets HPC Log Analytics: Scalable Approach to Understanding Systems at Extreme Scale

Today's high-performance computing (HPC) systems are heavily instrumented, generating logs containing information about abnormal events, such as critical conditions, faults, errors and failures, system resource utilization, and about the resource usage of user applications. These logs, once fully analyzed and correlated, can produce detailed information about the system health, root causes of failures, and analyze an application's interactions with the system, providing valuable insights to domain scientists and system administrators. However, processing HPC logs requires a deep understanding of hardware and software components at multiple layers of the system stack. Moreover, most log data is unstructured and voluminous, making it more difficult for system users and administrators to manually inspect the data. With rapid increases in the scale and complexity of HPC systems, log data processing is becoming a big data challenge. This paper introduces a HPC log data analytics framework that is based on a distributed NoSQL database technology, which provides scalability and high availability, and the Apache Spark framework for rapid in-memory processing of the log data. The analytics framework enables the extraction of a range of information about the system so that system administrators and end users alike can obtain necessary insights for their specific needs. We describe our experience with using this framework to glean insights from the log data about system behavior from the Titan supercomputer at the Oak Ridge National Laboratory.

Park, Byung↗

BESC public portal: an integrative analysis of a resequenced ethanol adapted Clostridium thermocellum mutant

The BioEnergy Science Center (BESC) is undertaking large experimental campaigns to understand the biosynthesis and biodegradation of biomass and to develop biofuel solutions. BESC is generating large volumes of diverse data, including genome sequences, omics data and assay results. The purpose of the BESC Knowledgebase is to serve as a centralized repository for experimentally generated data and to provide an integrated, interactive and user-friendly analysis framework. The Portal makes available tools for visualization, integration and analysis of data either produced by BESC or obtained from external resources.

09 BIOMASS FUELS↗

Shewanella knowledgebase: integration of the experimental data and computational predictions suggests a biological role for transcription of intergenic regions

Shewanellae are facultative gamma-proteobacteria whose remarkable respiratory versatility has resulted in interest in their utility for bioremediation of heavy metals and radionuclides and for energy generation in microbial fuel cells. Extensive experimental efforts over the last several years and the availability of 21 sequenced Shewanella genomes made it possible to collect and integrate a wealth of information on the genus into one public resource providing new avenues for making biological discoveries and for developing a system level understanding of the cellular processes. The Shewanella knowledgebase was established in 2005 to provide a framework for integrated genome-based studies on Shewanella ecophysiology. The present version of the knowledgebase provides access to a diverse set of experimental and genomic data along with tools for curation of genome annotations and visualization and integration of genomic data with experimental data. As a demonstration of the utility of this resource, we examined a single microarray data set from Shewanella oneidensis MR-1 for new insights into regulatory processes. The integrated analysis of the data predicted a new type of bacterial transcriptional regulation involving co-transcription of the intergenic region with the downstream gene and suggested a biological role for co-transcription that likely prevents the binding of a regulator of the upstream gene to the regulator binding site located in the intergenic region. Database URL: http://shewanella-knowledgebase.org:8080/Shewanella/ or http://spruce.ornl.gov:8080/Shewanella/

30 DIRECT ENERGY CONVERSION↗

Cazymes Analysis Toolkit (CAT): Webservice for searching and analyzing carbohydrateactive enzymes in a newly sequenced organism using CAZy database

The Carbohydrate-Active Enzyme (CAZy) database provides a rich set of manually annotated enzymes that degrade, modify, or create glycosidic bonds. Despite rich and invaluable information stored in the database, software tools utilizing this information for annotation of newly sequenced genomes by CAZy families are limited. We have employed two annotation approaches to fill the gap between manually curated high-quality protein sequences collected in the CAZy database and the growing number of other protein sequences produced by genome or metagenome sequencing projects. The first approach is based on a similarity search against the entire non-redundant sequences of the CAZy database. The second approach performs annotation using links or correspondences between the CAZy families and protein family domains. The links were discovered using the association rule learning algorithm applied to sequences from the CAZy database. The approaches complement each other and in combination achieved high specificity and sensitivity when cross-evaluated with the manually curated genomes of Clostridium thermocellum ATCC 27405 and Saccharophagus degradans 2-40. The capability of the proposed framework to predict the function of unknown protein domains (DUF) and of hypothetical proteins in the genome of Neurospora crassa is demonstrated. The framework is implemented as a Web service, the CAZymes Analysis Toolkit (CAT), and is available at http://cricket.ornl.gov/cgi-bin/cat.cgi.

59 BASIC BIOLOGICAL SCIENCES↗