244 Hits in 1.7 sec

Disease and Social Media in Post-Natural Disaster Recovery Philippines [article]

Lauren E Charles, Courtney D Corley
2021 medRxiv   pre-print
The Philippines is plagued with natural disasters and resulting precipitating factors for disease outbreaks. The developing country has a strong disease surveillance program during and post-disaster phases; however, latent disease contracted during these emergency situations emerges once the Filipinos return to their homes. Coined the social media capital of the world, the Philippines provides an opportunity to evaluate the potential of social media use in disease surveillance during the
more » ... covery period. By developing and defining a non-traditional method for enhancing detection of infectious diseases post-natural disaster recovery in the Philippines, this research aims to increase the resilience of affected developing countries through advanced passive disease surveillance with minimal cost and high impact. Methods: We collected 50 million geo-tagged tweets, weekly case counts for six diseases, and all-natural disasters from the Philippines between 2012 and 2013. We compared the predictive capability of various disease lexicon-based time series models (e.g., Twitter's BreakoutDetection, Autoregressive Integrated Moving Average with Explanatory Variable [ARIMAX], Multilinear regression, and Logistic regression) and document embeddings (Gensim's Doc2Vec). Results: The analyses show that the use of only tweets to predict disease outbreaks in the Philippines has varying results depending on which technique is applied, the disease type, and location. Overall, the most consistent predictive results were from the ARIMAX model which showed the significance in tweet value for prediction and a role of disaster in specific instances. Discussion: Overall, the use of disease/sick lexicon-filtered tweets as a predictor of disease in the Philippines appears promising. Due to the consistent and large increase use of Twitter within the country, it would be informative to repeat analysis on more recent years to confirm the top method for prediction. In addition, we suggest that a combination disease-specific model would produce the best results. The model would be one where the case counts of a disease are updated periodically along with the continuous monitoring of lexicon-based tweets plus or minus the time from disaster.
doi:10.1101/2021.03.22.21254137 fatcat:tmlniisfl5b3pchs7o5cdc62ru

Dynamic Input Structure and Network Assembly for Few-Shot Learning [article]

Nathan Hilliard, Nathan O. Hodas, Courtney D. Corley
2017 arXiv   pre-print
The ability to learn from a small number of examples has been a difficult problem in machine learning since its inception. While methods have succeeded with large amounts of training data, research has been underway in how to accomplish similar performance with fewer examples, known as one-shot or more generally few-shot learning. This technique has been shown to have promising performance, but in practice requires fixed-size inputs making it impractical for production systems where class sizes
more » ... can vary. This impedes training and the final utility of few-shot learning systems. This paper describes an approach to constructing and training a network that can handle arbitrary example sizes dynamically as the system is used.
arXiv:1708.06819v1 fatcat:tepsudknnfdtvnirb32v7phhh4

Disease Models for Event Prediction

Courtney D. Corley, Laura Pullum
2013 Online Journal of Public Health Informatics  
Objective The objective of this manuscript is to present a systematic review of biosurveillance models that operate on select agents and can forecast the occurrence of a disease event.
doi:10.5210/ojphi.v5i1.4589 fatcat:zbhtpjbtc5gpjouqju5qdcvjqe

Few-Shot Learning with Metric-Agnostic Conditional Embeddings [article]

Nathan Hilliard and Lawrence Phillips, Scott Howland, Artëm Yankov, Courtney D. Corley, Nathan O. Hodas
2018 arXiv   pre-print
Learning high quality class representations from few examples is a key problem in metric-learning approaches to few-shot learning. To accomplish this, we introduce a novel architecture where class representations are conditioned for each few-shot trial based on a target image. We also deviate from traditional metric-learning approaches by training a network to perform comparisons between classes rather than relying on a static metric comparison. This allows the network to decide what aspects of
more » ... each class are important for the comparison at hand. We find that this flexible architecture works well in practice, achieving state-of-the-art performance on the Caltech-UCSD birds fine-grained classification task.
arXiv:1802.04376v1 fatcat:c26dkcqjfjfypafbc35slyvep4

Medical and transmission vector vocabulary alignment with

William Smith, Alan Chappell, Courtney D. Corley
2015 International Conference on Biomedical Ontology  
diseaseTarget "Alpha-D-Mannose" "Dengue_fever,_protection_against" "Fucose" "Dengue_fever,_protection_against" Table 6 .  ... 
dblp:conf/icbo/SmithCC15 fatcat:piafgmoemnfihiemu3qigq6xui

Sharkzor: Interactive Deep Learning for Image Triage, Sort and Summary [article]

Meg Pirrung, Nathan Hilliard, Artëm Yankov, Nancy O'Brien, Paul Weidert, Courtney D Corley, Nathan O Hodas
2018 arXiv   pre-print
Correspondence to: Nathan O Hodas <>, Courtney D Corley <>. Proceedings of the 34 th International Conference on Machine Learning, Sydney, Australia, PMLR 70, 2017.  ... 
arXiv:1802.05316v1 fatcat:q6ry27aasbghrnegl3saiad5de

Prototypical Region Proposal Networks for Few-Shot Localization and Classification [article]

Elliott Skomski, Aaron Tuor, Andrew Avila, Lauren Phillips, Zachary New, Henry Kvinge, Courtney D. Corley, Nathan Hodas
2021 arXiv   pre-print
We define a feature map encoder f : R h×w×3 → R h ×w ×d to be a convolutional network that maps an image X to a feature map with d channels and possibly downsampled resolution, where h ≤ h, w ≤ w and d  ...  Squared Euclidean distances between the encoded query, q = f (x (q) ), and support centroids, d k = || q −s k || 2 2 , model a class membership probability distribution,ŷ = Softmax(−d).  ... 
arXiv:2104.03496v1 fatcat:dsbygyckkzdufoc3fpnbd5budi

Disease Prediction Models and Operational Readiness

Courtney D. Corley, Laura L. Pullum, David M. Hartley, Corey Benedum, Christine Noonan, Peter M. Rabinowitz, Mary J. Lancaster, Niko Speybroeck
2014 PLoS ONE  
The objective of this manuscript is to present a systematic review of biosurveillance models that operate on select agents and can forecast the occurrence of a disease event. We define a disease event to be a biological event with focus on the One Health paradigm. These events are characterized by evidence of infection and or disease condition. We reviewed models that attempted to predict a disease event, not merely its transmission dynamics and we considered models involving pathogens of
more » ... n as determined by the US National Select Agent Registry (as of June 2011). We searched commercial and government databases and harvested Google search results for eligible models, using terms and phrases provided by public health analysts relating to biosurveillance, remote sensing, risk assessments, spatial epidemiology, and ecological niche modeling. After removal of duplications and extraneous material, a core collection of 6,524 items was established, and these publications along with their abstracts are presented in a semantic wiki at As a result, we systematically reviewed 44 papers, and the results are presented in this analysis. We identified 44 models, classified as one or more of the following: event prediction (4), spatial (26), ecological niche (28) , diagnostic or clinical (6) , spread or response (9), and reviews (3) . The model parameters (e.g., etiology, climatic, spatial, cultural) and data sources (e.g., remote sensing, non-governmental organizations, expert opinion, epidemiological) were recorded and reviewed. A component of this review is the identification of verification and validation (V&V) methods applied to each model, if any V&V method was reported. All models were classified as either having undergone Some Verification or Validation method, or No Verification or Validation. We close by outlining an initial set of operational readiness level guidelines for disease prediction models based upon established Technology Readiness Level definitions.
doi:10.1371/journal.pone.0091989 pmid:24647562 pmcid:PMC3960139 fatcat:ayvzohfbzrh35aftm54vn5zt3u

Using Web and Social Media for Influenza Surveillance [chapter]

Courtney D. Corley, Diane J. Cook, Armin R. Mikler, Karan P. Singh
2010 Advances in Experimental Medicine and Biology  
Analysis of Google influenza-like-illness (ILI) search queries has shown a strongly correlated pattern with Centers for Disease Control (CDC) and Prevention seasonal ILI reporting data. Web and social media provide another resource to detect increases in ILI. This paper evaluates trends in blog posts that discuss influenza. Our key finding is that from 5th October 2008 to 31st January 2009, a high correlation exists between the frequency of posts, containing influenza keywords, per week and CDC influenza-like-illness surveillance data.
doi:10.1007/978-1-4419-5913-3_61 pmid:20865540 pmcid:PMC7123932 fatcat:2pommietvbe6zdkqskc3oaogb4

A Computational Framework to Study Public Health Epidemiology

Courtney D. Corley, Armin R. Mikler
2009 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing  
Our dynamic social network contact simulator described (DynSNIC) in Corley et al [8] considers social contacts in romantic and intimate settings.  ... 
doi:10.1109/ijcbs.2009.83 dblp:conf/ijcbs/CorleyM09 fatcat:hrbowwadd5cgzhh65vdtp2foze

Soda Pop: A Time-Series Clustering, Alarming and Disease Forecasting Application

Jeremiah Rounds, Lauren Charles-Smith, Courtney D. Corley
2017 Online Journal of Public Health Informatics  
ObjectiveTo introduce Soda Pop, an R/Shiny application designed to be adisease agnostic time-series clustering, alarming, and forecastingtool to assist in disease surveillance "triage, analysis and reporting"workflows within the Biosurveillance Ecosystem (BSVE) [1]. In thisposter, we highlight the new capabilities that are brought to the BSVEby Soda Pop with an emphasis on the impact of metholodogicaldecisions.IntroductionThe Biosurveillance Ecosystem (BSVE) is a biological andchemical threat
more » ... rveillance system sponsored by the Defense ThreatReduction Agency (DTRA). BSVE is intended to be user-friendly,multi-agency, cooperative, modular and threat agnostic platformfor biosurveillance [2]. In BSVE, a web-based workbench presentsthe analyst with applications (apps) developed by various DTRAfundedresearchers, which are deployed on-demand in the cloud(e.g., Amazon Web Services). These apps aim to address emergingneeds and refine capabilities to enable early warning of chemical andbiological threats for multiple users across local, state, and federalagencies.Soda Pop is an app developed by Pacific Northwest NationalLaboratory (PNNL) to meet the current needs of the BSVE forearly warning and detection of disease outbreaks. Aimed for use bya diverse set of analysts, the application is agnostic to data sourceand spatial scale enabling it to be generalizable across many diseasesand locations. To achieve this, we placed a particular emphasis onclustering and alerting of disease signals within Soda Pop withoutstrong prior assumptions on the nature of observed diseased counts.MethodsAlthough designed to be agnostic to the data source, Soda Pop wasinitially developed and tested on data summarizing Influenza-LikeIllness in military hospitals from collaboration with the Armed ForcesHealth Surveillance Branch. Currently, the data incorporated alsoincludes the CDC's National Notifiable Diseases Surveillance System(NNDSS) tables [3] and the WHO's Influenza A/B Influenza Data(Flunet) [4]. These data sources are now present in BSVE's Postgresdata storage for direct access.Soda Pop is designed to automate time-series tasks of datasummarization, exploration, clustering, alarming and forecasting.Built as an R/Shiny application, Soda Pop is founded on the powerfulstatistical tool R [5]. Where applicable, Soda Pop facilitates nonparametricseasonal decomposition of time-series; hierarchicalagglomerative clustering across reporting areas and between diseaseswithin reporting areas; and a variety of alarming techniques includingExponential Weighted Moving Average alarms and Early AberrationDetection [6].Soda Pop embeds these techniques within a user-interface designedto enhance an analyst's understanding of emerging trends in their dataand enables the inclusion of its graphical elements into their dossierfor further tracking and reporting. The ultimate goal of this softwareis to facilitate the discovery of unknown disease signals along withincreasing the speed of detection of unusual patterns within thesesignals.ConclusionsSoda Pop organizes common statistical disease surveillance tasksin a manner integrated with BSVE data source inputs and outputs.The app analyzes time-series disease data and supports a robust set ofclustering and alarming routines that avoid strong assumptions on thenature of observed disease counts. This attribute allows for flexibilityin the data source, spatial scale, and disease types making it useful toa wide range of analystsSoda Pop within the BSVE.KeywordsBSVE; Biosurveillance; R/Shiny; Clustering; AlarmingAcknowledgmentsThis work was supported by the Defense Threat Reduction Agency undercontract CB10082 with Pacific Northwest National LaboratoryReferences1. Dasey, Timothy, et al. "Biosurveillance Ecosystem (BSVE) WorkflowAnalysis." Online journal of public health informatics 5.1 (2013).2. Accessed 9/6/2016.3. Centers for Disease Control and Prevention. "National NotifiableDiseases Surveillance System (NNDSS)."4. World Health Organization. "FluNet." Global Influenza Surveillanceand Response System (GISRS).5. R Core Team (2016). R: A language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna, Austria.6. Salmon, Maëlle, et al. "Monitoring Count Time Series in R: AberrationDetection in Public Health Surveillance." Journal of StatisticalSoftware [Online], 70.10 (2016): 1 - 35.
doi:10.5210/ojphi.v9i1.7582 fatcat:4cyb7ucqqra6beuwytmu2lb64u

Uncovering the relationships between military community health and affects expressed in social media

Svitlana Volkova, Lauren E Charles, Josh Harrison, Courtney D Corley
2017 EPJ Data Science  
Military populations present a small, unique community whose mental and physical health impacts the security of the nation. Recent literature has explored social media's ability to enhance disease surveillance and characterize distinct communities with encouraging results. We present a novel analysis of the relationships between influenza-like illnesses (ILI) clinical data and affects (i.e., emotions and sentiments) extracted from social media around military facilities. Our analyses examine
more » ... differences in affects expressed by military and control populations, (2) affect changes over time by users, (3) differences in affects expressed during high and low ILI seasons, and (4) correlations and cross-correlations between ILI clinical visits and affects from an unprecedented scale -171M geo-tagged tweets across 31 global geolocations. Key findings include: Military and control populations differ in the way they express affects in social media over space and time. Control populations express more positive and less negative sentiments and less sadness, fear, disgust, and anger emotions than military. However, affects expressed in social media by both populations within the same area correlate similarly with ILI visits to military health facilities. We have identified potential responsible cofactors leading to location variability, e.g., region or state locale, military service type and/or the ratio of military to civilian populations. For most locations, ILI proportions positively correlate with sadness and neutral sentiment, which are the affects most often expressed during high ILI season. The ILI proportions negatively correlate with fear, disgust, surprise, and positive sentiment. These results are similar to the low ILI season where anger, surprise, and positive sentiment are highest. Finally, cross-correlation analysis shows that most affects lead ILI clinical visits, i.e. are predictive of ILI data, with affect-ILI leading intervals dependent on geolocation and affect type. Overall, information gained in this study exemplifies a usage of social media data to understand the correlation between psychological behavior and health in the military population and the potential for use of social media affects for prediction of ILI cases.
doi:10.1140/epjds/s13688-017-0102-z fatcat:tsxcylktdjeklkq3qadgawlyxq

Towards Influenza Surveillance in Military Populations Using Novel and Traditional Sources

Lauren Charles-Smith, Alexander Rittel, Umashanthi Pavalanathan, Courtney D. Corley
2016 Online Journal of Public Health Informatics  
doi:10.5210/ojphi.v8i1.6468 fatcat:rr6dmbztgbfr5dlmrtkwlln45a

Novel Analysis and Visualization of Chemical Events for Public Health Surveillance

Michael J. Henry, Lauren Charles-Smith, Kyungsik Han, Courtney D. Corley
2017 Online Journal of Public Health Informatics  
ObjectivePacific Northwest National Laboratory hosted an intern-basedweb application development contest in the summer of 2016 centeredaround developing novel chemical surveillance applications to aid inhealth situational awareness. Making up the three teams were threegraduate students (n=9) from various US schools majoring in non-public health domains, such as computer sicence and user design. Theinterns suc- cessfully developed three applications that demonstrateda value-add to chemical
more » ... llance—ChemAnalyzer (textanalytics), RetroSpect (retrospective analysis of chemical events),and ToxicBusters (geo-based trend analytics). These applicationswill be the basis for the first chemical surveillance application to beincorporated into the DTRA Biosurveillance Ecosystem (BSVE).IntroductionPacific Northwest National Laboratory (PNNL), on behalf theDefense Threat Reduction Agency (DTRA; project number CB10190),hosts an annual intern- based web app development contest. Previouscompetitions have focused on mobile biosurveillance applications.The 2016 competition pivoted away from biosurveillance to focus onaddressing challenges within the field of chemical surveillance andincreasing public health chemical situational awareness. The result ofthe app will be integrated within the DTRA BSVE.MethodsPNNL hosted nine graduate interns for a 10-week period inthe summer of 2016 as participants in a summer web applicationdevelopment contest. Students were drawn from such fields assoftware engineering and user experience and design and placedinto three teams of three students. The challenge presented to theinterns was to design and develop a fully-functional web applicationthat would address a critical need within the chemical surveillancecommunity. The interns developed their own ideas (vetted by PNNLand DTRA), discovered and inte- grated their own data sources,and produced their own visualizations and an- alytics, independentof any assistence outside of that provided in an advisory capacity.The competition end with a judging event with a panel of subjectmatter experts and cash awards were distributed to the teams.ResultsEach team produced a unique application. Although there wasmild overlap between some of the ideas, the applications weredeveloped independently and each reflected the unique contributionsof the teams. ChemAnalyzer is a text-analytics platform designedto facilitate more data- driven decision, given a corpus of text dataabout a chemical event. Their plat- form provided the ability toautomatically identify and highlight key words in documents relatedto chemical events. The keywords are drawn from an on- tologyinstalled with the system, as well as any user-identified keywords.The ChemAnalyzer team finished in third place. The RetroSpect teamdeveloped a visual analytic tool for performing retrospec- tive analysisand monitoring of chemical events. Their app provided the ability tosearch and analyze past events, as well as visualization of state andcounty information for the recorded chemical events. The RetroSpectteam finished in second place. The Toxicbusters team—the winnersof the competition—created a geo-based situational awareness toolfor tracking chemical events. Their app featured an updateable mapoverlay, search functionality for finding specific or related events,incident and city/state/national-level statistics and trends, as wellas news and social media integration based on keywords related tochemical surveillance.ConclusionsEach of the apps developed by the teams provides value to ananalyst tasked with monitoring chemical events. The apps integratedunique data sources to provides a full picture of a chemical event, andits effects upon the surrounding population. This integrated analyticsprovides a valuable benefit over existing workflows, where analystsmust monitor news, social, and other information sources manuallyfor real-time information. The apps developed by these interns aredesigned to enable identification and analysis of the incident asquickly as possible, allowing for more timely assessments of theincident and its impacts. The web app development contest provideda unique opportunity for students to learn about the emergingneeds in chemical surveillance as it relates to health sit- uationalawareness. Students were drawn from a variety of fields and weretasked with developing novel web apps addressing some of the mostpressing challenges in the field of chemical surveillance. The ideasgenerated by the students will help form the basis for future chemicalsurveillance application development to be integrated with the DTRABSVE.
doi:10.5210/ojphi.v9i1.7644 fatcat:unrv67loabbk3im53j4sm4daou

Forecasting influenza-like illness dynamics for military populations using neural networks and social media

Svitlana Volkova, Ellyn Ayton, Katherine Porterfield, Courtney D. Corley, Gerardo Chowell
2017 PLoS ONE  
(d) Neural network models learned from combined ILI and social media signals significantly outperform models that rely solely on ILI historical data, which adds to a great potential of alternative public  ... 
doi:10.1371/journal.pone.0188941 pmid:29244814 pmcid:PMC5731746 fatcat:nwgvpnmxlfd5dgzqoqswm6mvpq
« Previous Showing results 1 — 15 out of 244 results