Sciome LLC — Department of Health and Human Services SBIR Phase I: NIEHS

Description

Project SummaryMore thansystematic reviews are performed each year in the fields of environmental health and evidence basedmedicinewith each review requiringon averagebetween six months to one year of effort to completeOne of the mosttime consuming and repetitive aspects of this endeavor involves extraction of detailed information from a large numberof scientific documentsThe specific data items extracted differ among disciplinesbut within a given scientific domaincertain data points are extracted repeatedly for each review conductedResearch on use of natural language processingNLPfor extracting individual data elements has shown that it has the potential to greatly reduce the laborioustimeintensiveand repetitive nature of this stepHoweverthere is currently no integratedautomatic data extraction platformthat meets the needs of the systematic review communityWe propose a web based data extraction software platformspecifically designed for usage in the domain of systematic reviewBy combining multiple state of the art data extractionmethods utilizing NLPtext mining and machine learninginto a singleunified user interfacewe will thereby empowerthe end user with a powerful and novel tool for automating an otherwise arduous taskThe research we propose encompasses three specific aimsdevelop new data extraction models using deep learningand a new technique calleddata programmingdevelop a web based platform to semi automate the processdesign protocols and standards for packaging extraction models as software components and integrating work done byother research groups and vendorsIn the first aimwe will contribute novel data extraction modules designed andtrained specifically to extract data elements of interest to those conducting systematic reviews in the domain ofenvironmental healthFor this researchwe will employ state of the art machine learningNLP and text miningmethodologies to train and evaluate several novel extraction componentsIn our second aimwe will develop a webbased workbench which will allow users to upload scientific documents for automated data extractionOur system willalso be designed to allow for integration of data extraction approachescomponentsfrom other research groupsthusenabling end users to choose from a wide variety of advanced data extraction methodologies within one unified andintuitive software environmentIn our third aimwe will develop new protocols to standardize the inputs and outputsfor data extraction componentsThe resulting interfacewhich will enable seamless integration of third party extractioncomponents into the workbenchwill also facilitate the incorporation of feedback from users such that extractioncomponents can be continuously improved based on real time dataOur overarching goal is to translate emerging semi automated extraction technologies out of the lab and into practicalsoftware and to bring to market both the software itself as well as several premium data extraction componentsTheresults of the research conducted for Aimsrepresent the first step in this direction and will provide the foundation forfuture developmentsThese result will take us one step closer to the dream of creatingliving systematic reviewswhichare maintained using automated or semi automated methods and updated regularly as new evidence becomes available Project Narrative Systematic review is a formal process used widely in evidence based medicine and environmental health research to identifyassessand integrate the primary scientific literature with the goal of answering a specifictargeted question in pursuit of the current scientific consensusBy conducting research and development to build a flexibleextensible software system that automates the crucial and resource intensive process of extracting key data elements from scientific documentswe will make an important contribution toward ongoing efforts to automate systematic reviewThese efforts will serve to make systematic reviews both more efficient to produce and less expensive to maintaina result which will greatly accelerate the process by which scientific consensus is obtained in a variety of medical and health related disciplines having great public significance