BEGIN:VCALENDAR VERSION:2.0 PRODID:-//hacksw/handcal//NONSGML v1.0//EN METHOD:PUBLISH BEGIN:VEVENT DTSTAMP:20260421T170859Z DESCRIPTION:Click for Latest Location Information: http://edw2018.dataversi ty.net/sessionPop.cfm?confid=121&proposalid=9552\n

It seems everybody&rsq uo;s doing it: using machine learning, principally statistical-based text a nalysis, to try to induce knowledge out of large data sets comprised of uns tructured text documents from inside enterprises and from the Web. It works , sort of. Previously, knowledge engineers built most semantic models by ha nd, either individually or through group/crowd-sourced projects. It was har d, slow going. Neither approach is perfect; both have their pitfalls. Using them in combination -- a hybrid or blended approach -- is actually much be tter. This presentation will:

\n\n Highlight the strengths and challenges found in both ML and ontological-bas ed approaches \n Illustrate ways to use text-based ML to induce ‘rough order’ se mantic models, using examples from enterprise projects in healthcare and fr om social media \n Describe how ontologies can structure and augment statistical-based models at various steps in the process\n Survey available tools and other resources to use for hybrid modeling proje cts\n\n DTSTART:20180426T093000 SUMMARY:Stronger Together: Hybrid Knowledge Modeling Using Machine Learning + Ontologies DTEND:20180426T102959 LOCATION: See Description END:VEVENT END:VCALENDAR