BEGIN:VCALENDAR VERSION:2.0 PRODID:-//hacksw/handcal//NONSGML v1.0//EN METHOD:PUBLISH BEGIN:VEVENT DTSTAMP:20260310T212150Z DESCRIPTION:Click for Latest Location Information: http://edw2018.dataversi ty.net/sessionPop.cfm?confid=121&proposalid=9587\n

This presentation expl ains how to bridge the gap between Technical Metadata and Business Metadata using computational language. It is known that there are practical difficu lties in defining a well-formed Business Glossary. This is even more challe nging when the organization is around for a long time and has large or very large amounts of data. While there are many techniques available nowadays, much of the work, traditionally, requires manpower and time. In most cases , manpower also needs some level of specialization around the business know ledge involved. Our experience with establishing metadata repositories, bus iness glossaries, and the necessary architecture to successfully implement data lineage always stopped at the doors of human expertise and systems art ifacts loosely defined. To circumvent this problem, we decided to add compu tation language that helps discover aspects of natural language in data str uctures, annotations, and metadata. To validate our assumptions, we ar e building a prototype that enforces the ISO/IEC 11179 Data Standards. As t he field of artificial intelligence hits the leadership flavor of the year, we asked ourselves if we could start relatively small, but still focu s on the benefits of inferring from experience by building a machine l earning gadget that would work without specialized manpower and be of quali ty value for providing to our clients the needed data definitions and its p roperties properly mapped to data assets. Some of the aspects we will menti on are:

\n\n Problem statement\n Platform environment\n Agile data governance \n Automatic natural language understanding\n\n DTSTART:20180424T153000 SUMMARY:Deriving New Business Terms from Technical Metadata DTEND:20180424T162959 LOCATION: See Description END:VEVENT END:VCALENDAR