BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20200308T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20191103T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20200205T042754Z UID:3BBD0342-3B9B-4C50-97EB-8838393A676D DTSTART;TZID=America/New_York:20200204T173000 DTEND;TZID=America/New_York:20200204T183000 DESCRIPTION:In the emerging world of artificial intelligence algorithms and analytics\, curated and labeled data is often considered the “new oil. ” However\, extracting this “oil” is often expensive and fraught wit h difficulties. Specifically\, generation of labeled data can be labor int ensive and tedious. Crowd sourcing can distribute the labor and mitigate t he cost and tedium for some application scenarios. However\, for other app lications\, such as medicine\, the requirement of specialized knowledge an d skills can make crowd sourcing unviable and generation of labeled data i s therefore both expensive and limited\, in volume and accuracy\, by the a vailability of physicians’ time. The problem is even more acute for task s requiring sample level labeling of large\, high-resolution spatio-tempor al datasets\, for example\, for pixel-level labeling of images for medical image segmentation. In this talk\, through case studies\, we highlight ex amples where creative use of conventional machine learning\, computer visi on\, and image processing techniques allows us to efficiently generate lab eled data for new applications of deep learning. The examples particularly highlight that these conventional tools continue to be effective and usef ul and will therefore co-exist symbiotically with modern deep learning met hodologies.\n\nSpeaker(s): Gaurav Sharma\, \n\nAgenda: \nPizza served at 5 :15pm\, talk starts at 5:30pm\n\nRoom: IoT Lab in Erdle Commons\, 1st floo r\, Bldg: GLE (09)\, Rochester Institute of Technology\, Rochester\, New Y ork\, United States LOCATION:Room: IoT Lab in Erdle Commons\, 1st floor\, Bldg: GLE (09)\, Roch ester Institute of Technology\, Rochester\, New York\, United States ORGANIZER:rwpeec@rit.edu SEQUENCE:3 SUMMARY:Leveraging Old Tricks in A New World: Efficient Generation of Label ed Data for Deep Learning URL;VALUE=URI:https://events.vtools.ieee.org/m/213432 X-ALT-DESC:Description: <br /><p>In the emerging world of artificial intell igence algorithms and analytics\, curated and labeled data is often consid ered the &ldquo\;new oil.&rdquo\; However\, extracting this &ldquo\;oil&rd quo\; is often expensive and fraught with difficulties. Specifically\, gen eration of labeled data can be labor intensive and tedious. Crowd sourcing can distribute the labor and mitigate the cost and tedium for some applic ation scenarios. However\, for other applications\, such as medicine\, the requirement of specialized knowledge and skills can make crowd sourcing u nviable and generation of labeled data is therefore both expensive and lim ited\, in volume and accuracy\, by the availability of physicians&rsquo\; time. The problem is even more acute for tasks requiring sample level labe ling of large\, high-resolution spatio-temporal datasets\, for example\, f or pixel-level labeling of images for medical image segmentation. In this talk\, through case studies\, we highlight examples where creative use of conventional machine learning\, computer vision\, and image processing tec hniques allows us to efficiently generate labeled data for new application s of deep learning. The examples particularly highlight that these convent ional tools continue to be effective and useful and will therefore co-exis t symbiotically with modern deep learning methodologies.</p><br /><br />Ag enda: <br /><p>Pizza served at 5:15pm\, talk starts at 5:30pm</p> END:VEVENT END:VCALENDAR