BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:America/Los_Angeles
X-WR-TIMEZONE:America/Los_Angeles
BEGIN:VEVENT
UID:254@escience.washington.edu
DTSTART;TZID=America/Los_Angeles:20250220T163000
DTEND;TZID=America/Los_Angeles:20250220T172000
DTSTAMP:20250128T013316Z
URL:https://escience.washington.edu/events/uw-data-science-seminar-sheng-w
 ang/
SUMMARY:UW Data Science Seminar: Sheng Wang
DESCRIPTION:Please join us for the fourth talk in a special series in the U
 W Data Science Seminar featuring the AI@UW Seed Grant awardees. The AI@UW 
 Seed Grant projects partner domain researchers from an applied disciplinar
 y field with researchers specializing in AI theory and/or methodology. The
 se projects were supported by the eScience Institute in collaboration with
  the Office of Research\, the Paul G. Allen School of Computer Science &am
 p\; Engineering\, the Information School\, and the NSF Institute for Found
 ations of Data Science (IFDS).  This seminar will feature UW Computer Sci
 ence and Engineering Assistant Professor Sheng Wang on Thursday\, February
  20th from 4:30 to 5:20 p.m. PT.\n\n\n\nThe seminar will be held in Hitchc
 ock Hall 132 – Campus Map.\n"Generative AI for Multimodal Biomedicine"\
 nAbstract Biomedicine is inherently multimodal\, including imaging modali
 ties such as pathology\, CT\, MRI\, X-ray and ultrasounds\, as well as omi
 cs modality such as genomics\, epigenomics and transcriptomics. General do
 main multimodal approaches are not applicable to biomedicine because biome
 dical images are very different from general domain images\, thus necessit
 ating the development of modality-specific approaches. In this talk\, I wi
 ll introduce three recent works towards building multimodal biomedicine fo
 undation models. First\,  I will introduce GigaPath\, the whole-slide path
 ology foundation model that can handle gigapixel-level pathology images. G
 igaPath exploits a novel vision transformer architecture and achieves the 
 state-of-the-art results on 23 out of 26 cancer tasks\, including subtypin
 g and biomarker prediction. Next\, I will introduce OCTCube\, the first 3D
  OCT retinal imaging foundation model. OCTCube significantly outperformed 
 2D models on 27 out of 29 tasks\, including retinal disease prediction\, c
 ross-modality analysis\, cross-device generalization and systemic disease 
 prediction. Finally\, I will introduce BiomedParse\, a multi-modal foundat
 ion model that integrates 9 major biomedical imaging modalities by project
 ing all of them into the text space\, resulting in superior performance on
  segmentation\, detection\, and recognition\, paving the path for large-sc
 ale image-based biomedical discovery. I will conclude this task with discu
 ssion on how multi-modal generative AI can advance future medical applicat
 ions through multi-agent framework and integration with multi-omics datase
 ts.\n\nBiography: Sheng Wang is an assistant professor in the School of Co
 mputer Science and Engineering at the University of Washington Seattle. He
  obtained his B.S. degree in Computer Science from Peking University\, Ph.
 D. degree in Computer Science from University of Illinois at Urbana Champa
 ign\, and conducted postdoc training at Stanford School of Medicine. Sheng
  is currently interested in developing large-scale models for biomedical a
 pplications\, with a focus on digital pathology\, medical imaging foundati
 on models\, chromatin structure prediction\, and genomics-based drug disco
 very. His research has been published in top venues such as Nature\, Scien
 ce\, Nature Biotechnology\, Nature Methods\, Nature Machine Intelligence a
 nd The Lancet Oncology\, and used by major biomedical institutes\, includi
 ng Mayo Clinic\, Chan Zuckerberg Biohub\, UW Medicine\, and Providence gen
 omics.\n\nThe UW Data Science Seminar is an annual lecture series at the
  University of Washington that hosts scholars working across applied areas
  of data science\, such as the sciences\, engineering\, humanities and art
 s along with methodological areas in data science\, such as computer scien
 ce\, applied math and statistics. Our presenters come from all domain fiel
 ds and include occasional external speakers from regional partners\, gover
 nmental agencies and industry.\nThe 2024-2025 seminars will be held in per
 son\, and are free and open to the public.
ATTACH;FMTTYPE=image/jpeg:https://escience.washington.edu/wp-content/uploa
 ds/2025/01/swang_photo-scaled.jpeg
END:VEVENT
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:STANDARD
DTSTART:20241103T010000
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR