BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:TraCS Data Science Lab: LLM Mechanics in Research Using Clinic
 al Notes
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260910T043727Z
UID:tag:localist.com\,2008:EventInstance_53082463218265
DTSTART:20260616T163000Z
DTEND:20260616T173000Z
DESCRIPTION:LLM Mechanics in Research Using Clinical Notes\n\n \n\nLarge La
 nguage Models are increasingly used to analyze clinical notes\, but their 
 behavior can be difficult to understand without examining how prompts\, co
 ntext\, and outputs interact. In this seminar\, Deep Nag\, MS\, will intro
 duce the mechanics of using LLMs for clinical note research\, including ho
 w prompts are structured\, how clinical text is prepared for LLM input\, h
 ow outputs can be interpreted\, and where caution is warranted when applyi
 ng these methods in healthcare research. The session will focus on practic
 al examples from EHR-based research and will highlight both the promise an
 d limitations of LLMs in high-stakes clinical settings.\n\n \n\nSpeaker:\n
 Deep Nag\, MS\nResearch Data Scientist\nNC TraCS Institute\n\n \n\nRegiste
 r\n\n \n\nSeminars in the NC TraCS Data Science Lab Seminar Series cover a
  range of topics related to health care data science\, clinical data\, dat
 a engineering\, and working in these areas at UNC-Chapel Hill. This month'
 s seminar will be held via Zoom only.
LOCATION:
SUMMARY:TraCS Data Science Lab: LLM Mechanics in Research Using Clinical No
 tes
URL;VALUE=URI:https://calendar.unc.edu/event/tracs-data-science-lab-llm-mec
 hanics-in-research-using-clinical-notes
END:VEVENT
END:VCALENDAR