BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20261101T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260716T000555Z UID:6AF93FC4-411C-4B1E-85C6-A7A89EA0DA08 DTSTART;TZID=America/New_York:20260714T180000 DTEND;TZID=America/New_York:20260714T193000 DESCRIPTION:If you’ve ever wondered why your RAG pipeline still surfaces the wrong documents—even when everything seems set up correctly—you’ re not alone. Traditional approaches can struggle with domain-specific lan guage and evolving knowledge bases.\n\nThis session will introduce FLAIR ( Feedback Learning for Adaptive Information Retrieval)\, a lightweight fram ework that continuously improves retrieval quality using expert feedback\, without the need for expensive retraining.\n\nWhat to expect:\n- Real and synthesized signals to better understand user intent\n- Dynamic ranking t hat promotes relevant results in real time\n- Continuous improvement as yo ur knowledge base evolves\n\n📅 July 14\, 2026 | 6–8 PM\n📍 Microsof t NERD Center\, Cambridge\, MA\n\nThis is a great opportunity for anyone b uilding or maintaining RAG systems to explore a more adaptive\, feedback-d riven approach.\n#AI #RAG #MachineLearning #InformationRetrieval #Generati veAI #Microsoft #TechTalks\n\nREGISTER HERE: https://lnkd.in/efvnDuZw\n\nR oom: M Floor\, Bldg: NERD\, 1 Memorial Drive\, Cambridge\, Massachusetts\, United States LOCATION:Room: M Floor\, Bldg: NERD\, 1 Memorial Drive\, Cambridge\, Massac husetts\, United States ORGANIZER:cmiyachi@alum.mit.edu SEQUENCE:18 SUMMARY:The Self-Correcting RAG: Improving Retrieval via User Feedback Loop s URL;VALUE=URI:https://events.vtools.ieee.org/m/563618 X-ALT-DESC:Description: <br /><p dir="ltr">If you&rsquo\;ve ever wondered w hy your RAG pipeline still surfaces the wrong documents&mdash\;even when e verything seems set up correctly&mdash\;you&rsquo\;re not alone. Tradition al approaches can struggle with domain-specific language and evolving know ledge bases.</p>\n<p dir="ltr">This session will introduce FLAIR (Feedback Learning for Adaptive Information Retrieval)\, a lightweight framework th at continuously improves retrieval quality using expert feedback\, without the need for expensive retraining.</p>\n<p dir="ltr">What to expect:<br>- Real and synthesized signals to better understand user intent<br>- Dynami c ranking that promotes relevant results in real time<br>- Continuous impr ovement as your knowledge base evolves</p>\n<p dir="ltr">📅 July 14\, 20 26 | 6&ndash\;8 PM&nbsp\; <br>📍 Microsoft NERD Center\, Cambridge\, MA& nbsp\;</p>\n<p dir="ltr">This is a great opportunity for anyone building o r maintaining RAG systems to explore a more adaptive\, feedback-driven app roach.&nbsp\; <br><strong>#AI</strong> <strong>#RAG</strong> <strong>#Mach ineLearning</strong> <strong>#InformationRetrieval</strong> <strong>#Gener ativeAI</strong> <strong>#Microsoft</strong> <strong>#TechTalks</strong>&n bsp\;</p>\n<p dir="ltr">REGISTER HERE: <a href="https://lnkd.in/efvnDuZw"> https://lnkd.in/efvnDuZw</a>&nbsp\;</p> END:VEVENT END:VCALENDAR