BEGIN:VCALENDAR
VERSION:2.0
CALSCALE:GREGORIAN
PRODID:UW-Madison-Physics-Events
BEGIN:VEVENT
SEQUENCE:3
UID:UW-Physics-Event-9092
DTSTART:20250219T220000Z
DTEND:20250219T230000Z
DTSTAMP:20260413T184742Z
LAST-MODIFIED:20250220T184549Z
LOCATION:5280 CH & https://uwmadison.zoom.us/j/93056807183?pwd=bmRBTnF
 pQTZSYk1QSUVLb3BBY1M0QT09
SUMMARY:From Data to Discovery: Machine Learning and Real-Time AI at t
 he Frontier of Particle Physics\, NPAC (Nuclear/Particle/Astro/Cosmo) 
 Forum\, Dr. Abhijith Gandrakota\, Fermilab
DESCRIPTION:Uncovering beyond the Standard Model (BSM) physics near th
 e electroweak (EW) scale remains one of the most formidable challenges
  in high-energy physics\, where subtle signals are often obscured by o
 verwhelming background processes. In this talk\, I demonstrate how adv
 anced AI and machine learning techniques are revolutionizing the searc
 h for new physics at the CMS experiment at the Large Hadron Collider (
 LHC). I start by discussing the application of data scouting and Gauss
 ian Process regression for low-mass hadronic resonances\, which simpli
 fies the detection of slight excesses against smooth backgrounds. Reco
 gnizing the limitations imposed on these searches by the Trigger and D
 ata Acquisition systems\, we introduce AXOL1TL—an innovative anomaly
  detection method utilizing real-time AI tools implemented on FPGA har
 dware that achieves ultra-fast event processing in just 50 ns. Further
  advancements are presented through robust multi-background representa
 tion learning and attention-based methods for event filtering\, which 
 enhance both the efficiency and interpretability of our ML models at t
 he HL-LHC. Together\, these cutting-edge developments pave the way tow
 ard overcoming the challenges of identifying BSM signatures at the EW 
 scale\, opening new avenues for discovery in particle physics.
\n
URL:https://www.physics.wisc.edu/events/?id=9092
END:VEVENT
END:VCALENDAR