BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Iowa State University CALS LAS Web Team//sites.iastate.edu//EN
BEGIN:VEVENT
UID:20261005T212500-4994-www.cs.iastate.edu
DTSTART:20261005T212500Z
SEQUENCE:0
TRANSP:OPAQUE
DTEND:20261005T222500Z
LOCATION:Marston Hall\, Room 2200: https://www.fpm.iastate.edu/maps/buildin
 gs/view_plat.asp?id=1026
SUMMARY:CS Colloquium: Understanding Transformers
CLASS:PUBLIC
DESCRIPTION:Abstract:Transformers have become a powerful tool in modern mac
 hine learning\, but their behavior becomes increasingly difficult to under
 stand as they grow larger and deeper. In this talk\, I will use differenti
 al equations and continuum models to develop a mathematical view of transf
 ormers. The basic idea is to describe the collective behavior of many toke
 ns\, attention heads\, and layers rather than tracking each component indi
 vidually. This leads to a continuous model that captures both the forward 
 computation and the training process. I will discusshow these models relat
 e to finite transformers and what they can tell us about training and conv
 ergence.The talk will illustrate how differential equations can provide a 
 useful mathematical lens for understanding transformers and how they learn
 .\n\nMore information at: https://www.cs.iastate.edu/event/2026/cs-colloqu
 ium-understanding-transformers\n\nMarston Hall\, Room 2200: https://www.fp
 m.iastate.edu/maps/buildings/view_plat.asp?id=1026
DTSTAMP:20261007T071908Z
END:VEVENT
END:VCALENDAR