BEGIN:VCALENDAR
CALSCALE:GREGORIAN
VERSION:2.0
METHOD:PUBLISH
PRODID:-//Drupal iCal API//EN
X-WR-TIMEZONE:America/New_York
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
DTSTART:20070311T020000
TZNAME:EDT
TZOFFSETTO:-0400
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
DTSTART:20071104T020000
TZNAME:EST
TZOFFSETTO:-0500
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
SEQUENCE:1
X-APPLE-TRAVEL-ADVISORY-BEHAVIOR:AUTOMATIC
234141
20260403T141941Z
DTSTART;TZID=America/New_York:20260409T140000
DTEND;TZID=America/New_York:2
 0260409T145000
URL;TYPE=URI:https://www.wpi.edu/news/calendar/events/depar
 tment-mathematical-sciences-hj-gay-lecture-adrian-lewis-cornell-university
Department of Mathematical Sciences HJ Gay Lecture: Adrian Lewis, Cornell U
 niversity
\n\n\n      \n      \n\n\n\nDepartment of Mathematical Sciences\nAdrian Lewis, Cornell University\nThursday, April 9th, 2026\n2:
 00PM-2:50PM\nSalisbury Laboratories 104\n\nSpeaker: Adrian Lewis, Cornell 
 University\nTitle: First-order optimization without (much) geometry\nAbstr
 act: Contemporary optimization algorithms and their machine learning appli
 cations rely on derivative-like information for objectives. Unlike the cla
 ssical setting of smooth or convex functions on Euclidean space, complexit
 y analysis for modern algorithms must rely on milder or less familiar geom
 etric foundations. Objectives may be semi-algebraic, as in popular Kurdyka
 ­Lojasiewicz-style analyses, or (in a 2020 method of Zhang et al. ) merel
 y Lipschitz. Furthermore, rather than Euclidean space, the data and variab
 les may live in manifolds, like hyperbolic space or the positive-definite 
 matrices, or in more general nonpositively-curved geodesic spaces, like BH
 V-tree-space phylogenetic models. This talk, which relies on no optimizati
 on background, presents a sequence of short snapshots of this new arena.Jo
 int work with A. Goodwin, Siyu Kong, G. Lopez, A. Nicolae, and Tonghua Tia
 n.\n
END:VEVENT
END:VCALENDAR