BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
CALSCALE:GREGORIAN
PRODID:-//WordPress - MECv7.32.0//EN
X-ORIGINAL-URL:https://dcn.nat.fau.eu/
X-WR-CALNAME:
X-WR-CALDESC:FAU DCN-AvH. Chair for Dynamics, Control, Machine Learning and Numerics -Alexander von Humboldt Professorship
X-WR-TIMEZONE:Europe/Berlin
BEGIN:VTIMEZONE
TZID:Europe/Berlin
X-LIC-LOCATION:Europe/Berlin
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20260329T030000
RRULE:FREQ=YEARLY;BYMONTH=03;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20261025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=4SU
END:STANDARD
END:VTIMEZONE
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-PUBLISHED-TTL:PT1H
X-MS-OLK-FORCEINSPECTOROPEN:TRUE
BEGIN:VEVENT
CLASS:PUBLIC
UID:MEC-a522fbd52ff0d8e2c9faf085e7ec0966@dcn.nat.fau.eu
DTSTART;TZID=Europe/Berlin:20210521T111500
DTEND;TZID=Europe/Berlin:20210521T121500
DTSTAMP:20211020T083541Z
CREATED:20211020
LAST-MODIFIED:20240924
PRIORITY:5
SEQUENCE:1
TRANSP:OPAQUE
SUMMARY:Paris Bachelier Seminar: Turpike Control and Machine Learning
DESCRIPTION:Tomorrow May 21st. our Head Enrique Zuazua will talk on the Paris Bachelier Seminar about “Turpike Control and Machine Learning“, an event organized jointly with the Centre de Mathematiques Appliquées (Ecole Polytechnique), the CEREMADE (Université Paris Dauphine), the Centre de Recherche en Economie et Statistiques (CREST), the Laboratoire de Probabilités et modèles aleatoires (Universite de Paris VI- VII), the Laboratoire de Mathématiques et de Modélisation (Université d’Evry), and the Université de Marne La Vallée.\nAbstract. We first present the classical concept of turnpike control, assuring that, often, in long time-horizons, optimal controls and trajectories are essentially of a steady-state nature. We then show how it can be adapted to the models in Continuum Mechanics, leading a to substantial reduction of the computational cost. We then analyze Neural Ordinary Differential Equations (NODEs) from a control theoretical perspective to address some of the main challenges in Machine Learning and, in particular, data classification and Universal Approximation. More precisely, we adopt the perspective of the simultaneous control of systems of NODEs. We show how the turnpike principle can be adapted in this setting. We develop genuinely nonlinear and constructive methods allowing to estimate the complexity of the control strategies we develop. We also present the counterparts in the context of the control of neural transport equations, establishing a link towards optimal transport. Some open problems and perspective of research are also present.\nWHEN?\nFri. May 21, 2021 at 11:15H\nWHERE?\nVia zoom meeting\nSee the official page of the event\n
URL:https://dcn.nat.fau.eu/events/paris-bachelier-seminar-turpike-control-and-machine-learning/
CATEGORIES:EZuazua,Seminar/Talk
LOCATION:Worldwide
ATTACH;FMTTYPE=image/png:https://dcn.nat.fau.eu/wp-content/uploads/talk-EZuazua-21may2021.png
END:VEVENT
END:VCALENDAR