BEGIN:VCALENDAR
VERSION:2.0
CALSCALE:GREGORIAN
PRODID:UW-Madison-Physics-Events
BEGIN:VEVENT
SEQUENCE:0
UID:UW-Physics-Event-6427
DTSTART:20210421T160000Z
DTEND:20210421T171500Z
DTSTAMP:20230129T033543Z
LAST-MODIFIED:20210417T012555Z
LOCATION:Online Seminar: Please sign up for our mailing list at www.ph
ysicsmeetsml.org for zoom link
SUMMARY:Machine Learning for Calabi-Yau metrics\, Physics ∩ ML Semin
ar\, Fabian Ruehle\, CERN and Oxford
DESCRIPTION:String theory is a very promising candidate for a fundamen
tal theory of our universe. An interesting prediction of string theory
is that spacetime is ten-dimensional. Since we only observe four spac
etime dimensions\, the extra six dimensions are small and compact\, th
us evading detection. These extra six-dimensional spaces\, known as Ca
labi-Yau spaces\, are very special and elusive. They come with a speci
fic type of metric needed to make string theory consistent. While we k
now\, thanks to the heroic work of Calabi and Yau\, that this metric e
xists\, we neither know what it looks like nor how to construct it exp
licitly. Thinking of the metric as a function that satisfies three con
straints entering the Calabi-Yau theorem\, we can parameterize the met
ric as a neural network and formulate the problem as multiple continuo
us optimization tasks. I will show that this allows us to approximate
Calabi-Yau metrics to very high accuracy\, which will have important a
pplications in Physics and Mathematics.
URL:https://www.physics.wisc.edu/events/?id=6427
END:VEVENT
END:VCALENDAR