BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:Europe/Vienna
BEGIN:DAYLIGHT
DTSTART:20250330T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20241027T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260424T081433Z
UID:648ae4751ba1f228554167@ist.ac.at
DTSTART:20241219T130000
DTEND:20241219T150000
DESCRIPTION:Speaker: Richard Rimanyi\nhosted by Tamas Hausel\nAbstract: In 
 this talk we will enumerate the main reasons for a collection of matrices 
 multiplying to 0. Our motivation is Bayesian Learning Theory\, where one o
 f the goals is to progressively approximate an unknown distribution using 
 data generated from that distribution. A key component in this framework i
 s a function K (relative entropy)\, which is often highly singular. The in
 variants of the singularities of K (in the style of log canonical threshol
 d') are related to how well the Singular Learning Theory "generalizes"---o
 r\, in Machine Learning terms\, how efficiently the model can be trained. 
 Computing the singularity invariants in real-life scenarios of Machine Lea
 rning is notoriously difficult. In this talk\, we focus on an elementary e
 xample and compute the learning coefficients of Linear Neural Networks. Jo
 int work with S. P. Lehalleur.
LOCATION:Office Bldg West / Ground floor / Heinzel Seminar Room (I21.EG.101
 )\, ISTA
ORGANIZER:boosthui@ist.ac.at
SUMMARY:Richard Rimanyi: On a geometric problem of machine learning
URL:https://talks-calendar.ista.ac.at/events/5412
END:VEVENT
END:VCALENDAR