BEGIN:VCALENDAR
CALSCALE:GREGORIAN
PRODID:iCalendar-Ruby
VERSION:2.0
BEGIN:VEVENT
DESCRIPTION: How to use discrete mathematics and theoretical computer scien
 ce to understand neural networks? Guided by this question\, I will focus on
  neural networks with rectified linear unit (ReLU) activations\, a standard
  model and important building block in modern machine learning pipelines. T
 he functions represented by such networks are continuous and piecewise line
 ar. But how does the set of representable functions depend on the architect
 ure? And how difficult is it to train such networks to optimality? In my ta
 lk I will answer fundamental questions like these using methods from polyhe
 dral geometry\, combinatorial optimization\, and complexity theory. This st
 ream of research was started during my doctorate within " Facets of Complex
 ity " and carried much further since then. 
DTSTAMP:20240124T145700
DTSTART:20240129T141500
CLASS:PUBLIC
LOCATION:Freie Universität Berlin \n Institut für Informatik \n Takustr. 9 
 \n 14195 Berlin \n Seminar room 053 (ground floor)
SEQUENCE:0
SUMMARY:Christoph Hertrich (Frankfurt): (Old and New) Facets of Neural Netw
 ork Complexity
UID:138229135@/www.mi.fu-berlin.de
URL:https://www.mi.fu-berlin.de/en/facetsofcomplexity/monday/20240129-L-Her
 trich.html
END:VEVENT
END:VCALENDAR