BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Geometric Gaussian Processes: Viacheslav Borovitskiy\, ETH Zürich
  - Viacheslav Borovitskiy\, ETH Zürich
DTSTART:20230209T140000Z
DTEND:20230209T150000Z
UID:TALK197065@talks.cam.ac.uk
CONTACT:Kimberly Cole
DESCRIPTION:Gaussian processes (GPs) are often considered to be the gold s
 tandard in settings where well-calibrated predictive uncertainty is of utt
 er importance\, such as decision making.\n\nIt is important for applicatio
 ns to have a class of  “general purpose” GPs. Traditionally\, these ar
 e the stationary processes\, e.g. RBF or Matérn GPs\, at least for the us
 ual vectorial inputs. For non-vectorial inputs\, however\, there is often 
 no such class. This state of affairs hinders the use of GPs in a number of
  application areas ranging from robotics to drug design.\n\nIn this talk\,
  I will consider GPs taking inputs on a manifold\, on a node set of a grap
 h\, or in a discrete “space” of graphs. I will discuss a framework for
  defining the appropriate general purpose GPs\, as well as the analytic an
 d numerical techniques that make them tractable.\n\nBio:\nViacheslav Borov
 itskiy is a researcher interested in mathematically rich problems in machi
 ne learning. His works in the area received paper awards at the ICML & AIS
 TATS conferences.\n\nViacheslav obtained his PhD in the field of mathemati
 cs (harmonic analysis) from St. Petersburg Department of Steklov Mathemati
 cal Institute (PDMI RAS) in 2022.\n\nHaving received the ETH Zürich Postd
 octoral Fellowship\, he is now a postdoc at the Learning & Adaptive System
 s Group of ETH Zürich led by Prof. Andreas Krause.
LOCATION:CBL Seminar Room
END:VEVENT
END:VCALENDAR