Abstract

Gaussian processes (GPs) are often considered to be the gold standard in settings where well-calibrated predictive uncertainty is of utter importance, such as decision making.

It is important for applications to have a class of “general purpose” GPs. Traditionally, these are the stationary processes, e.g. RBF or Matérn GPs, at least for the usual 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.

In this talk, I will consider GPs taking inputs on a manifold, on a node set of a graph, or in a discrete “space” of graphs. I will discuss a framework for defining the appropriate general purpose GPs, as well as the analytic and numerical techniques that make them tractable.

Bio: Viacheslav Borovitskiy is a researcher interested in mathematically rich problems in machine learning. His works in the area received paper awards at the ICML & AISTATS conferences.

Viacheslav obtained his PhD in the field of mathematics (harmonic analysis) from St. Petersburg Department of Steklov Mathematical Institute (PDMI RAS ) in 2022.

Having received the ETH Z ürich Postdoctoral Fellowship, he is now a postdoc at the Learning & Adaptive Systems Group of ETH Z ürich led by Prof. Andreas Krause.