Graduate School and Research Center in Digital Sciences

Random feature expansions for deep Gaussian processes

Filippone, Maurizio

FIMI 2020, Workshop on Functional Inference and Machine Intelligence, 17-19 February 2020, Sophia Antipolis, France

Drawing meaningful conclusions on the way complex real life phenomena work and being able to predict the behavior of systems of interest require developing accurate and highly interpretable mathematical models whose parameters need to be estimated from observations. In modern applications, however, we are often challenged with the lack of such models, and even when these are available they are too computational demanding to be suitable for standard parameter optimization/inference methods. While probabilistic models based on Deep Gaussian Processes (DGPs) offer attractive tools to tackle these challenges in a principled way and to allow for a sound quantification of uncertainty, carrying out inference for these models poses huge computational challenges that arguably hinder their wide adoption. In this talk, I will present our contribution to the development of practical and scalable inference for DGPs, which can exploit distributed and GPU computing. In particular, I will introduce a formulation of DGPs based on random features that we infer using stochastic variational inference.  

Bibtex

Title:Random feature expansions for deep Gaussian processes
Type:Talk
Language:English
City:Sophia Antipolis
Country:FRANCE
Date:
Department:Data Science
Eurecom ref:6172
Copyright: © EURECOM. Personal use of this material is permitted. The definitive version of this paper was published in FIMI 2020, Workshop on Functional Inference and Machine Intelligence, 17-19 February 2020, Sophia Antipolis, France and is available at :
Bibtex: @talk{EURECOM+6172, year = {2020}, title = {{R}andom feature expansions for deep {G}aussian processes}, author = {{F}ilippone, {M}aurizio}, number = {EURECOM+6172}, month = {02}, institution = {Eurecom} address = {{S}ophia {A}ntipolis, {FRANCE}}, url = {http://www.eurecom.fr/publication/6172} }
See also: