Variational dependent multi-output Gaussian Process dynamical systems

  • Jing Zhao*
  • , Shiliang Sun
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

This paper presents a dependent multi-output Gaussian process (GP) for modeling complex dynamical systems. The outputs are dependent in this model, which is largely different from previous GP dynamical systems. We adopt convolved multi-output GPs to model the outputs, which are provided with a flexible multi-output covariance function. We adapt the variational inference method with inducing points for approximate posterior inference of latent variables. Conjugate gradient based optimization is used to solve parameters involved. Besides the temporal dependency, the proposed model also captures the dependency among outputs in complex dynamical systems. We evaluate the model on both synthetic and real-world data, and encouraging results are observed.

Original languageEnglish
Title of host publicationDiscovery Science - 17th International Conference, DS 2014, Proceedings
EditorsSašo Džeroski, Panče Panov, Dragi Kocev, Ljupčo Todorovski
PublisherSpringer Verlag
Pages350-361
Number of pages12
ISBN (Electronic)9783319118116
DOIs
StatePublished - 2014
Event17th International Conference on Discovery Science, DS 2014 - Bled, Slovenia
Duration: 8 Oct 201410 Oct 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8777
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference17th International Conference on Discovery Science, DS 2014
Country/TerritorySlovenia
CityBled
Period8/10/1410/10/14

Keywords

  • Dynamical system
  • Gaussian process
  • Multi-output modeling
  • Variational inference

Fingerprint

Dive into the research topics of 'Variational dependent multi-output Gaussian Process dynamical systems'. Together they form a unique fingerprint.

Cite this