Distributed robust Gaussian Process regression

Research output: Journal contributionsJournal articlesResearchpeer-review

Authors

We study distributed and robust Gaussian Processes where robustness is introduced by a Gaussian Process prior on the function values combined with a Student-t likelihood. The posterior distribution is approximated by a Laplace Approximation, and together with concepts from Bayesian Committee Machines, we efficiently distribute the computations and render robust GPs on huge data sets feasible. We provide a detailed derivation and report on empirical results. Our findings on real and artificial data show that our approach outperforms existing baselines in the presence of outliers by using all available data.
Original languageEnglish
JournalKnowledge and Information Systems
Volume55
Issue number2
Pages (from-to)415-435
Number of pages21
ISSN0219-1377
DOIs
Publication statusPublished - 01.05.2018

    Research areas

  • Business informatics - Distributed computation, Gaussian process regression, Laplace Approximation, Robust regression, Student-t likelihood

Recently viewed

Publications

  1. Continuous 3D scanning mode using servomotors instead of stepping motors in dynamic laser triangulation
  2. A decoupled MPC using a geometric approach and feedforward action for motion control in robotino
  3. Situated multiplying in primary school
  4. Dynamically changing sequencing rules with reinforcement learning in a job shop system with stochastic influences
  5. Long-term memory predictors of adult language learning at the interface between syntactic form and meaning
  6. Early Detection of Faillure in Conveyor Chain Systems by Wireless Sensor Node
  7. A Switching Cascade Sliding PID-PID Controllers Combined with a Feedforward and an MPC for an Actuator in Camless Internal Combustion Engines
  8. Transductive support vector machines for structured variables
  9. Selecting and Adapting Methods for Analysis and Design in Value-Sensitive Digital Social Innovation Projects: Toward Design Principles
  10. Accounting and Modeling as Design Metaphors for CEMIS
  11. Taking notes as a strategy for solving reality-based tasks in mathematics
  12. Special Issue The Discourse of Redundancy Introduction
  13. Towards a Global Script?
  14. Correlation of Microstructure and Local Mechanical Properties Along Build Direction for Multi-layer Friction Surfacing of Aluminum Alloys
  15. Multiphase-field modeling of temperature-driven intermetallic compound evolution in an Al-Mg system for application to solid-state joining processes
  16. Sliding-Mode-Based Input-Output Linearization of a Peltier Element for Ice Clamping Using a State and Disturbance Observer