We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Hypervolume-based Multi-objective Bayesian Optimization with Student-t Processes

Abstract: Student-$t$ processes have recently been proposed as an appealing alternative non-parameteric function prior. They feature enhanced flexibility and predictive variance. In this work the use of Student-$t$ processes are explored for multi-objective Bayesian optimization. In particular, an analytical expression for the hypervolume-based probability of improvement is developed for independent Student-$t$ process priors of the objectives. Its effectiveness is shown on a multi-objective optimization problem which is known to be difficult with traditional Gaussian processes.
Comments: 5 pages, 3 figures
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:1612.00393 [stat.ML]
  (or arXiv:1612.00393v1 [stat.ML] for this version)

Submission history

From: Joachim van der Herten [view email]
[v1] Thu, 1 Dec 2016 19:41:50 GMT (168kb)

Link back to: arXiv, form interface, contact.