eprintid: 10083559
rev_number: 20
eprint_status: archive
userid: 608
dir: disk0/10/08/35/59
datestamp: 2019-10-18 14:17:56
lastmod: 2021-09-28 22:14:56
status_changed: 2019-10-18 14:18:36
type: proceedings_section
metadata_visibility: show
creators_name: Wilson, JT
creators_name: Hutter, F
creators_name: Deisenroth, MP
title: Maximizing acquisition functions for Bayesian optimization
ispublished: pub
divisions: UCL
divisions: B04
divisions: C05
divisions: F48
note: This version is the version of record. For information on re-use, please refer to the publisher’s terms and conditions.
abstract: Bayesian optimization is a sample-efficient approach to global optimization that
relies on theoretically motivated value heuristics (acquisition functions) to guide
its search process. Fully maximizing acquisition functions produces the Bayes’
decision rule, but this ideal is difficult to achieve since these functions are frequently non-trivial to optimize. This statement is especially true when evaluating
queries in parallel, where acquisition functions are routinely non-convex, highdimensional, and intractable. We first show that acquisition functions estimated
via Monte Carlo integration are consistently amenable to gradient-based optimization. Subsequently, we identify a common family of acquisition functions, including EI and UCB, whose properties not only facilitate but justify use of greedy
approaches for their maximization.
date: 2018-12-08
date_type: published
publisher: Neural Information Processing Systems (NIPS)
official_url: https://papers.nips.cc/paper/8194-maximizing-acquisition-functions-for-bayesian-optimization.pdf
oa_status: green
full_text_type: pub
language: eng
primo: open
primo_central: open_green
verified: verified_manual
elements_id: 1705341
lyricists_name: Deisenroth, Marc
lyricists_id: MDEIS71
actors_name: Deisenroth, Marc
actors_id: MDEIS71
actors_role: owner
full_text_status: public
series: Neural Information Processing Systems (NIPS)
publication: Proceedings of the 32nd Conference on Neural Information Processing Systems (NIPS 2018)
volume: 31
place_of_pub: Montreal, QC, Canada
pages: 12
event_title: 32nd Conference on Neural Information Processing Systems (NIPS 2018), 3-8 December 2018, Montreal, QC, Canada
event_location: Montreal, CANADA
issn: 1049-5258
book_title: Proceedings of the 32nd Conference on Neural Information Processing Systems (NIPS 2018)
editors_name: Bengio, S
editors_name: Wallach, H
editors_name: Larochelle, H
editors_name: Grauman, K
editors_name: CesaBianchi, N
editors_name: Garnett, R
citation:        Wilson, JT;    Hutter, F;    Deisenroth, MP;      (2018)    Maximizing acquisition functions for Bayesian optimization.                     In: Bengio, S and Wallach, H and Larochelle, H and Grauman, K and CesaBianchi, N and Garnett, R, (eds.) Proceedings of the 32nd Conference on Neural Information Processing Systems (NIPS 2018).    Neural Information Processing Systems (NIPS): Montreal, QC, Canada.       Green open access   
 
document_url: https://discovery.ucl.ac.uk/id/eprint/10083559/1/1805.10196v2.pdf