AI Paper Atlas

Research map / Probabilistic and scientific modeling / Bayesian optimization

Bayesian Optimization with Cost-varying Variable Subsets

Sebastian Shenghong Tay, Chuan-Sheng Foo, Daisuke Urano, Richalynn Leong, Bryan Kian Hsiang Low

NeurIPS 2023 · Probabilistic and scientific modeling › Bayesian optimization · approach: Regret bounds for Bayesian optimization · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use UCB-CVS with cheap control sets for exploration instead of only the full-query control set (TS-PSQ, UCB-PSQ)UCB-CVS · GP sample, Hartmann, plant growth; cheap and moderate costssimple regretImproved
increase control set costs from cheap to expensiveUCB-CVS vs TS-PSQ · GP sample, Hartmann, plant growthsimple regret advantageHurt
use fixed plays per cost group (ETC-50, ETC-100) instead of cost-adaptive 4/c_j plays (ETC-Ada) (ablation)UCB-CVS ETC variant · expensive cost setsimple regretHurt
use control sets that are not subsets of each other (airfoil objective)TS-PSQ and UCB-PSQ versus UCB-CVS · Airfoil self-noisesimple regretMixed
increase variance of the random-variable distributions from 0.02 to 0.08 (ablation)UCB-CVS and baselines · 42 experimental settingslowest mean simple regret count: 11, 6, and 25Mixed
divide acquisition scores by cost for TS-PSQ and UCB-PSQ (ablation)TS-PSQ, UCB-PSQ · GP prior, moderate costssimple regret convergenceHurt
divide expected improvement by cost (ablation)EI-PSQ · GP prior, moderate costssimple regretNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy