AI Paper Atlas

Research map / Probabilistic and scientific modeling / Bayesian optimization

No-Regret Bayesian Optimization with Unknown Hyperparameters

Felix Berkenkamp, Angela P. Schoellig, Andreas Krause

JMLR 2019 · Probabilistic and scientific modeling › Bayesian optimization · approach: Acquisition function design · 50 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
add hyperparameter adaptation (A-GP-UCB) to MAP estimationGP-UCB with MAP · Synthetic example function (Gaussian kernel)cumulative regretImproved
use HMC hyperparameter estimation instead of adaptation (ablation)GP-UCB · Synthetic example function (Gaussian kernel)cumulative regretHurt
scale lengthscales aggressively by line search without lower bound (ablation)modified prior method · Synthetic example function (Gaussian kernel)cumulative regretHurt
misspecify hyperparameters in GP-UCB (wrong B0) (ablation)GP-UCB · Samples from a GP (B=4, B0=0.25)convergenceHurt
add hyperparameter adaptation when B0 and lengthscales are misspecifiedGP-UCB · Samples from a GP (B=4, B0=0.25, theta0=1)convergenceImproved
use one-step regret estimate instead of upper-bound regret estimator for h(t) (ablation)A-GP-UCB · Synthetic example functioneffective scaling noiseNo clear effect
add hyperparameter adaptation to MAP-based BOGP-UCB with MAP · MNIST logistic regression tuning (d=4)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