AI Paper Atlas

Research map / Probabilistic and scientific modeling / Bayesian optimization

Local Entropy Search over Descent Sequences for Bayesian Optimization

David Stenger, Armin Lindicke, Alexander von Rohr, Sebastian Trimpe

ICLR 2026 · Probabilistic and scientific modeling › Bayesian optimization · approach: Local and high-dimensional Bayesian optimization · 2 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
target descent sequence of local optimizer (LES) instead of trust region (TuRBO)TuRBO · GP samples within-model, high complexity, d=50median final objective: -7.1 -> -10.8Improved
use LES instead of global logEIlogEI · GP samples within-model, high complexity, d=50median final objective: -2.9 -> -10.8Improved
use LES instead of logEI at low dimensionlogEI · GP samples within-model, high complexity, d=5median final objective: -3.9 vs -2.8Hurt
use LES instead of logEIlogEI · GP samples within-model, medium complexity, d=50median final objective: -9.6 -> -10.4Improved
use LES instead of logEIlogEI · GP samples within-model, low complexity, d=50median final objective: -8.5 -> -8.8No clear effect
use LES instead of logEIlogEI · GP samples within-model, extremely low complexity, d=50median final objective: -3.0 -> -3.0No clear effect
use LES instead of TuRBO under MAP hyperparametersTuRBO · GP samples out-of-model, high complexity, d=50median final objective: -7.1 -> -7.8Improved
use LES instead of logEI-DSP under MAP hyperparameterslogEI-DSP · GP samples out-of-model, low complexity, d=50median final objective: -8.1 vs -8.5 for LES: LES betterHurt
use wrong, less complex hyperprior (logEI-DSP vs logEI)logEI · GP samples out-of-model, high complexity, d=50median final objective: -2.9 -> -4.1Improved
run LES on rover and Mopta08 tasksLES · Mopta08regretMixed
use LES with higher L and P (more accurate approximation)LES · GP samplesperformanceImproved
use local Thompson sampling or condition only on final point instead of full descent sequencesLES · GP samplesperformanceHurt

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