AI Paper Atlas

Research map / Probabilistic and scientific modeling / Bayesian optimization

ALAS: Additive Learnable Alpha-Stable Kernels for Flexible Bayesian Optimization

Weibo Huang, Cheng Hua

ICML 2026 · Probabilistic and scientific modeling › Bayesian optimization · approach: Local and high-dimensional 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
learn the stability parameter alpha instead of RBF fixed Gaussian smoothness (ablation)GP with 1D ALAS component · Weierstrass function (n=25, single run)average test PLL: -0.16 -> 0.23Improved
learn alpha on a smooth function sampled from an RBF GP (ablation)GP with 1D ALAS component · Smooth RBF-GP sample (n=25)average test PLL: RBF 2.791 vs ALAS 2.777No clear effect
use a single d-dimensional ALAS kernel with shared alpha instead of the separable ALAS-Sep (ablation)ALAS · Rosenbrock-10d, Robot-4d, Portfolio-5dlog optimality gapImproved
use separable ALAS-Sep with per-dimension alpha_j (ablation)ALAS · Weierstrass-3d, Levy-20d, Rastrigin-30dlog optimality gapImproved
learn an additional alpha parameter during early BO iterations (ablation)ALAS · several BO benchmarkslog optimality gapHurt
learn alpha on smoother benchmarks instead of a fixed Gaussian kernelALAS · Exponential-5d, Hartmann-3d/6dlog optimality gap: alpha converges to 1.96, 1.86, 1.92No 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