Rethinking Trust Region Bayesian Optimization in High Dimensions

2026-04-24

AI Summary

This work addresses the challenge of high-dimensional black-box optimization, where Trust Region Bayesian Optimization methods such as TuRBO suffer from degraded local Gaussian process (GP) models due to fixed length scales, leading to difficulties in balancing prior complexity with problem dimensionality. To overcome this limitation, the authors propose AdaScale-TuRBO, which adaptively scales the GP length scale jointly with respect to both the problem dimension and the trust region size, thereby preserving the geometric consistency of the kernel function. This adaptive strategy effectively mitigates overfitting or underfitting of the local GP surrogate in high dimensions. Empirical evaluations demonstrate that AdaScale-TuRBO consistently and significantly outperforms standard TuRBO and other state-of-the-art high-dimensional Bayesian optimization algorithms across multiple synthetic benchmarks and real-world trajectory planning tasks.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Intelligent Robots: Learning & Optimization for ROB
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Trust Region Bayesian Optimization (TuRBO) is an effective strategy for alleviating the curse of dimensionality in high-dimensional black-box optimization. However, inappropriate lengthscale design can cause the local Gaussian process (GP) model within the trust region to degenerate, leading to suboptimal performance in high dimensions. In this work, we show that TuRBO's local GP may remain either excessively complex or overly simple as the dimension $D$ and trust region side length $L$ vary. To address this issue, we propose a straightforward variant, AdaScale-TuRBO, which scales the GP lengthscale with both the problem dimension and trust region size, thereby preserving kernel geometry and maintaining consistent prior complexity. Empirically, we show that AdaScale-TuRBO can robustly outperform standard TuRBO and other popular high-dimensional BO methods on synthetic benchmarks and real-world trajectory planning tasks.

Problem

• high-dimensional optimization
• Bayesian optimization
• trust region
• Gaussian process
• lengthscale

Innovation

• Trust Region Bayesian Optimization
• Gaussian Process
• Lengthscale Adaptation
• High-dimensional Optimization
• AdaScale-TuRBO