analiza aiqbench
This research introduces RAMBO, a novel Bayesian Optimization method using Dirichlet Process Mixtures of Gaussian Processes to handle multi-regime search spaces where standard models fail due to uniform smoothness assumptions. By automatically discovering latent regimes with locally optimized hyperparameters and decomposing uncertainty into intra- and inter-regime components, the approach significantly outperforms existing baselines in complex applications like molecular conformation search and fusion reactor design. The work advances machine learning optimization techniques for heterogeneous landscapes without relying on pre-defined regime boundaries.
Regime-Adaptive Bayesian Optimization via Dirichlet Process Mixtures of Gaussian Processes
arXiv:2601.20043v2 Announce Type: replace Abstract: Standard Bayesian Optimization (BO) assumes uniform smoothness across the search space an assumption violated in multi-regime problems such as molecular conformation search through distinct energy basins or drug discovery across heterogeneous molecular scaffolds. A single GP either oversmooths sharp transitions or hallucinates noise in smooth regions, yielding miscalibrated uncertainty. We propose RAMBO, a Dirichlet Process Mixture of Gaussian Processes that automatically discovers latent regimes during optimization, each modeled by an independent GP with locally-optimized hyperparameters. We derive collapsed Gibbs sampling that analytically marginalizes latent functions for efficient inference, and introduce adaptive concentration parameter scheduling for coarse-to-fine regime discovery. Our acquisition functions decompose uncertainty into intra-regime and inter-regime components. Experiments on synthetic benchmarks and real-world applications, including molecular conformer optimization, virtual screening for drug discovery, and fusion reactor design, demonstrate consistent improvements over state-of-the-art baselines on multi-regime objectives.
Abstract:Standard Bayesian Optimization (BO) assumes uniform smoothness across the search space an assumption violated in multi-regime problems such as molecular conformation search through distinct energy basins or drug discovery across heterogeneous molecular scaffolds. A single GP either oversmooths sharp transitions or hallucinates noise in smooth regions, yielding miscalibrated uncertainty. We propose RAMBO, a Dirichlet Process Mixture of Gaussian Processes that automatically discovers latent regimes during optimization, each modeled by an independent GP with locally-optimized hyperparameters. We derive collapsed Gibbs sampling that analytically marginalizes latent functions for efficient inference, and introduce adaptive concentration parameter scheduling for coarse-to-fine regime discovery. Our acquisition functions decompose uncertainty into intra-regime and inter-regime components. Experiments on synthetic benchmarks and real-world applications, including molecular conformer optimization, virtual screening for drug discovery, and fusion reactor design, demonstrate consistent improvements over state-of-the-art baselines on multi-regime objectives.
Machine Learning (cs.LG); Machine Learning (stat.ML)
arXiv:2601.20043 [cs.LG]
(or arXiv:2601.20043v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2601.20043 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 232 cuvinte
Subiecte conexe
Measurement incompatibility in Bayesian multiparameter quantum estimation
Quantum Journal · 13 aug. 2026
Time-partitioning Gaussian Process Mixture Model
arXiv · 16 oct. 2026
Fisher Information based Stochastic Gradient Ascent for Online Learning of Dirichlet Process Mixture and Theory
arXiv · 28 iul. 2026
Joint Surrogate Learning of Objectives, Constraints, and Sensitivities for Efficient Multi-objective Optimization of Neural Dynamical Systems
arXiv · 30 sept. 2026