arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.20138cs.LG
收藏

基于流匹配的上下文批量黑盒优化

In-Context Batch Black-box Optimization via Flow Matching

ELLIS Institute Finland(芬兰ELLIS研究所);Department of Computer Science, Aalto University(阿尔托大学计算机系);Department of Computer Science, University of Manchester(曼彻斯特大学计算机系)

Quoc-Anh Hoang Nguyen, Xinyu Zhang, Daolang Huang, Julien Martinelli, Samuel Kaski

AI总结:

本文针对高成本并行评估场景下的批量黑盒优化难题,提出FMTR摊销式方法,通过条件流匹配训练后可高效生成批量,在合成与真实基准上性能优异且计算成本低。

摘要:

批量黑盒优化对可并行评估多个候选方案、但仅能负担少量评估轮次的高成本系统至关重要。在这种场景下,有效的批量构建需识别多个有前景的区域,同时避免冗余评估;而传统贝叶斯优化(Batch BO)随着批量规模增大,会需要愈发困难的批量选择流程。我们提出Flow Matching for Thompson-Sampling Refinement(FMTR),这是一种摊销式方法,可离线学习连续的、依赖历史的提议分布,从中可采样可变大小的批量。FMTR通过条件流匹配在Thompson采样的argmax上进行离线训练,这些argmax通过依赖视界的朗之万前推映射被细化到高价值区域,从而在优化视界上调整探索-利用权衡。在测试时,模型参数保持固定,批量生成既不需要针对特定任务的代理模型重拟合,也不需要数值获取优化。FMTR在合成及真实世界基准测试中,相较于批量BO和上下文基线方法取得了优异性能,同时大幅降低了批量选择的计算成本。

英文摘要:

Batch black-box optimization is important for expensive systems where many candidates can be evaluated in parallel, yet only a few rounds of evaluation are affordable. In this regime, effective batch construction must identify multiple promising regions while avoiding redundant evaluations, and conventional Bayesian optimization can require increasingly difficult batch-selection procedures as the batch size grows. We introduce Flow Matching for Thompson-Sampling Refinement (FMTR), an amortized approach that learns offline a continuous, history-conditioned proposal distribution from which variable-size batches can be sampled. FMTR is trained offline by conditional flow matching on Thompson-sampling argmaxes refined toward high-value regions via a horizon-dependent Langevin pushforward map, adapting the exploration-exploitation tradeoff over the optimization horizon. At test time, the model parameters remain fixed and batch generation requires neither task-specific surrogate refitting nor numerical acquisition optimization. FMTR achieves strong performance against batch BO and in-context baselines on synthetic and real-world benchmarks, while substantially reducing the computational cost of batch selection.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑