Parallel Predictive Entropy Search for Batch Global Optimization of Expensive Objective Functions
Amar Shah Department of Engineering Cambridge University、Zoubin Ghahramani Department of Engineering University of Cambridge
Abstract
We develop parallel predictive entropy search (PPES), a novel algorithm for Bayesian optimization of expensive black-box objective functions. At each iteration, PPES aims to select a batch of points which will maximize the information gain about the global maximizer of the objective. Well known strategies exist for suggesting a single evaluation point based on previous observations, while far fewer are known for selecting batches of points to evaluate in parallel. The few batch selection schemes that have been studied all resort to greedy methods to compute an optimal batch. To the best of our knowledge, PPES is the first non-greedy batch Bayesian optimization strategy. We demonstrate the benefit of this approach in optimization performance on both synthetic and real world applications, including problems in machine learning, rocket science and robotics.
中文速览
贝叶斯优化（Bayesian optimization）在实际工程中常常需要同时测试多个方案以节省时间，但现有的批量选点策略都依赖贪心算法——逐个填充候选点，容易陷入局部最优。本文提出"并行预测熵搜索"（Parallel Predictive Entropy Search，PPES），其核心思路是一次性选出整批评估点，使这批点能够最大化我们对目标函数全局最优位置的信息增益，从根本上避免了贪心决策。为了让这个本来难以计算的信息量变得可处理，作者利用互信息的对称性重写目标函数，再借助期望传播（Expectation Propagation）对关键的概率分布做高斯近似，从而得到实用的计算方案。在合成函数、机器学习超参数调优、火箭设计和机器人步态控制等多个真实任务上，PPES均优于现有的贪心批量方法，证明了非贪心批量选点策略的实际价值。
原文 arXiv:1511.07130;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1511.07130v1