📄 中文摘要
贝叶斯优化（BO）是一种强大的黑箱优化方法，适用于优化制造过程，但在处理高维多阶段系统时，其性能往往受到限制，尤其是在可以观察到中间输出的情况下。标准的贝叶斯优化将过程视为黑箱，忽略了中间观察和潜在的过程结构。部分可观察高斯过程网络（POGPN）将过程建模为有向无环图（DAG）。然而，当观察数据为高维状态空间时间序列时，使用中间观察会面临挑战。通过过程专家知识，可以从高维状态空间数据中提取低维潜在特征。提出了POGPN-JPSS框架，以解决这一问题。
联合参数与状态空间贝叶斯优化：利用过程专业知识加速制造优化
📄 English Summary
Joint Parameter and State-Space Bayesian Optimization: Using Process Expertise to Accelerate Manufacturing Optimization
Bayesian optimization (BO) is a powerful method for optimizing black-box manufacturing processes, yet its performance is often constrained when addressing high-dimensional multi-stage systems with observable intermediate outputs. Standard BO treats the process as a black box, neglecting intermediate observations and the underlying process structure. Partially Observable Gaussian Process Networks (POGPN) model the process as a Directed Acyclic Graph (DAG). However, utilizing intermediate observations becomes challenging when dealing with high-dimensional state-space time series. Process-expert knowledge can be leveraged to extract low-dimensional latent features from high-dimensional state-space data. The proposed POGPN-JPSS framework addresses this issue.