跳到正文
arXiv cs.AI· Angelo Ziletti·· 4 小时前AI 评分32

工业过程目标导向因果发现的合成基准发布

A Synthetic Benchmark for Target-Centric Causal Discovery in Industrial Processes

AI 导读

研究者提出一个面向工业过程目标导向因果发现的合成基准,其生成器是受批量生产启发的两层结构因果模型,由控制回路、延迟效应等组件构成。图谱围绕目标变量构建,各组件可独立设置,从而逐项改变识别难度。研究评测了从关联度量到效应估计的八种方法,结果显示随难度叠加所有方法均出现退化,且退化方式与其各自假设一致。

正文 · 原文

View PDF HTML (experimental)

Abstract:Causal discovery promises to find which recorded variables drive a process target such as product quality, yield or energy use. Process data with known causes are rare, and in existing benchmarks the difficulty of identifying the causes of a target is either fixed by the recorded data or set for the whole graph. We present a synthetic benchmark for target-centric causal discovery in industrial processes. Its generator is a two-level structural causal model inspired by batch production, assembled from components that each model one phenomenon of process data, such as a control loop or a delayed effect. The graph is built around the target: without sensor error, every non-cause can be distinguished from the direct causes by conditioning on them or by time order. Each component is set independently, so the difficulty can be varied one condition at a time. We evaluate eight methods, from association measures to effect estimation, with an error analysis by the role of the selected candidate; all degrade as the difficulties combine, each in a way consistent with its assumptions.
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.17556 [cs.LG]
  (or arXiv:2611.17556v1 [cs.LG] for this version)
  https://doi.org/10.48550/arXiv.2611.17556

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Angelo Ziletti [view email]
[v1] Thu, 12 Nov 2026 22:22:25 UTC (396 KB)

来源:arXiv cs.AI · arxiv.org