arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.03035stat.MLcs.LG
收藏

因果推断的广义贝叶斯方法

Generalized Bayes for Causal Inference

Munich Center for Machine Learning(慕尼黑机器学习中心)

Emil Javurek, Dennis Frauen, Yuxin Wang, Stefan Feuerriegel

AI总结:

本文提出了一种广义贝叶斯框架,用于因果推断,通过直接在因果估计量上放置先验并利用驱动识别的损失函数,实现对因果效应的完整不确定性量化。

摘要:

不确定性量化是因果机器学习许多应用的核心,但原理上为因果效应进行贝叶斯推断仍然具有挑战性。标准贝叶斯方法通常需要为数据生成过程指定概率模型,包括高维噪声成分,如倾向分数和结果回归。标准后验因此容易受到强建模选择的影响,包括复杂的先验 elicitation。在本文中,我们提出了一种用于因果推断的广义贝叶斯框架。我们的框架避免了显式的似然建模;相反,我们直接在因果估计量上放置先验,并通过驱动识别的损失函数更新这些估计量,从而得到因果效应的广义后验。因此,我们的框架将现有的基于损失的因果估计器转化为具有完整不确定性量化的估计器。我们的框架灵活且适用于广泛类型的因果估计量(例如,平均处理效应ATE、条件平均处理效应CATE)。进一步,我们的框架可以应用于最先进的因果机器学习流水线(例如,Neyman-正交元学习器)。对于Neyman-正交损失,我们证明广义后验收敛到其理想对应物,并且在第一阶段噪声估计误差下保持稳健。通过校准,我们因此在噪声估计器收敛于亚参数速率时仍能获得有效的频率学不确定性。实证上,我们展示了所提出框架在多种因果推断设置中提供校准的因果效应估计。据我们所知,这是第一个灵活的框架,用于构建因果机器学习的广义贝叶斯后验。

英文摘要:

Uncertainty quantification is central to many applications of causal machine learning, yet principled Bayesian inference for causal effects remains challenging. Standard Bayesian approaches typically require specifying a probabilistic model for the data-generating process, including high-dimensional nuisance components such as propensity scores and outcome regressions. Standard posteriors are thus vulnerable to strong modeling choices, including complex prior elicitation. In this paper, we propose a generalized Bayesian framework for causal inference. Our framework avoids explicit likelihood modeling; instead, we place priors directly on the causal estimands and update these using an identification-driven loss function, which yields generalized posteriors for causal effects. As a result, our framework turns existing loss-based causal estimators into estimators with full uncertainty quantification. Our framework is flexible and applicable to a broad range of causal estimands (e.g., ATE, CATE). Further, our framework can be applied on top of state-of-the-art causal machine learning pipelines (e.g., Neyman-orthogonal meta-learners). For Neyman-orthogonal losses, we show that the generalized posteriors converge to their oracle counterparts and remain robust to first-stage nuisance estimation error. With calibration, we thus obtain valid frequentist uncertainty even when nuisance estimators converge at slower-than-parametric rates. Empirically, we demonstrate that our proposed framework offers causal effect estimation with calibrated uncertainty across several causal inference settings. To the best of our knowledge, this is the first flexible framework for constructing generalized Bayesian posteriors for causal machine learning.

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。

↑