As an important problem of causal inference, we discuss the estimation of treatment effects (TEs) under unobserved confounding. Representing the confounder as a latent variable, we propose Intact-VAE, a new variant of variational autoencoder (VAE), motivated by the prognostic score that is sufficient for identifying TEs. Our VAE also naturally gives representation balanced for treatment groups, using its prior. Experiments on (semi-)synthetic datasets show state-of-the-art performance under diverse settings. Based on the identifiability of our model, further theoretical developments on identification and consistent estimation are also discussed. This paves the way towards principled causal effect estimation by deep neural networks.
翻译:作为重要的因果推断问题,我们讨论了在未观察到的混乱下对治疗效果的估计。我们以混凝土者为潜在变量,提出Intact-VAE,这是一个新的变异自动电解器变异变异变体(VAE),其动机是预测性分数足以确定TE。我们的VAE还自然地利用先前的数据,平衡了治疗组的代表性。关于(半)合成数据集的实验显示了不同环境中的最新性能。根据我们模型的可识别性,我们还讨论了关于识别和一致估算的进一步理论发展。这为深层神经网络有原则的因果关系估计铺平了道路。