Examples include understanding the effect of medications on a patient's health, or of teaching methods on a student's chance of graduation.

Using the predefined generative process for this dataset, we can also write P [ Z = z | X = x ]= P [ Z = z ] P [ X = x | Z = z ] P [ X = x ] = 0 . Within DGSE and DMSE, the latent variable is modeled as a Gaussian. While our method is an instance of generative models, we identify the following key differences: 1. We propose new generative model architectures that extend existing models (e.g., DSE, They are also simpler: they don't require auxiliary networks (e.g., like in CEV AE [ Appendix H.5.1 empirically shows that DMSE model compares favorably against CEV AE on synthetic We consider 100 replicates of this dataset, where the output is simulated according to setting'A ' of NPCI package [ The hidden layers have size of 20 units. The IHDP-Full Setting There are 25 input features in this experimental setting.

Although some simple models with unobserved con-founders are known to produce correct results [Angrist et al., 1996, Pearl et al., 2016, Miao et al.,

Although some simple models with unobserved con-founders are known to produce correct results [Angrist et al., 1996, Pearl et al., 2016, Miao et al.,

Using the predefined generative process for this dataset, we can also write P [ Z = z | X = x ]= P [ Z = z ] P [ X = x | Z = z ] P [ X = x ] = 0 . Within DGSE and DMSE, the latent variable is modeled as a Gaussian. While our method is an instance of generative models, we identify the following key differences: 1. We propose new generative model architectures that extend existing models (e.g., DSE, They are also simpler: they don't require auxiliary networks (e.g., like in CEV AE [ Appendix H.5.1 empirically shows that DMSE model compares favorably against CEV AE on synthetic We consider 100 replicates of this dataset, where the output is simulated according to setting'A ' of NPCI package [ The hidden layers have size of 20 units. The IHDP-Full Setting There are 25 input features in this experimental setting.

Determining causal effects of interventions onto outcomes from real-world, observational (non-randomized) data, e.g., treatment repurposing using electronic health records, is challenging due to underlying bias. Causal deep learning has improved over traditional techniques for estimating individualized treatment effects (ITE). We present the Doubly Robust Variational Information-theoretic Deep Adversarial Learning (DR-VIDAL), a novel generative framework that combines two joint models of treatment and outcome, ensuring an unbiased ITE estimation even when one of the two is misspecified. DR-VIDAL integrates: (i) a variational autoencoder (VAE) to factorize confounders into latent variables according to causal assumptions; (ii) an information-theoretic generative adversarial network (Info-GAN) to generate counterfactuals; (iii) a doubly robust block incorporating treatment propensities for outcome predictions. On synthetic and real-world datasets (Infant Health and Development Program, Twin Birth Registry, and National Supported Work Program), DR-VIDAL achieves better performance than other non-generative and generative methods. In conclusion, DR-VIDAL uniquely fuses causal assumptions, VAE, Info-GAN, and doubly robustness into a comprehensive, performant framework. Code is available at: https://github.com/Shantanu48114860/DR-VIDAL-AMIA-22 under MIT license.

Learning individual-level causal effects from observational data, such as inferring the most effective medication for a specific patient, is a problem of growing importance for policy makers. The most important aspect of inferring causal effects from observational data is the handling of confounders, factors that affect both an intervention and its outcome. A carefully designed observational study attempts to measure all important confounders. However, even if one does not have direct access to all confounders, there may exist noisy and uncertain measurement of proxies for confounders. We build on recent advances in latent variable modeling to simultaneously estimate the unknown latent space summarizing the confounders and the causal effect. Our method is based on Variational Autoencoders (VAE) which follow the causal structure of inference with proxies. We show our method is significantly more robust than existing methods, and matches the state-of-the-art on previous benchmarks focused on individual treatment effects.