Consistency of Neural Causal Partial Identification

TL;DR

This paper proves the consistency of neural causal models for partial identification in a general setting with continuous and categorical variables, emphasizing the importance of neural network design and Lipschitz regularization.

Contribution

It extends the theoretical guarantees of neural causal models to a broader setting, including continuous variables, and highlights the role of network architecture and regularization.

Findings

Proves consistency of NCMs with mixed variable types.

Shows Lipschitz regularization is crucial for asymptotic consistency.

Provides bounds on sample complexity and error in partial identification.

Abstract

Recent progress in Neural Causal Models (NCMs) showcased how identification and partial identification of causal effects can be automatically carried out via training of neural generative models that respect the constraints encoded in a given causal graph [Xia et al. 2022, Balazadeh et al. 2022]. However, formal consistency of these methods has only been proven for the case of discrete variables or only for linear causal models. In this work, we prove the consistency of partial identification via NCMs in a general setting with both continuous and categorical variables. Further, our results highlight the impact of the design of the underlying neural network architecture in terms of depth and connectivity as well as the importance of applying Lipschitz regularization in the training phase. In particular, we provide a counterexample showing that without Lipschitz regularization this method may not be asymptotically consistent. Our results are enabled by new results on the approximability of Structural Causal Models (SCMs) via neural generative models, together with an analysis of the sample complexity of the resulting architectures and how that translates into an error in the constrained optimization problem that defines the partial identification bounds.