Abstract:In many fields of scientific research and real-world applications, unbiased estimation of causal effects from non-experimental data is crucial for understanding the mechanism underlying the data and for decision-making on effective responses or interventions. A great deal of research has been conducted to address this challenging problem from different angles. For estimating causal effect in observational data, assumptions such as Markov condition, faithfulness, and causal sufficiency are always made. Under the assumptions, full knowledge, such as a set of covariates or an underlying causal graph, is typically required. A practical challenge is that in many applications, no such full knowledge or only some partial knowledge is available. In recent years, research has emerged to use search strategies based on graphical causal modelling to discover useful knowledge from data for causal effect estimation, with some mild assumptions, and has shown promise in tackling the practical challenge. In this survey, we review these data-driven methods on causal effect estimation for a single treatment with a single outcome of interest and focus on the challenges faced by data-driven causal effect estimation. We concisely summarise the basic concepts and theories that are essential for data-driven causal effect estimation using graphical causal modelling but are scattered around the literature. We identify and discuss the challenges faced by data-driven causal effect estimation and characterise the existing methods by their assumptions and the approaches to tackling the challenges. We analyse the strengths and limitations of the different types of methods and present an empirical evaluation to support the discussions. We hope this review will motivate more researchers to design better data-driven methods based on graphical causal modelling for the challenging problem of causal effect estimation.

DoWhy-GCM: An extension of DoWhy for causal inference in graphical causal models

DoWhy: Addressing Challenges in Expressing and Validating Causal Assumptions

causalgraph: A Python Package for Modeling, Persisting and Visualizing Causal Graphs Embedded in Knowledge Graphs

Mixed Graphical Models for Causal Analysis of Multi-modal Variables

Causalvis: Visualizations for Causal Inference

Deep Learning With DAGs

Graph Disentangle Causal Model: Enhancing Causal Inference in Networked Observational Data

Identification of Latent Variables From Graphical Model Residuals

BayesDAG: Gradient-Based Posterior Inference for Causal Discovery

Causal GNNs: A GNN-Driven Instrumental Variable Approach for Causal Inference in Networks

Continuous Bayesian Model Selection for Multivariate Causal Discovery

Dagma-DCE: Interpretable, Non-Parametric Differentiable Causal Discovery

Data-Driven Causal Effect Estimation Based on Graphical Causal Modelling: A Survey

Personalized Binomial DAGs Learning with Network Structured Covariates

Causality is all you need

Causal Inference for Knowledge Graph Completion

A Survey of Out-of-distribution Generalization for Graph Machine Learning from a Causal View

Causal Concept Graph Models: Beyond Causal Opacity in Deep Learning

When Graph Neural Network Meets Causality: Opportunities, Methodologies and An Outlook

Causal Discovery via Causal Star Graphs

OrphicX: A Causality-Inspired Latent Variable Model for Interpreting Graph Neural Networks