Causal Discovery from Incomplete Data: A Deep Learning Approach

Abstract

As systems are getting more autonomous with the development of artificial intelligence, it is important to discover the causal knowledge from observational sensory inputs. By encoding a series of cause-effect relations between events, causal networks can facilitate the prediction of effects from a given action and analyze their underlying data generation mechanism. However, missing data are ubiquitous in practical scenarios. Directly performing existing casual discovery algorithms on partially observed data may lead to the incorrect inference. To alleviate this issue, we proposed a deep learning framework, dubbed Imputated Causal Learning (ICL), to perform iterative missing data imputation and causal structure discovery. Through extensive simulations on both synthetic and real data, we show that ICL can outperform state-of-the-art methods under different missing data mechanisms.

References in corpus (10)

• Auto-Encoding Variational Bayes
• GAIN: Missing Data Imputation using Generative Adversarial Nets
• Distinguishing cause from effect using observational data: methods and benchmarks
• DAG-GNN: DAG Structure Learning with Graph Neural Networks
• Learning Bayesian Networks with the bnlearn R Package
• Identifiability of Causal Graphs using Functional Models
• Anomaly Detection and Redundancy Elimination of Big Sensor Data in Internet of Things
• Causal Discovery in the Presence of Missing Data
• Efficient Algorithms for Bayesian Network Parameter Learning from Incomplete Data
• Causal Discovery with Cascade Nonlinear Additive Noise Models

Cited by in corpus (4)

• Directed Acyclic Graphs With Tears
• On the Convergence of Continuous Constrained Optimization for Structure Learning
• iReason: Multimodal Commonsense Reasoning using Videos and Natural Language with Interpretability
• Identifiability of AMP chain graph models