Two Optimal Strategies for Active Learning of Causal Models from Interventional Data

Abstract

From observational data alone, a causal DAG is only identifiable up to Markov equivalence. Interventional data generally improves identifiability; however, the gain of an intervention strongly depends on the intervention target, that is, the intervened variables. We present active learning (that is, optimal experimental design) strategies calculating optimal interventions for two different learning goals. The first one is a greedy approach using single-vertex interventions that maximizes the number of edges that can be oriented after each intervention. The second one yields in polynomial time a minimum set of targets of arbitrary size that guarantees full identifiability. This second approach proves a conjecture of Eberhardt (2008) indicating the number of unbounded intervention targets which is sufficient and in the worst case necessary for full identifiability. In a simulation study, we compare our two active learning approaches to random interventions and an existing approach, and analyze the influence of estimation errors on the overall performance of active learning.

References in corpus (3)

• Identifiability of Causal Graphs using Functional Models
• Almost Optimal Intervention Sets for Causal Discovery
• On the Equivalence of Causal Models

Cited by in corpus (27)

• Causal Bandits: Learning Good Interventions via Causal Inference
• Causal Structure Learning: a Combinatorial Perspective
• Active Learning: Problem Settings and Recent Developments
• Learning Causal Graphs with Small Interventions
• Identifying Best Interventions through Online Importance Sampling
• Learning Neural Causal Models with Active Interventions
• Learning Causal Structures Using Regression Invariance
• Data-adaptive Active Sampling for Efficient Graph-Cognizant Classification
• ABCD-Strategy: Budgeted Experimental Design for Targeted Causal Structure Discovery
• Causality and Batch Reinforcement Learning: Complementary Approaches To Planning In Unknown Domains
• GeneDisco: A Benchmark for Experimental Design in Drug Discovery
• Active Learning for Decision-Making from Imbalanced Observational Data
• Active Invariant Causal Prediction: Experiment Selection through Stability
• Marginal integration for nonparametric causal inference
• Efficient Intervention Design for Causal Discovery with Latents
• Interventional Experiment Design for Causal Structure Learning
• Optimal experimental design via Bayesian optimization: active causal structure learning for Gaussian process networks
• Learning Bayesian Networks with Low Rank Conditional Probability Tables
• Active Structure Learning of Causal DAGs via Directed Clique Tree
• Optimal Experiment Design for Causal Discovery from Fixed Number of Experiments
• Intervention Efficient Algorithms for Approximate Learning of Causal Graphs
• Causal learning with sufficient statistics: an information bottleneck approach
• Causal Structure Learning: a Bayesian approach based on random graphs
• Combinatorial and algebraic perspectives on the marginal independence structure of Bayesian networks
• Matching a Desired Causal State via Shift Interventions
• Hierarchical Causal Bandit
• Joint estimation of causal effects from observational and intervention gene expression data