Meta-learners for Estimating Heterogeneous Treatment Effects using Machine Learning

Abstract

There is growing interest in estimating and analyzing heterogeneous treatment effects in experimental and observational studies. We describe a number of meta-algorithms that can take advantage of any supervised learning or regression method in machine learning and statistics to estimate the Conditional Average Treatment Effect (CATE) function. Meta-algorithms build on base algorithms---such as Random Forests (RF), Bayesian Additive Regression Trees (BART) or neural networks---to estimate the CATE, a function that the base algorithms are not designed to estimate directly. We introduce a new meta-algorithm, the X-learner, that is provably efficient when the number of units in one treatment group is much larger than in the other, and can exploit structural properties of the CATE function. For example, if the CATE function is linear and the response functions in treatment and control are Lipschitz continuous, the X-learner can still achieve the parametric rate under regularity conditions. We then introduce versions of the X-learner that use RF and BART as base learners. In extensive simulation studies, the X-learner performs favorably, although none of the meta-learners is uniformly the best. In two persuasion field experiments from political science, we demonstrate how our new X-learner can be used to target treatment regimes and to shed light on underlying mechanisms. A software package is provided that implements our methods.

References in corpus (1)

• Inference on a New Class of Sample Average Treatment Effects

Cited by in corpus (130)

• Causal machine learning for predicting treatment outcomes
• Discussion on "Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects" by Hahn, Murray and Carvalho
• Machine Learning Estimation of Heterogeneous Causal Effects: Empirical Monte Carlo Evidence
• Double Machine Learning based Program Evaluation under Unconfoundedness
• Quasi-Oracle Estimation of Heterogeneous Treatment Effects
• Response Transformation and Profit Decomposition for Revenue Uplift Modeling
• CausalML: Python Package for Causal Machine Learning
• Causality Learning: A New Perspective for Interpretable Machine Learning
• Applications of statistical causal inference in software engineering
• Conformal Inference of Counterfactuals and Individual Treatment Effects
• Machine Learning Estimation of Heterogeneous Treatment Effects with Instruments
• Generalization Bounds and Representation Learning for Estimation of Potential Outcomes and Causal Effects
• Nonparametric Estimation of Heterogeneous Treatment Effects: From Theory to Learning Algorithms
• Confounding-Robust Policy Improvement
• How well can a large language model explain business processes as perceived by users?
• DESCN: Deep Entire Space Cross Networks for Individual Treatment Effect Estimation
• Modern approaches for evaluating treatment effect heterogeneity from clinical trials and observational data
• Scientific Inference With Interpretable Machine Learning: Analyzing Models to Learn About Real-World Phenomena
• Targeting customers under response-dependent costs
• Experimental Evaluation of Individualized Treatment Rules
• Estimating and Mitigating the Congestion Effect of Curbside Pick-ups and Drop-offs: A Causal Inference Approach
• Free Lunch! Retrospective Uplift Modeling for Dynamic Promotions Recommendation within ROI Constraints
• Machine Learning Methods Economists Should Know About
• A Survey on Causal Inference
• Bridging the gap: Towards an Expanded Toolkit for AI-driven Decision-Making in the Public Sector
• Causal Network Motifs: Identifying Heterogeneous Spillover Effects in A/B Tests
• Transfer Learning for Estimating Causal Effects using Neural Networks
• Pareto Invariant Representation Learning for Multimedia Recommendation
• Estimating heterogeneous treatment effects with right-censored data via causal survival forests
• Treatment effect estimation with disentangled latent factors
• Estimating Heterogeneous Treatment Effects with Item-Level Outcome Data: Insights from Item Response Theory
• Estimating causal effects with optimization-based methods: A review and empirical comparison
• On Inductive Biases for Heterogeneous Treatment Effect Estimation
• Integrating Nearest Neighbors with Neural Network Models for Treatment Effect Estimation
• Counterfactual Prediction Under Outcome Measurement Error
• A Tree-based Model Averaging Approach for Personalized Treatment Effect Estimation from Heterogeneous Data Sources
• The Importance of Causality in Decision Making: A Perspective on Recommender Systems
• Adversarial Validation Approach to Concept Drift Problem in User Targeting Automation Systems at Uber
• Feature Selection Methods for Uplift Modeling and Heterogeneous Treatment Effect
• Causal Effect Inference for Structured Treatments
• Representation Balancing MDPs for Off-Policy Policy Evaluation
• Learning When-to-Treat Policies
• Heterogeneous Treatment Effect Estimation using machine learning for Healthcare application: tutorial and benchmark
• Treatment Allocation under Uncertain Costs
• Robust Estimation of Heterogeneous Treatment Effects using Electronic Health Record Data
• SurvITE: Learning Heterogeneous Treatment Effects from Time-to-Event Data
• Automated Video-EEG Analysis in Epilepsy Studies: Advances and Challenges
• What Makes Forest-Based Heterogeneous Treatment Effect Estimators Work?
• Inferring Individual Direct Causal Effects Under Heterogeneous Peer Influence
• Cross-Fitting and Averaging for Machine Learning Estimation of Heterogeneous Treatment Effects
• Minimax Optimal Nonparametric Estimation of Heterogeneous Treatment Effects
• Is it who you are or where you are? Accounting for compositional differences in cross-site treatment variation
• RealCause: Realistic Causal Inference Benchmarking
• Efficient Targeted Learning of Heterogeneous Treatment Effects for Multiple Subgroups
• A Look into Causal Effects under Entangled Treatment in Graphs: Investigating the Impact of Contact on MRSA Infection
• CausalMetaR: An R package for performing causally interpretable meta-analyses
• The WHY in Business Processes: Discovery of Causal Execution Dependencies
• Learning Treatment Effects in Panels with General Intervention Patterns
• Conditional Distributional Treatment Effect with Kernel Conditional Mean Embeddings and U-Statistic Regression
• Estimating Structural Target Functions using Machine Learning and Influence Functions
• A general framework for causal classification
• High-dimensional semi-supervised learning: in search for optimal inference of the mean
• Poincare: Recommending Publication Venues via Treatment Effect Estimation
• Nonparametric Heterogeneous Treatment Effect Estimation in Repeated Cross Sectional Designs
• Uplift Modeling for Multiple Treatments with Cost Optimization
• A Groupwise Approach for Inferring Heterogeneous Treatment Effects in Causal Inference
• Variable importance measures for heterogeneous treatment effects
• Flexible machine learning estimation of conditional average treatment effects: a blessing and a curse
• The wealth of nations and the health of populations: A quasi-experimental design of the impact of sovereign debt crises on child mortality
• Heterogeneous Effects of Software Patches in a Multiplayer Online Battle Arena Game
• Combining distributive ethics and causal Inference to make trade-offs between austerity and population health
• Recovery and inference of causal effects with sequential adjustment for confounding and attrition
• Federated Estimation of Causal Effects from Observational Data
• Affordable Uplift: Supervised Randomization in Controlled Experiments
• Stable discovery of interpretable subgroups via calibration in causal studies
• Group Average Treatment Effects for Observational Studies
• A unified survey of treatment effect heterogeneity modeling and uplift modeling
• Using Individualized Treatment Effects to Assess Treatment Effect Heterogeneity
• Machine learning in the social and health sciences
• Interpretable Assessment of Fairness During Model Evaluation
• Estimating Heterogeneous Causal Effects of High-Dimensional Treatments: Application to Conjoint Analysis
• Partial counterfactual identification and uplift modeling: theoretical results and real-world assessment
• Bridging or Breaking: Impact of Intergroup Interactions on Religious Polarization
• A Multi-resolution Theory for Approximating Infinite-$p$-Zero-$n$: Transitional Inference, Individualized Predictions, and a World Without Bias-Variance Trade-off
• A Semiparametric Approach to Model Effect Modification
• Assessment of Heterogeneous Treatment Effect Estimation Accuracy via Matching
• A Minimax Surrogate Loss Approach to Conditional Difference Estimation
• Heterogeneous Causal Learning for Effectiveness Optimization in User Marketing
• Learning Continuous Treatment Policy and Bipartite Embeddings for Matching with Heterogeneous Causal Effects
• Heterogeneous Treatment Effect Estimation through Deep Learning
• Improving pairwise comparison models using Empirical Bayes shrinkage
• Individualised Treatment Effects Estimation with Composite Treatments and Composite Outcomes
• Doubly Robust Direct Learning for Estimating Conditional Average Treatment Effect
• Split-Treatment Analysis to Rank Heterogeneous Causal Effects for Prospective Interventions
• Identify treatment effect patterns for personalised decisions
• Counterfactual Learning to Rank using Heterogeneous Treatment Effect Estimation
• UTBoost: Gradient Boosted Decision Trees for Uplift Modeling
• Combining Incomplete Observational and Randomized Data for Heterogeneous Treatment Effects
• Positivity-free Policy Learning with Observational Data
• A Deep Subgrouping Framework for Precision Drug Repurposing via Emulating Clinical Trials on Real-world Patient Data
• Estimating Causal Effects using a Multi-task Deep Ensemble
• A Twin Neural Model for Uplift
• Individualized prescriptive inference in ischaemic stroke
• Enhanced inference for distributions and quantiles of individual treatment effects in various experiments
• To do or not to do: cost-sensitive causal decision-making
• A causal fused lasso for interpretable heterogeneous treatment effects estimation
• A Loss-Function for Causal Machine-Learning
• Towards unique and unbiased causal effect estimation from data with hidden variables
• Treatment Targeting by AUUC Maximization with Generalization Guarantees
• Inherent Trade-offs in the Fair Allocation of Treatments
• Improved inference for doubly robust estimators of heterogeneous treatment effects
• A Causal Machine Learning Framework for Predicting Preventable Hospital Readmissions
• Finding and assessing treatment effect sweet spots in clinical trial data
• Efficient Balanced Treatment Assignments for Experimentation
• Estimating Potential Outcome Distributions with Collaborating Causal Networks
• Machine learning reveals how personalized climate communication can both succeed and backfire
• Wasserstein Random Forests and Applications in Heterogeneous Treatment Effects
• ADCB: An Alzheimer's disease benchmark for evaluating observational estimators of causal effects
• Guardrailed Uplift Targeting: A Causal Optimization Playbook for Marketing Strategy
• A Statistical Test for the Benefits of Personalizing Interventions
• Approaching an unknown communication system by latent space exploration and causal inference
• Estimation and Validation of Ratio-based Conditional Average Treatment Effects Using Observational Data
• Undersmoothing Causal Estimators with Generative Trees
• Neuroevolutionary Feature Representations for Causal Inference
• CausalNLP: A Practical Toolkit for Causal Inference with Text
• Finding Valid Adjustments under Non-ignorability with Minimal DAG Knowledge
• Identifiable Energy-based Representations: An Application to Estimating Heterogeneous Causal Effects
• Treatment effect estimation with Multilevel Regression and Poststratification
• Adaptive Multi-Source Causal Inference
• Selective Regression Under Fairness Criteria