mlsanityData to 14 Oct 2026

Combining Observational and Randomized Data for Estimating Heterogeneous Treatment Effects

arXiv 2202.12891 ↗First posted 25 Feb 2022stat.ML cs.LGcited by 32

Not on the map. Why?

Abstract (arXiv)

Estimating heterogeneous treatment effects is an important problem across many domains. In order to accurately estimate such treatment effects, one typically relies on data from observational studies or randomized experiments. Currently, most existing works rely exclusively on observational data, which is often confounded and, hence, yields biased estimates. While observational data is confounded, randomized data is unconfounded, but its sample size is usually too small to learn heterogeneous treatment effects. In this paper, we propose to estimate heterogeneous treatment effects by combining large amounts of observational data and small amounts of randomized data via representation learning. In particular, we introduce a two-step framework: first, we use observational data to learn a shared structure (in form of a representation); and then, we use randomized data to learn the data-specific structures. We analyze the finite sample properties of our framework and compare them to several natural baselines. As such, we derive conditions for when combining observational and randomized data is beneficial, and for when it is not. Based on this, we introduce a sample-efficient algorithm, called CorNet. We use extensive simulation studies to verify the theoretical properties of CorNet and multiple real-world datasets to demonstrate our method's superiority compared to existing methods.

Key facts. Cited by 32 papers in the map (data through 14 Oct 2026). Builds on 38 of its 94 references that are in the map. First posted on arXiv 25 Feb 2022.

Builds on

38 of 94 references.

References · 38

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
  3. 2015Domain-Adversarial Training of Neural Networkscited by 3,044
  4. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,917
  5. 2019Federated Learning: Challenges, Methods, and Future Directionscited by 1,538
  6. 2017A Survey on Multi-Task Learningcited by 840
  7. 2017An Overview of Multi-Task Learning in Deep Neural Networkscited by 803
  8. 2015Estimation and Inference of Heterogeneous Treatment Effects using Random Forestscited by 600
  9. 2017Causal Effect Inference with Deep Latent-Variable Modelscited by 410
  10. 2019Multi-Task Deep Neural Networks for Natural Language Understandingcited by 350
  11. 2016Learning Representations for Counterfactual Inferencecited by 329
  12. 2017Size-Independent Sample Complexity of Neural Networkscited by 290
26 more
  1. 2020Few-Shot Learning via Learning the Representation, Provablycited by 214
  2. 2015The Benefit of Multitask Representation Learningcited by 195
  3. 2020On the Theory of Transfer Learning: The Importance of Task Diversitycited by 186
  4. 2017Bayesian Inference of Individualized Treatment Effects using Multi-task Gaussian Processescited by 158
  5. 2021Nonparametric Estimation of Heterogeneous Treatment Effects: From Theory to Learning Algorithmscited by 144
  6. 2018The Blessings of Multiple Causescited by 104
  7. 2018Confounding-Robust Policy Improvementcited by 76
  8. 2019Time Series Deconfounder: Estimating Treatment Effects over Time in the Presence of Hidden Confounderscited by 75
  9. 2021On Inductive Biases for Heterogeneous Treatment Effect Estimationcited by 75
  10. 2018Removing Hidden Confounding by Experimental Groundingcited by 64
  11. 2018Interval Estimation of Individual-Level Causal Effects Under Unobserved Confoundingcited by 51
  12. 2018Learning Weighted Representations for Generalization Across Designscited by 41
  13. 2017Sensitivity analysis for inverse probability weighting estimators via the percentile bootstrapcited by 33
  14. 2019Semi-Supervised Learning by Augmented Distribution Alignmentcited by 30
  15. 2018Combining multiple observational data sources to estimate causal effectscited by 29
  16. 2020Combining Observational and Experimental Datasets Using Shrinkage Estimatorscited by 28
  17. 2020Using Experiments to Correct for Selection in Observational Studiescited by 28
  18. 2022Long-term Causal Inference Under Persistent Confounding via Data Combinationcited by 27
  19. 2021Estimating Average Treatment Effects via Orthogonal Regularizationcited by 25
  20. 2018Predicting with Proxies: Transfer Learning in High Dimensioncited by 24
  21. 2021Sequential Deconfounding for Causal Inference with Unobserved Confounderscited by 24
  22. 2021Adaptive Combination of Randomized and Observational Datacited by 22
  23. 2021Generalizing Off-Policy Learning under Sample Selection Biascited by 16
  24. 2021Deconfounding Temporal Autoencoder: Estimating Treatment Effects over Time Using Noisy Proxiescited by 16
  25. 2021AttDMM: An Attentive Deep Markov Model for Risk Scoring in Intensive Care Unitscited by 14
  26. 2021Generalizing Clinical Trials with Convex Hullscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 6; 2023: 5; 2024: 8; 2025: 9; 2026: 4.

32 papers.

Sort:

  1. 2024Causal machine learning for predicting treatment outcomescited by 74
  2. 2022Normalizing Flows for Interventional Density Estimationcited by 21
  3. 2022Transfer Learning on Heterogeneous Feature Spaces for Treatment Effects Estimationcited by 21
  4. 2024Prediction-powered Generalization of Causal Inferencescited by 15
  5. 2022Detecting hidden confounding in observational data using multiple environmentscited by 13
  6. 2024Meta-Learners for Partially-Identified Treatment Effects Across Multiple Environmentscited by 11
  7. 2022Estimating Conditional Average Treatment Effects with Missing Treatment Informationcited by 7
  8. 2024Constructing Confidence Intervals for Average Treatment Effects from Multiple Datasetscited by 6
  9. 2023Efficient and robust transfer learning of optimal individualized treatment regimes with right-censored survival datacited by 5
  10. 2024Minimax Regret Estimation for Generalizing Heterogeneous Treatment Effects with Multisite Datacited by 5
  11. 2022Heterogeneous Treatment Effect with Trained Kernels of the Nadaraya-Watson Regressioncited by 4
  12. 2025Improving the Generation and Evaluation of Synthetic Data for Downstream Medical Causal Inferencecited by 4
20 more
  1. 2023Transfer Learning for Causal Effect Estimationcited by 3
  2. 2023Causal Effect Estimation from Observational and Interventional Data Through Matrix Weighted Linear Estimatorscited by 3
  3. 2024Estimating Heterogeneous Treatment Effects by Combining Weak Instruments and Observational Datacited by 2
  4. 2025Long-term Causal Inference via Modeling Sequential Latent Confoundingcited by 2
  5. 2022BENK: The Beran Estimator with Neural Kernels for Estimating the Heterogeneous Treatment Effectcited by 1
  6. 2023Assessing Electricity Service Unfairness with Transfer Counterfactual Learningcited by 1
  7. 2023Estimating treatment effects from single-arm trials via latent-variable modelingcited by 1
  8. 2024Combining Incomplete Observational and Randomized Data for Heterogeneous Treatment Effectscited by 1
  9. 2025Robust estimation of heterogeneous treatment effects in randomized trials leveraging external datacited by 1
  10. 2026Observationally Informed Adaptive Causal Experimental Designcited by 1
  11. 2024Accounting for Missing Covariates in Heterogeneous Treatment Estimationcited by 0
  12. 2025A Partial Initialization Strategy to Mitigate the Overfitting Problem in CATE Estimation with Hidden Confoundingcited by 0
  13. 2025Deconfounded Warm-Start Thompson Sampling with Applications to Precision Medicinecited by 0
  14. 2025Uncovering Bias Mechanisms in Observational Studiescited by 0
  15. 2025Conditional Average Treatment Effect Estimation Under Hidden Confounderscited by 0
  16. 2025Mitigating Hidden Confounding by Progressive Confounder Imputation via Large Language Modelscited by 0
  17. 2025Augmenting Limited and Biased RCTs through Pseudo-Sample Matching-Based Observational Data Fusion Methodcited by 0
  18. 2026Testing Effect Homogeneity and Confounding in High-Dimensional Experimental and Observational Studiescited by 0
  19. 2026Orthogonal Learner for Estimating Heterogeneous Long-Term Treatment Effectscited by 0
  20. 2026Prediction-Powered Data Fusion for Treatment Effect Estimationcited by 0