Causal Effect Inference with Deep Latent-Variable Models

Abstract

Learning individual-level causal effects from observational data, such as inferring the most effective medication for a specific patient, is a problem of growing importance for policy makers. The most important aspect of inferring causal effects from observational data is the handling of confounders, factors that affect both an intervention and its outcome. A carefully designed observational study attempts to measure all important confounders. However, even if one does not have direct access to all confounders, there may exist noisy and uncertain measurement of proxies for confounders. We build on recent advances in latent variable modeling to simultaneously estimate the unknown latent space summarizing the confounders and the causal effect. Our method is based on Variational Autoencoders (VAE) which follow the causal structure of inference with proxies. We show our method is significantly more robust than existing methods, and matches the state-of-the-art on previous benchmarks focused on individual treatment effects.

Published as a conference paper at NIPS 2017

References in corpus (13)

• Auto-Encoding Variational Bayes
• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Auxiliary Deep Generative Models
• Edward: A library for probabilistic modeling, inference, and criticism
• Learning Representations for Counterfactual Inference
• Improving Variational Inference with Inverse Autoregressive Flow
• Learning mixtures of separated nonspherical Gaussians
• Operator Variational Inference
• Learning Mixtures of DAG Models
• Identifying Causal Effects With Proxy Variables of an Unmeasured Confounder
• Provable learning of Noisy-or Networks

Cited by in corpus (110)

• Causal Confusion in Imitation Learning
• Towards Realistic Individual Recourse and Actionable Explanations in Black-Box Decision Making Systems
• Deep Structural Causal Models for Tractable Counterfactual Inference
• Algorithmic recourse under imperfect causal knowledge: a probabilistic approach
• Amortized Causal Discovery: Learning to Infer Causal Graphs from Time-Series Data
• Learning Counterfactual Representations for Estimating Individual Dose-Response Curves
• On Feature Collapse and Deep Kernel Learning for Single Forward Pass Uncertainty
• Perfect Match: A Simple Method for Learning Representations For Counterfactual Inference With Neural Networks
• Adapting Neural Networks for the Estimation of Treatment Effects
• Learning Causal Semantic Representation for Out-of-Distribution Prediction
• Identifying Causal-Effect Inference Failure with Uncertainty-Aware Models
• Generalization Bounds and Representation Learning for Estimation of Potential Outcomes and Causal Effects
• On Multi-Cause Causal Inference with Unobserved Confounding: Counterexamples, Impossibility, and Alternatives
• A Survey on Causal Inference
• Multiple Causal Inference with Latent Confounding
• Using Embeddings to Correct for Unobserved Confounding in Networks
• Describing condensed matter from atomically resolved imaging data: from structure to generative and causal models
• VCNet and Functional Targeted Regularization For Learning Causal Effects of Continuous Treatments
• Causal Inference for Time series Analysis: Problems, Methods and Evaluation
• Sage: Leveraging ML to Diagnose Unpredictable Performance in Cloud Microservices
• Time Series Deconfounder: Estimating Treatment Effects over Time in the Presence of Hidden Confounders
• Disentangled Representation for Causal Mediation Analysis
• Confounding-Robust Policy Evaluation in Infinite-Horizon Reinforcement Learning
• The Causal-Neural Connection: Expressiveness, Learnability, and Inference
• Conditional out-of-sample generation for unpaired data using trVAE
• MetaCI: Meta-Learning for Causal Inference in a Heterogeneous Population
• Treatment effect estimation with disentangled latent factors
• On Inductive Biases for Heterogeneous Treatment Effect Estimation
• Estimation of Individual Treatment Effect in Latent Confounder Models via Adversarial Learning
• Invariant Representation Learning for Treatment Effect Estimation
• Counterfactual Reasoning for Fair Clinical Risk Prediction
• A Framework for Causal Discovery in non-intervenable systems
• Size-Invariant Graph Representations for Graph Classification Extrapolations
• Causal Influence Detection for Improving Efficiency in Reinforcement Learning
• The Medical Deconfounder: Assessing Treatment Effects with Electronic Health Records
• Conservative Policy Construction Using Variational Autoencoders for Logged Data with Missing Values
• Implementation of Tripartite Estimands Using Adherence Causal Estimators Under the Causal Inference Framework
• Robust Recursive Partitioning for Heterogeneous Treatment Effects with Uncertainty Quantification
• Causal Effect Inference for Structured Treatments
• Off-policy Evaluation in Infinite-Horizon Reinforcement Learning with Latent Confounders
• SurvITE: Learning Heterogeneous Treatment Effects from Time-to-Event Data
• Learning Individual Causal Effects from Networked Observational Data
• RealCause: Realistic Causal Inference Benchmarking
• MissDeepCausal: Causal Inference from Incomplete Data Using Deep Latent Variable Models
• Intact-VAE: Estimating Treatment Effects under Unobserved Confounding
• A Critical Look at the Consistency of Causal Estimation With Deep Latent Variable Models
• NCoRE: Neural Counterfactual Representation Learning for Combinations of Treatments
• CausalCity: Complex Simulations with Agency for Causal Discovery and Reasoning
• Conditional Distributional Treatment Effect with Kernel Conditional Mean Embeddings and U-Statistic Regression
• Copula-based Sensitivity Analysis for Multi-Treatment Causal Inference with Unobserved Confounding
• Poincare: Recommending Publication Venues via Treatment Effect Estimation
• Learning Matching Representations for Individualized Organ Transplantation Allocation
• A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution
• Environment Reconstruction with Hidden Confounders for Reinforcement Learning based Recommendation
• Counterfactual Evaluation of Treatment Assignment Functions with Networked Observational Data
• Reducing Selection Bias in Counterfactual Reasoning for Individual Treatment Effects Estimation
• Everything Has a Cause: Leveraging Causal Inference in Legal Text Analysis
• Sage: Using Unsupervised Learning for Scalable Performance Debugging in Microservices
• Federated Estimation of Causal Effects from Observational Data
• Causal Effect Variational Autoencoder with Uniform Treatment
• Deep Proxy Causal Learning and its Application to Confounded Bandit Policy Evaluation
• Estimating Individual Treatment Effects with Time-Varying Confounders
• A unified survey of treatment effect heterogeneity modeling and uplift modeling
• MultiMBNN: Matched and Balanced Causal Inference with Neural Networks
• Learning Latent Causal Structures with a Redundant Input Neural Network
• Counterfactual Cross-Validation: Stable Model Selection Procedure for Causal Inference Models
• CausCF: Causal Collaborative Filtering for RecommendationEffect Estimation
• Improving Fair Predictions Using Variational Inference In Causal Models
• Estimating Granger Causality with Unobserved Confounders via Deep Latent-Variable Recurrent Neural Network
• Causal World Models by Unsupervised Deconfounding of Physical Dynamics
• Balance Regularized Neural Network Models for Causal Effect Estimation
• Learning Continuous Treatment Policy and Bipartite Embeddings for Matching with Heterogeneous Causal Effects
• Representation Learning for Integrating Multi-domain Outcomes to Optimize Individualized Treatments
• Linked Causal Variational Autoencoder for Inferring Paired Spillover Effects
• Sequential Deconfounding for Causal Inference with Unobserved Confounders
• Counterfactual Situation Testing: From Single to Multidimensional Discrimination
• Doing Great at Estimating CATE? On the Neglected Assumptions in Benchmark Comparisons of Treatment Effect Estimators
• Causal Mediation Analysis with Hidden Confounders
• Learning Conditional Invariance through Cycle Consistency
• Counterfactual Prediction Under Selective Confounding
• Dependent Multi-Task Learning with Causal Intervention for Image Captioning
• Markdowns in E-Commerce Fresh Retail: A Counterfactual Prediction and Multi-Period Optimization Approach
• Bayesian causal inference via probabilistic program synthesis
• Dynamic Relational Inference in Multi-Agent Trajectories
• Towards unique and unbiased causal effect estimation from data with hidden variables
• Causal Autoregressive Flows
• High-Dimensional Feature Selection for Sample Efficient Treatment Effect Estimation
• SAG-VAE: End-to-end Joint Inference of Data Representations and Feature Relations
• Causal inference and machine learning approaches for evaluation of the health impacts of large-scale air quality regulations
• Treatment Targeting by AUUC Maximization with Generalization Guarantees
• Neural Networks for Learning Counterfactual G-Invariances from Single Environments
• Contrastive Mixture of Posteriors for Counterfactual Inference, Data Integration and Fairness
• Estimating Potential Outcome Distributions with Collaborating Causal Networks
• Adaptive Multi-Source Causal Inference
• Estimating Causal Effects With Partial Covariates For Clinical Interpretability
• A Kernel Test for Causal Association via Noise Contrastive Backdoor Adjustment
• Local search for efficient causal effect estimation
• Reliable Estimation of Individual Treatment Effect with Causal Information Bottleneck
• $\beta$-Intact-VAE: Identifying and Estimating Causal Effects under Limited Overlap
• Weighting-Based Treatment Effect Estimation via Distribution Learning
• Identification of Latent Variables From Graphical Model Residuals
• ADCB: An Alzheimer's disease benchmark for evaluating observational estimators of causal effects
• DoGR: Disaggregated Gaussian Regression for Reproducible Analysis of Heterogeneous Data
• VACA: Design of Variational Graph Autoencoders for Interventional and Counterfactual Queries
• Variational Auto-Encoder Architectures that Excel at Causal Inference
• Causal Modeling with Stochastic Confounders
• Cause-Effect Deep Information Bottleneck For Systematically Missing Covariates
• Causal Effect Estimation using Variational Information Bottleneck
• A Meta Learning Approach to Discerning Causal Graph Structure
• Estimating Individual Treatment Effects through Causal Populations Identification