Causal-BALD: Deep Bayesian Active Learning of Outcomes to Infer Treatment-Effects from Observational Data

Abstract

Estimating personalized treatment effects from high-dimensional observational data is essential in situations where experimental designs are infeasible, unethical, or expensive. Existing approaches rely on fitting deep models on outcomes observed for treated and control populations. However, when measuring individual outcomes is costly, as is the case of a tumor biopsy, a sample-efficient strategy for acquiring each result is required. Deep Bayesian active learning provides a framework for efficient data acquisition by selecting points with high uncertainty. However, existing methods bias training data acquisition towards regions of non-overlapping support between the treated and control populations. These are not sample-efficient because the treatment effect is not identifiable in such regions. We introduce causal, Bayesian acquisition functions grounded in information theory that bias data acquisition towards regions with overlapping support to maximize sample efficiency for learning personalized treatment effects. We demonstrate the performance of the proposed acquisition strategies on synthetic and semi-synthetic datasets IHDP and CMNIST and their extensions, which aim to simulate common dataset biases and pathologies.

arXiv comment24 pages, 8 Figures, 5 tables, NeurIPS 2021

References in corpus (18)

• Scikit-learn: Machine Learning in Python
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Spectral Normalization for Generative Adversarial Networks
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Deep Learning using Rectified Linear Units (ReLU)
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Tune: A Research Platform for Distributed Model Selection and Training
• Bayesian Active Learning for Classification and Preference Learning
• Scalable Variational Gaussian Process Classification
• BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learning
• Identifying Causal-Effect Inference Failure with Uncertainty-Aware Models
• Interval Estimation of Individual-Level Causal Effects Under Unobserved Confounding
• On Statistical Bias In Active Learning: How and When To Fix It
• Bayesian Nonparametric Causal Inference: Information Rates and Learning Algorithms
• Active Learning for Decision-Making from Imbalanced Observational Data
• Deconfounding Scores: Feature Representations for Causal Effect Estimation with Weak Overlap
• PowerEvaluationBALD: Efficient Evaluation-Oriented Deep (Bayesian) Active Learning with Stochastic Acquisition Functions