Graph Intervention Networks for Causal Effect Estimation

We address the estimation of conditional average treatment effects (CATEs) when treatments are graph-structured (e.g., molecular graphs of drugs). Given a weak condition on the effect, we propose a plug-in estimator that decomposes CATE estimation into separate, simpler optimization problems. Our estimator (a) isolates the causal estimands (reducing regularization bias), and (b) allows one to plug in arbitrary models for learning. In experiments with small-world and molecular graphs, we show that our approach outperforms prior approaches and is robust to varying selection biases. Our implementation is online 2.

[1]  Nathan Kallus,et al.  DeepMatch: Balancing Deep Covariate Representations for Causal Inference Using Adversarial Training , 2018, ICML.

[2]  P. Robinson.  ROOT-N-CONSISTENT SEMIPARAMETRIC REGRESSION , 1988 .

[3]  Z. Geng,et al.  Identifying Causal Effects With Proxy Variables of an Unmeasured Confounder. , 2016, Biometrika.

[4]  W. Bastiaan Kleijn,et al.  The HSIC Bottleneck: Deep Learning without Back-Propagation , 2019, AAAI.

[5]  Michael Pollmann.  Causal Inference for Spatial Treatments , 2020 .

[6]  Le Song,et al.  A Kernel Statistical Test of Independence , 2007, NIPS.

[7]  Matt J. Kusner,et al.  Proximal Causal Learning with Kernels: Two-Stage Estimation and Moment Restriction , 2021, ICML.

[8]  Joan Bruna,et al.  Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges , 2021, ArXiv.

[9]  Stefan Feuerriegel,et al.  Estimating Average Treatment Effects via Orthogonal Regularization , 2021, CIKM.

[10]  Uri Shalit,et al.  Estimating individual treatment effect: generalization bounds and algorithms , 2016, ICML.

[11]  Jared S. Murray,et al.  Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects (with Discussion) , 2020, 2108.02836.

[12]  Susan Athey,et al.  Recursive partitioning for heterogeneous causal effects , 2015, Proceedings of the National Academy of Sciences.

[13]  Jennifer L. Hill,et al.  Bayesian Nonparametric Modeling for Causal Inference , 2011 .

[14]  Chin-Wei Huang,et al.  RealCause: Realistic Causal Inference Benchmarking , 2020, ArXiv.

[15]  Ricardo Silva,et al.  Observational-Interventional Priors for Dose-Response Learning , 2016, NIPS.

[16]  Xinkun Nie,et al.  Quasi-oracle estimation of heterogeneous treatment effects , 2017, Biometrika.

[17]  Stefan Bauer,et al.  Learning Counterfactual Representations for Estimating Individual Dose-Response Curves , 2019, AAAI.

[18]  Fan Li.  Propensity Score Weighting for Causal Inference with Multiple Treatments , 2018 .

[19]  Nando de Freitas,et al.  Learning Deep Features in Instrumental Variable Regression , 2020, ICLR.

[20]  David Blei,et al.  Invariant Representation Learning for Treatment Effect Estimation , 2020, UAI.

[21]  Jean-Louis Reymond,et al.  Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17 , 2012, J. Chem. Inf. Model..

[22]  Mihaela van der Schaar,et al.  Nonparametric Estimation of Heterogeneous Treatment Effects: From Theory to Learning Algorithms , 2021, AISTATS.

[23]  S. Bhatt,et al.  Estimating the effects of non-pharmaceutical interventions on COVID-19 in Europe , 2020, Nature.

[24]  J. Pearl.  Causality: Models, Reasoning and Inference , 2000 .

[25]  Jan Eric Lenssen,et al.  Fast Graph Representation Learning with PyTorch Geometric , 2019, ArXiv.

[26]  Edward H. Kennedy.  Towards optimal doubly robust estimation of heterogeneous causal effects , 2020, 2004.14497.

[27]  Ioanna Manolopoulou,et al.  Estimating individual treatment effects using non‐parametric regression models: A review , 2020, Journal of the Royal Statistical Society: Series A (Statistics in Society).

[28]  M. Schaar,et al.  Estimating the Effects of Continuous-valued Interventions using Generative Adversarial Networks , 2020, NeurIPS.

[29]  Martin Grohe,et al.  Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks , 2018, AAAI.

[30]  Mihaela van der Schaar,et al.  Bayesian Inference of Individualized Treatment Effects using Multi-task Gaussian Processes , 2017, NIPS.

[31]  Uri Shalit,et al.  Learning Representations for Counterfactual Inference , 2016, ICML.

[32]  Qiang Liu,et al.  VCNet and Functional Targeted Regularization For Learning Causal Effects of Continuous Treatments , 2021, ICLR.

[33]  Uri Shalit,et al.  Quantifying Ignorance in Individual-Level Causal-Effect Estimates under Hidden Confounding , 2021, ICML.

[34]  Zhengyuan Zhou,et al.  Offline Multi-Action Policy Learning: Generalization and Optimization , 2018, Oper. Res..

[35]  Pavlo O. Dral,et al.  Quantum chemistry structures and properties of 134 kilo molecules , 2014, Scientific Data.

[36]  Tyler J. VanderWeele,et al.  Sensitivity Analysis in Observational Research: Introducing the E-Value , 2017, Annals of Internal Medicine.

[37]  Edward H Kennedy,et al.  Non‐parametric methods for doubly robust estimation of continuous treatment effects , 2015, Journal of the Royal Statistical Society. Series B, Statistical methodology.

[38]  Sören R. Künzel,et al.  Metalearners for estimating heterogeneous treatment effects using machine learning , 2017, Proceedings of the National Academy of Sciences.

[39]  E. J. Tchetgen Tchetgen,et al.  An Introduction to Proximal Causal Learning , 2020, medRxiv.

[40]  Gertjan J. Burghouts,et al.  Set Prediction without Imposing Structure as Conditional Density Estimation , 2020, ICLR.

[41]  Alexander D'Amour,et al.  Flexible Sensitivity Analysis for Observational Studies Without Observable Implications , 2018, Journal of the American Statistical Association.

[42]  Hisashi Kashima,et al.  GraphITE: Estimating Individual Effects of Graph-structured Treatments , 2020, CIKM.

[43]  David M. Blei,et al.  Adapting Neural Networks for the Estimation of Treatment Effects , 2019, NeurIPS.

[44]  Max Welling,et al.  Causal Effect Inference with Deep Latent-Variable Models , 2017, NIPS 2017.

[45]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[46]  Lawrence Carin,et al.  Reconsidering Generative Objectives For Counterfactual Reasoning , 2020, NeurIPS.

[47]  Y. Gal,et al.  Identifying Causal Effect Inference Failure with Uncertainty-Aware Models , 2020, NeurIPS.

[48]  Ilya Shpitser,et al.  Semi-Parametric Causal Sufficient Dimension Reduction Of High Dimensional Treatments , 2017 .

[49]  Stefan Wager,et al.  Estimation and Inference of Heterogeneous Treatment Effects using Random Forests , 2015, Journal of the American Statistical Association.

[50]  Max Welling,et al.  Modeling Relational Data with Graph Convolutional Networks , 2017, ESWC.

[51]  S. Athey,et al.  Generalized random forests , 2016, The Annals of Statistics.

[52]  Kevin Leyton-Brown,et al.  Deep IV: A Flexible Approach for Counterfactual Prediction , 2017, ICML.

[53]  Michael J. Lopez,et al.  Estimation of causal effects with multiple treatments: a review and new ideas , 2017, 1701.05132.

[54]  Walter Karlen,et al.  Perfect Match: A Simple Method for Learning Representations For Counterfactual Inference With Neural Networks , 2018, ArXiv.

[55]  S. Athey,et al.  Estimating Treatment Effects with Causal Forests: An Application , 2019, Observational Studies.

[56]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[57]  Joshua M. Stuart,et al.  The Cancer Genome Atlas Pan-Cancer analysis project , 2013, Nature Genetics.

[58]  Mihaela van der Schaar,et al.  Limits of Estimating Heterogeneous Treatment Effects: Guidelines for Practical Algorithm Design , 2018, ICML.

[59]  Aidong Zhang,et al.  Representation Learning for Treatment Effect Estimation from Observational Data , 2018, NeurIPS.

[60]  J. Robins,et al.  Double/Debiased Machine Learning for Treatment and Structural Parameters , 2017 .

[61]  D. Rubin,et al.  Causal Inference for Statistics, Social, and Biomedical Sciences: Sensitivity Analysis and Bounds , 2015 .

[62]  Arthur Gretton,et al.  Kernel Instrumental Variable Regression , 2019, NeurIPS.