This primer systematizes the emerging literature on causal inference using deep neural networks under the potential outcomes framework. It provides an intuitive introduction to building and optimizing custom deep learning models and shows how to adapt them to estimate/predict heterogeneous treatment effects. It also discusses ongoing work to extend causal inference to settings where confounding is nonlinear, time-varying, or encoded in text, networks, and images. To maximize accessibility, we also introduce prerequisite concepts from causal inference and deep learning. The primer differs from other treatments of deep learning and causal inference in its sharp focus on observational causal estimation, its extended exposition of key algorithms, and its detailed tutorials for implementing, training, and selecting among deep estimators in TensorFlow 2 and PyTorch.
Get full access to this article
View all access options for this article.
References
1.
AlaaAhmedVan Der SchaarMihaela. 2019. “Validating Causal Inference Models via Influence Functions.” In International Conference on Machine Learning, volume 36, pp. 191–201. Association for Computing Machinery.
AtanOnurJordonJamesVan Der SchaarMihaela. 2018. “Deep-Treat: Learning Optimal Personalized Treatments from Observational Data Using Neural Networks.” In Association for the Advancement of Artificial Intelligence Conference on Artificial Intelligence, volume 32, p. 2071–2078. Association for the Advancement of Artificial Intelligence.
4.
AtheySusanImbensGuido. 2016. “Recursive Partitioning for Heterogeneous Causal Effects.” Proceedings of the National Academy of Sciences113: 7353–60.
5.
AustinPeter C. 2011. “An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies.” Multivariate Behavioral Research46: 399–424.
6.
BengioYoshua. 2013. “Deep Learning of Representations: Looking Forward.” In International Conference on Statistical Language and Speech Processing, pp. 1–37. Association for Computing Machinery.
7.
BenkeserD.CaroneM.Van Der LaanMJ.GilbertP.B.. 2017. “Doubly Robust Nonparametric Inference on the Average Treatment Effect.” Biometrika104: 863–80.
8.
BenkeserDavidChambazAntoine. 2020. “A Ride in Targeted Learning Territory.” Journal de la société française de statistique161: 201–86.
9.
BicaIoanaAlaaAhmed M.JordonJamesvan der SchaarMihaela. 2020. “Estimating Counterfactual Treatment Outcomes Over Time through Adversarially Balanced Representations.” In International Conference on Learning Representations, volume 37. Association for Computing Machinery.
10.
BrandJennie EKochBernardXuJiahui. 2020. “Machine Learning.” In Sage Research Methods Foundations. SAGE.
11.
ChenHuigangHarinenTotteLeeJeong-YoonYungMikeZhaoZhenyu. 2020. “CausalML: Python Package for Causal Machine Learning.”.
12.
ChernozhukovVictorChetverikovDenisDemirerMertDufloEstherHansenChristianNeweyWhitneyRobinsJames. 2018. “Double/debiased Machine Learning for Treatment and Structural Parameters.” The Econometrics Journal21: C1–C68.
13.
ChernozhukovVictorNeweyWhitney KQuintas-MartinezVictorSyrgkanisVasilis. 2021. “Automatic debiased machine learning via neural nets for generalized linear regression.” arXiv preprint arXiv:2104.14737.
14.
ChernozhukovVictorNeweyWhitneyQuintas-MartinezVictor MSyrgkanisVasilis. 2022. “Riesznet and forestriesz: Automatic debiased machine learning with neural nets and random forests.” In International Conference on Machine Learning, pp. 3901–3914. PMLR.
15.
CrabbéJonathanCurthAliciaBicaIoanavan der SchaarMihaela. 2022. “Benchmarking heterogeneous treatment effect models through the lens of interpretability.” arXiv preprint arXiv:2206.08363.
16.
CristaliIrinaVeitchVictor. 2022. “Using Embeddings for Causal Estimation of Peer Influence in Social Networks.” ArXiv abs/2205.08033.
17.
CurthAliciaSvenssonDavidWeatherallJimvan der SchaarMihaela. 2021. “Really Doing Great at Estimating CATE? A Critical Look at ML Benchmarking Practices in Treatment Effect Estimation.” In Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks, edited by J. Vanschoren and S. Yeung, volume 1.
18.
CybenkoGeorge. 1989. “Approximation by Superpositions of a Sigmoidal Function.” Mathematics of Control, Signals and Systems5: 455.
19.
DevlinJacobChangMing-WeiLeeKentonToutanovaKristina. 2019. “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding.” In Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 4171–4186. Association for Computational Linguistics.
20.
FarrellMax HLiangTengyuanMisraSanjog. 2021. “Deep Neural Networks for Estimation and Inference.” Econometrica89: 181–213.
21.
FederAmirKeithKatherine AManzoorEmaadPryzantReidSridharDhanyaWood-DoughtyZachEisensteinJacobGrimmerJustinReichartRoiRobertsMargaret E, et al. 2021. “Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond.” arXiv preprint arXiv:2109.00725.
22.
FisherAaronKennedyEdward H. 2021. “Visually Communicating and Teaching Intuition for Influence Functions.” The American Statistician75: 162–72.
23.
FrauenDennisHattTobiasMelnychukValentynFeuerriegelStefan. 2022. “Estimating average causal effects from patient trajectories.” arXiv preprint arXiv:2203.01228.
24.
GilmerJustinSchoenholzSamuel SRileyPatrick FVinyalsOriolDahlGeorge E. 2017. “Neural message passing for quantum chemistry.” In International Conference on Machine Learning, volume 34, pp. 1263–1272. Association for Computing Machinery.
25.
GlynnAdam NQuinnKevin M. 2010. “An Introduction to the Augmented Inverse Propensity Weighted Estimator.” Political Analysis18: 36–56.
26.
GoldsteinAlexKapelnerAdamBleichJustinPitkinEmil. 2015. “Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation.” Journal of Computational and Graphical Statistics24: 44–65.
27.
GoldszmidtMoisésPearlJudea. 1996. “Qualitative Probabilities for Default Reasoning, Belief Revision, and Causal Modeling.” Artificial Intelligence84: 57–112.
GuiLinVeitchVictor. 2022. “Causal Estimation for Text Data with (Apparent) Overlap Violations.” arXiv preprint arXiv:2210.00079.
30.
GuoRuochengLiJundongLiuHuan. 2020. “Counterfactual Evaluation of Treatment Assignment Functions with Networked Observational Data.” In SIAM International Conference on Data Mining, pp. 271–279. Society for Industrial and Applied Mathematics.
31.
HastieT.TibshiraniR.FriedmanJ.H.. 2009. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer series in statistics. New York, NY: Springer.
32.
HeckKatherine EBravemanPaulaCubbinCatherineChávezGilberto FKielyJohn L. 2006. “Socioeconomic Status and Breastfeeding Initiation Among California Mothers.” Public health reports121: 51–9.
33.
HernánMiguel A.RobinsJames M.. 2020. Causal Inference: What If. 2020. Boca Raton: Chapman & Hall/CRC.
34.
HillJennifer L. 2011. “Bayesian Nonparametric Modeling for Causal Inference.” Journal of Computational and Graphical Statistics20: 217–40.
35.
HollandPaul W. 1986. “Statistics and Causal Inference.” Journal of the American statistical Association81: 945–60.
36.
ImbensGuido WRubinDonald B. 2015. Causal Inference in Statistics, Social, and Biomedical Sciences. New York, NY: Cambridge University Press.
37.
IoffeSergeySzegedyChristian. 2015. “Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift.” In International Conference on Machine Learning, volume 32, pp. 448–456. Association for Computing Machinery.
38.
JessonAndrewMindermannSörenGalYarinShalitUri. 2021. “Quantifying Ignorance in Individual-Level Causal-Effect Estimates under Hidden Confounding.” arXiv preprint arXiv:2103.04850.
39.
JohanssonFredrik DKallusNathanShalitUriSontagDavid. 2018. “Learning Weighted Representations for Generalization Across Designs.” Unpublished.
40.
JohanssonFredrik D.ShalitUriKallusNathanSontagDavid A.. 2020. “Generalization Bounds and Representation Learning for Estimation of Potential Outcomes and Causal Effects.” arXiv abs/2001.07426.
41.
JohanssonFredrik DShalitUriSontagDavid. 2016. “Learning Representations for Counterfactual Inference.” In International Conference on Machine Learning, volume 48. Association for Computing Machinery.
JooJungseockSteenFrancis FZhuSong-Chun. 2015. “Automated Facial Trait Judgment and Election Outcome Prediction: Social Dimensions of Face.” In International Conference on Computer Vision, pp. 3712–3720. IEEE.
44.
KeithKatherineJensenDavidO’ConnorBrendan. 2020. “Text and Causal Inference: A Review of Using Text to Remove Confounding from Causal Estimates.” In Annual Meeting of the Association for Computational Linguistics, volume 58, pp. 5332–5344, Online. Association for Computational Linguistics.
45.
KennedyEdward H. 2016. “Semiparametric Theory and Empirical Processes in Causal Inference.” In Statistical Causal Inferences and their Applications in Public Health Research, pp. 141–167. Springer.
46.
KennedyEdward H. 2020. “Towards optimal doubly robust estimation of heterogeneous causal effects.”.
47.
KingmaDiederik P.BaJimmy. 2015. “Adam: A Method for Stochastic Optimization.” In International Conference on Learning Representations, volume 3. OpenReview.
48.
KipfThomas NWellingMax. 2017. “Semi-supervised Classification with Graph Convolutional Networks.” In International Conference on Learning Representations, volume 5. OpenReview.
49.
KramerMichael SAboudFrancesMironovaElenaVanilovichIrinaPlattRobert WMatushLidiaIgumnovSergeiFombonneEricBogdanovichNataliaDucruetThierry, et al. 2008. “Breastfeeding and Child Cognitive Development: New Evidence From a Large Randomized Trial.” Archives of general psychiatry65: 578–84.
50.
KünzelSören RSekhonJasjeet SBickelPeter JYuBin. 2019. “Metalearners for Estimating Heterogeneous Treatment Effects Using Machine Learning.” Proceedings of the national academy of sciences116: 4156–65.
LiRuiHuStephanieLuMingyuUtsumiYuriaChakrabortyPrithwishSowDaby M.MadanPiyushLiJunGhalwashMohamedShahnZachLehmanLi-wei. 2021. “G-Net: a Recurrent Network Approach to G-Computation for Counterfactual Prediction Under a Dynamic Treatment Regime.” In Proceedings of Machine Learning for Health, edited by Subhrajit Roy, Stephen Pfohl, Emma Rocheteau, Girmaw Abebe Tadesse, Luis Oala, Fabian Falck, Yuyin Zhou, Liyue Shen, Ghada Zamzmi, Purity Mugambi, Ayah Zirikly, Matthew B. A. McDermott, and Emily Alsentzer, volume 158 of Proceedings of Machine Learning Research, pp. 282–299. PMLR.
53.
LimBryanAlaaAhmed Mvan der SchaarMihaela. 2018. “Forecasting Treatment Responses Over Time Using Recurrent Marginal Structural Networks.” In Neural Information Processing Systems, volume 18, pp. 7483–7493. Curran Associates Inc.
54.
LundbergScott MLeeSu-In. 2017. “A Unified Approach to Interpreting Model Predictions.” In Advances in Neural Information Processing Systems, edited by I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, volume 30. Curran Associates, Inc.
55.
McFowland IIIEdwardShaliziCosma Rohilla. 2023. “Estimating Causal Peer Influence in Homophilous Social Networks by Inferring Latent Locations.” Journal of the American Statistical Association118: 707–718.
56.
MelnychukValentynFrauenDennisFeuerriegelStefan. 2022. “Causal Transformer for Estimating Counterfactual Outcomes.” In Proceedings of the 39th International Conference on Machine Learning, edited by Kamalika Chaudhuri, Stefanie Jegelka, Le Song, Csaba Szepesvari, Gang Niu, and Sivan Sabato, volume 162 of Proceedings of Machine Learning Research, pp. 15293–15329. PMLR.
57.
MikolovTomasSutskeverIlyaChenKaiCorradoGregDeanJeffrey. 2013. “Distributed Representations of Words and Phrases and Their Compositionality.” In Neural Information Processing Systems, volume 26, p. 3111–3119. Curran Associates Inc.
NagpalChiragWeiDennisVinzamuriBhanukiranShekharMonicaBergerSara E.DasSubhroVarshneyKush R.. 2020. ‘I‘nterpretable Subgroup Discovery in Treatment Effect Estimation with Application to Opioid Prescribing Guidelines.” In Conference on Health, Inference, and Learning, p. 19–29. Association for Computing Machinery.
60.
NaimiAshley IMishlerAlan EKennedyEdward H. 2021. “Challenges in Obtaining Valid Causal Effect Estimates With Machine Learning Algorithms.” American Journal of Epidemiology192: 1536–44.
61.
NieXinkunWagerStefan. 2021. “Quasi-oracle Estimation of Heterogeneous Treatment Effects.” Biometrika108: 299–319.
62.
ParikhHarshVarjaoCarlosXuLouiseTchetgenEric Tchetgen. 2022. “Validating Causal Inference Methods.” In Proceedings of the 39th International Conference on Machine Learning, edited by Kamalika Chaudhuri, Stefanie Jegelka, Le Song, Csaba Szepesvari, Gang Niu, and Sivan Sabato, volume 162 of Proceedings of Machine Learning Research, pp. 17346–17358. PMLR.
63.
PearlJudea. 2009. Causality. New York, NY: Cambridge University Press.
64.
PryzantReidCardDallasJurafskyDanVeitchVictorSridharDhanya. 2021. “Causal Effects of Linguistic Properties.” In Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 4095–4109. Association for Computational Linguistics.
65.
RameyCraig TBryantDonna MWasikBarbara HSparlingJoseph JFendtKaye HLa VangeLisa M.. 1992. “Infant Health and Development Program for Low Birth Weight, Premature Infants: Program Elements, Family Participation, and Child Intelligence.” Pediatrics89: 454–65.
66.
RibeiroMarco TulioSinghSameerGuestrinCarlos. 2016. “‘ Why should i trust you?” Explaining the predictions of any classifier.’ In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, pp. 1135–1144.
67.
RobertsDaniel A.YaidaShoHaninBoris. 2022. The Principles of Deep Learning Theory. New York: Cambridge University Press. https://deeplearningtheory.com.
68.
RobinsJames. 1986. “A New Approach to Causal Inference in Mortality Studies with a Sustained Exposure Period—Application to Control of the Healthy Worker Survivor Effect.” Mathematical Modelling7: 1393–512.
69.
RobinsJames. 1987. “A Graphical Approach to the Identification and Estimation of Causal Parameters in Mortality Studies with Sustained Exposure Periods.” Journal of Chronic Diseases40: 139S–161S.
70.
RobinsJames M. 1994. “Correcting for Non-compliance in Randomized Trials Using Structural Nested Mean Models.” Communications in Statistics-Theory and Methods23: 2379–412.
71.
RobinsJamesHernánMiguel. 2008. “Estimation of the Causal Effects of Time-Varying Exposures.” In Longitudinal Data Analysis, Chapman & Hall/CRC Handbooks of Modern Statistical Methods, edited by Garett Fitzmaurice, Marie Davidian, Geert Verbeke, and Geert Molenberghs, pp. 553–99. Vancouver: Chapman and Hall/CRC
72.
RobinsJames MHernanMiguel AngelBrumbackBabette. 2000. “Marginal Structural Models and Causal Inference in Epidemiology.” Epidemiology11: 550–560.
73.
RosenbaumPaul RRubinDonald B. 1983. “The Central Role of the Propensity Score in Observational Studies for Causal Effects.” Biometrika70: 41–55.
74.
RubinDonald B. 1974. “Estimating Causal Effects of Treatments in Randomized and Non-randomized Studies.” Journal of Educational Psychology66: 688.
75.
SchnitzerMireille ELokJudith JGruberSusan. 2016. “Variable Selection for Confounder Control, Flexible Modeling and Collaborative Targeted Minimum Loss-based Estimation in Causal Inference.” The international journal of biostatistics12: 97–115.
76.
ShalitUriJohanssonFredrik DSontagDavid. 2017. “Estimating Individual Treatment Effect : Generalization Bounds and Algorithms.” In International Conference on Machine Learning. Association for Computing Machinery.
77.
ShaliziCosma RohillaThomasAndrew C.. 2011. “Homophily and Contagion Are Generically Confounded in Observational Social Network Studies.” Sociological Methods & Research40: 211–39.
78.
ShiClaudiaBleiDavid M.VeitchVictor. 2019. “Adapting Neural Networks for the Estimation of Treatment Effects.” In Proceedings of the 33rd International Conference on Neural Information Processing Systems. Red Hook, NY, USA: Curran Associates Inc., Article 225: 2507–17.
79.
SnoekJasperLarochelleHugoAdamsRyan P. 2012. “Practical Bayesian Optimization of Machine Learning Algorithms.” In Advances in Neural Information Processing Systems, edited by F. Pereira, C.J. Burges, L. Bottou, and K.Q. Weinberger, volume 25. Curran Associates, Inc.
80.
SrivastavaNitishHintonE. GeoffreyKrizhevskyAlexSutskeverIlyaSalakhutdinovRuslan. 2014. “Dropout: a Simple Way to Prevent Neural Networks From Overfitting.” Journal of Machine Learning Research15: 1929–58.
81.
StuartElizabeth A. 2010. “Matching Methods for Causal Inference: A Review and a Look Forward.” Statistical Science25: 1.
82.
SundararajanMukundTalyAnkurYanQiqi. 2017. “Axiomatic attribution for deep networks.” In International conference on machine learning, pp. 3319–3328. PMLR.
83.
TodorovAlexanderMandisodzaAnesu NGorenAmirHallCrystal C. 2005. “Inferences of Competence From Faces Predict Election Outcomes.” Science (New York, N.Y.)308: 1623–6.
84.
Van der LaanMark JRoseSherri. 2011. Targeted Learning: Causal Inference for Observational and Experimental Data. New York, NY: Springer Science & Business Media.
85.
VeitchVictorSridharDhanyaBleiDavid. 2020. “Adapting Text Embeddings for Causal Inference.” In Conference on Uncertainty in Artificial Intelligence, pp. 919–928. Association for Uncertainty in Artificial Intelligence.
86.
VeitchVictorWangYixinBleiDavid M.. 2019. “Using Embeddings to Correct for Unobserved Confounding in Networks.” In Proceedings of the 33rd International Conference on Neural Information Processing Systems. Red Hook, NY, USA: Curran Associates Inc., Article 1237: 13809–19.
87.
VeličkovićPetarCucurullGuillemCasanovaArantxaRomeroAdrianaLioPietroBengioYoshua. 2018. “Graph Attention Networks.” In International Conference for Learning Representations, volume 6. OpenReview.
88.
WagerStefanAtheySusan. 2018. “Estimation and Inference of Heterogeneous Treatment Effects Using Random Forests.” Journal of the American Statistical Association113: 1228–42.
89.
ZhangYaoBellotAlexisSchaarMihaela. 2020. “Learning Overlapping Representations for the Estimation of Individualized Treatment Effects.” In International Conference on Artificial Intelligence and Statistics, pp. 1005–1014. Association for Artificial Intelligence and Statistics.
90.
ZivichPaul NBreskinAlexander. 2021. “Machine Learning for Causal Inference: on the Use of Cross-fit Estimators.” Epidemiology (Cambridge, Mass.)32: 393.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
