{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,15]],"date-time":"2026-11-15T04:33:01Z","timestamp":1794717181323,"version":"4.9.1"},"reference-count":30,"publisher":"MDPI AG","issue":"2","license":[{"start":{"date-parts":[[2022,1,25]],"date-time":"2022-01-25T00:00:00Z","timestamp":1643068800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Entropy"],"abstract":"<jats:p>The estimation of average treatment effect (ATE) as a causal parameter is carried out in two steps, where in the first step, the treatment and outcome are modeled to incorporate the potential confounders, and in the second step, the predictions are inserted into the ATE estimators such as the augmented inverse probability weighting (AIPW) estimator. Due to the concerns regarding the non-linear or unknown relationships between confounders and the treatment and outcome, there has been interest in applying non-parametric methods such as machine learning (ML) algorithms instead. Some of the literature proposes to use two separate neural networks (NNs) where there is no regularization on the network\u2019s parameters except the stochastic gradient descent (SGD) in the NN\u2019s optimization. Our simulations indicate that the AIPW estimator suffers extensively if no regularization is utilized. We propose the normalization of AIPW (referred to as nAIPW) which can be helpful in some scenarios. nAIPW, provably, has the same properties as AIPW, that is, the double-robustness and orthogonality properties. Further, if the first-step algorithms converge fast enough, under regulatory conditions, nAIPW will be asymptotically normal. We also compare the performance of AIPW and nAIPW in terms of the bias and variance when small to moderate L1 regularization is imposed on the NNs.<\/jats:p>","DOI":"10.3390\/e24020179","type":"journal-article","created":{"date-parts":[[2022,1,25]],"date-time":"2022-01-25T20:38:58Z","timestamp":1643143138000},"page":"179","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":7,"title":["Normalized Augmented Inverse Probability Weighting with Neural Network Predictions"],"prefix":"10.3390","volume":"24","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0043-6325","authenticated-orcid":false,"given":"Mehdi","family":"Rostami","sequence":"first","affiliation":[{"name":"Dalla Lana School of Public Health, University of Toronto, 155 College st., Toronto, ON M5T 3M7, Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7177-8472","authenticated-orcid":false,"given":"Olli","family":"Saarela","sequence":"additional","affiliation":[{"name":"Dalla Lana School of Public Health, University of Toronto, 155 College st., Toronto, ON M5T 3M7, Canada"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2022,1,25]]},"reference":[{"key":"ref_1","unstructured":"Farrell, M.H., Liang, T., and Misra, S. (2018). Deep neural networks for estimation and inference: Application to causal effects and other semiparametric estimands. arXiv."},{"key":"ref_2","unstructured":"Goodfellow, I., Bengio, Y., Courville, A., and Bengio, Y. (2016). Deep Learning, MIT Press."},{"key":"ref_3","unstructured":"Cheng, X., Khomtchouk, B., Matloff, N., and Mohanty, P. (2018). Polynomial regression as an alternative to neural nets. arXiv."},{"key":"ref_4","unstructured":"Xu, B., Wang, N., Chen, T., and Li, M. (2015). Empirical evaluation of rectified activations in convolutional network. arXiv."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"2365","DOI":"10.1109\/TII.2020.3004461","article-title":"H_infinity Weighted Integral Event-Triggered Synchronization of Neural Networks With Mixed Delays","volume":"17","author":"Yan","year":"2020","journal-title":"IEEE Trans. Ind. Inform."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"846","DOI":"10.1080\/01621459.1994.10476818","article-title":"Estimation of regression coefficients when some regressors are not always observed","volume":"89","author":"Robins","year":"1994","journal-title":"J. Am. Stat. Assoc."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"2937","DOI":"10.1002\/sim.1903","article-title":"Stratification and weighting via the propensity score in estimation of causal treatment effects: A comparative study","volume":"23","author":"Lunceford","year":"2004","journal-title":"Stat. Med."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Van der Laan, M.J., and Rose, S. (2011). Targeted Learning: Causal Inference for Observational and Experimental Data, Springer Science & Business Media.","DOI":"10.1007\/978-1-4419-9782-1"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"C1","DOI":"10.1111\/ectj.12097","article-title":"Double\/debiased machine learning for treatment and structural parameters","volume":"21","author":"Chernozhukov","year":"2018","journal-title":"Econom. J."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"107","DOI":"10.1145\/3446776","article-title":"Understanding deep learning (still) requires rethinking generalization","volume":"64","author":"Zhang","year":"2021","journal-title":"Commun. ACM"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"1213","DOI":"10.1093\/aje\/kwr364","article-title":"Effects of adjusting for instrumental variables on bias and precision of effect estimates","volume":"174","author":"Myers","year":"2011","journal-title":"Am. J. Epidemiol."},{"key":"ref_12","unstructured":"D\u00edaz, I. (2018). Doubly robust estimators for the average treatment effect under positivity violations: Introducing the e-score. arXiv."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"663","DOI":"10.1080\/01621459.1952.10483446","article-title":"A generalization of sampling without replacement from a finite universe","volume":"47","author":"Horvitz","year":"1952","journal-title":"J. Am. Stat. Assoc."},{"key":"ref_14","first-page":"318","article-title":"Using multivariate matched sampling and regression adjustment to control bias in observational studies","volume":"74","author":"Rubin","year":"1979","journal-title":"J. Am. Stat. Assoc."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"41","DOI":"10.1093\/biomet\/70.1.41","article-title":"The central role of the propensity score in observational studies for causal effects","volume":"70","author":"Rosenbaum","year":"1983","journal-title":"Biometrika"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"33","DOI":"10.1080\/00031305.1985.10479383","article-title":"Constructing a control group using multivariate matched sampling methods that incorporate the propensity score","volume":"39","author":"Rosenbaum","year":"1985","journal-title":"Am. Stat."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"885","DOI":"10.1162\/REST_a_00431","article-title":"New evidence on the finite sample properties of propensity score reweighting and matching estimators","volume":"96","author":"Busso","year":"2014","journal-title":"Rev. Econ. Stat."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"315","DOI":"10.2307\/2998560","article-title":"On the role of the propensity score in efficient semiparametric estimation of average treatment effects","volume":"66","author":"Hahn","year":"1998","journal-title":"Econometrica"},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Hines, O., Dukes, O., Diaz-Ordaz, K., and Vansteelandt, S. (2021). Demystifying statistical learning based on efficient influence functions. arXiv.","DOI":"10.1080\/00031305.2021.2021984"},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"112","DOI":"10.1017\/S0266466617000056","article-title":"A general double robustness result for estimating average treatment effects","volume":"34","author":"Wooldridge","year":"2018","journal-title":"Econom. Theory"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"1096","DOI":"10.1080\/01621459.1999.10473862","article-title":"Adjusting for nonignorable drop-out using semiparametric nonresponse models","volume":"94","author":"Scharfstein","year":"1999","journal-title":"J. Am. Stat. Assoc."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Angrist, J.D., and Pischke, J.S. (2008). Mostly Harmless Econometrics: An Empiricist\u2019s Companion, Princeton University Press.","DOI":"10.2307\/j.ctvcm4j72"},{"key":"ref_23","unstructured":"Van der Vaart, A.W. (2000). Asymptotic Statistics, Cambridge University Press."},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Chernozhukov, V., Chetverikov, D., Demirer, M., Duflo, E., Hansen, C., and Newey, W.K. (2021, November 12). Double Machine Learning for Treatment and Causal Parameters; Technical Report, Cemmap Working Paper; 2016. Available online: https:\/\/ifs.org.uk\/uploads\/cemmap\/wps\/cwp491616.pdf.","DOI":"10.1920\/wp.cem.2016.4916"},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.jeconom.2015.06.017","article-title":"Robust inference on average treatment effects with possibly more covariates than observations","volume":"189","author":"Farrell","year":"2015","journal-title":"J. Econom."},{"key":"ref_26","unstructured":"Smucler, E., Rotnitzky, A., and Robins, J.M. (2019). A unifying approach for doubly-robust l1 regularized estimation of causal contrasts. arXiv."},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Friedman, J., Hastie, T., and Tibshirani, R. (2001). The Elements of Statistical Learning, Springer.","DOI":"10.1007\/978-0-387-21606-5"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"29","DOI":"10.1198\/000313002753631330","article-title":"The calculus of M-estimation","volume":"56","author":"Stefanski","year":"2002","journal-title":"Am. Stat."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"1303","DOI":"10.1080\/01621459.1994.10476870","article-title":"The stationary bootstrap","volume":"89","author":"Politis","year":"1994","journal-title":"J. Am. Stat. Assoc."},{"key":"ref_30","unstructured":"Kingma, D.P., and Ba, J. (2014). Adam: A method for stochastic optimization. arXiv."}],"container-title":["Entropy"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1099-4300\/24\/2\/179\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T22:07:27Z","timestamp":1760134047000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1099-4300\/24\/2\/179"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,1,25]]},"references-count":30,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2022,2]]}},"alternative-id":["e24020179"],"URL":"https:\/\/doi.org\/10.3390\/e24020179","relation":{},"ISSN":["1099-4300"],"issn-type":[{"value":"1099-4300","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,1,25]]}}}