{"type":"ScholarlyArticle","title":"Ancestral Instrument Method for Causal Inference without Complete Knowledge","authors":[{"name":"Debo Cheng","slug":"debo-cheng","given":"Debo","family":"Cheng","affiliation":"STEM , University of South Australia , Adelaide , SA , Australia"},{"name":"Jiuyong Li","slug":"jiuyong-li","given":"Jiuyong","family":"Li","affiliation":"STEM , University of South Australia , Adelaide , SA , Australia"},{"name":"Lin Liu","slug":"lin-liu","given":"Lin","family":"Liu","affiliation":"STEM , University of South Australia , Adelaide , SA , Australia"},{"name":"Jiji Zhang","slug":"jiji-zhang","given":"Jiji","family":"Zhang","affiliation":"Department of Religion and Philosophy , Hong Kong Baptist University , Hong Kong , China"},{"name":"Thuc Duy Le","slug":"thuc-duy-le","given":"Thuc Duy","family":"Le","affiliation":"STEM , University of South Australia , Adelaide , SA , Australia"},{"name":"Jixue Liu","slug":"jixue-liu","given":"Jixue","family":"Liu","affiliation":"STEM , University of South Australia , Adelaide , SA , Australia"}],"abstract":"Unobserved confounding is the main obstacle to causal effect estimation from observational data. Instrumental variables (IVs) are widely used for causal effect estimation when there exist latent confounders. With the standard IV method, when a given IV is valid, unbiased estimation can be obtained, but the validity requirement on a standard IV is strict and untestable. Conditional IVs have been proposed to relax the requirement of standard IVs by conditioning on a set of observed variables (known as a conditioning set for a conditional IV). However, the criterion for finding a conditioning set for a conditional IV needs a directed acyclic graph (DAG) representing the causal relationships of both observed and unobserved variables. This makes it challenging to discover a conditioning set directly from data. In this paper, by leveraging maximal ancestral graphs (MAGs) for causal inference with latent variables, we study the graphical properties of ancestral IVs, a type of conditional IVs using MAGs, and develop the theory to support data-driven discovery of the conditioning set for a given ancestral IV in data under the pretreatment variable assumption. Based on the theory, we develop an algorithm for unbiased causal effect estimation with a given ancestral IV and observational data. Extensive experiments on synthetic and real-world datasets demonstrate the performance of the algorithm in comparison with existing IV methods.","doi":"10.24963/ijcai.2022/671","url":"https://ijcai.org/proceedings/2022/671","pdf":"https://ijcai.org/proceedings/2022/0671.pdf","markdown":"https://ijcai.org/proceedings/2022/671.md","bibtex":"https://ijcai.org/proceedings/2022/bibtex/671","volume":{"slug":"2022","title":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2022","url":"https://ijcai.org/proceedings/2022/","year":2022,"location":"Vienna","dates":"23-29 July 2022","editors":"Luc De Raedt","isbn":"978-1-956792-00-3","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2022"},"section":"Uncertainty in AI","track":"Main Track","number":671,"pages":{"first":4843,"last":4849},"keywords":["Uncertainty in AI: Causality, Structural Causal Models and Causal Inference","Knowledge Representation and Reasoning: Causality"],"videos":[{"label":"Video #1","url":"https://slideslive.com/38986516","length_sec":88,"provider":"slideslive"},{"label":"Video #2","url":"https://slideslive.com/38984659","length_sec":351,"provider":"slideslive"}],"references":[{"position":1,"raw":"[Abadie, 2003] Alberto Abadie. Semiparametric instrumen- tal variable estimation of treatment response models. Jour- nal of Econometrics, 113(2):231-263, 2003.","title":"Semiparametric instrumental variable estimation of treatment response models","authors":"Alberto Abadie","year":"2003","venue":"Journal of Econometrics","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Angrist and Imbens, 1995] Joshua D Angrist and Guido W Imbens. Two-stage least squares estimation of average causal effects in models with variable treatment inten- sity. Journal of the American Statistical Association, 90(430):431-442, 1995.","title":"Two-stage least squares estimation of average causal effects in models with variable treatment intensity","authors":"D Joshua, Guido W Angrist, Imbens","year":"1995","venue":"Journal of the American Statistical Association","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Athey et al., 2019] Susan Athey, Julie Tibshirani, and Ste- fan Wager. Generalized random forests. The Annals of Statistics, 47(2):1148-1178, 2019.","title":"Generalized random forests","authors":"Susan Athey, Julie Tibshirani, Stefan Wager","year":"2019","venue":"The Annals of Statistics","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Brito and Pearl, 2002] Carlos Brito and Judea Pearl. Gener- alized instrumental variables. In UAI, pages 85-93, 2002.","title":"Carlos Brito and Judea Pearl. Generalized instrumental variables","authors":"Pearl Brito","year":"2002","venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Card, 1993] David Card. Using geographic variation in col- lege proximity to estimate the return to schooling. Econo- metrica, 69(9):1127-1160, 1993.","title":"Using geographic variation in college proximity to estimate the return to schooling","authors":"David Card","year":"1993","venue":"Econometrica","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Cheng et al., 2022] Debo Cheng, Jiuyong Li, et al. Ances- tral instrument method for causal inference without com- plete ksnowledge. arXiv preprint arXiv:2201.03810, 2022.","title":"Ancestral instrument method for causal inference without complete ksnowledge","authors":"Debo Cheng, Jiuyong Li","year":"2022","venue":null,"doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Chu et al., 2001] Tianjiao Chu, Richard Scheines, and Peter Spirtes. Semi-instrumental variables: a test for instrument admissibility. In UAI, pages 83-90, 2001.","title":"Semi-instrumental variables: a test for instrument admissibility","authors":"Tianjiao Chu, Richard Scheines, Peter Spirtes","year":"2001","venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Colombo et al., 2012] Diego Colombo, Marloes H Maathuis, et al. Learning high-dimensional directed acyclic graphs with latent and selection variables. The Annals of Statistics, 40(1):294-321, 2012.","title":"Learning high-dimensional directed acyclic graphs with latent and selection variables","authors":"Colombo","year":"2012","venue":"The Annals of Statistics","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Greene, 2003] William H Greene. Econometric Analysis. Pearson Education India, 2003.","title":"Econometric Analysis","authors":"William H Greene","year":"2003","venue":null,"doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Hartford et al., 2021] Jason S Hartford, Victor Veitch, et al. Valid causal inference with (some) invalid instruments. In ICML, pages 4096-4106. PMLR, 2021.","title":"Valid causal inference with (some) invalid instruments","authors":"Jason S Hartford, Victor Veitch","year":"2021","venue":"ICML","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Hernán and Robins, 2006] Miguel A Hernán and James M Robins. Instruments for causal inference: an epidemiolo- gist's dream? Epidemiology, 17(4):360-372, 2006.","title":"Instruments for causal inference: an epidemiologist's dream?","authors":"A Miguel, James M Hernán, Robins","year":"2006","venue":"Epidemiology","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Imbens and Rubin, 2015] Guido W Imbens and Donald B Rubin. Causal Inference in Statistics, Social, and Biomed- ical Sciences. Cambridge University Press, 2015.","title":"Causal Inference in Statistics, Social, and Biomedical Sciences","authors":"W Guido, Donald B Imbens, Rubin","year":"2015","venue":null,"doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Imbens, 2014] Guido W Imbens. Instrumental variables: An econometrician's perspective. Statistical Science, 29(3):323-358, 2014.","title":"Instrumental variables: An econometrician's perspective","authors":"Guido W Imbens","year":"2014","venue":"Statistical Science","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Kalisch et al., 2012] Markus Kalisch, Martin Mächler, et al. Causal inference using graphical models with the R pack- age pcalg. Journal of Statistical Software, 47(11):1-26, 2012.","title":"Causal inference using graphical models with the R package pcalg","authors":"Markus Kalisch, Martin Mächler","year":"2012","venue":"Journal of Statistical Software","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Kang et al., 2016] Hyunseung Kang, Anru Zhang, et al. Instrumental variables estimation with some invalid in- struments and its application to mendelian randomiza- tion. Journal of the American Statistical Association, 111(513):132-144, 2016.","title":"Instrumental variables estimation with some invalid instruments and its application to mendelian randomization","authors":"Kang","year":"2016","venue":"Journal of the American Statistical Association","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Kuroki and Cai, 2005] Manabu Kuroki and Zhihong Cai. Instrumental variable tests for directed acyclic graph mod- els. In AISTATS, pages 190-197, 2005.","title":"Instrumental variable tests for directed acyclic graph models","authors":"Manabu Kuroki, Zhihong Cai","year":"2005","venue":"AISTATS","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Maathuis et al., 2015] Marloes H Maathuis, Diego Colombo, et al. A generalized back-door criterion. The Annals of Statistics, 43(3):1060-1088, 2015.","title":"A generalized back-door criterion","authors":"Diego Marloes H Maathuis, Colombo","year":"2015","venue":"The Annals of Statistics","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Martens et al., 2006] Edwin P Martens, Wiebe R Pestman, et al. Instrumental variables: application and limitations. Epidemiology, 17(3):260-267, 2006.","title":"Instrumental variables: application and limitations","authors":"Edwin P Martens, Wiebe R Pestman","year":"2006","venue":"Epidemiology","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Martinussen and others, 2019] Torben Martinussen et al. Instrumental variables estimation under a structural cox model. Biostatistics, 20(1):65-79, 2019.","title":"Instrumental variables estimation under a structural cox model","authors":"Torben Martinussen, Martinussen","year":"2019","venue":"Biostatistics","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Pearl, 1995] Judea Pearl. On the testability of causal models with latent and instrumental variables. In UAI, pages 435- 443, 1995.","title":"Judea Pearl. On the testability of causal models with latent and instrumental variables","authors":"Pearl","year":"1995","venue":"UAI","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Pearl, 2009] Judea Pearl. Causality. Cambridge University Press, 2009.","title":"Judea Pearl. Causality","authors":"Pearl","year":"2009","venue":null,"doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Perković et al., 2018] Emilija Perković, Johannes Textor, and Markus Kalisch. Complete graphical characterization and construction of adjustment sets in markov equivalence classes of ancestral graphs. Journal of Machine Learning Research, 18:1-62, 2018.","title":"Complete graphical characterization and construction of adjustment sets in markov equivalence classes of ancestral graphs","authors":"Emilija Perković, Johannes Textor, Markus Kalisch","year":"2018","venue":"Journal of Machine Learning Research","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Richardson and Spirtes, 2002] Thomas Richardson and Pe- ter Spirtes. Ancestral graph markov models. The Annals of Statistics, 30(4):962-1030, 2002.","title":"Ancestral graph markov models","authors":"Thomas Richardson, Peter Spirtes","year":"2002","venue":"The Annals of Statistics","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Silva and Shimizu, 2017] Ricardo Silva and Shohei Shimizu. Learning instrumental variables with structural and non-gaussianity assumptions. Journal of Machine Learning Research, 18(120):1-49, 2017.","title":"Learning instrumental variables with structural and non-gaussianity assumptions","authors":"Shimizu Silva, Ricardo Silva, Shohei Shimizu","year":"2017","venue":"Journal of Machine Learning Research","doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Sjolander and Martinussen, 2019] Arvid Sjolander and Tor- ben Martinussen. Instrumental variable estimation with the R package ivtools. Epidemiologic Methods, 8(1), 2019.","title":"Instrumental variable estimation with the R package ivtools","authors":"Martinussen Sjolander, Arvid Sjolander, Torben Martinussen","year":"2019","venue":"Epidemiologic Methods","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Spirtes et al., 2000] Peter Spirtes, Clark N Glymour, et al. Causation, Prediction, and Search. MIT Press, 2000.","title":"Causation, Prediction, and Search","authors":"Peter Spirtes, Clark N Glymour","year":"2000","venue":null,"doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Vander Zander et al., 2015] B. Vander Zander, Johannes Textor, et al. Efficiently finding conditional instruments for causal inference. In IJCAI, pages 3243-3249, 2015.","title":"Efficiently finding conditional instruments for causal inference","authors":"B Vander Zander, Johannes Textor","year":"2015","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2015/457"},{"position":28,"raw":"[Verbeek, 2008] Marno Verbeek. A Guide to Modern Econo- metrics. John Wiley & Sons, 2008.","title":"A Guide to Modern Econometrics","authors":"Marno Verbeek","year":"2008","venue":null,"doi":null,"url":null,"internal_url":null},{"position":29,"raw":"[Zhang et al., 2020] Chi Zhang, Bryant Chen, et al. A simul- taneous discover-identify approach to causal inference in linear models. In AAAI, pages 10318-10325, 2020.","title":"A simultaneous discover-identify approach to causal inference in linear models","authors":"Zhang","year":"2020","venue":"AAAI","doi":null,"url":null,"internal_url":null},{"position":30,"raw":"[Zhang, 2008a] Jiji Zhang. Causal reasoning with ances- tral graphs. Journal of Machine Learning Research, 9(7):1437-1474, 2008.","title":"Causal reasoning with ancestral graphs","authors":"; Zhang, Jiji Zhang","year":"2008","venue":"Journal of Machine Learning Research","doi":null,"url":null,"internal_url":null},{"position":31,"raw":"[Zhang, 2008b] Jiji Zhang. On the completeness of orien- tation rules for causal discovery in the presence of la- tent confounders and selection bias. Artificial Intelligence, 172(16-17):1873-1896, 2008.","title":"On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias","authors":"Jiji Zhang, Zhang","year":"2008","venue":"Artificial Intelligence","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}