{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T01:06:52Z","timestamp":1788311212475,"version":"build-2803163510"},"reference-count":57,"publisher":"Oxford University Press (OUP)","issue":"3","funder":[{"DOI":"10.13039\/100000026","name":"NIDA","doi-asserted-by":"publisher","award":["F31DA052153"],"award-info":[{"award-number":["F31DA052153"]}],"id":[{"id":"10.13039\/100000026","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000026","name":"NIDA","doi-asserted-by":"publisher","award":["R01DA048096"],"award-info":[{"award-number":["R01DA048096"]}],"id":[{"id":"10.13039\/100000026","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"NSF","doi-asserted-by":"publisher","award":["1810829"],"award-info":[{"award-number":["1810829"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"NSF","doi-asserted-by":"publisher","award":["2113707"],"award-info":[{"award-number":["2113707"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000001","name":"NSF","doi-asserted-by":"publisher","award":["IIS-1718258"],"award-info":[{"award-number":["IIS-1718258"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100007297","name":"ONR","doi-asserted-by":"publisher","award":["N000142112841"],"award-info":[{"award-number":["N000142112841"]}],"id":[{"id":"10.13039\/100007297","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000025","name":"NIMH","doi-asserted-by":"publisher","award":["R01MH121099, R01MH124115"],"award-info":[{"award-number":["R01MH121099, R01MH124115"]}],"id":[{"id":"10.13039\/100000025","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000002","name":"NIH","doi-asserted-by":"publisher","award":["5KL2TR001420"],"award-info":[{"award-number":["5KL2TR001420"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,7,15]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>We consider a problem in multi-task learning (MTL) where multiple linear models are jointly trained on a collection of datasets (\u2018tasks\u2019). A key novelty of our framework is that it allows the sparsity pattern of regression coefficients and the values of non-zero coefficients to differ across tasks while still leveraging partially shared structure. Our methods encourage models to share information across tasks through separately encouraging (1) coefficient supports, and\/or (2) nonzero coefficient values to be similar. This allows models to borrow strength during variable selection even when non-zero coefficient values differ across tasks. We propose a novel mixed-integer programming formulation for our estimator. We develop custom scalable algorithms based on block coordinate descent and combinatorial local search to obtain high-quality (approximate) solutions for our estimator. Additionally, we propose a novel exact optimization algorithm to obtain globally optimal solutions. We investigate the theoretical properties of our estimators. We formally show how our estimators leverage the shared support information across tasks to achieve better variable selection performance. We evaluate the performance of our methods in simulations and two biomedical applications. Our proposed approaches appear to outperform other sparse MTL methods in variable selection and prediction accuracy. We provide the sMTL package on CRAN.<\/jats:p>","DOI":"10.1093\/jrsssb\/qkaf076","type":"journal-article","created":{"date-parts":[[2025,11,25]],"date-time":"2025-11-25T12:54:04Z","timestamp":1764075244000},"page":"1029-1059","source":"Crossref","is-referenced-by-count":1,"title":["Multi-task learning for sparsity pattern heterogeneity: statistical and computational perspectives"],"prefix":"10.1093","volume":"88","author":[{"given":"Kayhan","family":"Behdin","sequence":"first","affiliation":[{"name":"Operations Research Center, Massachusetts Institute of Technology , Cambridge, MA 02139 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gabriel","family":"Loewinger","sequence":"additional","affiliation":[{"name":"Machine Learning Core, National Institute of Mental Health , Bethesda, MD 20892 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kenneth T","family":"Kishida","sequence":"additional","affiliation":[{"name":"Department of Physiology and Pharmacology, Department of Neurosurgery, Wake Forest School of Medicine , Winston-Salem, NC 27101 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8783-5961","authenticated-orcid":false,"given":"Giovanni","family":"Parmigiani","sequence":"additional","affiliation":[{"name":"Department of Data Science, Dana Farber Cancer Institute , Boston, MA 02215 ,","place":["USA"]},{"name":"Department of Biostatistics, Harvard School of Public Health , Boston, MA 02215 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rahul","family":"Mazumder","sequence":"additional","affiliation":[{"name":"Operations Research Center, Massachusetts Institute of Technology , Cambridge, MA 02139 ,","place":["USA"]},{"name":"Sloan School of Management, Massachusetts Institute of Technology , Cambridge, MA 02139 ,","place":["USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2025,12,17]]},"reference":[{"key":"2026071507120621500_qkaf076-B1","doi-asserted-by":"crossref","first-page":"682","DOI":"10.1109\/ITW.2012.6404765","article-title":"The analog formulation of sparsity implies infinite divisibility and rules out bernoulli-gaussian priors","author":"Amini","year":"2012","journal-title":"2012 IEEE information theory workshop"},{"issue":"3","key":"2026071507120621500_qkaf076-B2","doi-asserted-by":"publisher","first-page":"327","DOI":"10.1214\/aoms\/1177729580","article-title":"Estimating linear restrictions on regression coefficients for multivariate normal distributions","volume":"22","author":"Anderson","year":"1951","journal-title":"The Annals of Mathematical Statistics"},{"key":"2026071507120621500_qkaf076-B3","article-title":"A spectral regularization framework for multi-task structure learning","author":"Argyriou","year":"2007","journal-title":"Advances in neural information processing systems"},{"issue":"5","key":"2026071507120621500_qkaf076-B4","doi-asserted-by":"publisher","first-page":"999","DOI":"10.1016\/j.neuron.2020.09.015","article-title":"Sub-second dopamine and serotonin signaling in human striatum during perceptual decision-making","volume":"108","author":"Bang","year":"2020","journal-title":"Neuron"},{"issue":"3","key":"2026071507120621500_qkaf076-B5","doi-asserted-by":"publisher","first-page":"1480","DOI":"10.1137\/120869778","article-title":"Sparsity constrained nonlinear optimization: Optimality conditions and algorithms","volume":"23","author":"Beck","year":"2013","journal-title":"SIAM Journal on Optimization"},{"issue":"1","key":"2026071507120621500_qkaf076-B6","doi-asserted-by":"publisher","first-page":"183","DOI":"10.1137\/080716542","article-title":"A fast iterative shrinkage-thresholding algorithm for linear inverse problems","volume":"2","author":"Beck","year":"2009","journal-title":"SIAM Journal on Imaging Sciences"},{"key":"2026071507120621500_qkaf076-B7","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2109.11142","article-title":"Sparse PCA: A new scalable estimator based on integer programming","author":"Behdin","year":"2021"},{"issue":"2","key":"2026071507120621500_qkaf076-B8","doi-asserted-by":"publisher","first-page":"813","DOI":"10.1214\/15-AOS1388","article-title":"Best subset selection via a modern optimization lens","volume":"44","author":"Bertsimas","year":"2016","journal-title":"Annals of Statistics"},{"issue":"1","key":"2026071507120621500_qkaf076-B9","doi-asserted-by":"publisher","first-page":"300","DOI":"10.1214\/18-AOS1804","article-title":"Sparse high-dimensional regression: Exact scalable algorithms and phase transitions","volume":"48","author":"Bertsimas","year":"2020","journal-title":"Annals of Statistics"},{"issue":"3","key":"2026071507120621500_qkaf076-B10","doi-asserted-by":"publisher","first-page":"731","DOI":"10.1111\/biom.v71.3","article-title":"The group exponential lasso for bi-level variable selection","volume":"71","author":"Breheny","year":"2015","journal-title":"Biometrics"},{"issue":"3","key":"2026071507120621500_qkaf076-B11","doi-asserted-by":"publisher","first-page":"369","DOI":"10.4310\/SII.2009.v2.n3.a10","article-title":"Penalized methods for bi-level variable selection","volume":"2","author":"Breheny","year":"2009","journal-title":"Statistics and its Interface"},{"issue":"2","key":"2026071507120621500_qkaf076-B12","doi-asserted-by":"publisher","first-page":"173","DOI":"10.1007\/s11222-013-9424-2","article-title":"Group descent algorithms for nonconvex penalized linear and logistic regression models with grouped predictors","volume":"25","author":"Breheny","year":"2015","journal-title":"Statistics and Computing"},{"issue":"4","key":"2026071507120621500_qkaf076-B13","doi-asserted-by":"publisher","first-page":"1723","DOI":"10.1214\/21-AOAS1456","article-title":"Bayesian multistudy factor analysis for high-throughput biological data","volume":"15","author":"De Vito","year":"2021","journal-title":"The Annals of Applied Statistics"},{"issue":"3","key":"2026071507120621500_qkaf076-B14","doi-asserted-by":"publisher","first-page":"307","DOI":"10.1007\/BF02592064","article-title":"An outer-approximation algorithm for a class of mixed-integer nonlinear programs","volume":"36","author":"Duran","year":"1986","journal-title":"Mathematical Programming"},{"key":"2026071507120621500_qkaf076-B15","first-page":"615","article-title":"Learning multiple tasks with kernel methods","volume":"6","author":"Evgeniou","year":"2005","journal-title":"Journal of Machine Learning Research"},{"key":"2026071507120621500_qkaf076-B16","author":"Evgeniou","year":"2004"},{"issue":"456","key":"2026071507120621500_qkaf076-B17","doi-asserted-by":"publisher","first-page":"1348","DOI":"10.1198\/016214501753382273","article-title":"Variable selection via nonconcave penalized likelihood and its oracle properties","volume":"96","author":"Fan","year":"2001","journal-title":"Journal of the American Statistical Association"},{"issue":"1","key":"2026071507120621500_qkaf076-B18","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18637\/jss.v033.i01","article-title":"Regularization paths for generalized linear models via coordinate descent","volume":"33","author":"Friedman","year":"2010","journal-title":"Journal of Statistical Software"},{"issue":"509","key":"2026071507120621500_qkaf076-B19","doi-asserted-by":"publisher","first-page":"435","DOI":"10.1080\/01621459.2014.993077","article-title":"Decoupling shrinkage and selection in bayesian linear models: A posterior summary perspective","volume":"110","author":"Hahn","year":"2015","journal-title":"Journal of the American Statistical Association"},{"issue":"5","key":"2026071507120621500_qkaf076-B20","doi-asserted-by":"publisher","first-page":"1517","DOI":"10.1287\/opre.2019.1919","article-title":"Fast best subset selection: Coordinate descent and local combinatorial optimization algorithms","volume":"68","author":"Hazimeh","year":"2020, 2022\/03\/14","journal-title":"Operations Research"},{"issue":"1","key":"2026071507120621500_qkaf076-B21","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1214\/21-AOS2155","article-title":"Grouped variable selection with discrete optimization: Computational and statistical perspectives","volume":"51","author":"Hazimeh","year":"2023","journal-title":"Annals of Statistics"},{"issue":"1-2","key":"2026071507120621500_qkaf076-B22","doi-asserted-by":"publisher","first-page":"347","DOI":"10.1007\/s10107-021-01712-4","article-title":"Sparse regression at scale: Branch-and-bound rooted in first-order optimization","volume":"196","author":"Hazimeh","year":"2020","journal-title":"Mathematical Programming"},{"key":"2026071507120621500_qkaf076-B23","author":"Hernandez-Lobato","year":"2015"},{"issue":"1","key":"2026071507120621500_qkaf076-B24","doi-asserted-by":"publisher","first-page":"55","DOI":"10.1080\/00401706.1970.10488634","article-title":"Ridge regression: Biased estimation for nonorthogonal problems","volume":"12","author":"Hoerl","year":"1970","journal-title":"Technometrics: A Journal of Statistics for the Physical, Chemical, and Engineering Sciences"},{"issue":"4","key":"2026071507120621500_qkaf076-B25","doi-asserted-by":"publisher","first-page":"481","DOI":"10.1214\/12-STS392","article-title":"A selective review of group selection in high-dimensional models","volume":"27","author":"Huang","year":"2012","journal-title":"Statistical Science"},{"issue":"3","key":"2026071507120621500_qkaf076-B26","doi-asserted-by":"publisher","first-page":"1095","DOI":"10.1214\/12-AOAS549","article-title":"Tree-guided group lasso for multi-response regression with structured sparsity, with an application to eQTL mapping","volume":"6","author":"Kim","year":"2012","journal-title":"The Annals of Applied Statistics"},{"issue":"1","key":"2026071507120621500_qkaf076-B27","doi-asserted-by":"publisher","first-page":"200","DOI":"10.1073\/pnas.1513619112","article-title":"Subsecond dopamine fluctuations in human striatum encode superposed error signals about actual and counterfactual reward","volume":"113","author":"Kishida","year":"2016","journal-title":"Proceedings of the National Academy of Sciences"},{"issue":"8","key":"2026071507120621500_qkaf076-B28","doi-asserted-by":"publisher","first-page":"e23291","DOI":"10.1371\/journal.pone.0023291","article-title":"Sub-second dopamine detection in human striatum","volume":"6","author":"Kishida","year":"2011","journal-title":"PLoS One"},{"key":"2026071507120621500_qkaf076-B29","author":"Liang","year":"2023"},{"key":"2026071507120621500_qkaf076-B30","author":"Liu","year":"2009"},{"issue":"4","key":"2026071507120621500_qkaf076-B31","doi-asserted-by":"publisher","first-page":"2145","DOI":"10.1214\/21-AOAS1574","article-title":"Hierarchical resampling for bagging in multistudy prediction with applications to human neurochemical sensing","volume":"16","author":"Loewinger","year":"2022","journal-title":"The Annals of Applied Statistics"},{"key":"2026071507120621500_qkaf076-B32","first-page":"361","article-title":"Multi-level lasso for sparse multi-task regression.","author":"Lozano","year":"2012","journal-title":"Proceedings of the 29th international conference on machine learning"},{"issue":"1","key":"2026071507120621500_qkaf076-B33","doi-asserted-by":"publisher","first-page":"129","DOI":"10.1287\/opre.2022.2276","article-title":"Subset selection with shrinkage: Sparse linear modeling when the SNR is low","volume":"71","author":"Mazumder","year":"2022","journal-title":"Operations Research"},{"key":"2026071507120621500_qkaf076-B34","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4899-2939-6","volume-title":"Subset selection in regression","author":"Miller","year":"1990"},{"issue":"3","key":"2026071507120621500_qkaf076-B35","doi-asserted-by":"publisher","first-page":"595","DOI":"10.1093\/biomet\/asw034","article-title":"Indirect multivariate response linear regression","volume":"103","author":"Molstad","year":"2016","journal-title":"Biometrika"},{"issue":"3","key":"2026071507120621500_qkaf076-B36","doi-asserted-by":"publisher","first-page":"612","DOI":"10.1080\/10618600.2020.1853551","article-title":"An explicit mean-covariance parameterization for multivariate response linear regression","volume":"30","author":"Molstad","year":"2021","journal-title":"Journal of Computational and Graphical Statistics"},{"key":"2026071507120621500_qkaf076-B37","doi-asserted-by":"publisher","article-title":"Random burst sensing of neurotransmitters","author":"Montague","DOI":"10.1101\/607077"},{"issue":"6","key":"2026071507120621500_qkaf076-B38","doi-asserted-by":"publisher","first-page":"1425","DOI":"10.1038\/npp.2017.304","article-title":"The protective action encoding of serotonin transients in the human brain","volume":"43","author":"Moran","year":"2018","journal-title":"Neuropsychopharmacology: Official Publication of the American College of Neuropsychopharmacology"},{"issue":"2","key":"2026071507120621500_qkaf076-B39","doi-asserted-by":"publisher","first-page":"227","DOI":"10.1137\/S0097539792240406","article-title":"Sparse approximate solutions to linear systems","volume":"24","author":"Natarajan","year":"1995","journal-title":"SIAM Journal on Computing"},{"issue":"3","key":"2026071507120621500_qkaf076-B40","doi-asserted-by":"publisher","first-page":"717","DOI":"10.1002\/asmb.v35.3","article-title":"Bayesian l0-regularized least squares","volume":"35","author":"Polson","year":"2019","journal-title":"Applied Stochastic Models in Business and Industry"},{"issue":"6","key":"2026071507120621500_qkaf076-B41","doi-asserted-by":"publisher","first-page":"3465","DOI":"10.1137\/090763184","article-title":"Trace norm regularization: Reformulations, algorithms, and multi-task learning","volume":"20","author":"Pong","year":"2010","journal-title":"SIAM Journal on Optimization"},{"issue":"531","key":"2026071507120621500_qkaf076-B42","doi-asserted-by":"publisher","first-page":"1125","DOI":"10.1080\/01621459.2019.1671197","article-title":"Modeling between-study heterogeneity for improved replicability in gene signature selection and clinical prediction","volume":"115","author":"Rashid","year":"2020","journal-title":"Journal of the American Statistical Association"},{"issue":"10","key":"2026071507120621500_qkaf076-B43","doi-asserted-by":"publisher","first-page":"6976","DOI":"10.1109\/TIT.2011.2165799","article-title":"Minimax rates of estimation for high-dimensional linear regression over \u2113q -balls","volume":"57","author":"Raskutti","year":"2011","journal-title":"IEEE Transactions on Information Theory"},{"key":"2026071507120621500_qkaf076-B44","author":"Reuther","year":"2018"},{"issue":"4","key":"2026071507120621500_qkaf076-B45","doi-asserted-by":"publisher","first-page":"947","DOI":"10.1198\/jcgs.2010.09188","article-title":"Sparse multivariate regression with covariance estimation","volume":"19","author":"Rothman","year":"2010","journal-title":"Journal of Computational and Graphical Statistics"},{"issue":"10","key":"2026071507120621500_qkaf076-B46","doi-asserted-by":"publisher","first-page":"4572","DOI":"10.1109\/TSP.2011.2160633","article-title":"From bernoulli-gaussian deconvolution to sparse signal restoration","volume":"59","author":"Soussen","year":"2011","journal-title":"IEEE Transactions on Signal Processing"},{"issue":"5","key":"2026071507120621500_qkaf076-B47","doi-asserted-by":"publisher","first-page":"2569","DOI":"10.1007\/s00429-015-1059-y","article-title":"Deep sparse multi-task learning for feature selection in Alzheimer\u2019s disease diagnosis","volume":"221","author":"Suk","year":"2016","journal-title":"Brain Structure & Function"},{"issue":"1","key":"2026071507120621500_qkaf076-B48","doi-asserted-by":"publisher","first-page":"267","DOI":"10.1111\/j.2517-6161.1996.tb02080.x","article-title":"Regression shrinkage and selection via the lasso","volume":"58","author":"Tibshirani","year":"1996","journal-title":"Journal of the Royal Statistical Society: Series B (Methodological)"},{"issue":"4","key":"2026071507120621500_qkaf076-B49","doi-asserted-by":"publisher","first-page":"1365","DOI":"10.1111\/biom.13517","article-title":"Integration of survival data from multiple studies","volume":"78","author":"Ventz","year":"2021","journal-title":"Biometrics"},{"key":"2026071507120621500_qkaf076-B50","doi-asserted-by":"crossref","DOI":"10.1017\/9781108627771","volume-title":"High-dimensional statistics: A non-asymptotic viewpoint","author":"Wainwright","year":"2019"},{"key":"2026071507120621500_qkaf076-B51","first-page":"2820","article-title":"Multiplicative multitask feature learning","volume":"17","author":"Wang","year":"2014","journal-title":"Journal of Machine Learning Research: JMLR"},{"issue":"9","key":"2026071507120621500_qkaf076-B52","doi-asserted-by":"publisher","first-page":"581","DOI":"10.1136\/medethics-2020-107102","article-title":"Machine learning in medicine: Should the pursuit of enhanced interpretability be abandoned?","volume":"48","author":"Yoon","year":"2021","journal-title":"Journal of Medical Ethics"},{"issue":"1","key":"2026071507120621500_qkaf076-B53","doi-asserted-by":"publisher","first-page":"31619","DOI":"10.1038\/srep31619","article-title":"Multitask learning improves prediction of cancer drug sensitivity","volume":"6","author":"Yuan","year":"2016","journal-title":"Scientific Reports"},{"issue":"1","key":"2026071507120621500_qkaf076-B54","doi-asserted-by":"publisher","first-page":"49","DOI":"10.1111\/j.1467-9868.2005.00532.x","article-title":"Model selection and estimation in regression with grouped variables","volume":"68","author":"Yuan","year":"2006","journal-title":"Journal of the Royal Statistical Society Series B"},{"issue":"5","key":"2026071507120621500_qkaf076-B55","doi-asserted-by":"publisher","first-page":"76","DOI":"10.1016\/j.neucom.2019.02.035","article-title":"Multi-task feature selection with sparse regularization to extract common and task-specific features","volume":"340","author":"Zhang","year":"2019","journal-title":"Neurocomputing"},{"issue":"12","key":"2026071507120621500_qkaf076-B56","doi-asserted-by":"publisher","first-page":"5586","DOI":"10.1109\/TKDE.2021.3070203","article-title":"A survey on multi-task learning","volume":"34","author":"Zhang","year":"2017","journal-title":"IEEE Transactions on Knowledge and Data Engineering"},{"key":"2026071507120621500_qkaf076-B57","article-title":"Probabilistic multi-task feature selection.","author":"Zhang","year":"2010","journal-title":"Advances in neural information processing systems"}],"container-title":["Journal of the Royal Statistical Society Series B: Statistical Methodology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/88\/3\/1029\/65926219\/qkaf076.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/88\/3\/1029\/65926219\/qkaf076.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,15]],"date-time":"2026-07-15T11:12:16Z","timestamp":1784113936000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jrsssb\/article\/88\/3\/1029\/8381812"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12,17]]},"references-count":57,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2025,12,17]]},"published-print":{"date-parts":[[2026,7,15]]}},"URL":"https:\/\/doi.org\/10.1093\/jrsssb\/qkaf076","relation":{},"ISSN":["1369-7412","1467-9868"],"issn-type":[{"value":"1369-7412","type":"print"},{"value":"1467-9868","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2026,7]]},"published":{"date-parts":[[2025,12,17]]}}}