{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,17]],"date-time":"2025-10-17T13:55:32Z","timestamp":1760709332060},"reference-count":62,"publisher":"MIT Press","issue":"8","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Neural Computation"],"published-print":{"date-parts":[[2017,8]]},"abstract":"<jats:p>Optimal control theory and machine learning techniques are combined to formulate and solve in closed form an optimal control formulation of online learning from supervised examples with regularization of the updates. The connections with the classical linear quadratic gaussian (LQG) optimal control problem, of which the proposed learning paradigm is a nontrivial variation as it involves random matrices, are investigated. The obtained optimal solutions are compared with the Kalman filter estimate of the parameter vector to be learned. It is shown that the proposed algorithm is less sensitive to outliers with respect to the Kalman estimate (thanks to the presence of the regularization term), thus providing smoother estimates with respect to time. The basic formulation of the proposed online learning framework refers to a discrete-time setting with a finite learning horizon and a linear model. Various extensions are investigated, including the infinite learning horizon and, via the so-called kernel trick, the case of nonlinear models.<\/jats:p>","DOI":"10.1162\/neco_a_00976","type":"journal-article","created":{"date-parts":[[2017,5,31]],"date-time":"2017-05-31T18:01:02Z","timestamp":1496253662000},"page":"2203-2291","source":"Crossref","is-referenced-by-count":5,"title":["LQG Online Learning"],"prefix":"10.1162","volume":"29","author":[{"given":"Giorgio","family":"Gnecco","sequence":"first","affiliation":[{"name":"DYSCO Research Unit, IMT School for Advanced Studies, Piazza S. Francesco, 19-55110 Lucca, Italy"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Alberto","family":"Bemporad","sequence":"additional","affiliation":[{"name":"DYSCO Research Unit, IMT School for Advanced Studies, Piazza S. Francesco, 19-55110 Lucca, Italy"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Marco","family":"Gori","sequence":"additional","affiliation":[{"name":"DIISM Department, University of Siena, Via Roma, 56-53100 Siena, Italy"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Marcello","family":"Sanguineti","sequence":"additional","affiliation":[{"name":"DIBRIS Department, University of Genoa, Via Opera Pia, 13-16145 Genova, Italy"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"281","reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1016\/j.automatica.2016.01.015"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.1109\/72.991413"},{"key":"B3","volume-title":"A linear systems primer","author":"Antsaklis P. J.","year":"2007"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1007\/978-0-8176-4757-5"},{"key":"B5","first-page":"2299","volume":"7","author":"Belkin M.","year":"2006","journal-title":"Journal of Machine Learning Research"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.1515\/9781400833344"},{"key":"B7","volume-title":"Dynamic programming and optimal control","volume":"1","author":"Bertsekas D. P.","year":"1995"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.1137\/S1052623494268522"},{"key":"B9","volume-title":"Neuro-dynamic programming","author":"Bertsekas D. P.","year":"1996"},{"key":"B10","first-page":"115","volume-title":"Dynamic programming and its applications","author":"Bertsekas D. P.","year":"1978"},{"key":"B11","volume-title":"Introduction to the mathematical and statistical foundations of econometrics","author":"Bierens H. J.","year":"2005"},{"key":"B12","volume-title":"Model predictive control","author":"Camacho E. F.","year":"2004"},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.1145\/1284680.1284681"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1016\/j.jprocont.2013.12.017"},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4757-3828-5"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511801389"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1016\/j.dcn.2010.12.001"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1007\/s102080010030"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.1007\/978-94-009-4828-0"},{"key":"B20","author":"De Palma D.","year":"2016","journal-title":"International Journal of Adaptive Control and Signal Processing"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.1162\/NECO_a_00406"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.1109\/9.362841"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.1007\/s10957-012-0118-2"},{"key":"B24","doi-asserted-by":"publisher","DOI":"10.1007\/s10589-013-9614-z"},{"key":"B25","first-page":"1217","volume-title":"Proceedings of the American Control Conference","author":"Gallieri M.","year":"2012"},{"key":"B26","doi-asserted-by":"publisher","DOI":"10.1109\/ECC.2015.7330911"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.1162\/NECO_a_00686"},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2014.2361866"},{"key":"B29","doi-asserted-by":"publisher","DOI":"10.1162\/NECO_a_00417"},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.1109\/TNSE.2015.2479086"},{"key":"B31","first-page":"746","volume":"46","author":"Gnecco G.","year":"2010","journal-title":"Journal of Optimization Theory and Applications"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2009.06.048"},{"key":"B33","volume-title":"Kalman filtering: Theory and practice using Matlab","author":"Grewal M.","year":"2001"},{"key":"B34","first-page":"29","volume-title":"Proceedings of the 4 IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing","author":"Jay E.","year":"2011"},{"key":"B35","volume":"3","author":"Khan J.","year":"2014","journal-title":"EURASIP Journal on Bioinformatics and Systems Biology"},{"key":"B36","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-0348-8155-5"},{"key":"B37","doi-asserted-by":"publisher","DOI":"10.3182\/20140824-6-ZA-1003.00119"},{"key":"B38","first-page":"852","volume":"13","author":"Li Y.","year":"2014","journal-title":"WSEAS Transactions on Mathematics"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.1177\/0954410015591614"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.1109\/TSP.2009.2022007"},{"key":"B41","volume-title":"Stochastic models, estimation, and control","volume":"3","author":"Maybeck P. S.","year":"1982"},{"key":"B42","doi-asserted-by":"publisher","DOI":"10.1017\/S1365100502010325"},{"key":"B43","doi-asserted-by":"publisher","DOI":"10.1515\/9781400835355"},{"key":"B44","volume-title":"Probability, random variables, and stochastic processes","author":"Papoulis A.","year":"1991"},{"key":"B45","volume-title":"La psychologie de l\u2019intelligence","author":"Piaget J.","year":"1961"},{"key":"B46","doi-asserted-by":"publisher","DOI":"10.1109\/IJCNN.2005.1556088"},{"key":"B47","first-page":"3413","volume":"12","author":"Recht B.","year":"2011","journal-title":"Journal of Machine Learning Research"},{"key":"B48","volume-title":"Real and complex analysis","author":"Rudin W.","year":"1987"},{"key":"B49","volume-title":"Fundamentals of adaptive filtering","author":"Sayed A.","year":"2003"},{"key":"B50","doi-asserted-by":"publisher","DOI":"10.1162\/089976698300017467"},{"key":"B51","doi-asserted-by":"publisher","DOI":"10.1561\/2200000018"},{"key":"B52","doi-asserted-by":"publisher","DOI":"10.1007\/s10208-004-0160-z"},{"key":"B53","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4471-0101-7"},{"key":"B54","doi-asserted-by":"publisher","DOI":"10.1002\/0471722138"},{"key":"B55","first-page":"161","volume-title":"Proceedings of the Yale Workshop on Adaptive and Learning Systems","author":"Sutton R.","year":"1992"},{"key":"B56","volume-title":"Reinforcement learning","author":"Sutton R. S.","year":"1998"},{"key":"B57","doi-asserted-by":"publisher","DOI":"10.1016\/S0893-6080(00)00077-0"},{"key":"B58","doi-asserted-by":"crossref","first-page":"267","DOI":"10.1111\/j.2517-6161.1996.tb02080.x","volume":"58","author":"Tibshirani R.","year":"1996","journal-title":"Journal of the Royal Statistical Society, Series B"},{"key":"B59","volume-title":"Optimal discrete control theory: The rational function structure model","author":"Vu K. M.","year":"2007"},{"key":"B60","doi-asserted-by":"publisher","DOI":"10.1109\/TAES.2012.6178086"},{"key":"B61","doi-asserted-by":"publisher","DOI":"10.1007\/s10208-006-0237-y"},{"key":"B62","first-page":"928","volume-title":"Proceedings of the 20 International Conference on Machine Learning","author":"Zinkevich M.","year":"2003"}],"container-title":["Neural Computation"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mitpressjournals.org\/doi\/pdf\/10.1162\/neco_a_00976","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,6,24]],"date-time":"2024-06-24T13:13:54Z","timestamp":1719234834000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/neco\/article\/29\/8\/2203-2291\/8310"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2017,8]]},"references-count":62,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2017,8]]}},"alternative-id":["10.1162\/neco_a_00976"],"URL":"https:\/\/doi.org\/10.1162\/neco_a_00976","relation":{},"ISSN":["0899-7667","1530-888X"],"issn-type":[{"value":"0899-7667","type":"print"},{"value":"1530-888X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2017,8]]}}}