{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T07:36:35Z","timestamp":1790840195849,"version":"4.1.0"},"reference-count":41,"publisher":"Informa UK Limited","issue":"535","content-domain":{"domain":["www.tandfonline.com"],"crossmark-restriction":true},"short-container-title":["Journal of the American Statistical Association"],"published-print":{"date-parts":[[2021,7,3]]},"DOI":"10.1080\/01621459.2020.1745812","type":"journal-article","created":{"date-parts":[[2020,3,20]],"date-time":"2020-03-20T11:39:49Z","timestamp":1584704389000},"page":"1507-1520","update-policy":"https:\/\/doi.org\/10.1080\/tandf_crossmark_01","source":"Crossref","is-referenced-by-count":11,"title":["Training Neural Networks as Learning Data-adaptive Kernels: Provable Representation and Approximation Benefits"],"prefix":"10.1080","volume":"116","author":[{"given":"Xialiang","family":"Dou","sequence":"first","affiliation":[{"name":"Department of Statistics, University of Chicago"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tengyuan","family":"Liang","sequence":"additional","affiliation":[{"name":"Booth School of Business, University of Chicago"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"301","published-online":{"date-parts":[[2020,4,23]]},"reference":[{"key":"CIT0001","unstructured":"Martin Anthony and Peter\u00a0L Bartlett.\n                      Neural network learning: Theoretical foundations\n                      . cambridge university press, 2009."},{"key":"CIT0002","unstructured":"Francis Bach. Breaking the curse of dimensionality with convex neural networks.\n                      Journal of Machine Learning Research\n                      , 18(19):1\u201353, 2017."},{"key":"CIT0003","doi-asserted-by":"publisher","DOI":"10.1214\/009053607000000631"},{"key":"CIT0004","unstructured":"Peter\u00a0L Bartlett, Philip\u00a0M Long, G\u00e1bor Lugosi, and Alexander Tsigler. Benign overfitting in linear regression.\n                      arXiv preprint arXiv:1906.11300\n                      , 2019."},{"key":"CIT0005","unstructured":"Mikhail Belkin, Daniel Hsu, Siyuan Ma, and Soumik Mandal. Reconciling modern machine learning and the bias-variance trade-off.\n                      arXiv preprint arXiv:1812.11118\n                      , 2018a."},{"key":"CIT0006","unstructured":"Mikhail Belkin, Siyuan Ma, and Soumik Mandal. To understand deep learning we need to understand kernel learning.\n                      arXiv preprint arXiv:1802.01396\n                      , 2018b."},{"key":"CIT0007","unstructured":"Yoshua Bengio, Nicolas\u00a0L Roux, Pascal Vincent, Olivier Delalleau, and Patrice Marcotte. Convex neural networks. In\n                      Advances in neural information processing systems\n                      , pages 123\u2013130, 2006."},{"key":"CIT0008","unstructured":"Bill Casselman. Essays in analysis. 2014."},{"key":"CIT0009","unstructured":"Lenaic Chizat and Francis Bach. A note on lazy training in supervised differentiable programming.\n                      arXiv preprint arXiv:1812.07956\n                      , 2018."},{"key":"CIT0010","unstructured":"Youngmin Cho and Lawrence\u00a0K Saul. Kernel methods for deep learning. In\n                      Advances in neural information processing systems\n                      , pages 342\u2013350, 2009."},{"key":"CIT0011","doi-asserted-by":"publisher","DOI":"10.1007\/BF02551274"},{"key":"CIT0012","unstructured":"Amit Daniely, Roy Frostig, and Yoram Singer. Toward deeper understanding of neural networks: The power of initialization and a dual view on expressivity. In\n                      Advances In Neural Information Processing Systems\n                      , pages 2253\u20132261, 2016."},{"key":"CIT0013","unstructured":"Simon\u00a0S Du, Xiyu Zhai, Barnabas Poczos, and Aarti Singh. Gradient descent provably optimizes over-parameterized neural networks.\n                      arXiv preprint arXiv:1810.02054\n                      , 2018."},{"key":"CIT0014","unstructured":"Max\u00a0H Farrell, Tengyuan Liang, and Sanjog Misra. Deep neural networks for estimation and inference: Application to causal effects and other semiparametric estimands.\n                      arXiv preprint arXiv:1809.09953\n                      , 2018."},{"key":"CIT0015","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1176345782"},{"key":"CIT0016","unstructured":"Behrooz Ghorbani, Song Mei, Theodor Misiakiewicz, and Andrea Montanari. Linearized two-layers neural networks in high dimension.\n                      arXiv preprint arXiv:1904.12191\n                      , 2019."},{"key":"CIT0017","unstructured":"Trevor Hastie, Andrea Montanari, Saharon Rosset, and Ryan\u00a0J Tibshirani. Surprises in high-dimensional ridgeless least squares interpolation.\n                      arXiv preprint arXiv:1903.08560\n                      , 2019."},{"key":"CIT0018","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(89)90020-8"},{"key":"CIT0019","unstructured":"Cong Huang, Gerald\u00a0HL Cheang, and Andrew\u00a0R Barron.\n                      Risk of penalized least squares, greedy selection and l1-penalization for flexible function libraries\n                      . PhD thesis, Yale University, 2008."},{"key":"CIT0020","unstructured":"Arthur Jacot, Franck Gabriel, and Cl\u00e9ment Hongler. Neural tangent kernel: Convergence and generalization in neural networks. In\n                      Advances in neural information processing systems\n                      , pages 8571\u20138580, 2018."},{"key":"CIT0021","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1176348546"},{"key":"CIT0022","doi-asserted-by":"publisher","DOI":"10.1137\/S0036141096303359"},{"key":"CIT0023","unstructured":"Frederic Koehler and Andrej Risteski. Representational power of relu networks and polynomial kernels: Beyond worst-case analysis.\n                      arXiv preprint arXiv:1805.11405\n                      , 2018."},{"key":"CIT0024","unstructured":"Tengyuan Liang and Alexander Rakhlin. Just interpolate: Kernel\u201d ridgeless\u201d regression can generalize.\n                      The Annals of Statistics, to appear\n                      , 2018."},{"key":"CIT0025","unstructured":"Tengyuan Liang, Alexander Rakhlin, and Xiyu Zhai. On the multiple descent of minimum-norm interpolants and restricted lower isometry of kernels. 2020."},{"key":"CIT0026","unstructured":"Siyuan Ma, Raef Bassily, and Mikhail Belkin. The power of interpolation: Understanding the effectiveness of sgd in modern over-parametrized learning.\n                      arXiv preprint arXiv:1712.06559\n                      , 2017."},{"key":"CIT0027","unstructured":"Hartmut Maennel, Olivier Bousquet, and Sylvain Gelly. Gradient descent quantizes relu network features.\n                      arXiv preprint arXiv:1803.08367\n                      , 2018."},{"key":"CIT0028","doi-asserted-by":"crossref","unstructured":"Song Mei, Andrea Montanari, and Phan-Minh Nguyen. A mean field view of the landscape of two-layers neural networks.\n                      arXiv preprint arXiv:1804.06561\n                      , 2018.","DOI":"10.1073\/pnas.1806579115"},{"key":"CIT0029","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1996.8.4.819"},{"key":"CIT0030","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1991.3.2.246"},{"key":"CIT0031","doi-asserted-by":"crossref","unstructured":"Tomaso Poggio, Hrushikesh Mhaskar, Lorenzo Rosasco, Brando Miranda, and Qianli Liao. Why and when can deep-but not shallow-networks avoid the curse of dimensionality: a review.\n                      International Journal of Automation and Computing\n                      , 14(5):503\u2013519, 2017.","DOI":"10.1007\/s11633-017-1054-2"},{"key":"CIT0032","unstructured":"Ali Rahimi and Benjamin Recht. Random features for large-scale kernel machines. In\n                      Advances in neural information processing systems\n                      , pages 1177\u20131184, 2008."},{"key":"CIT0033","unstructured":"Ali Rahimi and Benjamin Recht. Weighted sums of random kitchen sinks: Replacing minimization with randomization in learning. In\n                      Advances in neural information processing systems\n                      , pages 1313\u20131320, 2009."},{"key":"CIT0034","unstructured":"Alexander Rakhlin and Xiyu Zhai. Consistency of interpolation with laplace kernels is a high-dimensional phenomenon.\n                      arXiv preprint arXiv:1812.11167\n                      , 2018."},{"key":"CIT0035","unstructured":"Grant\u00a0M Rotskoff and Eric Vanden-Eijnden. Neural networks as interacting particle systems: Asymptotic convexity of the loss landscape and universal scaling of the approximation error.\n                      arXiv preprint arXiv:1805.00915\n                      , 2018."},{"key":"CIT0036","unstructured":"Alessandro Rudi and Lorenzo Rosasco. Generalization properties of learning with random features. In\n                      Advances in Neural Information Processing Systems\n                      , pages 3215\u20133225, 2017."},{"key":"CIT0037","doi-asserted-by":"publisher","DOI":"10.1016\/j.spa.2019.06.003"},{"key":"CIT0038","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1176345206"},{"key":"CIT0039","unstructured":"Vladimir Vapnik.\n                      Statistical learning theory. 1998\n                      , volume\u00a03. Wiley, New York, 1998."},{"key":"CIT0040","doi-asserted-by":"publisher","DOI":"10.1137\/1.9781611970128"},{"key":"CIT0041","unstructured":"Chiyuan Zhang, Samy Bengio, Moritz Hardt, Benjamin Recht, and Oriol Vinyals. Understanding deep learning requires rethinking generalization.\n                      arXiv preprint arXiv:1611.03530\n                      , 2016."}],"container-title":["Journal of the American Statistical Association"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.tandfonline.com\/doi\/pdf\/10.1080\/01621459.2020.1745812","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,8,1]],"date-time":"2023-08-01T11:58:34Z","timestamp":1690891114000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.tandfonline.com\/doi\/full\/10.1080\/01621459.2020.1745812"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,4,23]]},"references-count":41,"journal-issue":{"issue":"535","published-print":{"date-parts":[[2021,7,3]]}},"alternative-id":["10.1080\/01621459.2020.1745812"],"URL":"https:\/\/doi.org\/10.1080\/01621459.2020.1745812","relation":{},"ISSN":["0162-1459","1537-274X"],"issn-type":[{"value":"0162-1459","type":"print"},{"value":"1537-274X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020,4,23]]},"assertion":[{"value":"The publishing and review policy for this title is described in its Aims & Scope.","order":1,"name":"peerreview_statement","label":"Peer Review Statement"},{"value":"http:\/\/www.tandfonline.com\/action\/journalInformation?show=aimsScope&journalCode=uasa20","URL":"http:\/\/www.tandfonline.com\/action\/journalInformation?show=aimsScope&journalCode=uasa20","order":2,"name":"aims_and_scope_url","label":"Aim & Scope"},{"value":"2019-01-31","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2019-08-06","order":1,"name":"revised","label":"Revised","group":{"name":"publication_history","label":"Publication History"}},{"value":"2020-03-12","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2020-04-23","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}