{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,27]],"date-time":"2026-10-27T08:18:16Z","timestamp":1793089096155,"version":"4.5.0"},"reference-count":65,"publisher":"Institute of Mathematical Statistics","issue":"4","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Ann. Statist."],"published-print":{"date-parts":[[2025,8,1]]},"abstract":"<jats:p>Vintage factor analysis is one important type of factor analysis that aims to first find a low-dimensional representation of the original data, and then to seek a rotation such that the rotated low-dimensional representation is scientifically meaningful. The most widely used vintage factor analysis is the Principal Component Analysis (PCA) followed by the varimax rotation. Despite its popularity, little theoretical guarantee can be provided to date mainly because varimax rotation requires to solve a nonconvex optimization over the set of orthogonal matrices. In this paper, we propose a deflation varimax procedure that solves each row of an orthogonal matrix sequentially. In addition to its net computational gain and flexibility, we are able to fully establish theoretical guarantees for the proposed procedure in a broader context. Adopting this new deflation varimax as the second step after PCA, we further analyze this two-step procedure under a general class of factor models. Our results show that it estimates the factor loading matrix in the minimax optimal rate when the signal-to-noise-ratio (SNR) is moderate or large. In the low SNR regime, we offer possible improvement over using PCA and the deflation varimax when the additive noise under the factor model is structured. The modified procedure is shown to be minimax optimal in all SNR regimes. Our theory is valid for finite sample and allows the number of the latent factors to grow with the sample size as well as the ambient dimension to grow with, or even exceed, the sample size. Extensive simulation and real data analysis further corroborate our theoretical findings.<\/jats:p>","DOI":"10.1214\/25-aos2518","type":"journal-article","created":{"date-parts":[[2025,9,12]],"date-time":"2025-09-12T05:23:53Z","timestamp":1757654633000},"page":"1639-1666","source":"Crossref","is-referenced-by-count":3,"title":["Optimal vintage factor analysis with deflation varimax"],"prefix":"10.1214","volume":"53","author":[{"given":"Xin","family":"Bing","sequence":"first","affiliation":[{"name":"Department of Statistical Sciences, University of Toronto"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xin","family":"He","sequence":"additional","affiliation":[{"name":"School of Statistics and Data Science, Shanghai University of Finance and Economics"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dian","family":"Jin","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Rutgers University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuqian","family":"Zhang","sequence":"additional","affiliation":[{"name":"Department of Electrical and Computer Engineering, Rutgers University"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"108","reference":[{"key":"12","doi-asserted-by":"publisher","unstructured":"<small>BING, X.<\/small>, <small>BUNEA, F.<\/small>, <small>NING, Y.<\/small> and <small>WEGKAMP, M.<\/small> (2020). Adaptive estimation in structured factor models with applications to overlapping clustering. <i>Ann. Statist.<\/i> <b>48<\/b> 2055\u20132081.","DOI":"10.1214\/19-AOS1877"},{"key":"19","doi-asserted-by":"publisher","unstructured":"<small>Chandrasekaran, V.<\/small>, <small>Parrilo, P. A.<\/small> and <small>Willsky, A. S.<\/small> (2012). Latent variable graphical model selection via convex optimization. <i>Ann. Statist.<\/i> <b>40<\/b> 1935\u20131967.","DOI":"10.1214\/11-AOS949"},{"key":"40","doi-asserted-by":"crossref","unstructured":"<small>LAWLEY, D. N.<\/small> and <small>MAXWELL, A. E.<\/small> (1962). Factor analysis as a statistical method. <i>J. R. Stat. Soc<\/i>., <i>Ser. D, Stat.<\/i> <b>12<\/b> 209\u2013229.","DOI":"10.2307\/2986915"},{"key":"59","doi-asserted-by":"publisher","unstructured":"<small>VU, V. Q.<\/small> and <small>LEI, J.<\/small> (2013). Minimax sparse principal subspace estimation in high dimensions. <i>Ann. Statist.<\/i> <b>41<\/b> 2905\u20132947.","DOI":"10.1214\/13-AOS1151"},{"key":"15","doi-asserted-by":"publisher","unstructured":"<small>BUJA, A.<\/small> and <small>EYUBOGLU, N.<\/small> (1992). Remarks on parallel analysis. <i>Multivar. Behav. Res.<\/i> <b>27<\/b> 509\u2013540.","DOI":"10.1207\/s15327906mbr2704_2"},{"key":"29","doi-asserted-by":"publisher","unstructured":"<small>HSU, D.<\/small>, <small>KAKADE, S. M.<\/small> and <small>ZHANG, T.<\/small> (2011). Robust matrix decomposition with sparse corruptions. <i>IEEE Trans. Inf. Theory<\/i> <b>57<\/b> 7221\u20137234.","DOI":"10.1109\/TIT.2011.2158250"},{"key":"33","doi-asserted-by":"crossref","unstructured":"<small>HYV\u00c4RINEN, A.<\/small> and <small>OJA, E.<\/small> (2000). Independent component analysis: Algorithms and applications. <i>Neural Netw.<\/i> <b>13<\/b> 411\u2013430.","DOI":"10.1016\/S0893-6080(00)00026-5"},{"key":"64","doi-asserted-by":"publisher","unstructured":"<small>ZHANG, A. R.<\/small>, <small>CAI, T. T.<\/small> and <small>WU, Y.<\/small> (2022). Heteroskedastic PCA: Algorithm, optimality, and applications. <i>Ann. Statist.<\/i> <b>50<\/b> 53\u201380.","DOI":"10.1214\/21-aos2074"},{"key":"36","doi-asserted-by":"crossref","unstructured":"<small>Kaiser, H. F.<\/small> (1958). The varimax criterion for analytic rotation in factor analysis. <i>Psychometrika<\/i> <b>23<\/b> 187\u2013200.","DOI":"10.1007\/BF02289233"},{"key":"41","doi-asserted-by":"crossref","unstructured":"<small>LeCun, Y.<\/small>, <small>Bottou, L.<\/small>, <small>Bengio, Y.<\/small> and <small>Haffner, P.<\/small> (1998). Gradient-based learning applied to document recognition. <i>Proc. IEEE<\/i> <b>86<\/b> 2278\u20132324.","DOI":"10.1109\/5.726791"},{"key":"1","doi-asserted-by":"publisher","unstructured":"<small>Ahn, S. C.<\/small> and <small>Horenstein, A. R.<\/small> (2013). Eigenvalue ratio test for the number of factors. <i>Econometrica<\/i> <b>81<\/b> 1203\u20131227.","DOI":"10.3982\/ECTA8968"},{"key":"2","unstructured":"<small>Anandkumar, A.<\/small>, <small>Ge, R.<\/small>, <small>Hsu, D.<\/small>, <small>Kakade, S. M.<\/small> and <small>Telgarsky, M.<\/small> (2014). Tensor decompositions for learning latent variable models. <i>J. Mach. Learn. Res.<\/i> <b>15<\/b> 2773\u20132832."},{"key":"8","doi-asserted-by":"publisher","unstructured":"<small>Bai, J.<\/small> and <small>Li, K.<\/small> (2012). Statistical analysis of factor models of high dimension. <i>Ann. Statist.<\/i> <b>40<\/b> 436\u2013465.","DOI":"10.1214\/11-AOS966"},{"key":"9","doi-asserted-by":"publisher","unstructured":"<small>Bai, J.<\/small> and <small>Ng, S.<\/small> (2002). Determining the number of factors in approximate factor models. <i>Econometrica<\/i> <b>70<\/b> 191\u2013221.","DOI":"10.1111\/1468-0262.00273"},{"key":"17","doi-asserted-by":"publisher","unstructured":"<small>Cand\u00e8s, E. J.<\/small>, <small>Li, X.<\/small>, <small>Ma, Y.<\/small> and <small>Wright, J.<\/small> (2011). Robust principal component analysis? <i>J. ACM<\/i> <b>58<\/b> Art. 11, 37.","DOI":"10.1145\/1970392.1970395"},{"key":"21","doi-asserted-by":"publisher","unstructured":"<small>Fan, J.<\/small>, <small>Liao, Y.<\/small> and <small>Mincheva, M.<\/small> (2013). Large covariance estimation by thresholding principal orthogonal complements. <i>J. R. Stat. Soc. Ser. B. Stat. Methodol.<\/i> <b>75<\/b> 603\u2013680.","DOI":"10.1111\/rssb.12016"},{"key":"28","doi-asserted-by":"crossref","unstructured":"<small>Hotelling, H.<\/small> (1933). Analysis of a complex of statistical variables into principal components. <i>J. Educ. Psychol.<\/i> <b>24<\/b> 417.","DOI":"10.1037\/h0071325"},{"key":"39","doi-asserted-by":"publisher","unstructured":"<small>Lam, C.<\/small> and <small>Yao, Q.<\/small> (2012). Factor modeling for high-dimensional time series: Inference for the number of factors. <i>Ann. Statist.<\/i> <b>40<\/b> 694\u2013726.","DOI":"10.1214\/12-AOS970"},{"key":"7","doi-asserted-by":"publisher","unstructured":"<small>Bai, J.<\/small> (2003). Inferential theory for factor models of large dimensions. <i>Econometrica<\/i> <b>71<\/b> 135\u2013171.","DOI":"10.1111\/1468-0262.00392"},{"key":"3","unstructured":"<small>ANANDKUMAR, A.<\/small>, <small>GE, R.<\/small> and <small>JANZAMIN, M.<\/small> (2014). Guaranteed non-orthogonal tensor decomposition via alternating Rank-1 updates. Preprint. Available at arXiv:<a href=\"https:\/\/arxiv.org\/abs\/1402.5180\">1402.5180<\/a>."},{"key":"4","unstructured":"<small>ANDERSON, T. W.<\/small> and <small>RUBIN, H.<\/small> (1956). Statistical inference in factor analysis. In <i>Proceedings of the Third Berkeley Symposium on Mathematical Statistics and Probability<\/i>, 1954<i>\u2013<\/i>1955, <i>Vol. V<\/i> 111\u2013150. Univ. California Press, Berkeley, CA."},{"key":"5","unstructured":"<small>ARORA, S.<\/small>, <small>GE, R.<\/small>, <small>MOITRA, A.<\/small> and <small>SACHDEVA, S.<\/small> (2012). Provable ICA with unknown Gaussian noise, with implications for Gaussian mixtures and autoencoders. <i>Adv. Neural Inf. Process. Syst.<\/i> <b>25<\/b>."},{"key":"6","doi-asserted-by":"publisher","unstructured":"<small>AUDDY, A.<\/small> and <small>YUAN, M.<\/small> (2025). Large-dimensional independent component analysis: Statistical optimality and computational tractability. <i>Ann. Statist.<\/i> <b>53<\/b>.","DOI":"10.1214\/24-aos2419"},{"key":"10","doi-asserted-by":"crossref","unstructured":"<small>BAI, J.<\/small>, <small>NG, S.<\/small> et al. (2008). Large dimensional factor analysis. <i>Found Trends Econom.<\/i> <b>3<\/b> 89\u2013163.","DOI":"10.1561\/0800000002"},{"key":"11","unstructured":"<small>BAI, Y.<\/small>, <small>JIANG, Q.<\/small> and <small>SUN, J.<\/small> (2018). Subgradient descent learns orthogonal dictionaries. Preprint. Available at arXiv:<a href=\"https:\/\/arxiv.org\/abs\/1810.10702\">1810.10702<\/a>."},{"key":"13","unstructured":"<small>BING, X.<\/small>, <small>BUNEA, F.<\/small>, <small>STRIMAS-MACKEY, S.<\/small> and <small>WEGKAMP, M.<\/small> (2021). Prediction under latent factor regression: Adaptive PCR, interpolating predictors and beyond. <i>J. Mach. Learn. Res.<\/i> <b>22<\/b> Paper No. 177, 50."},{"key":"14","doi-asserted-by":"publisher","unstructured":"<small>BING, X.<\/small>, <small>BUNEA, F.<\/small> and <small>WEGKAMP, M.<\/small> (2023). Detecting approximate replicate components of a high-dimensional random vector with latent structure. <i>Bernoulli<\/i> <b>29<\/b> 1368\u20131391.","DOI":"10.3150\/22-bej1502"},{"key":"16","doi-asserted-by":"publisher","unstructured":"<small>BUNEA, F.<\/small>, <small>GIRAUD, C.<\/small>, <small>LUO, X.<\/small>, <small>ROYER, M.<\/small> and <small>VERZELEN, N.<\/small> (2020). Model assisted variable clustering: Minimax-optimal recovery and algorithms. <i>Ann. Statist.<\/i> <b>48<\/b> 111\u2013137.","DOI":"10.1214\/18-AOS1794"},{"key":"18","doi-asserted-by":"crossref","unstructured":"<small>CARROLL, J. B.<\/small> (1953). An analytical solution for approximating simple structure in factor analysis. <i>Psychometrika<\/i> <b>18<\/b> 23\u201338.","DOI":"10.1007\/BF02289025"},{"key":"20","unstructured":"<small>DOUGLAS, S. C.<\/small> (2003). On the convergence behavior of the fastICA algorithm. In 4<i>th International Symposium on Independent Component Analysis and Blind Signal Separation<\/i>."},{"key":"22","doi-asserted-by":"crossref","unstructured":"<small>FERGUSON, G. A.<\/small> (1954). On learning and human ability. <i>Can. J. Psychol.<\/i> <b>8<\/b> 95.","DOI":"10.1037\/h0083598"},{"key":"23","doi-asserted-by":"crossref","unstructured":"<small>FIELD, D. J.<\/small> (1994). What is the goal of sensory coding? <i>Neural Comput.<\/i> <b>6<\/b> 559\u2013601.","DOI":"10.1162\/neco.1994.6.4.559"},{"key":"24","doi-asserted-by":"crossref","unstructured":"<small>FIORI, A. M.<\/small> and <small>ZENGA, M.<\/small> (2009). Karl Pearson and the origin of kurtosis. <i>Int. Stat. Rev.<\/i> <b>77<\/b> 40\u201350.","DOI":"10.1111\/j.1751-5823.2009.00076.x"},{"key":"25","doi-asserted-by":"crossref","unstructured":"<small>BING, X.<\/small>, <small>HE, X.<\/small>, <small>JIN, D.<\/small> and <small>ZHANG, Y.<\/small> (2025). Supplement to \u201cOptimal vintage factor analysis with deflation varimax.\u201d <a href=\"https:\/\/doi.org\/10.1214\/25-AOS2518SUPPA\">https:\/\/doi.org\/10.1214\/25-AOS2518SUPPA<\/a>, <a href=\"https:\/\/doi.org\/10.1214\/25-AOS2518SUPPB\">https:\/\/doi.org\/10.1214\/25-AOS2518SUPPB<\/a>","DOI":"10.1214\/25-AOS2518SUPPB"},{"key":"26","unstructured":"<small>HOLZINGER, K. J.<\/small> and <small>HARMAN, H. H.<\/small> (1941). <i>Factor Analysis. A Synthesis of Factorial Methods<\/i>. Univ. Chicago Press, Chicago, IL."},{"key":"27","unstructured":"<small>HORST, P.<\/small> (1965). <i>Factor Analysis of Data Matrices<\/i>. Holt, Rinehart &amp; Winston, New York\u2013Toronto\u2013London."},{"key":"30","unstructured":"<small>ANANDKUMAR, A.<\/small>, <small>GE, R.<\/small> and <small>JANZAMIN, M.<\/small> (2015). Learning overcomplete latent variable models through tensor methods. In <i>Conference on Learning Theory<\/i> 36\u2013112. PMLR."},{"key":"31","doi-asserted-by":"publisher","unstructured":"<small>HYVARINEN, A.<\/small> (1997). One-unit contrast functions for independent component analysis: A statistical analysis. In <i>Neural Networks for Signal Processing VII. Proceedings of the<\/i> 1997 <i>IEEE Signal Processing Society Workshop<\/i> 388\u2013397.","DOI":"10.1109\/NNSP.1997.622420"},{"key":"32","doi-asserted-by":"crossref","unstructured":"<small>HYV\u00c4RINEN, A.<\/small> and <small>OJA, E.<\/small> (1997). A fast fixed-point algorithm for independent component analysis. <i>Neural Comput.<\/i> <b>9<\/b> 1483\u20131492.","DOI":"10.1162\/neco.1997.9.7.1483"},{"key":"34","doi-asserted-by":"publisher","unstructured":"<small>JIN, D.<\/small>, <small>BING, X.<\/small> and <small>ZHANG, Y.<\/small> (2023). Unique sparse decomposition of low rank matrices. <i>IEEE Trans. Inf. Theory<\/i> <b>69<\/b> 2452\u20132484.","DOI":"10.1109\/tit.2022.3223230"},{"key":"35","doi-asserted-by":"publisher","unstructured":"<small>JUNG, A.<\/small>, <small>ELDAR, Y. C.<\/small> and <small>G\u00d6RTZ, N.<\/small> (2016). On the minimax risk of dictionary learning. <i>IEEE Trans. Inf. Theory<\/i> <b>62<\/b> 1501\u20131515.","DOI":"10.1109\/TIT.2016.2517006"},{"key":"37","doi-asserted-by":"publisher","unstructured":"<small>KAISER, H. F.<\/small> (1959). Computer program for varimax rotation in factor analysis. <i>Educ. Psychol. Meas.<\/i> <b>19<\/b> 413\u2013420.","DOI":"10.1177\/001316445901900314"},{"key":"38","doi-asserted-by":"crossref","unstructured":"<small>KAISER, H. F.<\/small> and <small>RICE, J.<\/small> (1974). Little jiffy, mark IV. <i>Educ. Psychol. Meas.<\/i> <b>34<\/b> 111\u2013117.","DOI":"10.1177\/001316447403400115"},{"key":"42","unstructured":"<small>MULAIK, S. A.<\/small> (2010). <i>Foundations of Factor Analysis<\/i>, 2nd ed. <i>Statistics in the Social and Behavioral Sciences Series<\/i>. CRC Press, Boca Raton, FL."},{"key":"43","doi-asserted-by":"publisher","unstructured":"<small>NEUHAUS, J. O.<\/small> and <small>WRIGLEY, C.<\/small> (1954). The Quartimax Method. <i>Br. J. Math. Stat. Psychol.<\/i> <b>7<\/b> 81\u201391.","DOI":"10.1111\/j.2044-8317.1954.tb00147.x"},{"key":"44","doi-asserted-by":"publisher","unstructured":"<small>NORDHAUSEN, K.<\/small> and <small>OJA, H.<\/small> (2018). Independent component analysis: A statistical perspective. <i>Wiley Interdiscip. Rev<\/i>.: <i>Comput. Stat.<\/i> <b>10<\/b> e1440, 23.","DOI":"10.1002\/wics.1440"},{"key":"45","doi-asserted-by":"publisher","unstructured":"<small>OJA, E.<\/small> and <small>YUAN, Z.<\/small> (2006). The FastICA algorithm revisited: Convergence analysis. <i>IEEE Trans. Neural Netw.<\/i> <b>17<\/b> 1370\u20131381.","DOI":"10.1109\/TNN.2006.880980"},{"key":"46","doi-asserted-by":"publisher","unstructured":"<small>OLLILA, E.<\/small> (2010). The deflation-based FastICA estimator: Statistical analysis revisited. <i>IEEE Trans. Signal Process.<\/i> <b>58<\/b> 1527\u20131541.","DOI":"10.1109\/TSP.2009.2036072"},{"key":"47","doi-asserted-by":"crossref","unstructured":"<small>OLSHAUSEN, B. A.<\/small> and <small>FIELD, D. J.<\/small> (1996). Emergence of simple-cell receptive field properties by learning a sparse code for natural images. <i>Nature<\/i> <b>381<\/b> 607\u2013609.","DOI":"10.1038\/381607a0"},{"key":"48","unstructured":"<small>QU, Q.<\/small>, <small>ZHAI, Y.<\/small>, <small>LI, X.<\/small>, <small>ZHANG, Y.<\/small> and <small>ZHU, Z.<\/small> (2020). Geometric analysis of nonconvex optimization landscapes for overcomplete learning. In <i>International Conference on Learning Representations<\/i>."},{"key":"49","doi-asserted-by":"publisher","unstructured":"<small>ROHE, K.<\/small> and <small>ZENG, M.<\/small> (2023). Vintage factor analysis with Varimax performs statistical inference. <i>J. R. Stat. Soc. Ser. B. Stat. Methodol.<\/i> <b>85<\/b> 1037\u20131060.","DOI":"10.1093\/jrsssb\/qkad029"},{"key":"50","doi-asserted-by":"crossref","unstructured":"<small>SAUNDERS, D. R.<\/small> (1953). An analytic method for rotation to orthogonal simple structure. <i>ETS Research Bulletin Series<\/i> <b>1953<\/b> i\u201328.","DOI":"10.1002\/j.2333-8504.1953.tb00890.x"},{"key":"51","doi-asserted-by":"publisher","unstructured":"<small>SHERIN, R. J.<\/small> (1966). A matrix formulation of Kaiser\u2019s varimax criterion. <i>Psychometrika<\/i> <b>31<\/b> 535\u2013538.","DOI":"10.1007\/BF02289522"},{"key":"52","unstructured":"<small>SPIELMAN, D. A.<\/small>, <small>WANG, H.<\/small> and <small>WRIGHT, J.<\/small> (2012). Exact recovery of sparsely-used dictionaries. In <i>Conference on Learning Theory<\/i>."},{"key":"53","doi-asserted-by":"publisher","unstructured":"<small>STOCK, J. H.<\/small> and <small>WATSON, M. W.<\/small> (2012). Generalized shrinkage methods for forecasting using many predictors. <i>J. Bus. Econom. Statist.<\/i> <b>30<\/b> 481\u2013493.","DOI":"10.1080\/07350015.2012.715956"},{"key":"54","doi-asserted-by":"publisher","unstructured":"<small>SUN, J.<\/small>, <small>QU, Q.<\/small> and <small>WRIGHT, J.<\/small> (2017). Complete dictionary recovery over the sphere I: Overview and the geometric picture. <i>IEEE Trans. Inf. Theory<\/i> <b>63<\/b> 853\u2013884.","DOI":"10.1109\/TIT.2016.2632162"},{"key":"55","doi-asserted-by":"crossref","unstructured":"<small>THURSTONE, L. L.<\/small> (1935). <i>The Vectors of Mind<\/i>: <i>Multiple-Factor Analysis for the Isolation of Primary Traits<\/i>. Univ. Chicago Press, Chicago.","DOI":"10.1037\/10018-000"},{"key":"56","doi-asserted-by":"crossref","unstructured":"<small>TICHAVSK\u00dd, P.<\/small>, <small>KOLDOVSK\u00dd, Z.<\/small> and <small>OJA, E.<\/small> (2006). Performance analysis of the FastICA algorithm and Cram\u00e9r\u2013Rao bounds for linear independent component analysis. <i>IEEE Trans. Signal Process.<\/i> <b>54<\/b> 1189\u20131203.","DOI":"10.1109\/TSP.2006.870561"},{"key":"57","unstructured":"<small>THURSTONE, L. L.<\/small> (1947). <i>Multiple Factor Analysis<\/i>. Univ. Chicago Press, Chicago."},{"key":"60","doi-asserted-by":"publisher","unstructured":"<small>WEI, T.<\/small> (2015). A convergence and asymptotic analysis of the generalized symmetric FastICA algorithm. <i>IEEE Trans. Signal Process.<\/i> <b>63<\/b> 6445\u20136458.","DOI":"10.1109\/TSP.2015.2468686"},{"key":"61","doi-asserted-by":"publisher","unstructured":"<small>XUE, Y.<\/small>, <small>LAU, V. K. N.<\/small> and <small>CAI, S.<\/small> (2021). Efficient sparse coding using hierarchical Riemannian pursuit. <i>IEEE Trans. Signal Process.<\/i> <b>69<\/b> 4069\u20134084.","DOI":"10.1109\/TSP.2021.3093769"},{"key":"62","unstructured":"<small>ZHAI, Y.<\/small>, <small>MEHTA, H.<\/small>, <small>ZHOU, Z.<\/small> and <small>MA, Y.<\/small> (2020a). Understanding l4-based dictionary learning: Interpretation, stability, and robustness. In <i>International Conference on Learning Representations<\/i>."},{"key":"63","unstructured":"<small>ZHAI, Y.<\/small>, <small>YANG, Z.<\/small>, <small>LIAO, Z.<\/small>, <small>WRIGHT, J.<\/small> and <small>MA, Y.<\/small> (2020b). Complete dictionary learning via <math xmlns=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" overflow=\"scroll\" id=\"eq_1\"><msup><mrow><mi>\u2113<\/mi><\/mrow><mrow><mn>4<\/mn><\/mrow><\/msup><\/math>-norm maximization over the orthogonal group. <i>J. Mach. Learn. Res.<\/i> <b>21<\/b> Paper No. 165, 68."},{"key":"65","doi-asserted-by":"publisher","unstructured":"<small>ZHANG, Y.<\/small>, <small>KUO, H.-W.<\/small> and <small>WRIGHT, J.<\/small> (2020). Structured local optima in sparse blind deconvolution. <i>IEEE Trans. Inf. Theory<\/i> <b>66<\/b> 419\u2013452.","DOI":"10.1109\/TIT.2019.2940657"},{"key":"58","unstructured":"<small>VEMPALA, S. S.<\/small> and <small>XIAO, Y.<\/small> (2011). Structure from local optima: Learning subspace juntas via higher order PCA. Preprint. Available at arXiv:<a href=\"https:\/\/arxiv.org\/abs\/1108.3329\">1108.3329<\/a>."}],"container-title":["The Annals of Statistics"],"original-title":[],"deposited":{"date-parts":[[2026,10,7]],"date-time":"2026-10-07T16:37:58Z","timestamp":1791391078000},"score":1,"resource":{"primary":{"URL":"https:\/\/projecteuclid.org\/journals\/annals-of-statistics\/volume-53\/issue-4\/Optimal-vintage-factor-analysis-with-deflation-varimax\/10.1214\/25-AOS2518.full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8,1]]},"references-count":65,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2025,8,1]]}},"URL":"https:\/\/doi.org\/10.1214\/25-aos2518","relation":{},"ISSN":["0090-5364","2168-8966"],"issn-type":[{"value":"0090-5364","type":"print"},{"value":"2168-8966","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,8,1]]}}}