{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T22:13:44Z","timestamp":1777587224446,"version":"3.51.4"},"reference-count":38,"publisher":"American Mathematical Society (AMS)","issue":"347","license":[{"start":{"date-parts":[[2024,10,27]],"date-time":"2024-10-27T00:00:00Z","timestamp":1729987200000},"content-version":"am","delay-in-days":366,"URL":"https:\/\/www.ams.org\/publications\/copyright-and-permissions"}],"funder":[{"DOI":"10.13039\/501100012456","name":"National Social Science Fund of China","doi-asserted-by":"publisher","award":["12201022"],"award-info":[{"award-number":["12201022"]}],"id":[{"id":"10.13039\/501100012456","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Math. Comp."],"abstract":"

\n In this paper, we focus on the nonlinear least squares problem:\n \n \n \n \n \n min<\/mml:mo>\n \n \n x<\/mml:mi>\n <\/mml:mrow>\n \n \u2208\n \n <\/mml:mo>\n \n \n H<\/mml:mi>\n <\/mml:mrow>\n d<\/mml:mi>\n <\/mml:msup>\n <\/mml:mrow>\n <\/mml:munder>\n \n \u2016\n \n <\/mml:mo>\n \n |\n \n <\/mml:mo>\n A<\/mml:mi>\n \n x<\/mml:mi>\n <\/mml:mrow>\n \n |\n \n <\/mml:mo>\n \n \u2212\n \n <\/mml:mo>\n \n b<\/mml:mi>\n <\/mml:mrow>\n \n \u2016\n \n <\/mml:mo>\n <\/mml:mrow>\n \\min _{{\\boldsymbol {x}}\\in \\mathbb {H}^d}\\|\\lvert A{\\boldsymbol {x}}\\rvert -{\\boldsymbol {b}}\\|<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n where\n \n \n \n \n A<\/mml:mi>\n \n \u2208\n \n <\/mml:mo>\n \n \n H<\/mml:mi>\n <\/mml:mrow>\n \n m<\/mml:mi>\n \n \u00d7\n \n <\/mml:mo>\n d<\/mml:mi>\n <\/mml:mrow>\n <\/mml:msup>\n <\/mml:mrow>\n A\\in \\mathbb {H}^{m\\times d}<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n ,\n \n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n \n \u2208\n \n <\/mml:mo>\n \n \n R<\/mml:mi>\n <\/mml:mrow>\n m<\/mml:mi>\n <\/mml:msup>\n <\/mml:mrow>\n {\\boldsymbol {b}}\\in \\mathbb {R}^m<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n with\n \n \n \n \n \n H<\/mml:mi>\n <\/mml:mrow>\n \n \u2208\n \n <\/mml:mo>\n \n {<\/mml:mo>\n \n R<\/mml:mi>\n <\/mml:mrow>\n ,<\/mml:mo>\n \n C<\/mml:mi>\n <\/mml:mrow>\n }<\/mml:mo>\n <\/mml:mrow>\n <\/mml:mrow>\n \\mathbb {H}\\in \\left \\{\\mathbb {R},\\mathbb {C}\\right \\}<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n and consider the uniqueness and stability of solutions. This problem arises in applications such as phase retrieval and absolute value rectification neural networks. While several results have been developed to characterize the uniqueness and stability of solutions when\n \n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n =<\/mml:mo>\n \n |\n \n <\/mml:mo>\n A<\/mml:mi>\n \n \n x<\/mml:mi>\n <\/mml:mrow>\n 0<\/mml:mn>\n <\/mml:msub>\n \n |\n \n <\/mml:mo>\n <\/mml:mrow>\n {\\boldsymbol {b}}=\\lvert A{\\boldsymbol {x}}_0\\rvert<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n for some\n \n \n \n \n \n \n x<\/mml:mi>\n <\/mml:mrow>\n 0<\/mml:mn>\n <\/mml:msub>\n \n \u2208\n \n <\/mml:mo>\n \n \n H<\/mml:mi>\n <\/mml:mrow>\n d<\/mml:mi>\n <\/mml:msup>\n <\/mml:mrow>\n {\\boldsymbol {x}}_0\\in \\mathbb {H}^d<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n , no existing results address the case where\n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n {\\boldsymbol {b}}<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n is arbitrary. In this paper, we investigate the uniqueness and stability of solutions for the more general case where\n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n {\\boldsymbol {b}}<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n is not necessarily equal to\n \n \n \n \n \n |\n \n <\/mml:mo>\n A<\/mml:mi>\n \n \n x<\/mml:mi>\n <\/mml:mrow>\n 0<\/mml:mn>\n <\/mml:msub>\n \n |\n \n <\/mml:mo>\n <\/mml:mrow>\n \\lvert A{\\boldsymbol {x}}_0\\rvert<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n for any\n \n \n \n \n \n \n x<\/mml:mi>\n <\/mml:mrow>\n 0<\/mml:mn>\n <\/mml:msub>\n \n \u2208\n \n <\/mml:mo>\n \n \n H<\/mml:mi>\n <\/mml:mrow>\n d<\/mml:mi>\n <\/mml:msup>\n <\/mml:mrow>\n {\\boldsymbol {x}}_0\\in \\mathbb {H}^d<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n . We prove that for any matrix\n \n \n \n \n A<\/mml:mi>\n \n \u2208\n \n <\/mml:mo>\n \n \n H<\/mml:mi>\n <\/mml:mrow>\n \n m<\/mml:mi>\n \n \u00d7\n \n <\/mml:mo>\n d<\/mml:mi>\n <\/mml:mrow>\n <\/mml:msup>\n <\/mml:mrow>\n A\\in \\mathbb {H}^{m\\times d}<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n , there is always a vector\n \n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n \n \u2208\n \n <\/mml:mo>\n \n \n R<\/mml:mi>\n <\/mml:mrow>\n m<\/mml:mi>\n <\/mml:msup>\n <\/mml:mrow>\n {\\boldsymbol {b}}\\in \\mathbb {R}^m<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n for which the solution to the nonlinear least squares problem is not unique. However, we show that such \u201cbad\u201d vectors\n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n {\\boldsymbol {b}}<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n are negligible in practice; specifically, if\n \n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n \n \u2208\n \n <\/mml:mo>\n \n \n R<\/mml:mi>\n <\/mml:mrow>\n \n m<\/mml:mi>\n <\/mml:msubsup>\n <\/mml:mrow>\n {\\boldsymbol {b}}\\in \\mathbb {R}_{ }^m<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n does not lie in some measure zero set, then the solution is unique. Furthermore, we establish certain conditions under which the solution is guaranteed to be unique. Regarding the stability of solutions, we prove that the solution is not uniformly stable. However, if we restrict the vectors\n \n \n \n \n b<\/mml:mi>\n <\/mml:mrow>\n {\\boldsymbol {b}}<\/mml:annotation>\n <\/mml:semantics>\n <\/mml:math>\n <\/inline-formula>\n to a convex set where the solution to the least squares problem is unique, then the solution becomes stable. To the best of our knowledge, our results represent the first theoretical results of the uniqueness and stability of solutions for the nonlinear least squares problem.\n <\/p>","DOI":"10.1090\/mcom\/3918","type":"journal-article","created":{"date-parts":[[2023,10,11]],"date-time":"2023-10-11T10:21:50Z","timestamp":1697019710000},"page":"1247-1264","source":"Crossref","is-referenced-by-count":2,"title":["Uniqueness and stability for the solution of a nonlinear least squares problem"],"prefix":"10.1090","volume":"93","author":[{"given":"Meng","family":"Huang","sequence":"first","affiliation":[]},{"given":"Zhiqiang","family":"Xu","sequence":"additional","affiliation":[]}],"member":"14","published-online":{"date-parts":[[2023,10,27]]},"reference":[{"issue":"3","key":"1","doi-asserted-by":"publisher","first-page":"1895","DOI":"10.1137\/16M1071481","article-title":"Phase retrieval in the general setting of continuous frames for Banach spaces","volume":"49","author":"Alaifari, Rima","year":"2017","journal-title":"SIAM J. Math. Anal.","ISSN":"https:\/\/id.crossref.org\/issn\/0036-1410","issn-type":"print"},{"issue":"3","key":"2","doi-asserted-by":"publisher","first-page":"345","DOI":"10.1016\/j.acha.2005.07.001","article-title":"On signal reconstruction without phase","volume":"20","author":"Balan, Radu","year":"2006","journal-title":"Appl. Comput. Harmon. Anal.","ISSN":"https:\/\/id.crossref.org\/issn\/1063-5203","issn-type":"print"},{"issue":"1","key":"3","doi-asserted-by":"publisher","first-page":"106","DOI":"10.1016\/j.acha.2013.10.002","article-title":"Saving phase: injectivity and stability for phase retrieval","volume":"37","author":"Bandeira, Afonso S.","year":"2014","journal-title":"Appl. Comput. Harmon. Anal.","ISSN":"https:\/\/id.crossref.org\/issn\/1063-5203","issn-type":"print"},{"key":"4","doi-asserted-by":"crossref","unstructured":"O. Bunk, A. Diaz, F. Pfeiffer, C. David, B. Schmitt, D. K. Satapathy, and J. F. V. D. Veen, Diffractive imaging for periodic samples: retrieving one-dimensional concentration profiles across microfluidic channels, Acta Crystallogr. Sect. A 63 (2007), no. 4, 306\u2013314.","DOI":"10.1107\/S0108767307021903"},{"key":"5","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1016\/j.acha.2022.01.002","article-title":"Solving phase retrieval with random initial guess is nearly as good as by spectral initialization","volume":"58","author":"Cai, Jian-Feng","year":"2022","journal-title":"Appl. Comput. Harmon. Anal.","ISSN":"https:\/\/id.crossref.org\/issn\/1063-5203","issn-type":"print"},{"key":"6","doi-asserted-by":"publisher","first-page":"63","DOI":"10.1090\/btran\/12","article-title":"Phase retrieval in infinite-dimensional Hilbert spaces","volume":"3","author":"Cahill, Jameson","year":"2016","journal-title":"Trans. Amer. Math. Soc. Ser. B"},{"key":"7","series-title":"Graduate Texts in Mathematics","isbn-type":"print","volume-title":"Nonsmooth analysis and control theory","volume":"178","author":"Clarke, F. H.","year":"1998","ISBN":"https:\/\/id.crossref.org\/isbn\/0387983368"},{"key":"8","doi-asserted-by":"crossref","unstructured":"R. Collobert and J. Weston, A unified architecture for natural language processing: deep neural networks with multitask learning, In Proceedings of the 25th international conference on Machine learning, 2008, pp. 160\u2013167.","DOI":"10.1145\/1390156.1390177"},{"issue":"2","key":"9","doi-asserted-by":"publisher","first-page":"346","DOI":"10.1016\/j.acha.2014.06.005","article-title":"An algebraic characterization of injectivity in phase retrieval","volume":"38","author":"Conca, Aldo","year":"2015","journal-title":"Appl. Comput. Harmon. Anal.","ISSN":"https:\/\/id.crossref.org\/issn\/1063-5203","issn-type":"print"},{"key":"10","series-title":"CMS Books in Mathematics\/Ouvrages de Math\\'{e}matiques de la SMC","isbn-type":"print","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4684-9298-9","volume-title":"Best approximation in inner product spaces","volume":"7","author":"Deutsch, Frank","year":"2001","ISBN":"https:\/\/id.crossref.org\/isbn\/0387951563"},{"key":"11","unstructured":"S. S. Du, X. Zhai, B. Poczos, and A. Singh, Gradient descent provably optimizes over-parameterized neural networks, International Conference on Learning Representations, 2019."},{"key":"12","unstructured":"C. Fienup and J. Dainty, Phase Retrieval and Image Reconstruction for Astronomy, Image recovery: theory and application, 1987, pp. 231\u2013275."},{"key":"13","doi-asserted-by":"crossref","unstructured":"J. R. Fienup, Phase retrieval algorithms: a comparison, Appl. Opt. 21 (1982), no. 15, 2758\u20132769.","DOI":"10.1364\/AO.21.002758"},{"key":"14","doi-asserted-by":"publisher","first-page":"5427","DOI":"10.1109\/TSP.2020.3022817","article-title":"Perturbed amplitude flow for phase retrieval","volume":"68","author":"Gao, Bing","year":"2020","journal-title":"IEEE Trans. Signal Process.","ISSN":"https:\/\/id.crossref.org\/issn\/1053-587X","issn-type":"print"},{"key":"15","unstructured":"R. W. Gerchberg, A practical algorithm for the determination of phase from image and diffraction plane pictures, Optik 35 (1972), 237\u2013246."},{"key":"16","unstructured":"S. Goel, V. Kanade, A. Klivans, and J. Thaler, Reliably learning the relu in polynomial time, In Conference on Learning Theory, 2017, pp. 1004\u20131042."},{"key":"17","series-title":"Adaptive Computation and Machine Learning","isbn-type":"print","volume-title":"Deep learning","author":"Goodfellow, Ian","year":"2016","ISBN":"https:\/\/id.crossref.org\/isbn\/9780262035613"},{"issue":"2","key":"18","doi-asserted-by":"publisher","first-page":"301","DOI":"10.1137\/19M1256865","article-title":"Phase retrieval: uniqueness and stability","volume":"62","author":"Grohs, Philipp","year":"2020","journal-title":"SIAM Rev.","ISSN":"https:\/\/id.crossref.org\/issn\/1095-7200","issn-type":"print"},{"key":"19","unstructured":"P. Hand, O. Leong, and V. Voroninski, Phase retrieval under a generative prior, In Advances in Neural Information Processing Systems, 2018, pp. 9136\u20139146."},{"key":"20","doi-asserted-by":"crossref","unstructured":"R. W. Harrison, Phase problem in crystallography, JOSA A 10 (1993), no. 5, 1046\u20131055.","DOI":"10.1364\/JOSAA.10.001046"},{"key":"21","unstructured":"E. Hazan, K. Levy, and S. Shalev-Shwartz, Beyond convexity: Stochastic quasi-convex optimization, In Advances in Neural Information Processing Systems Vol. 28, 2015, pp. 1594\u20131602."},{"issue":"12","key":"22","doi-asserted-by":"publisher","first-page":"7967","DOI":"10.1109\/TIT.2020.2983562","article-title":"The estimation performance of nonlinear least squares for phase retrieval","volume":"66","author":"Huang, Meng","year":"2020","journal-title":"IEEE Trans. Inform. Theory","ISSN":"https:\/\/id.crossref.org\/issn\/0018-9448","issn-type":"print"},{"key":"23","doi-asserted-by":"crossref","unstructured":"K. Jarrett, K. Kavukcuoglu, M. Ranzato, and Y. LeCun, What is the best multi-stage architecture for object recognition? In 2009 IEEE 12th international conference on computer vision, 2009, pp. 2146\u20132153.","DOI":"10.1109\/ICCV.2009.5459469"},{"key":"24","unstructured":"S. M. Kakade, V. Kanade, O. Shamir, and A. Kalai, Efficient learning of generalized linear and single index models with isotonic regression, In Advances in Neural Information Processing Systems, 2011, pp. 927\u2013935."},{"key":"25","unstructured":"A. T. Kalai and R. Sastry, The isotron algorithm: high-dimensional isotonic regression, In COLT, 2009."},{"issue":"6","key":"26","doi-asserted-by":"publisher","first-page":"2747","DOI":"10.1214\/17-AOS1637","article-title":"The landscape of empirical risk for nonconvex losses","volume":"46","author":"Mei, Song","year":"2018","journal-title":"Ann. Statist.","ISSN":"https:\/\/id.crossref.org\/issn\/0090-5364","issn-type":"print"},{"key":"27","doi-asserted-by":"crossref","unstructured":"J. Miao, T. Ishikawa, Q. Shen, and T. Earnest, Extending x-ray crystallography to allow the imaging of noncrystalline materials, cells, and single protein complexes, Annu. Rev. Phys. Chem. 59 (2008), 387\u2013410.","DOI":"10.1146\/annurev.physchem.59.032607.093642"},{"key":"28","doi-asserted-by":"crossref","unstructured":"R. P. Millane, Phase retrieval in crystallography and optics, JOSA A 7 (1990), no. 3, 394\u2013411.","DOI":"10.1364\/JOSAA.7.000394"},{"issue":"18","key":"29","doi-asserted-by":"publisher","first-page":"4814","DOI":"10.1109\/TSP.2015.2448516","article-title":"Phase retrieval using alternating minimization","volume":"63","author":"Netrapalli, Praneeth","year":"2015","journal-title":"IEEE Trans. Signal Process.","ISSN":"https:\/\/id.crossref.org\/issn\/1053-587X","issn-type":"print"},{"key":"30","doi-asserted-by":"crossref","unstructured":"S. Oymak and M. Soltanolkotabi, Towards moderate overparameterization: global convergence guarantees for training shallow neural networks, IEEE J. Selected Areas Inform. Theory 1 (2020), no. 1, 84\u2013105.","DOI":"10.1109\/JSAIT.2020.2991332"},{"key":"31","doi-asserted-by":"crossref","unstructured":"F. Shamshad and A. Ahmed, Compressed sensing based robust phase retrieval via deep generative priors, IEEE Sensors J., 2020.","DOI":"10.1109\/JSEN.2020.3018751"},{"key":"32","unstructured":"M. Soltanolkotabi, Learning relus via gradient descent, In Advances in Neural Information processing systems, 2017, pp. 2007\u20132017."},{"issue":"2","key":"33","doi-asserted-by":"publisher","first-page":"386","DOI":"10.1016\/j.acha.2015.06.004","article-title":"A strong restricted isometry property, with an application to phaseless compressed sensing","volume":"40","author":"Voroninski, Vladislav","year":"2016","journal-title":"Appl. Comput. Harmon. Anal.","ISSN":"https:\/\/id.crossref.org\/issn\/1063-5203","issn-type":"print"},{"key":"34","doi-asserted-by":"publisher","first-page":"41","DOI":"10.1080\/713817747","article-title":"The question of phase retrieval in optics","volume":"10","author":"Walther, Adriaan","year":"1963","journal-title":"Optica Acta","ISSN":"https:\/\/id.crossref.org\/issn\/0030-3909","issn-type":"print"},{"issue":"2","key":"35","doi-asserted-by":"publisher","first-page":"773","DOI":"10.1109\/TIT.2017.2756858","article-title":"Solving systems of random quadratic equations via truncated amplitude flow","volume":"64","author":"Wang, Gang","year":"2018","journal-title":"IEEE Trans. Inform. Theory","ISSN":"https:\/\/id.crossref.org\/issn\/0018-9448","issn-type":"print"},{"issue":"2","key":"36","doi-asserted-by":"publisher","first-page":"423","DOI":"10.1016\/j.acha.2017.09.003","article-title":"Generalized phase retrieval: measurement number, matrix recovery and beyond","volume":"47","author":"Wang, Yang","year":"2019","journal-title":"Appl. Comput. Harmon. Anal.","ISSN":"https:\/\/id.crossref.org\/issn\/1063-5203","issn-type":"print"},{"key":"37","doi-asserted-by":"crossref","unstructured":"G. Xu, H. Wu, and Y. Shi, Structural design of convolutional neural networks for steganalysis, IEEE Signal Process. Lett. 23 (2016), no. 5, 708\u2013712.","DOI":"10.1109\/LSP.2016.2548421"},{"key":"38","unstructured":"H. Zhang and Y. Liang, Reshaped wirtinger flow for solving quadratic system of equations, In Advances in Neural Information Processing Systems, 1986, pp. 2622\u20132630."}],"container-title":["Mathematics of Computation"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.ams.org\/mcom\/2024-93-347\/S0025-5718-2023-03918-7\/mcom3918_AM.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/www.ams.org\/mcom\/2024-93-347\/S0025-5718-2023-03918-7\/S0025-5718-2023-03918-7.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,22]],"date-time":"2026-04-22T05:16:15Z","timestamp":1776834975000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ams.org\/mcom\/2024-93-347\/S0025-5718-2023-03918-7\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,10,27]]},"references-count":38,"journal-issue":{"issue":"347","published-print":{"date-parts":[[2024,5]]}},"alternative-id":["S0025-5718-2023-03918-7"],"URL":"https:\/\/doi.org\/10.1090\/mcom\/3918","archive":["CLOCKSS","Portico"],"relation":{},"ISSN":["1088-6842","0025-5718"],"issn-type":[{"value":"1088-6842","type":"electronic"},{"value":"0025-5718","type":"print"}],"subject":[],"published":{"date-parts":[[2023,10,27]]}}}