{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,7,30]],"date-time":"2025-07-30T15:49:20Z","timestamp":1753890560487,"version":"3.41.2"},"reference-count":34,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,1,15]],"date-time":"2025-01-15T00:00:00Z","timestamp":1736899200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100006502","name":"Defense Sciences Office, DARPA","doi-asserted-by":"publisher","award":["HR0011-22-9-0077"],"award-info":[{"award-number":["HR0011-22-9-0077"]}],"id":[{"id":"10.13039\/100006502","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Comput. Sci."],"abstract":"<jats:p>Manifold representation learning holds great promise for theoretical understanding and characterization of deep neural networks' behaviors through the lens of geometries. However, data scarcity remains a major challenge in manifold analysis especially for data and applications with real-world complexity. To address this issue, we propose manifold representation meta-learning (MRML) based on autoencoders to recover the underlying manifold structures without uniformly or densely sampled data. Specifically, we adopt episodic training, following model agnostic meta-learning, to meta-learn autoencoders that are generalizable to unseen samples specifically corresponding to regions with low-sampling density. We demonstrate the effectiveness of MRML via empirical experiments on LineMOD, a dataset curated for 6-D object pose estimation. We also apply topological metrics based on persistent homology and neighborhood graphs for quantitative assessment of manifolds reconstructed by MRML. In comparison to state-of-the-art baselines, our proposed approach demonstrates improved manifold reconstruction better matching the data manifold by preserving prominent topological features and relative proximity of samples.<\/jats:p>","DOI":"10.3389\/fcomp.2024.1255517","type":"journal-article","created":{"date-parts":[[2025,1,15]],"date-time":"2025-01-15T06:10:29Z","timestamp":1736921429000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Recovering manifold representations via unsupervised meta-learning"],"prefix":"10.3389","volume":"6","author":[{"given":"Yunye","family":"Gong","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jiachen","family":"Yao","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ruyi","family":"Lian","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xiao","family":"Lin","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chao","family":"Chen","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ajay","family":"Divakaran","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yi","family":"Yao","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2025,1,15]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1948","DOI":"10.1080\/03610918.2021.1894335","article-title":"Comparison of persistence diagrams","volume":"52","author":"Agami","year":"2020","journal-title":"Commun. Stat. Simulat. Comput"},{"key":"B2","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s41109-019-0179-3","article-title":"Persistence homology of networks: methods and applications","volume":"4","author":"Aktas","year":"2019","journal-title":"Appl. Netw. Sci"},{"key":"B3","article-title":"\u201cLearning to learn by gradient descent by gradient descent,\u201d","author":"Andrychowicz","year":"2016","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B4","article-title":"Autoencoders","author":"Bank","year":"2020","journal-title":"arXiv preprint arXiv:2003.05991"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1798","DOI":"10.1109\/TPAMI.2013.50","article-title":"Representation learning: a review and new perspectives","volume":"35","author":"Bengio","year":"2013","journal-title":"IEEE Trans. Patt. Analy. Mach. Intell"},{"key":"B6","doi-asserted-by":"publisher","first-page":"263","DOI":"10.1145\/1064092.1064133","article-title":"\u201cStability of persistence diagrams,\u201d","author":"Cohen-Steiner","year":"2005","journal-title":"Proceedings of the Twenty-First Annual Symposium on Computational Geometry"},{"key":"B7","doi-asserted-by":"publisher","first-page":"7381","DOI":"10.1109\/TPAMI.2022.3222104","article-title":"Geometry regularized autoencoders","volume":"45","author":"Duque","year":"2022","journal-title":"IEEE Trans. Patt. Analy. Mach. Intell"},{"key":"B8","doi-asserted-by":"publisher","first-page":"257","DOI":"10.1090\/conm\/453\/08802","article-title":"Persistent homology \u2013 a survey","volume":"453","author":"Edelsbrunner","year":"2008","journal-title":"Contemp. Mathem"},{"key":"B9","first-page":"1126","article-title":"\u201cModel-agnostic meta-learning for fast adaptation of deep networks,\u201d","volume-title":"International Conference on Machine Learning","author":"Finn","year":"2017"},{"key":"B10","article-title":"Isometric autoencoders","author":"Gropp","year":"2020","journal-title":"arXiv preprint arXiv:2006.09289"},{"key":"B11","first-page":"548","article-title":"\u201cModel based training, detection and pose estimation of texture-less 3D objects in heavily cluttered scenes,\u201d","volume-title":"Computer Vision-ACCV 2012: 11th Asian Conference on Computer Vision, Daejeon, Korea, November 5-9, 2012, Revised Selected Papers, Part I 11","author":"Hinterstoisser","year":"2012"},{"key":"B12","doi-asserted-by":"publisher","first-page":"5149","DOI":"10.1109\/TPAMI.2021.3079209","article-title":"Meta-learning in neural networks: a survey","volume":"44","author":"Hospedales","year":"2022","journal-title":"IEEE Trans. Patt. Analy. Mach. Intell"},{"key":"B13","article-title":"Unsupervised learning via meta-learning","author":"Hsu","year":"2019","journal-title":"arXiv preprint arXiv:1810.02334"},{"key":"B14","article-title":"Adam: a method for stochastic optimization","author":"Kingma","year":"2015","journal-title":"arXiv preprint arXiv:1412.6980"},{"key":"B15","article-title":"\u201cMeta-GMVAE: mixture of gaussian vaes for unsupervised meta-learning,\u201d","author":"Lee","year":"2021","journal-title":"International Conference on Learning Representations"},{"key":"B16","first-page":"536","article-title":"\u201cNeighborhood reconstructing autoencoders,\u201d","author":"Lee","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B17","article-title":"\u201cRegularized autoencoders for isometric representation learning,\u201d","author":"Lee","year":"2022","journal-title":"ICLR"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v32i1.11596","article-title":"\u201cLearning to generalize: meta-learning for domain generalization,\u201d","author":"Li","year":"2018","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1446","DOI":"10.1109\/ICCV.2019.00153","article-title":"\u201cEpisodic training for domain generalization,\u201d","author":"Li","year":"2019","journal-title":"Proceedings of the IEEE\/CVF International Conference on Computer Vision"},{"key":"B20","doi-asserted-by":"publisher","first-page":"84","DOI":"10.1016\/j.neucom.2013.09.055","article-title":"Autoencoder for words","volume":"139","author":"Liou","year":"2014","journal-title":"Neurocomputing"},{"key":"B21","doi-asserted-by":"publisher","first-page":"1013","DOI":"10.1109\/CVPR46437.2021.00107","article-title":"\u201cFEDDG: federated domain generalization on medical image segmentation via episodic learning in continuous frequency space,\u201d","author":"Liu","year":"2021","journal-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition"},{"key":"B22","doi-asserted-by":"publisher","first-page":"124007","DOI":"10.1088\/0266-5611\/27\/12\/124007","article-title":"Probability measures on the space of persistence diagrams","volume":"27","author":"Mileyko","year":"2011","journal-title":"Inverse Problems"},{"key":"B23","first-page":"7045","article-title":"\u201cTopological autoencoders,\u201d","volume-title":"International Conference on Machine Learning","author":"Moor","year":"2020"},{"key":"B24","article-title":"Geometric autoencoders-what you see is what you decode","author":"Nazari","year":"2023","journal-title":"arXiv preprint arXiv:2306.17638"},{"journal-title":"Five balltree construction algorithms","year":"1989","author":"Omohundro","key":"B25"},{"key":"B26","article-title":"Persistent-homology-based machine learning and its applications - a survey","author":"Pun","year":"2018","journal-title":"arXiv preprint arXiv:1811.00252"},{"key":"B27","article-title":"\u201cWitness autoencoder: shaping the latent space with witness complexes,\u201d","author":"Sch\u00f6nenberger","year":"2020","journal-title":"NeurIPS 2020 Workshop TDA and Beyond"},{"key":"B28","article-title":"Chart auto-encoders for manifold structured data","author":"Schonsheck","year":"2019","journal-title":"arXiv preprint arXiv:1912.10094"},{"key":"B29","first-page":"30","article-title":"\u201cPrototypical networks for few-shot learning,\u201d","author":"Snell","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B30","doi-asserted-by":"publisher","first-page":"2319","DOI":"10.1126\/science.290.5500.2319","article-title":"A global geometric framework for nonlinear dimensionality reduction","volume":"290","author":"Tenenbaum","year":"2000","journal-title":"Science"},{"key":"B31","first-page":"2579","article-title":"Visualizing data using t-SNE","volume":"9","author":"van der Maaten","year":"2008","journal-title":"JMLR"},{"key":"B32","doi-asserted-by":"publisher","first-page":"75","DOI":"10.1007\/s10472-021-09761-3","article-title":"Topological measurement of deep neural networks using persistent homology","volume":"90","author":"Watanabe","year":"2022","journal-title":"Ann. Mathem. Artif. Intell"},{"key":"B33","doi-asserted-by":"publisher","first-page":"e0228728","DOI":"10.1371\/journal.pone.0228728","article-title":"Metrics for graph comparison: a practitioner's guide","volume":"15","author":"Wills","year":"2020","journal-title":"PLoS ONE"},{"key":"B34","doi-asserted-by":"publisher","first-page":"6404","DOI":"10.1609\/aaai.v34i04.6111","article-title":"\u201cMeta-amortized variational inference and learning,\u201d","author":"Wu","year":"2020","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"}],"container-title":["Frontiers in Computer Science"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2024.1255517\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,1,15]],"date-time":"2025-01-15T06:10:35Z","timestamp":1736921435000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2024.1255517\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,1,15]]},"references-count":34,"alternative-id":["10.3389\/fcomp.2024.1255517"],"URL":"https:\/\/doi.org\/10.3389\/fcomp.2024.1255517","relation":{},"ISSN":["2624-9898"],"issn-type":[{"type":"electronic","value":"2624-9898"}],"subject":[],"published":{"date-parts":[[2025,1,15]]},"article-number":"1255517"}}