{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,10]],"date-time":"2026-03-10T19:29:47Z","timestamp":1773170987307,"version":"3.50.1"},"reference-count":99,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"4","license":[{"start":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T00:00:00Z","timestamp":1775001600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T00:00:00Z","timestamp":1775001600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T00:00:00Z","timestamp":1775001600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Gift Donation from Cisco"},{"name":"Kempner Institute Research Fellowship"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2026,4]]},"DOI":"10.1109\/tpami.2025.3645586","type":"journal-article","created":{"date-parts":[[2025,12,22]],"date-time":"2025-12-22T18:41:22Z","timestamp":1766428882000},"page":"4539-4552","source":"Crossref","is-referenced-by-count":0,"title":["Unsupervised Representation Learning From Sparse Transformation Analysis"],"prefix":"10.1109","volume":"48","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-1573-5643","authenticated-orcid":false,"given":"Yue","family":"Song","sequence":"first","affiliation":[{"name":"Computing and Mathematical Sciences, Caltech, Pasadena, CA, USA"}]},{"given":"T. Anderson","family":"Keller","sequence":"additional","affiliation":[{"name":"Kempner Institute for the Study of Natural and Artificial Intelligence, Harvard University, Cambridge, MA, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9127-1989","authenticated-orcid":false,"given":"Yisong","family":"Yue","sequence":"additional","affiliation":[{"name":"Computing and Mathematical Sciences, Caltech, Pasadena, CA, USA"}]},{"given":"Pietro","family":"Perona","sequence":"additional","affiliation":[{"name":"Computing and Mathematical Sciences, Caltech, Pasadena, CA, USA"}]},{"given":"Max","family":"Welling","sequence":"additional","affiliation":[{"name":"Amsterdam Machine Learning Lab, University of Amsterdam, Amsterdam, The Netherlands"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.7551\/mitpress\/9780262518420.003.0013"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1016\/S0042-6989(97)00169-7"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1016\/0165-1684(94)90029-9"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1162\/089976602317318938"},{"key":"ref5","article-title":"Towards nonlinear disentanglement in natural data with temporal sparse coding","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Klindt","year":"2021"},{"key":"ref6","first-page":"2990","article-title":"Group equivariant convolutional networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Cohen","year":"2016"},{"key":"ref7","article-title":"Towards a definition of disentangled representations","author":"Higgins","year":"2018"},{"key":"ref8","article-title":"Transformation properties of learned visual representations","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Cohen","year":"2015"},{"key":"ref9","article-title":"Steerable CNNs","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Cohen","year":"2017"},{"key":"ref10","first-page":"8867","article-title":"Equivariant diffusion for molecule generation in 3D","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Hoogeboom","year":"2022"},{"key":"ref11","first-page":"28585","article-title":"Topographic VAEs learn equivariant capsules","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Keller","year":"2021"},{"key":"ref12","first-page":"49761","article-title":"Flow factorzied representation learning","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Song","year":"2023"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.1988.10478694"},{"key":"ref14","first-page":"7360","article-title":"Semi-supervised StyleGAN for disentanglement learning","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Nie","year":"2020"},{"key":"ref15","first-page":"1","article-title":"The multi-agent behavior dataset: Mouse dyadic social interactions","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Sun","year":"2021"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.350"},{"key":"ref17","first-page":"2180","article-title":"InfoGAN: Interpretable representation learning by information maximizing generative adversarial nets","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Chen","year":"2016"},{"key":"ref18","article-title":"Beta-VAE: Learning basic visual concepts with a constrained variational framework","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Higgins","year":"2016"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00584"},{"key":"ref20","article-title":"On the \u2018steerability\u2019 of generative adversarial networks","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Jahanian","year":"2020"},{"key":"ref21","first-page":"9786","article-title":"Unsupervised discovery of interpretable directions in the GAN latent space","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Voynov","year":"2020"},{"key":"ref22","first-page":"9841","article-title":"GANSpace: Discovering interpretable GAN controls","volume-title":"Proc. Neural Inf. Process. Syst.","author":"H\u00e4rk\u00f6nen","year":"2020"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58607-2_40"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58539-6_35"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.00158"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00665"},{"key":"ref27","first-page":"16648","article-title":"Low-rank subspaces in GANs","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Zhu","year":"2021"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00633"},{"key":"ref29","first-page":"27612","article-title":"Region-based semantic factorization in GANs","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Zhu","year":"2022"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2022.3228979"},{"key":"ref31","article-title":"PandA: Unsupervised learning of parts and appearances in the feature maps of GANs","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Oldfield","year":"2023"},{"key":"ref32","first-page":"32288","article-title":"Latent traversals in generative models as potential flows","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Song","year":"2023"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00709"},{"key":"ref34","article-title":"Deep unsupervised clustering with Gaussian mixture variational autoencoders","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Dilokthanakul","year":"2016"},{"key":"ref35","first-page":"708","article-title":"Learning disentangled joint continuous and discrete representations","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Dupont","year":"2018"},{"key":"ref36","article-title":"Variational inference of disentangled latent concepts from unlabeled observations","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Kumar","year":"2018"},{"key":"ref37","first-page":"2649","article-title":"Disentangling by factorising","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Kim","year":"2018"},{"key":"ref38","first-page":"2615","article-title":"Isolating sources of disentanglement in variational autoencoders","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Chen","year":"2018"},{"key":"ref39","first-page":"3091","article-title":"Learning discrete and continuous factors of data via alternating disentanglement","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Jeong","year":"2019"},{"key":"ref40","first-page":"13434","article-title":"ODE2VAE: Deep generative second order odes with Bayesian neural networks","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Yildiz","year":"2019"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00794"},{"key":"ref42","first-page":"8655","article-title":"ControlVAE: Controllable variational autoencoder","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Shao","year":"2020"},{"key":"ref43","first-page":"6348","article-title":"Weakly-supervised disentanglement without compromises","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Locatello","year":"2020"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i10.21387"},{"key":"ref45","article-title":"DAVA: Disentangling adversarial variational autoencoder","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Estermann","year":"2023"},{"key":"ref46","article-title":"Diffusion models already have a semantic latent space","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Kwon","year":"2023"},{"key":"ref47","first-page":"24129","article-title":"Understanding the latent space of diffusion models through the lens of Riemannian geometry","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Park","year":"2023"},{"key":"ref48","first-page":"69130","article-title":"DisDiff: Unsupervised disentanglement of diffusion probabilistic models","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Yang","year":"2023"},{"key":"ref49","first-page":"36336","article-title":"InfoDiffusion: Representation learning using information maximizing diffusion models","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Wang","year":"2023"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.7551\/mitpress\/1754.001.0001"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511803161"},{"key":"ref52","first-page":"4114","article-title":"Challenging common assumptions in the unsupervised learning of disentangled representations","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Locatello","year":"2019"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.00947"},{"key":"ref54","article-title":"Learning neural causal models from unknown interventions","author":"Ke","year":"2019"},{"key":"ref55","first-page":"2093","article-title":"Disentangling mixtures of unknown causal interventions","volume-title":"Proc. Uncertainty Artif. Intell.","author":"Kumar","year":"2021"},{"key":"ref56","first-page":"77098","article-title":"Causal temporal representation learning with nonstationary sparse transition","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"37","author":"Song","year":"2024"},{"key":"ref57","first-page":"2892","article-title":"Equivariance through parameter-sharing","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Ravanbakhsh","year":"2017"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.758"},{"key":"ref59","first-page":"7366","article-title":"Deep scale-spaces: Equivariance over scale","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Worrall","year":"2019"},{"key":"ref60","first-page":"3165","article-title":"Generalizing convolutional neural networks for equivariance to lie groups on arbitrary continuous data","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Finzi","year":"2020"},{"key":"ref61","first-page":"1586","article-title":"Learning to convolve: A generalized weight-tying approach","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Diaconu","year":"2019"},{"key":"ref62","first-page":"2359","article-title":"Variational autoencoder with learned latent structure","volume-title":"Int. Conf. Artif. Intell. Statist.","author":"Connor","year":"2021"},{"key":"ref63","first-page":"1876","article-title":"Unsupervised learning of disentangled and interpretable representations from sequential data","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Hsu","year":"2017"},{"key":"ref64","first-page":"4417","article-title":"Unsupervised learning of disentangled representations from video","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Denton","year":"2017"},{"key":"ref65","article-title":"Decomposing motion and content for natural video sequence prediction","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Villegas","year":"2017"},{"key":"ref66","first-page":"5670","article-title":"Disentangled sequential autoencoder","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Li","year":"2018"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00165"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00657"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58592-1_7"},{"key":"ref70","first-page":"305","article-title":"Disentangled representations for sequence data using information bottleneck principle","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Yamada","year":"2020"},{"key":"ref71","first-page":"10105","article-title":"Contrastively disentangled sequential variational autoencoder","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Bai","year":"2021"},{"key":"ref72","article-title":"Disentangled recurrent Wasserstein autoencoder","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Han","year":"2021"},{"key":"ref73","first-page":"8700","article-title":"Decoupling local and global representations of time series","volume-title":"Proc. Int. Conf. Artif. Intell. Statist.","author":"Tonekaboni","year":"2022"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v37i8.26111"},{"key":"ref75","article-title":"Multifactor sequential disentanglement via structured Koopman autoencoders","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Berman","year":"2023"},{"key":"ref76","first-page":"3539","article-title":"Sequential disentanglement by extracting static information from a single sequence element","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Nimrod","year":"2024"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1016\/j.jcp.2018.10.045"},{"key":"ref78","article-title":"Learning neural PDE solvers with convergence guarantees","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Hsieh","year":"2019"},{"key":"ref79","article-title":"Message passing neural PDE solvers","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Brandstetter","year":"2022"},{"key":"ref80","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2759"},{"key":"ref81","article-title":"Competitive physics informed networks","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Zeng","year":"2023"},{"key":"ref82","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/acb416"},{"key":"ref83","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1842"},{"key":"ref84","first-page":"6840","article-title":"Denoising diffusion probabilistic models","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Ho","year":"2020"},{"key":"ref85","article-title":"Denoising diffusion implicit models","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Song","year":"2021"},{"key":"ref86","article-title":"Score-based generative modeling through stochastic differential equations","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Song","year":"2021"},{"key":"ref87","first-page":"15379","article-title":"Hamiltonian neural networks","volume-title":"Proc. Neural Inf. Process. Syst.","author":"Greydanus","year":"2019"},{"key":"ref88","article-title":"Hamiltonian generative networks","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Toth","year":"2020"},{"key":"ref89","article-title":"The MNIST database of handwritten digits","author":"LeCun","year":"1998"},{"key":"ref90","first-page":"1530","article-title":"Variational inference with normalizing flows","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Rezende","year":"2015"},{"key":"ref91","first-page":"6572","article-title":"Neural ordinary differential equations","volume-title":"Proc. 32nd Int. Conf. Neural Inf. Process. Syst.","author":"Chen","year":"2019"},{"key":"ref92","first-page":"25","article-title":"\u00dcber integrale der hydrodynamischen Gleichungen, welche den Wirbelbewegungen entsprechen","volume":"55","author":"Helmholtz","year":"1858","journal-title":"J. Reine Angew. Math."},{"key":"ref93","doi-asserted-by":"publisher","DOI":"10.1080\/14786446708639824"},{"key":"ref94","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4612-1029-0"},{"key":"ref95","article-title":"Categorical reparameterization with gumbel-softmax","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Jang","year":"2017"},{"key":"ref96","article-title":"Auto-encoding variational Bayes","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Kingma","year":"2014"},{"key":"ref97","doi-asserted-by":"publisher","DOI":"10.1007\/s002110050002"},{"key":"ref98","article-title":"3D shapes dataset","author":"Burgess","year":"2018"},{"key":"ref99","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.63720"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/34\/11424231\/11310844.pdf?arnumber=11310844","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,10]],"date-time":"2026-03-10T01:34:53Z","timestamp":1773106493000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11310844\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4]]},"references-count":99,"journal-issue":{"issue":"4"},"URL":"https:\/\/doi.org\/10.1109\/tpami.2025.3645586","relation":{},"ISSN":["0162-8828","2160-9292","1939-3539"],"issn-type":[{"value":"0162-8828","type":"print"},{"value":"2160-9292","type":"electronic"},{"value":"1939-3539","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4]]}}}