{"database":"biostudies-literature","file_versions":[],"scores":null,"additional":{"submitter":["Huguet G"],"funding":["NICHD NIH HHS","NIGMS NIH HHS"],"pagination":["29705-29718"],"full_dataset_link":["https://www.ebi.ac.uk/biostudies/studies/S-EPMC10312391"],"repository":["biostudies-literature"],"omics_type":["Unknown"],"volume":["35"],"pubmed_abstract":["We present a method called Manifold Interpolating Optimal-Transport Flow (MIOFlow) that learns stochastic, continuous population dynamics from static snapshot samples taken at sporadic timepoints. MIOFlow combines dynamic models, manifold learning, and optimal transport by training neural ordinary differential equations (Neural ODE) to interpolate between static population snapshots as penalized by optimal transport with manifold ground distance. Further, we ensure that the flow follows the geometry by operating in the latent space of an autoencoder that we call a geodesic autoencoder (GAE). In GAE the latent space distance between points is regularized to match a novel multiscale geodesic distance on the data manifold that we define. We show that this method is superior to normalizing flo"],"journal":["Advances in neural information processing systems"],"pubmed_title":["Manifold Interpolating Optimal-Transport Flows for Trajectory Inference."],"pmcid":["PMC10312391"],"funding_grant_id":["R01 HD100035","R01 GM130847"],"pubmed_authors":["Huguet G","Kuchroo M","Krishnaswamy S","Tong A","Fasina O","Magruder DS","Wolf G"],"additional_accession":[]},"is_claimable":false,"name":"Manifold Interpolating Optimal-Transport Flows for Trajectory Inference.","description":"We present a method called Manifold Interpolating Optimal-Transport Flow (MIOFlow) that learns stochastic, continuous population dynamics from static snapshot samples taken at sporadic timepoints. MIOFlow combines dynamic models, manifold learning, and optimal transport by training neural ordinary differential equations (Neural ODE) to interpolate between static population snapshots as penalized by optimal transport with manifold ground distance. Further, we ensure that the flow follows the geometry by operating in the latent space of an autoencoder that we call a geodesic autoencoder (GAE). In GAE the latent space distance between points is regularized to match a novel multiscale geodesic distance on the data manifold that we define. We show that this method is superior to normalizing flo","dates":{"release":"2022-01-01T00:00:00Z","publication":"2022 Dec","modification":"2025-04-04T14:08:10.767Z","creation":"2025-04-04T14:08:10.767Z"},"accession":"S-EPMC10312391","cross_references":{"pubmed":["37397786"]}}