Scalable conditional deep inverse Rosenblatt transports using tensor trains and gradient-based dimension reduction - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Article Dans Une Revue Journal of Computational Physics Année : 2023

Scalable conditional deep inverse Rosenblatt transports using tensor trains and gradient-based dimension reduction

Résumé

We present a novel offline-online method to mitigate the computational burden of the characterization of posterior random variables in statistical learning. In the offline phase, the proposed method learns the joint law of the parameter random variables and the observable random variables in the tensor-train (TT) format. In the online phase, the resulting order-preserving conditional transport can characterize the posterior random variables given newly observed data in real time. Compared with the state-of-the-art normalizing flow techniques, the proposed method relies on function approximation and is equipped with a thorough performance analysis. The function approximation perspective also allows us to further extend the capability of transport maps in challenging problems with high-dimensional observations and high-dimensional parameters. On the one hand, we present novel heuristics to reorder and/or reparametrize the variables to enhance the approximation power of TT. On the other hand, we integrate the TT-based transport maps and the parameter reordering/reparametrization into layered compositions to further improve the performance of the resulting transport maps. We demonstrate the efficiency of the proposed method on various statistical learning tasks in ordinary differential equations (ODEs) and partial differential equations (PDEs).

Dates et versions

hal-03527999 , version 1 (17-01-2022)

Licence

Paternité

Identifiants

Citer

Tiangang Cui, Sergey Dolgov, Olivier Zahm. Scalable conditional deep inverse Rosenblatt transports using tensor trains and gradient-based dimension reduction. Journal of Computational Physics, 2023, pp.112103. ⟨10.1016/j.jcp.2023.112103⟩. ⟨hal-03527999⟩
59 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More