Stochastic Latent Twins: Learning Stochastic Solution Operators

DOI10.48550/arXiv.2610.21084MaRDI QIDQ7395209arXivFDO

Authors Matthias Chung, Jack Michael Solomon, Deepanshu Verma

Publication date 15 October 2026

Published in arXiv

Tldr We introduce Stochastic Latent Twins, which learn stochastic solution operators for physics-informed uncertainty prediction by separating representation from transition and learning the distribution-valued evolution map directly, enabling direct probabilistic prediction without recursive time stepping or generative modeling.

Full work available at URL https://doi.org/10.48550/arXiv.2610.21084

arXiv classification

cs.CEmath.NA

Cites work

• Calculation of the Wasserstein Distance Between Probability Distributions on the Line
• Deterministic Nonperiodic Flow
• Matrix Analysis
• On a Formula for the L2 Wasserstein Metric between Measures on Euclidean and Hilbert Spaces
• On the Bures-Wasserstein distance between positive definite matrices
• Optimal Transport
• Scalable Gradients for Stochastic Differential Equations
• Scientific computing with MATLAB and Octave
• Scikit-learn: machine learning in Python
• Stochastic Equations in Infinite Dimensions
• Strictly Proper Scoring Rules, Prediction, and Estimation
• The Differentiation of Pseudo-Inverses and Nonlinear Least Squares Problems Whose Variables Separate
• Time reversal of diffusions

This page was built for publication: Stochastic Latent Twins: Learning Stochastic Solution Operators

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7395209)