Search Results for author:
Found 8 papers, 0 papers with code
An autoencoder compression approach for accelerating large-scale inverse problems
In this work, we propose a close-to-ideal scalable compression approach using autoencoders to eliminate the need for checkpointing and substantial memory storage, thereby reducing both the time-to-solution and memory requirements.
Layerwise Sparsifying Training and Sequential Learning Strategy for Neural Architecture Adaptation
We derive the necessary conditions for trainability of a newly added layer and analyze the role of manifold regularization.
A Model-Constrained Tangent Slope Learning Approach for Dynamical Systems
Real-time accurate solutions of large-scale complex dynamical systems are in critical need for control, optimization, uncertainty quantification, and decision-making in practical engineering and science applications, especially digital twin applications.
A Unified and Constructive Framework for the Universality of Neural Networks
This paper is the first effort to provide a unified and constructive framework for the universality of a large class of activation functions including most of existing ones.
TNet: A Model-Constrained Tikhonov Network Approach for Inverse Problems
This is the strength of DL but also one of its key limitations when applied to science and engineering problems in which underlying physical properties and desired accuracy need to be achieved.
A Multilevel Block Preconditioner for the HDG Trace System Applied to Incompressible Resistive MHD
Additionally for the upper block a preliminary study of an alternate nodal block system solver based on a multilevel approximate nested dissection is presented.
Numerical Analysis Numerical Analysis Computational Physics Plasma Physics
Accelerating PDE-constrained Inverse Solutions with Deep Learning and Reduced Order Models
In this paper, we propose to enlarge the validity of ROMs and hence improve the accuracy outside the reduced subspaces by incorporating a data-driven ML technique.
Solving Bayesian Inverse Problems via Variational Autoencoders
Further, this framework possesses an inherent adaptive optimization property that emerges through the learning of the posterior uncertainty.
