Continuous Methods : Hamiltonian Domain Translation

07/08/2022

∙

This paper proposes a novel approach to domain translation. Leveraging established parallels between generative models and dynamical systems, we propose a reformulation of the Cycle-GAN architecture. By embedding our model with a Hamiltonian structure, we obtain a continuous, expressive and most importantly invertible generative model for domain translation.

READ FULL TEXT

Continuous Methods : Hamiltonian Domain Translation

Separable Hamiltonian Neural Networks

Discrete gradient methods for irreversible port-Hamiltonian systems

Hamiltonian Generative Networks

Improved a posteriori Error Bounds for Reduced port-Hamiltonian Systems

Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data

Continuous-Stage Runge-Kutta approximation to Differential Problems

Deep Ptych: Subsampled Fourier Ptychography using Generative Priors

Continuous Methods : Hamiltonian Domain Translation

Related Research

Separable Hamiltonian Neural Networks

Discrete gradient methods for irreversible port-Hamiltonian systems

Hamiltonian Generative Networks

Improved a posteriori Error Bounds for Reduced port-Hamiltonian Systems

Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data

Continuous-Stage Runge-Kutta approximation to Differential Problems

Deep Ptych: Subsampled Fourier Ptychography using Generative Priors