Deep learning with differential Gaussian process flows

10/09/2018
by   Pashupati Hegde, et al.
16

We propose a novel deep learning paradigm of differential flows that learn a stochastic differential equation transformations of inputs prior to a standard classification or regression function. The key property of differential Gaussian processes is the warping of inputs through infinitely deep, but infinitesimal, differential fields, that generalise discrete layers into a dynamical system. We demonstrate state-of-the-art results that exceed the performance of deep Gaussian processes and neural networks

READ FULL TEXT

page 1

page 2

page 3

page 4

research
09/05/2022

Nonparametric estimation of linear multiplier in SDEs driven by general Gaussian processes

We investigate the asymptotic properties of a kernel-type nonparametric ...
research
06/24/2022

Physically Consistent Learning of Conservative Lagrangian Systems with Gaussian Processes

This paper proposes a physically consistent Gaussian Process (GP) enabli...
research
11/10/2020

Learning ODE Models with Qualitative Structure Using Gaussian Processes

Recent advances in learning techniques have enabled the modelling of dyn...
research
02/07/2021

Latent Map Gaussian Processes for Mixed Variable Metamodeling

Gaussian processes (GPs) are ubiquitously used in sciences and engineeri...
research
03/12/2018

Learning unknown ODE models with Gaussian processes

In conventional ODE modelling coefficients of an equation driving the sy...
research
06/27/2022

Supernova Light Curves Approximation based on Neural Network Models

Photometric data-driven classification of supernovae becomes a challenge...
research
12/13/2018

High dimensional inference for the structural health monitoring of lock gates

Locks and dams are critical pieces of inland waterways. However, many co...

Please sign up or login with your details

Forgot password? Click here to reset