Learning finite-dimensional coding schemes with nonlinear reconstruction maps

12/23/2018
by   Jaeho Lee, et al.
0

This paper generalizes the Maurer--Pontil framework of finite-dimensional lossy coding schemes to the setting where a high-dimensional random vector is mapped to an element of a compact set in a lower-dimensional Euclidean space, and the reconstruction map belongs to a given class of nonlinear maps. We establish a connection to approximate generative modeling under structural constraints using the tools from the theory of optimal transportation. Next, we consider the problem of learning a coding scheme on the basis of a finite collection of training samples and present generalization bounds that hold with high probability. We then illustrate the general theory in the setting where the reconstruction maps are implemented by deep neural nets.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
01/03/2016

Dimensionality-Dependent Generalization Bounds for k-Dimensional Coding Schemes

The k-dimensional coding schemes refer to a collection of methods that a...
research
04/03/2022

Fitting an immersed submanifold to data via Sussmann's orbit theorem

This paper describes an approach for fitting an immersed submanifold of ...
research
08/21/2013

Invertibility and Robustness of Phaseless Reconstruction

This paper is concerned with the question of reconstructing a vector in ...
research
04/15/2021

Approximate and discrete Euclidean vector bundles

We introduce ε-approximate versions of the notion of Euclidean vector bu...
research
02/03/2010

K-Dimensional Coding Schemes in Hilbert Spaces

This paper presents a general coding method where data in a Hilbert spac...
research
11/08/2022

Multiple Packing: Lower Bounds via Infinite Constellations

We study the problem of high-dimensional multiple packing in Euclidean s...
research
12/09/2022

Max filtering with reflection groups

Given a finite-dimensional real inner product space V and a finite subgr...

Please sign up or login with your details

Forgot password? Click here to reset