Saved in:
Bibliographic Details
Main Authors: De la Fuente, Alfredo, Singh, Saurabh, Ballé, Johannes
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2402.15345
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914689748828160
author De la Fuente, Alfredo
Singh, Saurabh
Ballé, Johannes
author_facet De la Fuente, Alfredo
Singh, Saurabh
Ballé, Johannes
contents We introduce a lightweight, flexible and end-to-end trainable probability density model parameterized by a constrained Fourier basis. We assess its performance at approximating a range of multi-modal 1D densities, which are generally difficult to fit. In comparison to the deep factorized model introduced in [1], our model achieves a lower cross entropy at a similar computational budget. In addition, we also evaluate our method on a toy compression task, demonstrating its utility in learned compression.
format Preprint
id arxiv_https___arxiv_org_abs_2402_15345
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Fourier Basis Density Model
De la Fuente, Alfredo
Singh, Saurabh
Ballé, Johannes
Machine Learning
We introduce a lightweight, flexible and end-to-end trainable probability density model parameterized by a constrained Fourier basis. We assess its performance at approximating a range of multi-modal 1D densities, which are generally difficult to fit. In comparison to the deep factorized model introduced in [1], our model achieves a lower cross entropy at a similar computational budget. In addition, we also evaluate our method on a toy compression task, demonstrating its utility in learned compression.
title Fourier Basis Density Model
topic Machine Learning
url https://arxiv.org/abs/2402.15345