Saved in:
Bibliographic Details
Main Author: Turinici, Gabriel
Format: Preprint
Published: 2022
Subjects:
Online Access:https://arxiv.org/abs/2202.09573
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914683891482624
author Turinici, Gabriel
author_facet Turinici, Gabriel
contents The decoder-based machine learning generative algorithms such as Generative Adversarial Networks (GAN), Variational Auto-Encoders (VAE), Transformers show impressive results when constructing objects similar to those in a training ensemble. However, the generation of new objects builds mainly on the understanding of the hidden structure of the training dataset followed by a sampling from a multi-dimensional normal variable. In particular each sample is independent from the others and can repeatedly propose same type of objects. To cure this drawback we introduce a kernel-based measure quantization method that can produce new objects from a given target measure by approximating it as a whole and even staying away from elements already drawn from that distribution. This ensures a better diversity of the produced objects. The method is tested on classic machine learning benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2202_09573
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Diversity in deep generative models and generative AI
Turinici, Gabriel
Computer Vision and Pattern Recognition
Artificial Intelligence
I.2; G.3
The decoder-based machine learning generative algorithms such as Generative Adversarial Networks (GAN), Variational Auto-Encoders (VAE), Transformers show impressive results when constructing objects similar to those in a training ensemble. However, the generation of new objects builds mainly on the understanding of the hidden structure of the training dataset followed by a sampling from a multi-dimensional normal variable. In particular each sample is independent from the others and can repeatedly propose same type of objects. To cure this drawback we introduce a kernel-based measure quantization method that can produce new objects from a given target measure by approximating it as a whole and even staying away from elements already drawn from that distribution. This ensures a better diversity of the produced objects. The method is tested on classic machine learning benchmarks.
title Diversity in deep generative models and generative AI
topic Computer Vision and Pattern Recognition
Artificial Intelligence
I.2; G.3
url https://arxiv.org/abs/2202.09573