MAGAN: A Meta-Analysis for Generative Adversarial Networks’ Latent Space

Frederic Rizk, Rodrigue Rizk, Dominick Rizk, Chee-Hung Henry Chu, Chee-Hung Henry Chu

2023

Abstract

Generative Adversarial Networks (GANs) are an emerging class of deep neural networks that has sparked considerable interest in the field of unsupervised learning because of its exceptional data generation performance. Nevertheless, the GAN’s latent space that represents the core of these generative models has not been studied in depth in terms of its effect on the generated image space. In this paper, we propose and evaluate MAGAN, an algorithm for Meta-Analysis for GANs’ latent space. GAN-derived synthetic images are also evaluated in terms of their efficiency in complementing the data training, where the produced output is employed for data augmentation, mitigating the labeled data scarcity. The results suggest that GANs may be used as a parameter-controlled data generator for data-driven augmentation. The quantitative findings show that MAGAN can correctly trace the relationship between the arithmetic adjustments in the latent space and their effects on the output in the image space. We empirically determine the parameter ε for each class such that the latent space is insensitive to a shift of ε×σ from the mean vector, where σ is the standard deviation of a particular class.

Paper Citation

in Harvard Style

Rizk F., Rizk R., Rizk D. and Henry Chu C. (2023). MAGAN: A Meta-Analysis for Generative Adversarial Networks’ Latent Space. In Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-626-2, pages 488-494. DOI: 10.5220/0011771900003411

in Bibtex Style

@conference{icpram23,
author={Frederic Rizk and Rodrigue Rizk and Dominick Rizk and Chee-Hung Henry Chu},
title={MAGAN: A Meta-Analysis for Generative Adversarial Networks’ Latent Space},
booktitle={Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2023},
pages={488-494},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011771900003411},
isbn={978-989-758-626-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - MAGAN: A Meta-Analysis for Generative Adversarial Networks’ Latent Space
SN - 978-989-758-626-2
AU - Rizk F.
AU - Rizk R.
AU - Rizk D.
AU - Henry Chu C.
PY - 2023
SP - 488
EP - 494
DO - 10.5220/0011771900003411