ORIGINAL RESEARCH article
Front. Artif. Intell.
Sec. Machine Learning and Artificial Intelligence
This article is part of the Research TopicEthical Artificial Intelligence: Methods and ApplicationsView all articles
Do Generative Models Learn Rare Generative Factors?
Provisionally accepted
- University of Edinburgh, Edinburgh, United Kingdom
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Generative models are becoming a promising tool in AI alongside discriminative learning. Several models have been proposed to learn in an unsupervised fashion the corresponding generative factors, namely the latent variables critical for capturing the full spectrum of data variability. Diffusion Models (DMs), Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs) are of particular interest due to their impressive ability to generate highly realistic data. Through a systematic empirical study, this paper delves into the intricate challenge of how DMs, GANs and VAEs internalize and replicate rare generative factors. Our findings reveal a pronounced tendency towards memorisation of these factors. We study the reasons for this memorisation and demonstrate that strategies such as spectral decoupling can mitigate this issue to a certain extent
Keywords: Generative factors, latent variables, Diffusion Models (DMs), generative adversarial networks (GANs), variational autoencoders (VAEs), rare generative factors, Rare generative factors (RGFs)
Received: 01 Sep 2025; Accepted: 28 Oct 2025.
Copyright: © 2025 Haider, Moroshko, Xue and Tsaftaris. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Fasih Haider, fasih.haider@ed.ac.uk
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.