On the Stability of Iterative Retraining of Generative Models on their own Data

arXiv:2310.00429v4 Announce Type: replace
Abstract: Deep generative models have made tremendous progress in modeling complex data, often exhibiting generation quality that surpasses a typical human's ability to discern the authenticity of samples. Undeniably, a key driver of this success is enabled by the massive amounts of web-scale data consumed by these models. Due to these models' striking performance and ease of availability, the web will inevitably be increasingly populated with synthetic content. Such a fact directly implies that future iterations …

abstract arxiv authenticity cs.lg data deep generative models driver generative generative models human iterative key massive modeling progress quality retraining samples scale stability stat.ml success type web