Diffusion models power leading generative AI, but when and how they memorize training data, especially on low-dimensional manifolds, remains unclear. We find memorization emerges gradually, not abruptly: as data become scarce, diffusion models experience a smooth collapse where their capacity to vary across independent directions diminishes. Measuring latent dimensionality via the learned score field, we reveal how generative behavior increasingly centers on a few examples while other variations "freeze out". We propose a geometric memorization theory, showing that salient features collapse first, then finer details, leading to near point-wise replication. This mirrors physical systems condensing into a few low-energy configurations. Our theoretical predictions align with both synthetic and real data, identifying geometric memorization as a distinct phase between generalization and exact copying.
@article{arxiv.2410.08727,
title = {Losing dimensions: Geometric memorization in generative diffusion},
author = {Beatrice Achilli and Enrico Ventura and Gianluigi Silvestri and Bao Pham and Gabriel Raya and Dmitry Krotov and Carlo Lucibello and Luca Ambrogioni},
journal= {arXiv preprint arXiv:2410.08727},
year = {2026}
}