Google Research· ResearchResearch paper

Towards demystifying the creativity of diffusion models

Google Research published a paper showing that the creativity of diffusion models is a mathematical consequence of neural networks learning a smoothed version of the score function, which causes interpolation between training data points.

From the source

We show that a diffusion model’s creativity (its ability to generate novel data, rather than just memorize its training set) is a mathematical consequence of neural networks learning a "smoothed" version of the score function, driving the model to interpolate between training data points along the hidden data manifold.

research.google

Announced

15 Jul 2026 18:06 UTC

Detected