Diffusion-based generative models with state-of-the-art likelihoods on image density estimation benchmarks

Home

Discover

Bookmarks

Keep Up With Latest Trending Papers. Computer Science, AI and Machine Learning and more.Subscribe

Bookmarked byAung Aung and Opensourcist

CC BY

Source arXiv

Machine Learning

Stats Machine Learning

Diffusion-based generative models with state-of-the-art likelihoods on image density estimation benchmarks

Variational Diffusion Models

We introduce a family of diffusion-based generative models that obtain state-of-the-art likelihoods on image density estimation benchmarks, outperforming autoregressive models that have dominated these benchmarks for many years with often significantly faster optimization.

We show that the variational lower bound simplifies to a remarkably short expression in terms of the signal-to-noise ratio of the diffused data, thereby improving our theoretical understanding of this model class.

We also show that the continuous-time variational lower bound is invariant to the noise schedule, except for the signal-to-noise ratio at its endpoints.

Combining these advances with architectural improvements, we obtain state-of-the-art likelihoods on image density estimation benchmarks, outperforming autoregressive models that have dominated these benchmarks for many years with often significantly faster optimization.

Authors

Diederik P. Kingma, Tim Salimans, Ben Poole, Jonathan Ho

◐ Latest Activity

◐ Recommended Follows