Diffusing Innovation
Another recent innovation this book hasn’t completely covered is the success of diffusion models in generative applications. As you briefly learned in What Is the State of the Art?, latent diffusion is the process of progressively denoising over long time series. This process can be used to start from an image of random noise to generate incredibly realistic images.
At their core, diffusion models are based on a process of iterative refinement. Initially, they start with a pattern of random noise. Over successive iterations, this noise is gradually shaped into a coherent output, whether that be an image, text, or another form of data. While their most popular application is in generative art, diffusion models have been ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access