BREAKTHROUGH – Los Alamos Generates AI Pictures Without Using a Random Picture to Start

Los Alamos national lab made an artificial intelligence framework called Blackout Diffusion generates images from a completely empty picture. The machine-learning algorithm, unlike other generative diffusion models, does not require initiating a random seed to get started.

The results are comparable to DALL-E or Midjourney, but require fewer computational resources than these models.

Diffusion models create samples similar to the data they are trained on. They work by taking an image and repeatedly adding noise until the image is unrecognizable. Throughout the process the model tries to learn how to revert it back to its original state.

An important unique aspect of Blackout Diffusion is the space it works in. Existing generative diffusion models work in continuous spaces, meaning the space they work in is dense and infinite. However, working in continuous spaces limits their potential for scientific applications. Blackout Diffusion means that entirely new types of images can be created by AI.

Arxiv – Blackout Diffusion: Generative Diffusion Models in Discrete-State Spaces

They believe that with further work, discrete-space Diffusion Models may be found that are far more computationally efficient than Gaussian Diffusion Models, because the state space is far smaller.