What Is Diffusion in AI and How Does It Work?
Understand the core concept behind generative AI models
This year, we witnessed the rise of text-to-image AI as advancements significantly improved the quality of output AI models can produce.
Controversial AI solutions such as Stable Diffusion and OpenAI’s DALL-E have been integrated into websites like Canva or DeviantArt to provide advanced creative tools, customized branding, and even generate product ideas.
Yet, the technology behind these systems, known as diffusion, has the potential to do more than create art. Much more. Some research groups are using diffusion to digitize music, DNA sequences, and develop new drugs, and the tech has improved at a staggering pace over the past few years — so this is only the beginning.
What exactly is diffusion, and why is it such a significant advancement, you might ask? Let’s take a look at the origins of diffusion and how it’s evolved to become the powerful technological driver it is today.
How diffusion came about
Remember the hysteria around deepfakes a few years back — such as the famous “ deeptom” TikTok experiment?
These apps used AI to insert people’s faces (or their entire bodies) into existing images and videos…
