Generative AI enables users to generate new content based on a variety of inputs — text, images, sounds, animation, 3D models, or other types of data. Unlike traditional AI that classifies or predicts, generative AI creates.
<aside> 💡
AI Chatbots use generative AI to provide responses based on a single interaction. A person makes a query (request for information) and the chatbot uses Natural Language Processing (NLP) to reply.
</aside>
Generative AI uses neural networks to identify patterns and structures within existing data, then uses those learned patterns to generate new and original content.
One of the breakthroughs with generative AI is the ability to leverage unsupervised or semi-supervised learning for training. This allows organisations to use large amounts of unlabelled data to create foundation models — large-scale base models that can be adapted for multiple tasks. Examples include GPT-3 (text generation) and Stable Diffusion (image generation).
Improving Diffusion Models as an Alternative To GANs, Part 1 | NVIDIA Technical Blog
There are multiple types of generative models. Each has different strengths, and combining the best attributes of each leads to even more powerful models.
Also known as denoising diffusion probabilistic models (DDPMs), diffusion models are generative models that determine vectors in latent space through a two-step process during training: