The new model, dubbed Genie, can create playable environments from a single image prompts.
For better or for worse, generative AIs continue to evolve by leaps and bounds with each passing day, a fact that has once again been proven by Google's DeepMind team with the reveal of Genie, a new AI-powered model capable of creating entire games from just a single image prompt.
Trained without any action labels on a large dataset of publicly available Internet videos, Genie can turn any image, whether it's a real-world photograph, a sketch, an AI-generated image, or a painting, into a simplistic 2D platformer, with the team noting that this approach is versatile and applicable across various domains. Moreover, developers highlight that this new model opens the door for future AI agents to be trained "in a never-ending curriculum of new, generated worlds."
"At 11B parameters, Genie can be considered a foundation world model," comments the team. "It is comprised of a spatiotemporal video tokenizer, an autoregressive dynamics model, and a simple and scalable latent action model.
Genie enables users to act in the generated environments on a frame-by-frame basis despite training without any ground-truth action labels or other domain-specific requirements typically found in the world model literature. Further, the resulting learned latent action space facilitates training agents to imitate behaviors from unseen videos, opening the path for training generalist agents of the future."
"Genie also has implications for training generalist agents. Previous works have shown that game environments can be an effective testbed for developing AI agents, but we are often limited by the number of games available.
With Genie, our future AI agents can be trained in a never-ending curriculum of new, generated worlds. In our paper, we have a proof of concept that the latent actions learned by Genie can transfer to real human-designed environments, but this is just scratching the surface of what may be possible in the future."
Learn more about Genie here and don't forget to join our 80 Level Talent platform and our Telegram channel, follow us on Instagram, Twitter, and LinkedIn, where we share breakdowns, the latest news, awesome artworks, and more.