Google introduced its artificial intelligence that can create 3D games!
See Full Size
Able to model many real-world effects
Genie 2, developed by DeepMind, is not a game engine; instead, it’s a diffusion model that produces images as the player (a human or other AI agent) moves through the world the software simulates. Genie 2 can reveal details about the environment while producing frames. This gives him ability to model effects such as water, smoke, light and physics It makes money. The model can render scenes from first-person, third-person or isometric perspectives.
Genie 2 simulated can remember parts of a scene even after they are out of the player’s field of vision and can reconstruct these items correctly once they become visible again.
However, Genie 2 has some limitations. DeepMind, the model up to 60 seconds He says he can produce coherent scenes. The sample videos published are much shorter, approximately 10 to 20 seconds long. Also, when Genie 2 needs to maintain the illusion of a coherent world for longer, Image distortion occurs and image quality decreases. DeepMind did not share details about how it trained Genie 2, other than stating that it was based on a large-scale video dataset.
It is not currently possible for Genie 2, which is in the experimental phase, to be available in the near future, but radically change the game world It has potential. For now, the company sees the model as a tool to train and evaluate other AI agents, including its own SIMA algorithm. It will also serve as a tool that artists and designers can use to quickly prototype and test their ideas. DeepMind in the future He expects world models like Genie 2 to play an important role on the path to artificial general intelligence.
This news our mobile application Download using
You can read it whenever you want (even offline):
