Google’s DeepMind division has developed an artificial intelligence model, Genie, which is capable of transforming images into video games. It allows you to create game worlds for platformers in just a few simple steps.
The Genie model is relatively small, with 11 billion parameters. It is trained on more than 200,000 hours of video of people playing through two-dimensional platformers. Such games are quite formulaic, so it is not surprising that Genie figured out the mechanics and physics of actions associated with them. The training proved to be quite effective even though the video streams did not contain information about when a button or control was pressed.
As a result, Genie takes a single image (photo, sketch, or AI-generated image) and turns it into a playable game that responds to the user’s controls. In fact, the image is transformed into a rudimentary interactive environment in one step.
I am really excited to reveal what @GoogleDeepMind’s Open Endedness Team has been up to ????. We introduce Genie ????, a foundation world model trained exclusively from Internet videos that can generate infinite variations of action-controlling 2D worlds given image prompts. pic.twitter.com/TnQ8uv81wc
— Tim Rocktäschel (@_rockt) February 26, 2024
So far, you shouldn’t expect a high-quality model of the games created. Genie is a research project, not a final product. The model was trained on video with an ultra-low resolution of 160×90 pixels and a frequency of only 10 frames per second. As such, it generates similarly low-resolution “games” that run for only 16 seconds at 1 fps.
However, the basic concept has been tested, and there is every indication that Genie greatly improves scaling. This only requires using longer, higher-resolution videos and additional processing power.
Source: newatlas
The competition of ITS authors is ongoing. Write an article about game development, gaming and gaming devices and win a professional Logitech G923 Racing Wheel or one of the low-profile Logitech G815 LIGHTSYNC RGB Mechanical Gaming Keyboard!
