Nvidia researchers use AI to re-create Pac-Man from scratch – Software – News | Instant News

Nvidia researchers achieved an interesting AI achievement to be associated with 40th Pac-Man’s birthday arrives at an arcade in Japan. Ahead of this classic pill chewing ghosts on the anniversary of the labyrinth game, researchers training a powerful new AI model called Nvidia GameGAN with game footage. The result: GameGAN produced a fully functional Pac-Man version from the start.

To prepare GameGAN for its game duplication tasks, Nvidia researchers fed AI 50,000 episodes of the Pac-Man gameplay. AI processes this data and issues duplicate, fully functional games, without reference to the original game engine. “That means that even without understanding the basic rules of the game, AI can create games with convincing results,” Nvidia confirmed.

HEXUS has published news about GAN before. In AI, GAN refers to the Generative Hostility Network. Nvidia briefly explained that GAN consists of two competing neural networks, a generator and a discriminator. GAN will help minimize human training inputs for AI because it pairs a creative generator with an algorithm that quality checks output – differentiator. “You can think of this as competition between counterfeiters and police,” the word GAN’s pioneer, Ian Goodfellow, is now a research scientist at Google.

A fully functional Pac-Man game is Gim “The first research to emulate a gaming machine using a GAN-based neural network,” claimed Seung-Wook Kim, an Nvidia researcher and lead author on the project.

Publisher Pacco Nam Bandai is quite staggered by the release of GameGAN, reports the Nvidia blog. The publisher ponders whether GameGAN can see practical use in the gaming industry. For example it will be used for “Speed ​​up the creative process of developing new level layouts, characters and even games”.

Later this year you will be able to try the Nvidia AI tribute to Pac-Man through AI Playground, where there are a number of interactive demos of Nvidia AI technology available.

Of course this research might apply outside the gaming environment. Nvidia went on to talk about other AI projects he was working on such as autonomous robots, simulators, and autonomous cars where GameGAN AI might apply.

Other recent technology warnings include: Java press 25, Microsoft Solitaire press 30, and WordPress 17 years old tomorrow.

image source

to request modification Contact us at Here or [email protected]