On Pac-Man’s 40th Anniversary, NVIDIA AI Recreates The Iconic Game From Scratch

by Darshit Singh 3 years ago Views 2448

Nvidia says it will be releasing the recreated game online in the near future which is playable by humans. Without any coding and pre-rendered images for the software to draw, the AI model was simply shown the game in action after which it was recreated frame by frame from this small set of information

On Pac-Man’s 40th Anniversary, NVIDIA AI Recreates
                                                                                                        Image credit - Wikimedia Commons

Nvidia is known for its powerful graphics cards, but the company has been doing some serious work into artificial intelligence, too. For its latest venture, Nvidia researchers taught an AI system to recreate the game of Pac-Man from scratch simply by watching it being played.


Nvidia says it will be releasing the recreated game online in the near future which is playable by humans. Without any coding and pre-rendered images for the software to draw, the AI model was simply shown the game in action after which it was recreated frame by frame from this small set of information.

In a blog post, NVIDIA said that “even without understanding a game’s fundamental rules, AI can recreate the game with convincing results.” 

NVIDIA’s AI program which created iT is called NVIDIA GameGAN where GAN stands for Generative Adversarial Network and is a common structure used in machine learning.

Though the game developed by the GameGAN is not perfect, it is blurry and has other drawbacks but the basic dynamics of the game are the same: eat pellets, avoid ghosts, and try not to die.

“We wanted to see whether the AI could learn the rules of an environment just by looking at the screenplay of an agent moving through the game. And it did.” said Seung-Wook Kim, an NVIDIA researcher and lead author on the project.

After this feat, game developers could automatically generate layouts for new game levels, and AI researchers can easily develop simulator systems for training robots, which as of now is a cumbersome and time-consuming process.

Explaining the bigger picture by citing an example of a camera installed inside a car “We could eventually have an AI that can learn to mimic the rules of driving, the laws of physics, just by watching videos and seeing agents take actions in an environment,” said Sanja Fidler, director of NVIDIA’s Toronto research lab. “GameGAN is the first step toward that.”

Latest Videos

Latest Videos

Facebook Feed