Bandai Namco is celebrating Pac-Man’s 40th birthday all year long. But, Friday is technically the big day: Namco began publicly testing Pac-Man in Tokyo arcades on May 22, 1980. A lot has changed in the intervening four decades. Including, of course, the capabilities of computers. Artificial intelligence has advanced to the point of being able to drive cars and produce reasonably convincing “deepfakes” in both audio and video. Now it’s easy to understand how video games work just by watching them being played.
Nvidia Research announced Friday that it has produced a new iteration of Pac-Man. It was generated entirely by AI. The company built an AI model that was able to create a fully functional playable version. A version consisting of the seminal 8-bit arcade game without access to the underlying game engine. With no innate understanding of Pac-Man’s gameplay, the AI “trained” by watching sessions of Pac-Man — the official version from Bandai Namco — to learn the game’s rules and mechanics.
“We trained this artificial intelligence on 50,000 episodes of Pac-Man being played, without the AI actually seeing any of the code or anything — just seeing pixels coming out of the game engine,” said Rev Lebaredian, vice president of simulation technology at Nvidia, in a media briefing earlier this week. “It observed it just like a human might.”
The AI model in question is known as Nvidia GameGAN. It relies on generative adversarial networks (GAN). It’s a common system in machine learning that pits two neural networks against each other for applications such as AI-generated images. Meanwhile, GameGAN is the first GAN to be able to reproduce a video game on its own, according to Nvidia.
“This is the first research to emulate a game engine using GAN-based neural networks,” said Seung-Wook Kim, an Nvidia researcher and the project lead for GameGAN, in an Nvidia blog post. “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.”
Nvidia’s researchers gave GameGAN only two inputs: the footage of the Pac-Man play sessions paired with data on the keystrokes which is used to control the game. The training took place over four days on an Nvidia DGX system, one of the company’s AI workstations, using four Nvidia Quadro GV100 GPUs.
By observing the gameplay in the 50,000 “episodes” of Pac-Man, GameGAN learned how the game works. It figured out that Pac-Man moves around the maze while he can’t travel through walls. They learned that the ghosts chase Pac-Man, and that the game ends if one touches him. Also, that the ghosts turn blue when Pac-Man eats a power pellet, and that the pellet allows him to eat the ghosts.
The sessions in question were themselves played by an AI agent, not by humans. This ultimately resulted in the GameGAN version of Pac-Man being a somewhat inaccurate representation of the real thing. That’s because the AI agent playing the game was too good at it: “The Pac-Man almost never dies,” explained Sanja Fidler, director of Nvidia’s Toronto research lab and a co-author on the GameGAN project, during a briefing. “So the learned GameGAN that reproduces this game has this bias of never killing Pac-Man.”
What that means in practice is that if you’re playing the GameGAN version of Pac-Man, and you make a move that would ordinarily result in Pac-Man’s death, the AI goes out of its way to avoid that outcome. Sometimes he would even break the rules of the game to do so. The environment of the game might be affected by this.
Nvidia really believes that GameGAN could have all kinds of real-world applications that would help people like game developers.
“We’re going to be applying this not just to 2D classic games like this, but also to modern 3D-style games, and even things that aren’t really games,” said Lebaredian. “We can see a road to much more complex simulators that are created from this fundamental idea.”
Lebaredian explained that GameGAN could be useful in developing an AI tool that assists artists with asset generation. This is some of the grunt work of game development.
Imagine being able to train an AI on the visual style and “rules” of a game world, and having it produce new art assets that make sense in the context of that world. Even procedural generation requires a lot of initial work to set up. GameGAN, said Lebaredian, is “potentially a way to short-circuit some of that work.”
“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,” Fidler said in the Nvidia blog post. “GameGAN is the first step toward that.”
Nvidia plans to publicly release the GameGAN-generated version of Pac-Man this summer.
GameSpot’s Play For Charity Streams Will Feature Big Name Guests
In early May we announced Play For All. А GameSpot event. It takes the exciting news, interviews, previews, and analysis of gaming’s biggest summer announcements, and combines it with a fundraising effort to help healthcare workers. The world at large is engaging in social distancing. But let’s not forget people who are putting themselves at risk to help others every single day. The goal is to bring together the community to celebrate video games and enjoy all the exciting new announcements. Meanwhile also raising money for an important cause.
To achieve this, organizers had to reach out to a bunch of their friends from across the industry and convinced them to come and hang out. To play games, and have some fun with the rest. These charity streams will deliver hours of entertainment. Also, since they’re partnering with Direct Relief to fundraise, players will have the chance to watch along from the comfort of your home. Some will even get the chance to donate to support the cause or spread the word.
Play For All kicks off on June 1, and they will have daily streams taking place from 12 PM – 2 PM PT/3 PM – 5 PM ET/8 PM – 10 PM BST. As events continue to be announced we’ll all adjust our schedule. You’ll be able to watch all the streams on GameSpot.com, Twitch, YouTube, Facebook, and Twitter.
We’re going to kick things off with a crossover with pals at Giant Bomb. A stream that’ll have an assortment of folks from both sites playing games together and no doubt causing a bit of internet mischief. Then, over the next few weeks you’ll be able to see streams from Ben Hanson and the MinnMax crew; writer, host, actor, producer extraordinaire Alanah Pearce; the voice of Assassin’s Creed Origins’ Bayek and Aya, Abubakar Salim and Alex Wilton Regan; the Blood God of Bloodborne, HeyZeusHeresToast; Troy Baker aka the voice of every video game character we
love; and accomplished documentary maker and gaming’s greatest Irishman, NoClip’s Danny O’Dwyer.
We will also have Chad Michael Collins, the voice of Alex in Call of Duty: Modern Warfare; Roger Craig Smith and Justine Huxley, the voices of Apex Legends’ Mirage and Wattson; the return of Skyrim Mods with former GameSpot hosts Seb and Cam; Joe Zieja, the voice of Fire Emblem: Three Houses’ Claude; actor, host, and producer Naomi Kyle; the What’s Good Games crew; and chicken wing fan and DC comics writer, Kinda Funny’s Greg Miller.
Many more guests are about to be announced over the coming days. Of course, GameSpot is also working closely with publishers and developers to bring you up-to-date coverage on the hottest games of the summer.