Live data from Hacker News

Pac-Man recreated with a GAN trained on 50k game episodes

blogs.nvidia.com

1–10 of 154 posts

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#5
post #3

For those confused, this is just a recreation of an animated image of pacman, not a playable recreation of the actual game. Impressive still, but not exactly what's lead to be believed by the title.

It's not quite as playable as the full pacman game, but it is more than an animation since it's conditioned on the action taken

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#6
post #3

For those confused, this is just a recreation of an animated image of pacman, not a playable recreation of the actual game. Impressive still, but not exactly what's lead to be believed by the title.

Is it? says the model keep track of user input to generate appropriate frames.

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#9
Interesting to explore how this can be used. You still need to develop a game with an actual engine first, to generate all the samples for training the AI how to generate frames without an engine.

But maybe if training sample size can be considerably reduced, game making can open up to far less technical people. Instead of developing with an engine, game makers could show the AI "when I press this button, change the scene like this", and the AI could try to figure out the rules from given samples.

Re: Pac-Man recreated with a GAN trained on 50k game episodes

#10

Thought experiment: Is my brain also generating reality without a physics engine and is filling in details that it thinks should be there? If true, what are the advantages/disadvantages of it?

Yes, absolutely. You can catch a baseball because your brain predicts where it will go before your eyes see its actual movement. This may be a good place to start learning more: https://en.wikipedia.org/wiki/Direct_and_indirect_realism
Post reply on HN